diff --git "a/trainer_state.json" "b/trainer_state.json"
--- "a/trainer_state.json"
+++ "b/trainer_state.json"
@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.665983186385073,
+  "epoch": 2.17859339758048,
   "eval_steps": 500,
-  "global_step": 65000,
+  "global_step": 85000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -390007,13 +390007,120013 @@
       "learning_rate": 8.697664399072166e-07,
       "loss": 0.6416,
       "step": 65000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.697390051629402e-07,
+      "loss": 0.647,
+      "step": 65001
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.697115705183984e-07,
+      "loss": 0.5986,
+      "step": 65002
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.696841359736127e-07,
+      "loss": 0.6035,
+      "step": 65003
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.696567015286037e-07,
+      "loss": 0.6846,
+      "step": 65004
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.696292671833927e-07,
+      "loss": 0.5737,
+      "step": 65005
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.696018329380003e-07,
+      "loss": 0.6299,
+      "step": 65006
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.695743987924482e-07,
+      "loss": 0.6592,
+      "step": 65007
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.695469647467567e-07,
+      "loss": 0.6426,
+      "step": 65008
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.695195308009473e-07,
+      "loss": 0.6465,
+      "step": 65009
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.69492096955041e-07,
+      "loss": 0.6621,
+      "step": 65010
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.694646632090582e-07,
+      "loss": 0.6924,
+      "step": 65011
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.694372295630204e-07,
+      "loss": 0.7881,
+      "step": 65012
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.694097960169484e-07,
+      "loss": 0.7549,
+      "step": 65013
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.693823625708636e-07,
+      "loss": 0.5713,
+      "step": 65014
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.693549292247864e-07,
+      "loss": 0.6377,
+      "step": 65015
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.693274959787383e-07,
+      "loss": 0.6455,
+      "step": 65016
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.6930006283274e-07,
+      "loss": 0.6494,
+      "step": 65017
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.692726297868128e-07,
+      "loss": 0.5,
+      "step": 65018
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.692451968409777e-07,
+      "loss": 0.7271,
+      "step": 65019
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.692177639952555e-07,
+      "loss": 0.6025,
+      "step": 65020
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.691903312496669e-07,
+      "loss": 0.519,
+      "step": 65021
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.691628986042335e-07,
+      "loss": 0.7061,
+      "step": 65022
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.691354660589758e-07,
+      "loss": 0.5186,
+      "step": 65023
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.691080336139153e-07,
+      "loss": 0.6206,
+      "step": 65024
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.690806012690725e-07,
+      "loss": 0.7734,
+      "step": 65025
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.690531690244689e-07,
+      "loss": 0.4861,
+      "step": 65026
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.690257368801251e-07,
+      "loss": 0.4944,
+      "step": 65027
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.689983048360625e-07,
+      "loss": 0.7686,
+      "step": 65028
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68970872892302e-07,
+      "loss": 0.668,
+      "step": 65029
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68943441048864e-07,
+      "loss": 0.5969,
+      "step": 65030
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.689160093057703e-07,
+      "loss": 0.751,
+      "step": 65031
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.688885776630414e-07,
+      "loss": 0.5332,
+      "step": 65032
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.688611461206987e-07,
+      "loss": 0.4998,
+      "step": 65033
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.688337146787628e-07,
+      "loss": 0.7998,
+      "step": 65034
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.688062833372549e-07,
+      "loss": 0.5742,
+      "step": 65035
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.687788520961961e-07,
+      "loss": 0.5532,
+      "step": 65036
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.687514209556072e-07,
+      "loss": 0.4434,
+      "step": 65037
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.687239899155095e-07,
+      "loss": 0.6855,
+      "step": 65038
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68696558975924e-07,
+      "loss": 0.5771,
+      "step": 65039
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68669128136871e-07,
+      "loss": 0.71,
+      "step": 65040
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.686416973983722e-07,
+      "loss": 0.6445,
+      "step": 65041
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.686142667604484e-07,
+      "loss": 0.6973,
+      "step": 65042
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.685868362231206e-07,
+      "loss": 0.5564,
+      "step": 65043
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.685594057864096e-07,
+      "loss": 0.6377,
+      "step": 65044
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68531975450337e-07,
+      "loss": 0.7119,
+      "step": 65045
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.685045452149231e-07,
+      "loss": 0.5596,
+      "step": 65046
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.684771150801895e-07,
+      "loss": 0.54,
+      "step": 65047
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.684496850461567e-07,
+      "loss": 0.8135,
+      "step": 65048
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.684222551128464e-07,
+      "loss": 0.707,
+      "step": 65049
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.683948252802787e-07,
+      "loss": 0.6523,
+      "step": 65050
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68367395548475e-07,
+      "loss": 0.6992,
+      "step": 65051
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.683399659174566e-07,
+      "loss": 0.604,
+      "step": 65052
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.683125363872442e-07,
+      "loss": 0.7021,
+      "step": 65053
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.682851069578585e-07,
+      "loss": 0.6807,
+      "step": 65054
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.682576776293211e-07,
+      "loss": 0.6987,
+      "step": 65055
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.682302484016526e-07,
+      "loss": 0.8018,
+      "step": 65056
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.682028192748742e-07,
+      "loss": 0.5146,
+      "step": 65057
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.681753902490068e-07,
+      "loss": 0.625,
+      "step": 65058
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.681479613240718e-07,
+      "loss": 0.6196,
+      "step": 65059
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.681205325000896e-07,
+      "loss": 0.6797,
+      "step": 65060
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.680931037770812e-07,
+      "loss": 0.5229,
+      "step": 65061
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.680656751550681e-07,
+      "loss": 0.6504,
+      "step": 65062
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.68038246634071e-07,
+      "loss": 0.6572,
+      "step": 65063
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.680108182141109e-07,
+      "loss": 0.7393,
+      "step": 65064
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.679833898952088e-07,
+      "loss": 0.4739,
+      "step": 65065
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.679559616773859e-07,
+      "loss": 0.6016,
+      "step": 65066
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.679285335606628e-07,
+      "loss": 0.7754,
+      "step": 65067
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.67901105545061e-07,
+      "loss": 0.6621,
+      "step": 65068
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.678736776306015e-07,
+      "loss": 0.7822,
+      "step": 65069
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.678462498173048e-07,
+      "loss": 0.7734,
+      "step": 65070
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.678188221051919e-07,
+      "loss": 0.6855,
+      "step": 65071
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.677913944942843e-07,
+      "loss": 0.5786,
+      "step": 65072
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.677639669846027e-07,
+      "loss": 0.7646,
+      "step": 65073
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.677365395761681e-07,
+      "loss": 0.7256,
+      "step": 65074
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.677091122690015e-07,
+      "loss": 0.7383,
+      "step": 65075
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.67681685063124e-07,
+      "loss": 0.5359,
+      "step": 65076
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.676542579585564e-07,
+      "loss": 0.626,
+      "step": 65077
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.676268309553202e-07,
+      "loss": 0.7495,
+      "step": 65078
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.675994040534362e-07,
+      "loss": 0.6885,
+      "step": 65079
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.675719772529247e-07,
+      "loss": 0.752,
+      "step": 65080
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.675445505538076e-07,
+      "loss": 0.627,
+      "step": 65081
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.675171239561052e-07,
+      "loss": 0.6758,
+      "step": 65082
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.674896974598391e-07,
+      "loss": 0.4558,
+      "step": 65083
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.674622710650299e-07,
+      "loss": 0.793,
+      "step": 65084
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.67434844771699e-07,
+      "loss": 0.6367,
+      "step": 65085
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.674074185798668e-07,
+      "loss": 0.5898,
+      "step": 65086
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.67379992489555e-07,
+      "loss": 0.6394,
+      "step": 65087
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.673525665007839e-07,
+      "loss": 0.7124,
+      "step": 65088
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.673251406135755e-07,
+      "loss": 0.6362,
+      "step": 65089
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.672977148279495e-07,
+      "loss": 0.6646,
+      "step": 65090
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.672702891439279e-07,
+      "loss": 0.5166,
+      "step": 65091
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.67242863561531e-07,
+      "loss": 0.502,
+      "step": 65092
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.672154380807804e-07,
+      "loss": 0.7412,
+      "step": 65093
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.671880127016963e-07,
+      "loss": 0.8213,
+      "step": 65094
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.671605874243008e-07,
+      "loss": 0.5532,
+      "step": 65095
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.671331622486141e-07,
+      "loss": 0.8242,
+      "step": 65096
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.671057371746575e-07,
+      "loss": 0.6904,
+      "step": 65097
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.670783122024519e-07,
+      "loss": 0.6758,
+      "step": 65098
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.670508873320186e-07,
+      "loss": 0.7324,
+      "step": 65099
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.670234625633781e-07,
+      "loss": 0.7812,
+      "step": 65100
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.669960378965515e-07,
+      "loss": 0.6357,
+      "step": 65101
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.669686133315601e-07,
+      "loss": 0.7598,
+      "step": 65102
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.669411888684244e-07,
+      "loss": 0.5481,
+      "step": 65103
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.66913764507166e-07,
+      "loss": 0.5483,
+      "step": 65104
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.668863402478055e-07,
+      "loss": 0.5106,
+      "step": 65105
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.668589160903641e-07,
+      "loss": 0.6562,
+      "step": 65106
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.668314920348625e-07,
+      "loss": 0.6611,
+      "step": 65107
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.668040680813222e-07,
+      "loss": 0.4812,
+      "step": 65108
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.66776644229764e-07,
+      "loss": 0.6494,
+      "step": 65109
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.667492204802087e-07,
+      "loss": 0.7021,
+      "step": 65110
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.667217968326771e-07,
+      "loss": 0.4702,
+      "step": 65111
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.666943732871908e-07,
+      "loss": 0.5146,
+      "step": 65112
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.666669498437701e-07,
+      "loss": 0.6704,
+      "step": 65113
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.666395265024367e-07,
+      "loss": 0.7109,
+      "step": 65114
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.666121032632111e-07,
+      "loss": 0.708,
+      "step": 65115
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.665846801261146e-07,
+      "loss": 0.5938,
+      "step": 65116
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.66557257091168e-07,
+      "loss": 0.5854,
+      "step": 65117
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.665298341583924e-07,
+      "loss": 0.7012,
+      "step": 65118
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.665024113278092e-07,
+      "loss": 0.6592,
+      "step": 65119
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.664749885994384e-07,
+      "loss": 0.6338,
+      "step": 65120
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.664475659733018e-07,
+      "loss": 0.6338,
+      "step": 65121
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.664201434494199e-07,
+      "loss": 0.6558,
+      "step": 65122
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.663927210278144e-07,
+      "loss": 0.5688,
+      "step": 65123
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.663652987085052e-07,
+      "loss": 0.4788,
+      "step": 65124
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.663378764915146e-07,
+      "loss": 0.6821,
+      "step": 65125
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.663104543768626e-07,
+      "loss": 0.613,
+      "step": 65126
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.662830323645705e-07,
+      "loss": 0.5601,
+      "step": 65127
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.662556104546596e-07,
+      "loss": 0.6318,
+      "step": 65128
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.662281886471508e-07,
+      "loss": 0.6504,
+      "step": 65129
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.662007669420644e-07,
+      "loss": 0.6865,
+      "step": 65130
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.661733453394222e-07,
+      "loss": 0.5979,
+      "step": 65131
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.661459238392448e-07,
+      "loss": 0.4546,
+      "step": 65132
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.661185024415533e-07,
+      "loss": 0.5518,
+      "step": 65133
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.660910811463688e-07,
+      "loss": 0.6738,
+      "step": 65134
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.660636599537121e-07,
+      "loss": 0.5977,
+      "step": 65135
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.660362388636042e-07,
+      "loss": 0.749,
+      "step": 65136
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.660088178760665e-07,
+      "loss": 0.6206,
+      "step": 65137
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.659813969911194e-07,
+      "loss": 0.5498,
+      "step": 65138
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.659539762087846e-07,
+      "loss": 0.6797,
+      "step": 65139
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.659265555290824e-07,
+      "loss": 0.584,
+      "step": 65140
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.658991349520341e-07,
+      "loss": 0.5933,
+      "step": 65141
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.658717144776607e-07,
+      "loss": 0.7451,
+      "step": 65142
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.658442941059832e-07,
+      "loss": 0.7539,
+      "step": 65143
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.658168738370222e-07,
+      "loss": 0.5215,
+      "step": 65144
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.657894536707995e-07,
+      "loss": 0.6162,
+      "step": 65145
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.657620336073353e-07,
+      "loss": 0.6748,
+      "step": 65146
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.65734613646651e-07,
+      "loss": 0.5461,
+      "step": 65147
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.657071937887675e-07,
+      "loss": 0.6191,
+      "step": 65148
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.656797740337063e-07,
+      "loss": 0.5245,
+      "step": 65149
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.656523543814876e-07,
+      "loss": 0.5347,
+      "step": 65150
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.656249348321325e-07,
+      "loss": 0.7627,
+      "step": 65151
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.655975153856625e-07,
+      "loss": 0.5513,
+      "step": 65152
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.65570096042098e-07,
+      "loss": 0.6758,
+      "step": 65153
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.655426768014607e-07,
+      "loss": 0.7998,
+      "step": 65154
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.655152576637708e-07,
+      "loss": 0.6963,
+      "step": 65155
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.654878386290499e-07,
+      "loss": 0.4182,
+      "step": 65156
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.654604196973186e-07,
+      "loss": 0.5981,
+      "step": 65157
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.654330008685983e-07,
+      "loss": 0.667,
+      "step": 65158
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.654055821429101e-07,
+      "loss": 0.792,
+      "step": 65159
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.653781635202742e-07,
+      "loss": 0.6313,
+      "step": 65160
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.653507450007119e-07,
+      "loss": 0.6348,
+      "step": 65161
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.653233265842446e-07,
+      "loss": 0.7285,
+      "step": 65162
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.652959082708929e-07,
+      "loss": 0.6797,
+      "step": 65163
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.652684900606779e-07,
+      "loss": 0.5225,
+      "step": 65164
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.652410719536207e-07,
+      "loss": 0.5396,
+      "step": 65165
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.652136539497423e-07,
+      "loss": 0.7949,
+      "step": 65166
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.651862360490634e-07,
+      "loss": 0.6836,
+      "step": 65167
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.651588182516054e-07,
+      "loss": 0.7949,
+      "step": 65168
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.651314005573894e-07,
+      "loss": 0.5884,
+      "step": 65169
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.651039829664355e-07,
+      "loss": 0.6816,
+      "step": 65170
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.650765654787655e-07,
+      "loss": 0.6299,
+      "step": 65171
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.650491480944e-07,
+      "loss": 0.5454,
+      "step": 65172
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.650217308133603e-07,
+      "loss": 0.6494,
+      "step": 65173
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.649943136356672e-07,
+      "loss": 0.6953,
+      "step": 65174
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.649668965613419e-07,
+      "loss": 0.6172,
+      "step": 65175
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.649394795904051e-07,
+      "loss": 0.6685,
+      "step": 65176
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.649120627228781e-07,
+      "loss": 0.6006,
+      "step": 65177
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.648846459587814e-07,
+      "loss": 0.6562,
+      "step": 65178
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.648572292981371e-07,
+      "loss": 0.6436,
+      "step": 65179
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.648298127409646e-07,
+      "loss": 0.5664,
+      "step": 65180
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.64802396287286e-07,
+      "loss": 0.6416,
+      "step": 65181
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.647749799371218e-07,
+      "loss": 0.7295,
+      "step": 65182
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.647475636904934e-07,
+      "loss": 0.7363,
+      "step": 65183
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.647201475474213e-07,
+      "loss": 0.6265,
+      "step": 65184
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.64692731507927e-07,
+      "loss": 0.7588,
+      "step": 65185
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.64665315572031e-07,
+      "loss": 0.6455,
+      "step": 65186
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.646378997397548e-07,
+      "loss": 0.7744,
+      "step": 65187
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.646104840111191e-07,
+      "loss": 0.5938,
+      "step": 65188
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.64583068386145e-07,
+      "loss": 0.6831,
+      "step": 65189
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.645556528648535e-07,
+      "loss": 0.5388,
+      "step": 65190
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.645282374472651e-07,
+      "loss": 0.6543,
+      "step": 65191
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.645008221334016e-07,
+      "loss": 0.7549,
+      "step": 65192
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.644734069232832e-07,
+      "loss": 0.6777,
+      "step": 65193
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.644459918169315e-07,
+      "loss": 0.7578,
+      "step": 65194
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.644185768143671e-07,
+      "loss": 0.5239,
+      "step": 65195
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.643911619156114e-07,
+      "loss": 0.8271,
+      "step": 65196
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.643637471206851e-07,
+      "loss": 0.7207,
+      "step": 65197
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.643363324296091e-07,
+      "loss": 0.4932,
+      "step": 65198
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.643089178424051e-07,
+      "loss": 0.5,
+      "step": 65199
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.642815033590931e-07,
+      "loss": 0.6396,
+      "step": 65200
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.642540889796943e-07,
+      "loss": 0.7061,
+      "step": 65201
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.642266747042302e-07,
+      "loss": 0.8193,
+      "step": 65202
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.641992605327211e-07,
+      "loss": 0.6973,
+      "step": 65203
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.641718464651888e-07,
+      "loss": 0.4221,
+      "step": 65204
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.641444325016536e-07,
+      "loss": 0.5898,
+      "step": 65205
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.641170186421371e-07,
+      "loss": 0.6011,
+      "step": 65206
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.640896048866595e-07,
+      "loss": 0.6167,
+      "step": 65207
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.640621912352426e-07,
+      "loss": 0.6934,
+      "step": 65208
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.640347776879071e-07,
+      "loss": 0.4792,
+      "step": 65209
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.640073642446737e-07,
+      "loss": 0.749,
+      "step": 65210
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.639799509055637e-07,
+      "loss": 0.7188,
+      "step": 65211
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.639525376705979e-07,
+      "loss": 0.6855,
+      "step": 65212
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.639251245397973e-07,
+      "loss": 0.5947,
+      "step": 65213
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63897711513183e-07,
+      "loss": 0.6045,
+      "step": 65214
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.638702985907762e-07,
+      "loss": 0.5991,
+      "step": 65215
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.638428857725975e-07,
+      "loss": 0.7754,
+      "step": 65216
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.638154730586679e-07,
+      "loss": 0.6201,
+      "step": 65217
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.637880604490089e-07,
+      "loss": 0.6968,
+      "step": 65218
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.637606479436411e-07,
+      "loss": 0.7612,
+      "step": 65219
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.637332355425852e-07,
+      "loss": 0.6118,
+      "step": 65220
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.637058232458627e-07,
+      "loss": 0.7383,
+      "step": 65221
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.636784110534942e-07,
+      "loss": 0.7021,
+      "step": 65222
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63650998965501e-07,
+      "loss": 0.71,
+      "step": 65223
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.636235869819038e-07,
+      "loss": 0.6348,
+      "step": 65224
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.635961751027239e-07,
+      "loss": 0.5669,
+      "step": 65225
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63568763327982e-07,
+      "loss": 0.7119,
+      "step": 65226
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.635413516576995e-07,
+      "loss": 0.6523,
+      "step": 65227
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.635139400918968e-07,
+      "loss": 0.748,
+      "step": 65228
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.634865286305956e-07,
+      "loss": 0.6934,
+      "step": 65229
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.634591172738162e-07,
+      "loss": 0.6367,
+      "step": 65230
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.634317060215798e-07,
+      "loss": 0.6682,
+      "step": 65231
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.634042948739078e-07,
+      "loss": 0.5786,
+      "step": 65232
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.633768838308206e-07,
+      "loss": 0.4668,
+      "step": 65233
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.633494728923394e-07,
+      "loss": 0.752,
+      "step": 65234
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.633220620584853e-07,
+      "loss": 0.4934,
+      "step": 65235
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63294651329279e-07,
+      "loss": 0.5688,
+      "step": 65236
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63267240704742e-07,
+      "loss": 0.5431,
+      "step": 65237
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.632398301848948e-07,
+      "loss": 0.7705,
+      "step": 65238
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63212419769759e-07,
+      "loss": 0.6812,
+      "step": 65239
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.631850094593547e-07,
+      "loss": 0.6982,
+      "step": 65240
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.631575992537035e-07,
+      "loss": 0.6572,
+      "step": 65241
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.631301891528262e-07,
+      "loss": 0.6113,
+      "step": 65242
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.631027791567438e-07,
+      "loss": 0.6807,
+      "step": 65243
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.630753692654773e-07,
+      "loss": 0.6548,
+      "step": 65244
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.630479594790476e-07,
+      "loss": 0.562,
+      "step": 65245
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.63020549797476e-07,
+      "loss": 0.5793,
+      "step": 65246
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.629931402207831e-07,
+      "loss": 0.5708,
+      "step": 65247
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.629657307489902e-07,
+      "loss": 0.5752,
+      "step": 65248
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.629383213821183e-07,
+      "loss": 0.5757,
+      "step": 65249
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.629109121201881e-07,
+      "loss": 0.4954,
+      "step": 65250
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.628835029632205e-07,
+      "loss": 0.6523,
+      "step": 65251
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.628560939112368e-07,
+      "loss": 0.6333,
+      "step": 65252
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.628286849642577e-07,
+      "loss": 0.5276,
+      "step": 65253
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.628012761223047e-07,
+      "loss": 0.7168,
+      "step": 65254
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.627738673853982e-07,
+      "loss": 0.748,
+      "step": 65255
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.627464587535596e-07,
+      "loss": 0.7236,
+      "step": 65256
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.627190502268095e-07,
+      "loss": 0.7734,
+      "step": 65257
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.626916418051693e-07,
+      "loss": 0.5679,
+      "step": 65258
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.626642334886602e-07,
+      "loss": 0.7119,
+      "step": 65259
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.62636825277302e-07,
+      "loss": 0.667,
+      "step": 65260
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.62609417171117e-07,
+      "loss": 0.6523,
+      "step": 65261
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.625820091701254e-07,
+      "loss": 0.5962,
+      "step": 65262
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.625546012743485e-07,
+      "loss": 0.7051,
+      "step": 65263
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.625271934838071e-07,
+      "loss": 0.6091,
+      "step": 65264
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.624997857985226e-07,
+      "loss": 0.6602,
+      "step": 65265
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.624723782185153e-07,
+      "loss": 0.6533,
+      "step": 65266
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.624449707438069e-07,
+      "loss": 0.6973,
+      "step": 65267
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.624175633744179e-07,
+      "loss": 0.6582,
+      "step": 65268
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.623901561103699e-07,
+      "loss": 0.561,
+      "step": 65269
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.623627489516829e-07,
+      "loss": 0.5933,
+      "step": 65270
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.623353418983788e-07,
+      "loss": 0.6504,
+      "step": 65271
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.623079349504777e-07,
+      "loss": 0.5825,
+      "step": 65272
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.622805281080014e-07,
+      "loss": 0.6533,
+      "step": 65273
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.622531213709704e-07,
+      "loss": 0.709,
+      "step": 65274
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.622257147394061e-07,
+      "loss": 0.6533,
+      "step": 65275
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.621983082133288e-07,
+      "loss": 0.7188,
+      "step": 65276
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.621709017927605e-07,
+      "loss": 0.5627,
+      "step": 65277
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.621434954777212e-07,
+      "loss": 0.6826,
+      "step": 65278
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.621160892682326e-07,
+      "loss": 0.5017,
+      "step": 65279
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.620886831643153e-07,
+      "loss": 0.4847,
+      "step": 65280
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.6206127716599e-07,
+      "loss": 0.7085,
+      "step": 65281
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.620338712732785e-07,
+      "loss": 0.7461,
+      "step": 65282
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.620064654862008e-07,
+      "loss": 0.521,
+      "step": 65283
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.619790598047788e-07,
+      "loss": 0.6172,
+      "step": 65284
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.619516542290329e-07,
+      "loss": 0.7109,
+      "step": 65285
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.619242487589844e-07,
+      "loss": 0.7832,
+      "step": 65286
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.618968433946539e-07,
+      "loss": 0.7979,
+      "step": 65287
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.61869438136063e-07,
+      "loss": 0.6309,
+      "step": 65288
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.618420329832323e-07,
+      "loss": 0.6533,
+      "step": 65289
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.618146279361829e-07,
+      "loss": 0.6279,
+      "step": 65290
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.617872229949352e-07,
+      "loss": 0.5122,
+      "step": 65291
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.61759818159511e-07,
+      "loss": 0.4038,
+      "step": 65292
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.617324134299307e-07,
+      "loss": 0.6504,
+      "step": 65293
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.617050088062158e-07,
+      "loss": 0.4727,
+      "step": 65294
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.616776042883868e-07,
+      "loss": 0.8057,
+      "step": 65295
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.61650199876465e-07,
+      "loss": 0.6465,
+      "step": 65296
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.616227955704712e-07,
+      "loss": 0.5759,
+      "step": 65297
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.615953913704267e-07,
+      "loss": 0.7354,
+      "step": 65298
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.615679872763524e-07,
+      "loss": 0.5568,
+      "step": 65299
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.615405832882686e-07,
+      "loss": 0.7588,
+      "step": 65300
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.615131794061972e-07,
+      "loss": 0.7354,
+      "step": 65301
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.614857756301584e-07,
+      "loss": 0.7783,
+      "step": 65302
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.614583719601739e-07,
+      "loss": 0.6035,
+      "step": 65303
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.614309683962642e-07,
+      "loss": 0.6885,
+      "step": 65304
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.614035649384506e-07,
+      "loss": 0.665,
+      "step": 65305
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.613761615867539e-07,
+      "loss": 0.6147,
+      "step": 65306
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.613487583411947e-07,
+      "loss": 0.4941,
+      "step": 65307
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.613213552017951e-07,
+      "loss": 0.5933,
+      "step": 65308
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.612939521685749e-07,
+      "loss": 0.5742,
+      "step": 65309
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.612665492415555e-07,
+      "loss": 0.7861,
+      "step": 65310
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.612391464207581e-07,
+      "loss": 0.564,
+      "step": 65311
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.612117437062033e-07,
+      "loss": 0.6299,
+      "step": 65312
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.611843410979125e-07,
+      "loss": 0.6465,
+      "step": 65313
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.611569385959062e-07,
+      "loss": 0.5488,
+      "step": 65314
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.611295362002059e-07,
+      "loss": 0.4443,
+      "step": 65315
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.611021339108321e-07,
+      "loss": 0.4495,
+      "step": 65316
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.610747317278062e-07,
+      "loss": 0.6211,
+      "step": 65317
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.610473296511493e-07,
+      "loss": 0.7051,
+      "step": 65318
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.610199276808816e-07,
+      "loss": 0.7061,
+      "step": 65319
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.609925258170248e-07,
+      "loss": 0.5918,
+      "step": 65320
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.609651240595994e-07,
+      "loss": 0.4585,
+      "step": 65321
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.609377224086267e-07,
+      "loss": 0.6392,
+      "step": 65322
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.609103208641277e-07,
+      "loss": 0.6768,
+      "step": 65323
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.60882919426123e-07,
+      "loss": 0.6111,
+      "step": 65324
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.608555180946341e-07,
+      "loss": 0.6221,
+      "step": 65325
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.608281168696816e-07,
+      "loss": 0.6895,
+      "step": 65326
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.608007157512867e-07,
+      "loss": 0.5298,
+      "step": 65327
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.607733147394706e-07,
+      "loss": 0.6528,
+      "step": 65328
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.607459138342534e-07,
+      "loss": 0.5713,
+      "step": 65329
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.60718513035657e-07,
+      "loss": 0.7168,
+      "step": 65330
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.606911123437017e-07,
+      "loss": 0.5977,
+      "step": 65331
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.606637117584092e-07,
+      "loss": 0.3945,
+      "step": 65332
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.606363112797997e-07,
+      "loss": 0.4766,
+      "step": 65333
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.606089109078949e-07,
+      "loss": 0.7148,
+      "step": 65334
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.605815106427152e-07,
+      "loss": 0.5747,
+      "step": 65335
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.60554110484282e-07,
+      "loss": 0.7578,
+      "step": 65336
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.605267104326159e-07,
+      "loss": 0.4846,
+      "step": 65337
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.604993104877386e-07,
+      "loss": 0.5859,
+      "step": 65338
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.604719106496703e-07,
+      "loss": 0.5474,
+      "step": 65339
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.60444510918432e-07,
+      "loss": 0.7422,
+      "step": 65340
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.604171112940451e-07,
+      "loss": 0.7119,
+      "step": 65341
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.603897117765305e-07,
+      "loss": 0.4556,
+      "step": 65342
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.603623123659088e-07,
+      "loss": 0.6006,
+      "step": 65343
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.603349130622015e-07,
+      "loss": 0.5962,
+      "step": 65344
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.603075138654291e-07,
+      "loss": 0.6353,
+      "step": 65345
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.602801147756132e-07,
+      "loss": 0.8916,
+      "step": 65346
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.60252715792774e-07,
+      "loss": 0.6556,
+      "step": 65347
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.602253169169333e-07,
+      "loss": 0.5037,
+      "step": 65348
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.601979181481113e-07,
+      "loss": 0.6631,
+      "step": 65349
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.601705194863294e-07,
+      "loss": 0.5537,
+      "step": 65350
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 8.601431209316084e-07,
+      "loss": 0.6846,
+      "step": 65351
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.601157224839694e-07,
+      "loss": 0.6045,
+      "step": 65352
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.600883241434336e-07,
+      "loss": 0.5967,
+      "step": 65353
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.600609259100213e-07,
+      "loss": 0.6196,
+      "step": 65354
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.600335277837543e-07,
+      "loss": 0.5154,
+      "step": 65355
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.60006129764653e-07,
+      "loss": 0.5449,
+      "step": 65356
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.599787318527387e-07,
+      "loss": 0.6411,
+      "step": 65357
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.599513340480325e-07,
+      "loss": 0.51,
+      "step": 65358
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.599239363505548e-07,
+      "loss": 0.5151,
+      "step": 65359
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.598965387603269e-07,
+      "loss": 0.7236,
+      "step": 65360
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.598691412773698e-07,
+      "loss": 0.7295,
+      "step": 65361
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.598417439017044e-07,
+      "loss": 0.8066,
+      "step": 65362
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.598143466333518e-07,
+      "loss": 0.6221,
+      "step": 65363
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.597869494723328e-07,
+      "loss": 0.7002,
+      "step": 65364
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.597595524186687e-07,
+      "loss": 0.6245,
+      "step": 65365
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.5973215547238e-07,
+      "loss": 0.748,
+      "step": 65366
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.597047586334881e-07,
+      "loss": 0.6409,
+      "step": 65367
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.596773619020141e-07,
+      "loss": 0.5007,
+      "step": 65368
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.596499652779783e-07,
+      "loss": 0.5303,
+      "step": 65369
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.596225687614022e-07,
+      "loss": 0.7129,
+      "step": 65370
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.595951723523065e-07,
+      "loss": 0.4945,
+      "step": 65371
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.595677760507125e-07,
+      "loss": 0.7383,
+      "step": 65372
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.595403798566408e-07,
+      "loss": 0.5996,
+      "step": 65373
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.595129837701127e-07,
+      "loss": 0.7178,
+      "step": 65374
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.594855877911489e-07,
+      "loss": 0.5396,
+      "step": 65375
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.594581919197708e-07,
+      "loss": 0.5659,
+      "step": 65376
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.594307961559988e-07,
+      "loss": 0.7314,
+      "step": 65377
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.594034004998548e-07,
+      "loss": 0.9614,
+      "step": 65378
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.593760049513586e-07,
+      "loss": 0.5249,
+      "step": 65379
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.593486095105319e-07,
+      "loss": 0.6865,
+      "step": 65380
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.593212141773952e-07,
+      "loss": 0.5312,
+      "step": 65381
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.592938189519702e-07,
+      "loss": 0.6665,
+      "step": 65382
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.592664238342771e-07,
+      "loss": 0.5269,
+      "step": 65383
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.592390288243374e-07,
+      "loss": 0.665,
+      "step": 65384
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.592116339221718e-07,
+      "loss": 0.6904,
+      "step": 65385
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.591842391278015e-07,
+      "loss": 0.4327,
+      "step": 65386
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.591568444412471e-07,
+      "loss": 0.6377,
+      "step": 65387
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.591294498625304e-07,
+      "loss": 0.6973,
+      "step": 65388
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.591020553916715e-07,
+      "loss": 0.5996,
+      "step": 65389
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.590746610286915e-07,
+      "loss": 0.6611,
+      "step": 65390
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.590472667736117e-07,
+      "loss": 0.6924,
+      "step": 65391
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.590198726264528e-07,
+      "loss": 0.5288,
+      "step": 65392
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.589924785872361e-07,
+      "loss": 0.6597,
+      "step": 65393
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.589650846559822e-07,
+      "loss": 0.571,
+      "step": 65394
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.589376908327123e-07,
+      "loss": 0.6338,
+      "step": 65395
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.589102971174476e-07,
+      "loss": 0.6013,
+      "step": 65396
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.588829035102084e-07,
+      "loss": 0.4756,
+      "step": 65397
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.588555100110164e-07,
+      "loss": 0.5566,
+      "step": 65398
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.588281166198921e-07,
+      "loss": 0.5303,
+      "step": 65399
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.588007233368564e-07,
+      "loss": 0.4475,
+      "step": 65400
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.587733301619309e-07,
+      "loss": 0.7129,
+      "step": 65401
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.587459370951358e-07,
+      "loss": 0.4858,
+      "step": 65402
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.587185441364927e-07,
+      "loss": 0.6675,
+      "step": 65403
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.58691151286022e-07,
+      "loss": 0.5747,
+      "step": 65404
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.586637585437452e-07,
+      "loss": 0.7744,
+      "step": 65405
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.58636365909683e-07,
+      "loss": 0.6279,
+      "step": 65406
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.586089733838567e-07,
+      "loss": 0.5269,
+      "step": 65407
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.585815809662869e-07,
+      "loss": 0.5266,
+      "step": 65408
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.585541886569945e-07,
+      "loss": 0.5752,
+      "step": 65409
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.585267964560007e-07,
+      "loss": 0.6187,
+      "step": 65410
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.584994043633265e-07,
+      "loss": 0.6729,
+      "step": 65411
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.584720123789929e-07,
+      "loss": 0.5024,
+      "step": 65412
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.584446205030206e-07,
+      "loss": 0.6924,
+      "step": 65413
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.584172287354307e-07,
+      "loss": 0.5498,
+      "step": 65414
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.583898370762444e-07,
+      "loss": 0.5116,
+      "step": 65415
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.583624455254821e-07,
+      "loss": 0.627,
+      "step": 65416
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.583350540831655e-07,
+      "loss": 0.6758,
+      "step": 65417
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.583076627493156e-07,
+      "loss": 0.6211,
+      "step": 65418
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.582802715239524e-07,
+      "loss": 0.6504,
+      "step": 65419
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.582528804070978e-07,
+      "loss": 0.7959,
+      "step": 65420
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.582254893987721e-07,
+      "loss": 0.3722,
+      "step": 65421
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.581980984989969e-07,
+      "loss": 0.6511,
+      "step": 65422
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.581707077077927e-07,
+      "loss": 0.5479,
+      "step": 65423
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.581433170251809e-07,
+      "loss": 0.7783,
+      "step": 65424
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.58115926451182e-07,
+      "loss": 0.7773,
+      "step": 65425
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.580885359858174e-07,
+      "loss": 0.6553,
+      "step": 65426
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.580611456291079e-07,
+      "loss": 0.8027,
+      "step": 65427
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.580337553810745e-07,
+      "loss": 0.543,
+      "step": 65428
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.580063652417382e-07,
+      "loss": 0.6484,
+      "step": 65429
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.579789752111195e-07,
+      "loss": 0.6211,
+      "step": 65430
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.579515852892401e-07,
+      "loss": 0.6846,
+      "step": 65431
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.579241954761206e-07,
+      "loss": 0.4543,
+      "step": 65432
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.578968057717818e-07,
+      "loss": 0.7607,
+      "step": 65433
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.578694161762451e-07,
+      "loss": 0.6494,
+      "step": 65434
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.57842026689531e-07,
+      "loss": 0.5728,
+      "step": 65435
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.57814637311661e-07,
+      "loss": 0.6543,
+      "step": 65436
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.577872480426556e-07,
+      "loss": 0.6318,
+      "step": 65437
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.577598588825364e-07,
+      "loss": 0.5293,
+      "step": 65438
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.577324698313236e-07,
+      "loss": 0.5244,
+      "step": 65439
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.577050808890383e-07,
+      "loss": 0.6953,
+      "step": 65440
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.57677692055702e-07,
+      "loss": 0.6816,
+      "step": 65441
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.57650303331335e-07,
+      "loss": 0.7148,
+      "step": 65442
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.57622914715959e-07,
+      "loss": 0.5825,
+      "step": 65443
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.575955262095943e-07,
+      "loss": 0.6392,
+      "step": 65444
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.575681378122623e-07,
+      "loss": 0.6826,
+      "step": 65445
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.575407495239837e-07,
+      "loss": 0.4434,
+      "step": 65446
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.575133613447798e-07,
+      "loss": 0.6025,
+      "step": 65447
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.574859732746715e-07,
+      "loss": 0.5632,
+      "step": 65448
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.574585853136795e-07,
+      "loss": 0.5796,
+      "step": 65449
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.574311974618247e-07,
+      "loss": 0.5713,
+      "step": 65450
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.574038097191285e-07,
+      "loss": 0.5786,
+      "step": 65451
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.573764220856115e-07,
+      "loss": 0.8379,
+      "step": 65452
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.573490345612948e-07,
+      "loss": 0.624,
+      "step": 65453
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.573216471461994e-07,
+      "loss": 0.6504,
+      "step": 65454
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.572942598403463e-07,
+      "loss": 0.71,
+      "step": 65455
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.572668726437563e-07,
+      "loss": 0.6909,
+      "step": 65456
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.572394855564507e-07,
+      "loss": 0.6982,
+      "step": 65457
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.572120985784504e-07,
+      "loss": 0.6182,
+      "step": 65458
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.571847117097759e-07,
+      "loss": 0.6431,
+      "step": 65459
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.571573249504487e-07,
+      "loss": 0.4893,
+      "step": 65460
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.571299383004893e-07,
+      "loss": 0.5732,
+      "step": 65461
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.571025517599193e-07,
+      "loss": 0.4791,
+      "step": 65462
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.570751653287589e-07,
+      "loss": 0.8975,
+      "step": 65463
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.570477790070297e-07,
+      "loss": 0.6787,
+      "step": 65464
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.570203927947524e-07,
+      "loss": 0.5728,
+      "step": 65465
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.56993006691948e-07,
+      "loss": 0.6082,
+      "step": 65466
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.569656206986376e-07,
+      "loss": 0.5669,
+      "step": 65467
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.569382348148424e-07,
+      "loss": 0.6953,
+      "step": 65468
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.569108490405824e-07,
+      "loss": 0.6489,
+      "step": 65469
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.568834633758795e-07,
+      "loss": 0.416,
+      "step": 65470
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.568560778207541e-07,
+      "loss": 0.7061,
+      "step": 65471
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.568286923752274e-07,
+      "loss": 0.5879,
+      "step": 65472
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.568013070393205e-07,
+      "loss": 0.5801,
+      "step": 65473
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.567739218130544e-07,
+      "loss": 0.7969,
+      "step": 65474
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.567465366964497e-07,
+      "loss": 0.5161,
+      "step": 65475
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.567191516895277e-07,
+      "loss": 0.4927,
+      "step": 65476
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.566917667923092e-07,
+      "loss": 0.6865,
+      "step": 65477
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.566643820048156e-07,
+      "loss": 0.6445,
+      "step": 65478
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.566369973270672e-07,
+      "loss": 0.6875,
+      "step": 65479
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.566096127590851e-07,
+      "loss": 0.6353,
+      "step": 65480
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.565822283008906e-07,
+      "loss": 0.5498,
+      "step": 65481
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.565548439525043e-07,
+      "loss": 0.5438,
+      "step": 65482
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.565274597139476e-07,
+      "loss": 0.5098,
+      "step": 65483
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.565000755852409e-07,
+      "loss": 0.6582,
+      "step": 65484
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.564726915664059e-07,
+      "loss": 0.6221,
+      "step": 65485
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.56445307657463e-07,
+      "loss": 0.6147,
+      "step": 65486
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.564179238584332e-07,
+      "loss": 0.5881,
+      "step": 65487
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.563905401693379e-07,
+      "loss": 0.7588,
+      "step": 65488
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.563631565901978e-07,
+      "loss": 0.7305,
+      "step": 65489
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.563357731210334e-07,
+      "loss": 0.7246,
+      "step": 65490
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.563083897618663e-07,
+      "loss": 0.5483,
+      "step": 65491
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.562810065127172e-07,
+      "loss": 0.6714,
+      "step": 65492
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.562536233736071e-07,
+      "loss": 0.8213,
+      "step": 65493
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.562262403445571e-07,
+      "loss": 0.6626,
+      "step": 65494
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.56198857425588e-07,
+      "loss": 0.6953,
+      "step": 65495
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.561714746167208e-07,
+      "loss": 0.5601,
+      "step": 65496
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.561440919179765e-07,
+      "loss": 0.6108,
+      "step": 65497
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.561167093293764e-07,
+      "loss": 0.7461,
+      "step": 65498
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.560893268509407e-07,
+      "loss": 0.7148,
+      "step": 65499
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.56061944482691e-07,
+      "loss": 0.4834,
+      "step": 65500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.560345622246478e-07,
+      "loss": 0.6143,
+      "step": 65501
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.560071800768325e-07,
+      "loss": 0.7119,
+      "step": 65502
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.559797980392657e-07,
+      "loss": 0.623,
+      "step": 65503
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.559524161119688e-07,
+      "loss": 0.8037,
+      "step": 65504
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.559250342949624e-07,
+      "loss": 0.543,
+      "step": 65505
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.558976525882675e-07,
+      "loss": 0.6289,
+      "step": 65506
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.558702709919052e-07,
+      "loss": 0.7314,
+      "step": 65507
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.558428895058967e-07,
+      "loss": 0.6084,
+      "step": 65508
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.558155081302623e-07,
+      "loss": 0.6011,
+      "step": 65509
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.557881268650236e-07,
+      "loss": 0.6862,
+      "step": 65510
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.55760745710201e-07,
+      "loss": 0.6943,
+      "step": 65511
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.557333646658158e-07,
+      "loss": 0.3956,
+      "step": 65512
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.55705983731889e-07,
+      "loss": 0.6123,
+      "step": 65513
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.556786029084415e-07,
+      "loss": 0.6821,
+      "step": 65514
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.556512221954941e-07,
+      "loss": 0.6865,
+      "step": 65515
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.556238415930682e-07,
+      "loss": 0.7334,
+      "step": 65516
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.555964611011843e-07,
+      "loss": 0.6904,
+      "step": 65517
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.555690807198638e-07,
+      "loss": 0.572,
+      "step": 65518
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.555417004491274e-07,
+      "loss": 0.6602,
+      "step": 65519
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.555143202889959e-07,
+      "loss": 0.6182,
+      "step": 65520
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.554869402394904e-07,
+      "loss": 0.6714,
+      "step": 65521
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.554595603006321e-07,
+      "loss": 0.6445,
+      "step": 65522
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.554321804724415e-07,
+      "loss": 0.6143,
+      "step": 65523
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.554048007549401e-07,
+      "loss": 0.6008,
+      "step": 65524
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.553774211481482e-07,
+      "loss": 0.4453,
+      "step": 65525
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.553500416520876e-07,
+      "loss": 0.7305,
+      "step": 65526
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.553226622667784e-07,
+      "loss": 0.6875,
+      "step": 65527
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.552952829922426e-07,
+      "loss": 0.7051,
+      "step": 65528
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.552679038285002e-07,
+      "loss": 0.7412,
+      "step": 65529
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.552405247755724e-07,
+      "loss": 0.8369,
+      "step": 65530
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.552131458334803e-07,
+      "loss": 0.7559,
+      "step": 65531
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.551857670022449e-07,
+      "loss": 0.6709,
+      "step": 65532
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.551583882818872e-07,
+      "loss": 0.7305,
+      "step": 65533
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.551310096724278e-07,
+      "loss": 0.5571,
+      "step": 65534
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.551036311738882e-07,
+      "loss": 0.5283,
+      "step": 65535
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.550762527862887e-07,
+      "loss": 0.5432,
+      "step": 65536
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.550488745096512e-07,
+      "loss": 0.4685,
+      "step": 65537
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.55021496343996e-07,
+      "loss": 0.6431,
+      "step": 65538
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.549941182893441e-07,
+      "loss": 0.7139,
+      "step": 65539
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.549667403457163e-07,
+      "loss": 0.5513,
+      "step": 65540
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.549393625131341e-07,
+      "loss": 0.5947,
+      "step": 65541
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.549119847916177e-07,
+      "loss": 0.6807,
+      "step": 65542
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.54884607181189e-07,
+      "loss": 0.6104,
+      "step": 65543
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.548572296818681e-07,
+      "loss": 0.5947,
+      "step": 65544
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.548298522936766e-07,
+      "loss": 0.3641,
+      "step": 65545
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.54802475016635e-07,
+      "loss": 0.5718,
+      "step": 65546
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.547750978507647e-07,
+      "loss": 0.6123,
+      "step": 65547
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.547477207960867e-07,
+      "loss": 0.7046,
+      "step": 65548
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.547203438526213e-07,
+      "loss": 0.605,
+      "step": 65549
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.546929670203899e-07,
+      "loss": 0.7061,
+      "step": 65550
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.546655902994133e-07,
+      "loss": 0.4492,
+      "step": 65551
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.546382136897127e-07,
+      "loss": 0.5557,
+      "step": 65552
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.546108371913089e-07,
+      "loss": 0.4873,
+      "step": 65553
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.545834608042229e-07,
+      "loss": 0.5664,
+      "step": 65554
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.545560845284756e-07,
+      "loss": 0.5605,
+      "step": 65555
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.545287083640882e-07,
+      "loss": 0.8115,
+      "step": 65556
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.545013323110812e-07,
+      "loss": 0.5542,
+      "step": 65557
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.544739563694765e-07,
+      "loss": 0.3495,
+      "step": 65558
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.544465805392937e-07,
+      "loss": 0.6758,
+      "step": 65559
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.544192048205548e-07,
+      "loss": 0.4863,
+      "step": 65560
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.543918292132803e-07,
+      "loss": 0.6279,
+      "step": 65561
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.543644537174913e-07,
+      "loss": 0.5916,
+      "step": 65562
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.543370783332085e-07,
+      "loss": 0.387,
+      "step": 65563
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.543097030604534e-07,
+      "loss": 0.6704,
+      "step": 65564
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.542823278992465e-07,
+      "loss": 0.7988,
+      "step": 65565
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.542549528496091e-07,
+      "loss": 0.5358,
+      "step": 65566
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.542275779115617e-07,
+      "loss": 0.5364,
+      "step": 65567
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.542002030851261e-07,
+      "loss": 0.6299,
+      "step": 65568
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.541728283703223e-07,
+      "loss": 0.7822,
+      "step": 65569
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.541454537671716e-07,
+      "loss": 0.6885,
+      "step": 65570
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.541180792756952e-07,
+      "loss": 0.6216,
+      "step": 65571
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.540907048959136e-07,
+      "loss": 0.7305,
+      "step": 65572
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.540633306278484e-07,
+      "loss": 0.687,
+      "step": 65573
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.540359564715198e-07,
+      "loss": 0.6689,
+      "step": 65574
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.540085824269495e-07,
+      "loss": 0.5029,
+      "step": 65575
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.53981208494158e-07,
+      "loss": 0.4651,
+      "step": 65576
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.539538346731663e-07,
+      "loss": 0.6792,
+      "step": 65577
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.539264609639958e-07,
+      "loss": 0.6035,
+      "step": 65578
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.538990873666669e-07,
+      "loss": 0.6362,
+      "step": 65579
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.538717138812004e-07,
+      "loss": 0.4749,
+      "step": 65580
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.53844340507618e-07,
+      "loss": 0.5825,
+      "step": 65581
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.5381696724594e-07,
+      "loss": 0.5264,
+      "step": 65582
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.537895940961878e-07,
+      "loss": 0.6372,
+      "step": 65583
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.537622210583819e-07,
+      "loss": 0.5732,
+      "step": 65584
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.537348481325439e-07,
+      "loss": 0.5913,
+      "step": 65585
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.537074753186941e-07,
+      "loss": 0.6738,
+      "step": 65586
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.536801026168541e-07,
+      "loss": 0.6982,
+      "step": 65587
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.536527300270446e-07,
+      "loss": 0.5977,
+      "step": 65588
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.536253575492861e-07,
+      "loss": 0.7109,
+      "step": 65589
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.535979851836002e-07,
+      "loss": 0.8467,
+      "step": 65590
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.535706129300073e-07,
+      "loss": 0.6719,
+      "step": 65591
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.535432407885289e-07,
+      "loss": 0.6826,
+      "step": 65592
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.535158687591855e-07,
+      "loss": 0.6279,
+      "step": 65593
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.534884968419984e-07,
+      "loss": 0.5757,
+      "step": 65594
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.534611250369885e-07,
+      "loss": 0.7764,
+      "step": 65595
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.534337533441765e-07,
+      "loss": 0.561,
+      "step": 65596
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.534063817635836e-07,
+      "loss": 0.604,
+      "step": 65597
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.533790102952309e-07,
+      "loss": 0.4473,
+      "step": 65598
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.533516389391388e-07,
+      "loss": 0.7051,
+      "step": 65599
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.533242676953289e-07,
+      "loss": 0.6455,
+      "step": 65600
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.532968965638217e-07,
+      "loss": 0.5713,
+      "step": 65601
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.532695255446383e-07,
+      "loss": 0.5649,
+      "step": 65602
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.532421546377996e-07,
+      "loss": 0.6924,
+      "step": 65603
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.532147838433267e-07,
+      "loss": 0.8027,
+      "step": 65604
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.531874131612406e-07,
+      "loss": 0.7021,
+      "step": 65605
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.531600425915621e-07,
+      "loss": 0.5129,
+      "step": 65606
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.531326721343125e-07,
+      "loss": 0.5364,
+      "step": 65607
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.53105301789512e-07,
+      "loss": 0.6992,
+      "step": 65608
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.530779315571823e-07,
+      "loss": 0.5879,
+      "step": 65609
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.530505614373438e-07,
+      "loss": 0.625,
+      "step": 65610
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.530231914300179e-07,
+      "loss": 0.6357,
+      "step": 65611
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.529958215352255e-07,
+      "loss": 0.6733,
+      "step": 65612
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.529684517529871e-07,
+      "loss": 0.6333,
+      "step": 65613
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.529410820833243e-07,
+      "loss": 0.5618,
+      "step": 65614
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.529137125262575e-07,
+      "loss": 0.6436,
+      "step": 65615
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.528863430818081e-07,
+      "loss": 0.6416,
+      "step": 65616
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.528589737499971e-07,
+      "loss": 0.6396,
+      "step": 65617
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.528316045308447e-07,
+      "loss": 0.7676,
+      "step": 65618
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.528042354243727e-07,
+      "loss": 0.6899,
+      "step": 65619
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.527768664306016e-07,
+      "loss": 0.5659,
+      "step": 65620
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.527494975495526e-07,
+      "loss": 0.7754,
+      "step": 65621
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.527221287812462e-07,
+      "loss": 0.5343,
+      "step": 65622
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.526947601257042e-07,
+      "loss": 0.6855,
+      "step": 65623
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.526673915829466e-07,
+      "loss": 0.4601,
+      "step": 65624
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.526400231529951e-07,
+      "loss": 0.7324,
+      "step": 65625
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.526126548358701e-07,
+      "loss": 0.7041,
+      "step": 65626
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.525852866315933e-07,
+      "loss": 0.7217,
+      "step": 65627
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.525579185401849e-07,
+      "loss": 0.8193,
+      "step": 65628
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.525305505616661e-07,
+      "loss": 0.6572,
+      "step": 65629
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.525031826960577e-07,
+      "loss": 0.6064,
+      "step": 65630
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.524758149433811e-07,
+      "loss": 0.4502,
+      "step": 65631
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.524484473036567e-07,
+      "loss": 0.7617,
+      "step": 65632
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.52421079776906e-07,
+      "loss": 0.6831,
+      "step": 65633
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.523937123631494e-07,
+      "loss": 0.5571,
+      "step": 65634
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.523663450624084e-07,
+      "loss": 0.6543,
+      "step": 65635
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.523389778747035e-07,
+      "loss": 0.5029,
+      "step": 65636
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.523116108000563e-07,
+      "loss": 0.627,
+      "step": 65637
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.522842438384872e-07,
+      "loss": 0.6348,
+      "step": 65638
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.522568769900168e-07,
+      "loss": 0.6143,
+      "step": 65639
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.522295102546668e-07,
+      "loss": 0.7305,
+      "step": 65640
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.522021436324578e-07,
+      "loss": 0.7803,
+      "step": 65641
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.521747771234109e-07,
+      "loss": 0.6895,
+      "step": 65642
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.521474107275468e-07,
+      "loss": 0.5796,
+      "step": 65643
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.521200444448869e-07,
+      "loss": 0.4736,
+      "step": 65644
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.520926782754516e-07,
+      "loss": 0.5659,
+      "step": 65645
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.520653122192624e-07,
+      "loss": 0.6426,
+      "step": 65646
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.520379462763402e-07,
+      "loss": 0.5715,
+      "step": 65647
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.520105804467055e-07,
+      "loss": 0.5625,
+      "step": 65648
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.519832147303793e-07,
+      "loss": 0.7227,
+      "step": 65649
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.519558491273829e-07,
+      "loss": 0.6523,
+      "step": 65650
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.51928483637737e-07,
+      "loss": 0.6392,
+      "step": 65651
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.519011182614627e-07,
+      "loss": 0.667,
+      "step": 65652
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.518737529985808e-07,
+      "loss": 0.6299,
+      "step": 65653
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.518463878491126e-07,
+      "loss": 0.6396,
+      "step": 65654
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.518190228130785e-07,
+      "loss": 0.748,
+      "step": 65655
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.517916578905e-07,
+      "loss": 0.708,
+      "step": 65656
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.517642930813981e-07,
+      "loss": 0.6792,
+      "step": 65657
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.51736928385793e-07,
+      "loss": 0.502,
+      "step": 65658
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.517095638037063e-07,
+      "loss": 0.7568,
+      "step": 65659
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.516821993351586e-07,
+      "loss": 0.7358,
+      "step": 65660
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.516548349801711e-07,
+      "loss": 0.6797,
+      "step": 65661
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.516274707387646e-07,
+      "loss": 0.6123,
+      "step": 65662
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.516001066109604e-07,
+      "loss": 0.8477,
+      "step": 65663
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.515727425967788e-07,
+      "loss": 0.5391,
+      "step": 65664
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.515453786962415e-07,
+      "loss": 0.8867,
+      "step": 65665
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.515180149093686e-07,
+      "loss": 0.5217,
+      "step": 65666
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.514906512361823e-07,
+      "loss": 0.8174,
+      "step": 65667
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.514632876767023e-07,
+      "loss": 0.5137,
+      "step": 65668
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.514359242309501e-07,
+      "loss": 0.665,
+      "step": 65669
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.514085608989463e-07,
+      "loss": 0.7578,
+      "step": 65670
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.513811976807125e-07,
+      "loss": 0.6357,
+      "step": 65671
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.51353834576269e-07,
+      "loss": 0.686,
+      "step": 65672
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.513264715856372e-07,
+      "loss": 0.7549,
+      "step": 65673
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.512991087088377e-07,
+      "loss": 0.6445,
+      "step": 65674
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.512717459458919e-07,
+      "loss": 0.5786,
+      "step": 65675
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.512443832968202e-07,
+      "loss": 0.6152,
+      "step": 65676
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.512170207616444e-07,
+      "loss": 0.7627,
+      "step": 65677
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.511896583403846e-07,
+      "loss": 0.5471,
+      "step": 65678
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.511622960330617e-07,
+      "loss": 0.6943,
+      "step": 65679
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.511349338396974e-07,
+      "loss": 0.6348,
+      "step": 65680
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.511075717603117e-07,
+      "loss": 0.7188,
+      "step": 65681
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.510802097949266e-07,
+      "loss": 0.6738,
+      "step": 65682
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.510528479435623e-07,
+      "loss": 0.7314,
+      "step": 65683
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.510254862062401e-07,
+      "loss": 0.5681,
+      "step": 65684
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.50998124582981e-07,
+      "loss": 0.5032,
+      "step": 65685
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.509707630738053e-07,
+      "loss": 0.793,
+      "step": 65686
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.509434016787351e-07,
+      "loss": 0.7231,
+      "step": 65687
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.509160403977904e-07,
+      "loss": 0.6035,
+      "step": 65688
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.508886792309921e-07,
+      "loss": 0.6367,
+      "step": 65689
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.508613181783618e-07,
+      "loss": 0.7822,
+      "step": 65690
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.508339572399199e-07,
+      "loss": 0.8281,
+      "step": 65691
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.508065964156878e-07,
+      "loss": 0.6826,
+      "step": 65692
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.507792357056862e-07,
+      "loss": 0.7715,
+      "step": 65693
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.50751875109936e-07,
+      "loss": 0.5767,
+      "step": 65694
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.507245146284581e-07,
+      "loss": 0.4414,
+      "step": 65695
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.506971542612738e-07,
+      "loss": 0.6846,
+      "step": 65696
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.506697940084041e-07,
+      "loss": 0.5933,
+      "step": 65697
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.506424338698692e-07,
+      "loss": 0.6855,
+      "step": 65698
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.506150738456907e-07,
+      "loss": 0.5044,
+      "step": 65699
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.505877139358892e-07,
+      "loss": 0.5564,
+      "step": 65700
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.50560354140486e-07,
+      "loss": 0.6382,
+      "step": 65701
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.505329944595019e-07,
+      "loss": 0.7695,
+      "step": 65702
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.505056348929575e-07,
+      "loss": 0.6475,
+      "step": 65703
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.504782754408743e-07,
+      "loss": 0.7002,
+      "step": 65704
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.504509161032729e-07,
+      "loss": 0.6143,
+      "step": 65705
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.504235568801745e-07,
+      "loss": 0.6655,
+      "step": 65706
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.503961977716002e-07,
+      "loss": 0.9082,
+      "step": 65707
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.503688387775703e-07,
+      "loss": 0.7812,
+      "step": 65708
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.503414798981061e-07,
+      "loss": 0.6719,
+      "step": 65709
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.503141211332285e-07,
+      "loss": 0.3951,
+      "step": 65710
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.502867624829585e-07,
+      "loss": 0.6562,
+      "step": 65711
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.50259403947317e-07,
+      "loss": 0.917,
+      "step": 65712
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.502320455263251e-07,
+      "loss": 0.6733,
+      "step": 65713
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.502046872200035e-07,
+      "loss": 0.6543,
+      "step": 65714
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.501773290283734e-07,
+      "loss": 0.6597,
+      "step": 65715
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.501499709514556e-07,
+      "loss": 0.5615,
+      "step": 65716
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.501226129892714e-07,
+      "loss": 0.6213,
+      "step": 65717
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.500952551418412e-07,
+      "loss": 0.5127,
+      "step": 65718
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.500678974091862e-07,
+      "loss": 0.501,
+      "step": 65719
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.500405397913271e-07,
+      "loss": 0.6943,
+      "step": 65720
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.500131822882852e-07,
+      "loss": 0.5752,
+      "step": 65721
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.499858249000811e-07,
+      "loss": 0.6191,
+      "step": 65722
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.499584676267362e-07,
+      "loss": 0.6233,
+      "step": 65723
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.49931110468271e-07,
+      "loss": 0.9092,
+      "step": 65724
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.499037534247068e-07,
+      "loss": 0.6777,
+      "step": 65725
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.498763964960644e-07,
+      "loss": 0.7246,
+      "step": 65726
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.49849039682365e-07,
+      "loss": 0.645,
+      "step": 65727
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.498216829836291e-07,
+      "loss": 0.5859,
+      "step": 65728
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.497943263998775e-07,
+      "loss": 0.5981,
+      "step": 65729
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.497669699311318e-07,
+      "loss": 0.6299,
+      "step": 65730
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.497396135774125e-07,
+      "loss": 0.687,
+      "step": 65731
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.497122573387407e-07,
+      "loss": 0.6606,
+      "step": 65732
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.496849012151372e-07,
+      "loss": 0.5479,
+      "step": 65733
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.496575452066234e-07,
+      "loss": 0.7656,
+      "step": 65734
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.496301893132195e-07,
+      "loss": 0.3424,
+      "step": 65735
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.496028335349471e-07,
+      "loss": 0.6836,
+      "step": 65736
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.495754778718272e-07,
+      "loss": 0.5796,
+      "step": 65737
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.4954812232388e-07,
+      "loss": 0.8076,
+      "step": 65738
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.49520766891127e-07,
+      "loss": 0.5713,
+      "step": 65739
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.49493411573589e-07,
+      "loss": 0.6582,
+      "step": 65740
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 8.494660563712869e-07,
+      "loss": 0.6177,
+      "step": 65741
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.494387012842418e-07,
+      "loss": 0.7397,
+      "step": 65742
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.494113463124743e-07,
+      "loss": 0.4883,
+      "step": 65743
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.49383991456006e-07,
+      "loss": 0.6221,
+      "step": 65744
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.493566367148573e-07,
+      "loss": 0.6016,
+      "step": 65745
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.493292820890493e-07,
+      "loss": 0.6245,
+      "step": 65746
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.493019275786033e-07,
+      "loss": 0.7891,
+      "step": 65747
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.492745731835394e-07,
+      "loss": 0.5923,
+      "step": 65748
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.492472189038792e-07,
+      "loss": 0.6826,
+      "step": 65749
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.492198647396433e-07,
+      "loss": 0.6387,
+      "step": 65750
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.491925106908531e-07,
+      "loss": 0.7275,
+      "step": 65751
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.491651567575289e-07,
+      "loss": 0.832,
+      "step": 65752
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.491378029396923e-07,
+      "loss": 0.3721,
+      "step": 65753
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.491104492373638e-07,
+      "loss": 0.7461,
+      "step": 65754
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.490830956505645e-07,
+      "loss": 0.6357,
+      "step": 65755
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.490557421793153e-07,
+      "loss": 0.6133,
+      "step": 65756
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.490283888236378e-07,
+      "loss": 0.5947,
+      "step": 65757
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.490010355835516e-07,
+      "loss": 0.5068,
+      "step": 65758
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.489736824590786e-07,
+      "loss": 0.6885,
+      "step": 65759
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.489463294502394e-07,
+      "loss": 0.8184,
+      "step": 65760
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.489189765570551e-07,
+      "loss": 0.707,
+      "step": 65761
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.488916237795465e-07,
+      "loss": 0.8125,
+      "step": 65762
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.488642711177347e-07,
+      "loss": 0.5701,
+      "step": 65763
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.488369185716405e-07,
+      "loss": 0.5371,
+      "step": 65764
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.488095661412852e-07,
+      "loss": 0.7168,
+      "step": 65765
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.487822138266891e-07,
+      "loss": 0.6265,
+      "step": 65766
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.487548616278741e-07,
+      "loss": 0.46,
+      "step": 65767
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.487275095448602e-07,
+      "loss": 0.5259,
+      "step": 65768
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.487001575776685e-07,
+      "loss": 0.8232,
+      "step": 65769
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.486728057263203e-07,
+      "loss": 0.6543,
+      "step": 65770
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.486454539908362e-07,
+      "loss": 0.6768,
+      "step": 65771
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.486181023712375e-07,
+      "loss": 0.5343,
+      "step": 65772
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.485907508675448e-07,
+      "loss": 0.6372,
+      "step": 65773
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.485633994797794e-07,
+      "loss": 0.8125,
+      "step": 65774
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.485360482079621e-07,
+      "loss": 0.6606,
+      "step": 65775
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.485086970521135e-07,
+      "loss": 0.5942,
+      "step": 65776
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.484813460122553e-07,
+      "loss": 0.4692,
+      "step": 65777
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.484539950884077e-07,
+      "loss": 0.6426,
+      "step": 65778
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.484266442805916e-07,
+      "loss": 0.6533,
+      "step": 65779
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.483992935888287e-07,
+      "loss": 0.5964,
+      "step": 65780
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.48371943013139e-07,
+      "loss": 0.5068,
+      "step": 65781
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.483445925535444e-07,
+      "loss": 0.6855,
+      "step": 65782
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.483172422100651e-07,
+      "loss": 0.5957,
+      "step": 65783
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.482898919827224e-07,
+      "loss": 0.4897,
+      "step": 65784
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.48262541871537e-07,
+      "loss": 0.7129,
+      "step": 65785
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.482351918765303e-07,
+      "loss": 0.457,
+      "step": 65786
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.48207841997723e-07,
+      "loss": 0.6646,
+      "step": 65787
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.481804922351357e-07,
+      "loss": 0.7295,
+      "step": 65788
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.481531425887898e-07,
+      "loss": 0.6733,
+      "step": 65789
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.481257930587058e-07,
+      "loss": 0.6631,
+      "step": 65790
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.48098443644905e-07,
+      "loss": 0.7627,
+      "step": 65791
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.480710943474084e-07,
+      "loss": 0.6816,
+      "step": 65792
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.480437451662365e-07,
+      "loss": 0.7012,
+      "step": 65793
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.480163961014108e-07,
+      "loss": 0.7454,
+      "step": 65794
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.479890471529516e-07,
+      "loss": 0.7319,
+      "step": 65795
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.479616983208805e-07,
+      "loss": 0.4592,
+      "step": 65796
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.479343496052183e-07,
+      "loss": 0.5879,
+      "step": 65797
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.479070010059854e-07,
+      "loss": 0.7861,
+      "step": 65798
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.478796525232034e-07,
+      "loss": 0.7061,
+      "step": 65799
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.478523041568926e-07,
+      "loss": 0.6807,
+      "step": 65800
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.478249559070746e-07,
+      "loss": 0.6729,
+      "step": 65801
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.477976077737698e-07,
+      "loss": 0.6299,
+      "step": 65802
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.477702597569996e-07,
+      "loss": 0.7627,
+      "step": 65803
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.477429118567845e-07,
+      "loss": 0.6201,
+      "step": 65804
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.477155640731458e-07,
+      "loss": 0.5421,
+      "step": 65805
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.476882164061041e-07,
+      "loss": 0.6455,
+      "step": 65806
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.47660868855681e-07,
+      "loss": 0.7451,
+      "step": 65807
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.476335214218967e-07,
+      "loss": 0.6064,
+      "step": 65808
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.476061741047725e-07,
+      "loss": 0.71,
+      "step": 65809
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.475788269043289e-07,
+      "loss": 0.6836,
+      "step": 65810
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.475514798205875e-07,
+      "loss": 0.5649,
+      "step": 65811
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.475241328535687e-07,
+      "loss": 0.6572,
+      "step": 65812
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.474967860032938e-07,
+      "loss": 0.8047,
+      "step": 65813
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.474694392697834e-07,
+      "loss": 0.5486,
+      "step": 65814
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.47442092653059e-07,
+      "loss": 0.8057,
+      "step": 65815
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.474147461531408e-07,
+      "loss": 0.7227,
+      "step": 65816
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.473873997700506e-07,
+      "loss": 0.6875,
+      "step": 65817
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.473600535038086e-07,
+      "loss": 0.562,
+      "step": 65818
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.473327073544358e-07,
+      "loss": 0.48,
+      "step": 65819
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.473053613219535e-07,
+      "loss": 0.8115,
+      "step": 65820
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.472780154063824e-07,
+      "loss": 0.6079,
+      "step": 65821
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.472506696077435e-07,
+      "loss": 0.7256,
+      "step": 65822
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.472233239260576e-07,
+      "loss": 0.3335,
+      "step": 65823
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.471959783613461e-07,
+      "loss": 0.7129,
+      "step": 65824
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.471686329136293e-07,
+      "loss": 0.7266,
+      "step": 65825
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.471412875829287e-07,
+      "loss": 0.7212,
+      "step": 65826
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.471139423692652e-07,
+      "loss": 0.6304,
+      "step": 65827
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.470865972726593e-07,
+      "loss": 0.7144,
+      "step": 65828
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.47059252293132e-07,
+      "loss": 0.8047,
+      "step": 65829
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.470319074307044e-07,
+      "loss": 0.6807,
+      "step": 65830
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.470045626853974e-07,
+      "loss": 0.5903,
+      "step": 65831
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.469772180572322e-07,
+      "loss": 0.6826,
+      "step": 65832
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.469498735462293e-07,
+      "loss": 0.7783,
+      "step": 65833
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.469225291524099e-07,
+      "loss": 0.6826,
+      "step": 65834
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.468951848757948e-07,
+      "loss": 0.7026,
+      "step": 65835
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.468678407164052e-07,
+      "loss": 0.6343,
+      "step": 65836
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.468404966742622e-07,
+      "loss": 0.5322,
+      "step": 65837
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.468131527493858e-07,
+      "loss": 0.7344,
+      "step": 65838
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.467858089417978e-07,
+      "loss": 0.5209,
+      "step": 65839
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.467584652515185e-07,
+      "loss": 0.6013,
+      "step": 65840
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.467311216785695e-07,
+      "loss": 0.6602,
+      "step": 65841
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.467037782229714e-07,
+      "loss": 0.5405,
+      "step": 65842
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.466764348847451e-07,
+      "loss": 0.8154,
+      "step": 65843
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.466490916639116e-07,
+      "loss": 0.6768,
+      "step": 65844
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.46621748560492e-07,
+      "loss": 0.7129,
+      "step": 65845
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.465944055745068e-07,
+      "loss": 0.7412,
+      "step": 65846
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.465670627059778e-07,
+      "loss": 0.4684,
+      "step": 65847
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.465397199549249e-07,
+      "loss": 0.5708,
+      "step": 65848
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.465123773213696e-07,
+      "loss": 0.519,
+      "step": 65849
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.464850348053326e-07,
+      "loss": 0.7178,
+      "step": 65850
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.464576924068349e-07,
+      "loss": 0.7012,
+      "step": 65851
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.464303501258975e-07,
+      "loss": 0.5361,
+      "step": 65852
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.464030079625414e-07,
+      "loss": 0.5293,
+      "step": 65853
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.463756659167874e-07,
+      "loss": 0.7793,
+      "step": 65854
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.463483239886565e-07,
+      "loss": 0.5889,
+      "step": 65855
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.463209821781695e-07,
+      "loss": 0.6807,
+      "step": 65856
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.462936404853479e-07,
+      "loss": 0.6089,
+      "step": 65857
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.462662989102121e-07,
+      "loss": 0.5347,
+      "step": 65858
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.462389574527826e-07,
+      "loss": 0.7021,
+      "step": 65859
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.462116161130814e-07,
+      "loss": 0.688,
+      "step": 65860
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.461842748911285e-07,
+      "loss": 0.6167,
+      "step": 65861
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.461569337869454e-07,
+      "loss": 0.8242,
+      "step": 65862
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.461295928005528e-07,
+      "loss": 0.6328,
+      "step": 65863
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.461022519319719e-07,
+      "loss": 0.6709,
+      "step": 65864
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.460749111812233e-07,
+      "loss": 0.6533,
+      "step": 65865
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.460475705483278e-07,
+      "loss": 0.6709,
+      "step": 65866
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.460202300333073e-07,
+      "loss": 0.4055,
+      "step": 65867
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.459928896361816e-07,
+      "loss": 0.4274,
+      "step": 65868
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.45965549356972e-07,
+      "loss": 0.5371,
+      "step": 65869
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.459382091956996e-07,
+      "loss": 0.533,
+      "step": 65870
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.45910869152385e-07,
+      "loss": 0.5486,
+      "step": 65871
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.458835292270496e-07,
+      "loss": 0.6455,
+      "step": 65872
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.458561894197138e-07,
+      "loss": 0.7168,
+      "step": 65873
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.458288497303992e-07,
+      "loss": 0.6631,
+      "step": 65874
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.45801510159126e-07,
+      "loss": 0.5669,
+      "step": 65875
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.457741707059158e-07,
+      "loss": 0.7261,
+      "step": 65876
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.457468313707895e-07,
+      "loss": 0.4729,
+      "step": 65877
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.457194921537673e-07,
+      "loss": 0.6562,
+      "step": 65878
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.456921530548707e-07,
+      "loss": 0.6934,
+      "step": 65879
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.456648140741204e-07,
+      "loss": 0.4998,
+      "step": 65880
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.456374752115376e-07,
+      "loss": 0.7197,
+      "step": 65881
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.456101364671431e-07,
+      "loss": 0.397,
+      "step": 65882
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.455827978409578e-07,
+      "loss": 0.7109,
+      "step": 65883
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.455554593330025e-07,
+      "loss": 0.7432,
+      "step": 65884
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.455281209432984e-07,
+      "loss": 0.7656,
+      "step": 65885
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.455007826718663e-07,
+      "loss": 0.6316,
+      "step": 65886
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.454734445187274e-07,
+      "loss": 0.5894,
+      "step": 65887
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.454461064839021e-07,
+      "loss": 0.6636,
+      "step": 65888
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.454187685674118e-07,
+      "loss": 0.7104,
+      "step": 65889
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.453914307692768e-07,
+      "loss": 0.689,
+      "step": 65890
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.45364093089519e-07,
+      "loss": 0.6587,
+      "step": 65891
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.453367555281584e-07,
+      "loss": 0.7256,
+      "step": 65892
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.453094180852166e-07,
+      "loss": 0.5913,
+      "step": 65893
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.452820807607139e-07,
+      "loss": 0.6611,
+      "step": 65894
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.45254743554672e-07,
+      "loss": 0.6387,
+      "step": 65895
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.452274064671112e-07,
+      "loss": 0.5737,
+      "step": 65896
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.452000694980531e-07,
+      "loss": 0.645,
+      "step": 65897
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.451727326475179e-07,
+      "loss": 0.5923,
+      "step": 65898
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.451453959155268e-07,
+      "loss": 0.5029,
+      "step": 65899
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.451180593021006e-07,
+      "loss": 0.6602,
+      "step": 65900
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.450907228072606e-07,
+      "loss": 0.749,
+      "step": 65901
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.450633864310272e-07,
+      "loss": 0.6133,
+      "step": 65902
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.450360501734219e-07,
+      "loss": 0.5879,
+      "step": 65903
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.450087140344653e-07,
+      "loss": 0.6436,
+      "step": 65904
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.449813780141784e-07,
+      "loss": 0.6641,
+      "step": 65905
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.449540421125824e-07,
+      "loss": 0.6038,
+      "step": 65906
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.449267063296977e-07,
+      "loss": 0.5698,
+      "step": 65907
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.448993706655455e-07,
+      "loss": 0.6953,
+      "step": 65908
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.448720351201466e-07,
+      "loss": 0.5757,
+      "step": 65909
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.448446996935222e-07,
+      "loss": 0.6079,
+      "step": 65910
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.44817364385693e-07,
+      "loss": 0.6514,
+      "step": 65911
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.447900291966801e-07,
+      "loss": 0.467,
+      "step": 65912
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.447626941265041e-07,
+      "loss": 0.6265,
+      "step": 65913
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.447353591751865e-07,
+      "loss": 0.6084,
+      "step": 65914
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.447080243427476e-07,
+      "loss": 0.7295,
+      "step": 65915
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.446806896292094e-07,
+      "loss": 0.7178,
+      "step": 65916
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.446533550345912e-07,
+      "loss": 0.6299,
+      "step": 65917
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.446260205589151e-07,
+      "loss": 0.7373,
+      "step": 65918
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.445986862022014e-07,
+      "loss": 0.5796,
+      "step": 65919
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.445713519644717e-07,
+      "loss": 0.6152,
+      "step": 65920
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.445440178457463e-07,
+      "loss": 0.7344,
+      "step": 65921
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.445166838460467e-07,
+      "loss": 0.8486,
+      "step": 65922
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.444893499653931e-07,
+      "loss": 0.7109,
+      "step": 65923
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.444620162038073e-07,
+      "loss": 0.665,
+      "step": 65924
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.444346825613094e-07,
+      "loss": 0.6753,
+      "step": 65925
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.444073490379212e-07,
+      "loss": 0.8916,
+      "step": 65926
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.443800156336629e-07,
+      "loss": 0.6406,
+      "step": 65927
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.443526823485555e-07,
+      "loss": 0.873,
+      "step": 65928
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.443253491826204e-07,
+      "loss": 0.4657,
+      "step": 65929
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.442980161358779e-07,
+      "loss": 0.6357,
+      "step": 65930
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.442706832083494e-07,
+      "loss": 0.6914,
+      "step": 65931
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.442433504000555e-07,
+      "loss": 0.5171,
+      "step": 65932
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.442160177110175e-07,
+      "loss": 0.7236,
+      "step": 65933
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.441886851412559e-07,
+      "loss": 0.5576,
+      "step": 65934
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.441613526907922e-07,
+      "loss": 0.6636,
+      "step": 65935
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.441340203596472e-07,
+      "loss": 0.3677,
+      "step": 65936
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.441066881478414e-07,
+      "loss": 0.5791,
+      "step": 65937
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.440793560553956e-07,
+      "loss": 0.5996,
+      "step": 65938
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.440520240823313e-07,
+      "loss": 0.6567,
+      "step": 65939
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.44024692228669e-07,
+      "loss": 0.4696,
+      "step": 65940
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.4399736049443e-07,
+      "loss": 0.7158,
+      "step": 65941
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.439700288796349e-07,
+      "loss": 0.6396,
+      "step": 65942
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.439426973843051e-07,
+      "loss": 0.5474,
+      "step": 65943
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.439153660084608e-07,
+      "loss": 0.6758,
+      "step": 65944
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.438880347521235e-07,
+      "loss": 0.917,
+      "step": 65945
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.438607036153143e-07,
+      "loss": 0.6865,
+      "step": 65946
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.438333725980534e-07,
+      "loss": 0.7637,
+      "step": 65947
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.438060417003623e-07,
+      "loss": 0.5361,
+      "step": 65948
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.437787109222614e-07,
+      "loss": 0.667,
+      "step": 65949
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.437513802637722e-07,
+      "loss": 0.8057,
+      "step": 65950
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.437240497249153e-07,
+      "loss": 0.4927,
+      "step": 65951
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.436967193057118e-07,
+      "loss": 0.6924,
+      "step": 65952
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.436693890061824e-07,
+      "loss": 0.6826,
+      "step": 65953
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.436420588263484e-07,
+      "loss": 0.5601,
+      "step": 65954
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.436147287662303e-07,
+      "loss": 0.5735,
+      "step": 65955
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.435873988258495e-07,
+      "loss": 0.6855,
+      "step": 65956
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.435600690052262e-07,
+      "loss": 0.5713,
+      "step": 65957
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.43532739304382e-07,
+      "loss": 0.8066,
+      "step": 65958
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.435054097233375e-07,
+      "loss": 0.6001,
+      "step": 65959
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.434780802621138e-07,
+      "loss": 0.5854,
+      "step": 65960
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.434507509207316e-07,
+      "loss": 0.5854,
+      "step": 65961
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.434234216992121e-07,
+      "loss": 0.6191,
+      "step": 65962
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.433960925975759e-07,
+      "loss": 0.7012,
+      "step": 65963
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.433687636158443e-07,
+      "loss": 0.8008,
+      "step": 65964
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.433414347540378e-07,
+      "loss": 0.6216,
+      "step": 65965
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.43314106012178e-07,
+      "loss": 0.6094,
+      "step": 65966
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.432867773902852e-07,
+      "loss": 0.5339,
+      "step": 65967
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.432594488883803e-07,
+      "loss": 0.5676,
+      "step": 65968
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.432321205064847e-07,
+      "loss": 0.4907,
+      "step": 65969
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.432047922446187e-07,
+      "loss": 0.7432,
+      "step": 65970
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.431774641028039e-07,
+      "loss": 0.8018,
+      "step": 65971
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.431501360810609e-07,
+      "loss": 0.6226,
+      "step": 65972
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.431228081794104e-07,
+      "loss": 0.6348,
+      "step": 65973
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.430954803978739e-07,
+      "loss": 0.6611,
+      "step": 65974
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.430681527364714e-07,
+      "loss": 0.4231,
+      "step": 65975
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.430408251952251e-07,
+      "loss": 0.834,
+      "step": 65976
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.430134977741549e-07,
+      "loss": 0.5742,
+      "step": 65977
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.42986170473282e-07,
+      "loss": 0.3323,
+      "step": 65978
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.429588432926274e-07,
+      "loss": 0.4312,
+      "step": 65979
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.42931516232212e-07,
+      "loss": 0.7266,
+      "step": 65980
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.429041892920567e-07,
+      "loss": 0.644,
+      "step": 65981
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.428768624721823e-07,
+      "loss": 0.5713,
+      "step": 65982
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.428495357726101e-07,
+      "loss": 0.6748,
+      "step": 65983
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.428222091933606e-07,
+      "loss": 0.6143,
+      "step": 65984
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.427948827344549e-07,
+      "loss": 0.7437,
+      "step": 65985
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.427675563959145e-07,
+      "loss": 0.5461,
+      "step": 65986
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.427402301777591e-07,
+      "loss": 0.6492,
+      "step": 65987
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.427129040800105e-07,
+      "loss": 0.5713,
+      "step": 65988
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.426855781026895e-07,
+      "loss": 0.7061,
+      "step": 65989
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.426582522458165e-07,
+      "loss": 0.5791,
+      "step": 65990
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.426309265094133e-07,
+      "loss": 0.7354,
+      "step": 65991
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.426036008935e-07,
+      "loss": 0.6318,
+      "step": 65992
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.425762753980982e-07,
+      "loss": 0.6929,
+      "step": 65993
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.425489500232283e-07,
+      "loss": 0.7354,
+      "step": 65994
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.425216247689115e-07,
+      "loss": 0.6606,
+      "step": 65995
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.42494299635169e-07,
+      "loss": 0.6494,
+      "step": 65996
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.424669746220208e-07,
+      "loss": 0.7563,
+      "step": 65997
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.424396497294887e-07,
+      "loss": 0.7139,
+      "step": 65998
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.424123249575931e-07,
+      "loss": 0.7549,
+      "step": 65999
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.423850003063553e-07,
+      "loss": 0.7969,
+      "step": 66000
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.42357675775796e-07,
+      "loss": 0.5791,
+      "step": 66001
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.423303513659362e-07,
+      "loss": 0.7388,
+      "step": 66002
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.423030270767967e-07,
+      "loss": 0.5034,
+      "step": 66003
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.422757029083987e-07,
+      "loss": 0.4434,
+      "step": 66004
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.422483788607627e-07,
+      "loss": 0.6104,
+      "step": 66005
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.422210549339104e-07,
+      "loss": 0.6357,
+      "step": 66006
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.421937311278618e-07,
+      "loss": 0.5605,
+      "step": 66007
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.421664074426383e-07,
+      "loss": 0.5095,
+      "step": 66008
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.421390838782605e-07,
+      "loss": 0.6338,
+      "step": 66009
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.421117604347499e-07,
+      "loss": 0.6089,
+      "step": 66010
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.420844371121267e-07,
+      "loss": 0.5093,
+      "step": 66011
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.420571139104123e-07,
+      "loss": 0.5938,
+      "step": 66012
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.420297908296275e-07,
+      "loss": 0.501,
+      "step": 66013
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.420024678697932e-07,
+      "loss": 0.6057,
+      "step": 66014
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.419751450309304e-07,
+      "loss": 0.6611,
+      "step": 66015
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.419478223130603e-07,
+      "loss": 0.7593,
+      "step": 66016
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.419204997162032e-07,
+      "loss": 0.6162,
+      "step": 66017
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.418931772403801e-07,
+      "loss": 0.7637,
+      "step": 66018
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.418658548856124e-07,
+      "loss": 0.6479,
+      "step": 66019
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.418385326519205e-07,
+      "loss": 0.7754,
+      "step": 66020
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.418112105393258e-07,
+      "loss": 0.5217,
+      "step": 66021
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.417838885478488e-07,
+      "loss": 0.7217,
+      "step": 66022
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.417565666775108e-07,
+      "loss": 0.5674,
+      "step": 66023
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.417292449283322e-07,
+      "loss": 0.5073,
+      "step": 66024
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.417019233003344e-07,
+      "loss": 0.7095,
+      "step": 66025
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.416746017935386e-07,
+      "loss": 0.5786,
+      "step": 66026
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.41647280407965e-07,
+      "loss": 0.6226,
+      "step": 66027
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.416199591436344e-07,
+      "loss": 0.7871,
+      "step": 66028
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.415926380005686e-07,
+      "loss": 0.5569,
+      "step": 66029
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.415653169787876e-07,
+      "loss": 0.6582,
+      "step": 66030
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.415379960783131e-07,
+      "loss": 0.6318,
+      "step": 66031
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.415106752991655e-07,
+      "loss": 0.834,
+      "step": 66032
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.414833546413659e-07,
+      "loss": 0.6475,
+      "step": 66033
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.414560341049351e-07,
+      "loss": 0.5002,
+      "step": 66034
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.414287136898943e-07,
+      "loss": 0.7021,
+      "step": 66035
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.414013933962645e-07,
+      "loss": 0.7158,
+      "step": 66036
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.413740732240659e-07,
+      "loss": 0.8291,
+      "step": 66037
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.413467531733203e-07,
+      "loss": 0.8701,
+      "step": 66038
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.413194332440477e-07,
+      "loss": 0.8115,
+      "step": 66039
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.412921134362698e-07,
+      "loss": 0.5533,
+      "step": 66040
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.412647937500071e-07,
+      "loss": 0.7354,
+      "step": 66041
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.412374741852808e-07,
+      "loss": 0.6123,
+      "step": 66042
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.412101547421116e-07,
+      "loss": 0.5691,
+      "step": 66043
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.411828354205205e-07,
+      "loss": 0.4631,
+      "step": 66044
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.411555162205284e-07,
+      "loss": 0.7295,
+      "step": 66045
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.411281971421566e-07,
+      "loss": 0.624,
+      "step": 66046
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.411008781854249e-07,
+      "loss": 0.7725,
+      "step": 66047
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.410735593503555e-07,
+      "loss": 0.5879,
+      "step": 66048
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.410462406369683e-07,
+      "loss": 0.585,
+      "step": 66049
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.410189220452851e-07,
+      "loss": 0.6143,
+      "step": 66050
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.409916035753262e-07,
+      "loss": 0.5781,
+      "step": 66051
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.409642852271128e-07,
+      "loss": 0.667,
+      "step": 66052
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.409369670006655e-07,
+      "loss": 0.627,
+      "step": 66053
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.409096488960058e-07,
+      "loss": 0.6431,
+      "step": 66054
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.40882330913154e-07,
+      "loss": 0.5371,
+      "step": 66055
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.408550130521316e-07,
+      "loss": 0.5312,
+      "step": 66056
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.408276953129591e-07,
+      "loss": 0.6001,
+      "step": 66057
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.408003776956573e-07,
+      "loss": 0.542,
+      "step": 66058
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.407730602002475e-07,
+      "loss": 0.7197,
+      "step": 66059
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.407457428267503e-07,
+      "loss": 0.5889,
+      "step": 66060
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.407184255751869e-07,
+      "loss": 0.5576,
+      "step": 66061
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.406911084455781e-07,
+      "loss": 0.7285,
+      "step": 66062
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.406637914379445e-07,
+      "loss": 0.5215,
+      "step": 66063
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.406364745523076e-07,
+      "loss": 0.5459,
+      "step": 66064
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.406091577886879e-07,
+      "loss": 0.6438,
+      "step": 66065
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.405818411471067e-07,
+      "loss": 0.7832,
+      "step": 66066
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.405545246275845e-07,
+      "loss": 0.6777,
+      "step": 66067
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.405272082301422e-07,
+      "loss": 0.6025,
+      "step": 66068
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.404998919548012e-07,
+      "loss": 0.6206,
+      "step": 66069
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.404725758015817e-07,
+      "loss": 0.563,
+      "step": 66070
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.404452597705053e-07,
+      "loss": 0.6172,
+      "step": 66071
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.404179438615924e-07,
+      "loss": 0.5684,
+      "step": 66072
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.403906280748643e-07,
+      "loss": 0.7031,
+      "step": 66073
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.403633124103417e-07,
+      "loss": 0.6899,
+      "step": 66074
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.403359968680457e-07,
+      "loss": 0.5874,
+      "step": 66075
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.403086814479972e-07,
+      "loss": 0.5771,
+      "step": 66076
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.402813661502166e-07,
+      "loss": 0.6836,
+      "step": 66077
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.402540509747256e-07,
+      "loss": 0.6426,
+      "step": 66078
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.402267359215445e-07,
+      "loss": 0.627,
+      "step": 66079
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.401994209906944e-07,
+      "loss": 0.6416,
+      "step": 66080
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.401721061821965e-07,
+      "loss": 0.6777,
+      "step": 66081
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.401447914960711e-07,
+      "loss": 0.5054,
+      "step": 66082
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.401174769323398e-07,
+      "loss": 0.5032,
+      "step": 66083
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.400901624910228e-07,
+      "loss": 0.8281,
+      "step": 66084
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.400628481721419e-07,
+      "loss": 0.5892,
+      "step": 66085
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.400355339757176e-07,
+      "loss": 0.6904,
+      "step": 66086
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.400082199017703e-07,
+      "loss": 0.5845,
+      "step": 66087
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.399809059503216e-07,
+      "loss": 0.6016,
+      "step": 66088
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.399535921213918e-07,
+      "loss": 0.8779,
+      "step": 66089
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.399262784150027e-07,
+      "loss": 0.6318,
+      "step": 66090
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.398989648311742e-07,
+      "loss": 0.7192,
+      "step": 66091
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.39871651369928e-07,
+      "loss": 0.6133,
+      "step": 66092
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.398443380312845e-07,
+      "loss": 0.6372,
+      "step": 66093
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.398170248152651e-07,
+      "loss": 0.8135,
+      "step": 66094
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.397897117218902e-07,
+      "loss": 0.5625,
+      "step": 66095
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.397623987511813e-07,
+      "loss": 0.5361,
+      "step": 66096
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.397350859031588e-07,
+      "loss": 0.7041,
+      "step": 66097
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.397077731778437e-07,
+      "loss": 0.6562,
+      "step": 66098
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.396804605752569e-07,
+      "loss": 0.75,
+      "step": 66099
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.396531480954195e-07,
+      "loss": 0.5483,
+      "step": 66100
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.396258357383523e-07,
+      "loss": 0.5396,
+      "step": 66101
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.395985235040762e-07,
+      "loss": 0.7368,
+      "step": 66102
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.39571211392612e-07,
+      "loss": 0.584,
+      "step": 66103
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.39543899403981e-07,
+      "loss": 0.6904,
+      "step": 66104
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.395165875382036e-07,
+      "loss": 0.6787,
+      "step": 66105
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.394892757953014e-07,
+      "loss": 0.667,
+      "step": 66106
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.394619641752947e-07,
+      "loss": 0.7773,
+      "step": 66107
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.394346526782043e-07,
+      "loss": 0.446,
+      "step": 66108
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.394073413040518e-07,
+      "loss": 0.6357,
+      "step": 66109
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.393800300528573e-07,
+      "loss": 0.4493,
+      "step": 66110
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.393527189246424e-07,
+      "loss": 0.6094,
+      "step": 66111
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.393254079194276e-07,
+      "loss": 0.2987,
+      "step": 66112
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.392980970372341e-07,
+      "loss": 0.4878,
+      "step": 66113
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.392707862780825e-07,
+      "loss": 0.3899,
+      "step": 66114
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.392434756419939e-07,
+      "loss": 0.9658,
+      "step": 66115
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.392161651289896e-07,
+      "loss": 0.6318,
+      "step": 66116
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.391888547390899e-07,
+      "loss": 0.6523,
+      "step": 66117
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.391615444723154e-07,
+      "loss": 0.6841,
+      "step": 66118
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.39134234328688e-07,
+      "loss": 0.2665,
+      "step": 66119
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.391069243082278e-07,
+      "loss": 0.627,
+      "step": 66120
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.390796144109563e-07,
+      "loss": 0.6099,
+      "step": 66121
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.390523046368938e-07,
+      "loss": 0.5317,
+      "step": 66122
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.390249949860619e-07,
+      "loss": 0.625,
+      "step": 66123
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.389976854584809e-07,
+      "loss": 0.6812,
+      "step": 66124
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.389703760541722e-07,
+      "loss": 0.6311,
+      "step": 66125
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.389430667731566e-07,
+      "loss": 0.6553,
+      "step": 66126
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.389157576154547e-07,
+      "loss": 0.5874,
+      "step": 66127
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.388884485810875e-07,
+      "loss": 0.5151,
+      "step": 66128
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.38861139670076e-07,
+      "loss": 0.6777,
+      "step": 66129
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.388338308824412e-07,
+      "loss": 0.5867,
+      "step": 66130
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.388065222182038e-07,
+      "loss": 0.528,
+      "step": 66131
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 8.387792136773851e-07,
+      "loss": 0.5457,
+      "step": 66132
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.387519052600054e-07,
+      "loss": 0.7285,
+      "step": 66133
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.387245969660863e-07,
+      "loss": 0.5693,
+      "step": 66134
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.386972887956481e-07,
+      "loss": 0.6587,
+      "step": 66135
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.386699807487124e-07,
+      "loss": 0.6113,
+      "step": 66136
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.386426728252994e-07,
+      "loss": 0.5913,
+      "step": 66137
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.386153650254302e-07,
+      "loss": 0.5908,
+      "step": 66138
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.385880573491257e-07,
+      "loss": 0.6157,
+      "step": 66139
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.385607497964071e-07,
+      "loss": 0.5874,
+      "step": 66140
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.385334423672948e-07,
+      "loss": 0.7627,
+      "step": 66141
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.385061350618104e-07,
+      "loss": 0.7139,
+      "step": 66142
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.384788278799741e-07,
+      "loss": 0.4014,
+      "step": 66143
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.384515208218075e-07,
+      "loss": 0.6377,
+      "step": 66144
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.384242138873308e-07,
+      "loss": 0.6016,
+      "step": 66145
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.383969070765657e-07,
+      "loss": 0.5752,
+      "step": 66146
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.383696003895324e-07,
+      "loss": 0.4167,
+      "step": 66147
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.383422938262519e-07,
+      "loss": 0.6504,
+      "step": 66148
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.383149873867455e-07,
+      "loss": 0.7451,
+      "step": 66149
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.382876810710336e-07,
+      "loss": 0.8438,
+      "step": 66150
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.382603748791377e-07,
+      "loss": 0.5781,
+      "step": 66151
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.382330688110784e-07,
+      "loss": 0.6553,
+      "step": 66152
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.382057628668763e-07,
+      "loss": 0.3699,
+      "step": 66153
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.381784570465529e-07,
+      "loss": 0.5273,
+      "step": 66154
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.381511513501285e-07,
+      "loss": 0.7139,
+      "step": 66155
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.381238457776248e-07,
+      "loss": 0.6323,
+      "step": 66156
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.38096540329062e-07,
+      "loss": 0.7036,
+      "step": 66157
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.380692350044612e-07,
+      "loss": 0.6826,
+      "step": 66158
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.380419298038433e-07,
+      "loss": 0.6895,
+      "step": 66159
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.380146247272291e-07,
+      "loss": 0.7539,
+      "step": 66160
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.379873197746399e-07,
+      "loss": 0.7383,
+      "step": 66161
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.379600149460962e-07,
+      "loss": 0.6455,
+      "step": 66162
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.379327102416191e-07,
+      "loss": 0.7148,
+      "step": 66163
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.379054056612295e-07,
+      "loss": 0.6001,
+      "step": 66164
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.378781012049483e-07,
+      "loss": 0.5344,
+      "step": 66165
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.378507968727967e-07,
+      "loss": 0.4023,
+      "step": 66166
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.378234926647948e-07,
+      "loss": 0.6943,
+      "step": 66167
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.377961885809642e-07,
+      "loss": 0.6016,
+      "step": 66168
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.377688846213256e-07,
+      "loss": 0.428,
+      "step": 66169
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.377415807858998e-07,
+      "loss": 0.4912,
+      "step": 66170
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.377142770747079e-07,
+      "loss": 0.5151,
+      "step": 66171
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.376869734877706e-07,
+      "loss": 0.709,
+      "step": 66172
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.376596700251092e-07,
+      "loss": 0.5737,
+      "step": 66173
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.37632366686744e-07,
+      "loss": 0.6118,
+      "step": 66174
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.376050634726964e-07,
+      "loss": 0.5527,
+      "step": 66175
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.375777603829875e-07,
+      "loss": 0.5674,
+      "step": 66176
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.375504574176372e-07,
+      "loss": 0.707,
+      "step": 66177
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.375231545766674e-07,
+      "loss": 0.4482,
+      "step": 66178
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.374958518600984e-07,
+      "loss": 0.6357,
+      "step": 66179
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.374685492679515e-07,
+      "loss": 0.5498,
+      "step": 66180
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.374412468002474e-07,
+      "loss": 0.4805,
+      "step": 66181
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.374139444570072e-07,
+      "loss": 0.6133,
+      "step": 66182
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.373866422382514e-07,
+      "loss": 0.3939,
+      "step": 66183
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.373593401440014e-07,
+      "loss": 0.5399,
+      "step": 66184
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.373320381742777e-07,
+      "loss": 0.6865,
+      "step": 66185
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.373047363291018e-07,
+      "loss": 0.6992,
+      "step": 66186
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.37277434608494e-07,
+      "loss": 0.6079,
+      "step": 66187
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.372501330124754e-07,
+      "loss": 0.5,
+      "step": 66188
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.372228315410666e-07,
+      "loss": 0.6904,
+      "step": 66189
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.37195530194289e-07,
+      "loss": 0.6973,
+      "step": 66190
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.371682289721631e-07,
+      "loss": 0.6562,
+      "step": 66191
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.371409278747104e-07,
+      "loss": 0.4868,
+      "step": 66192
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.371136269019509e-07,
+      "loss": 0.7241,
+      "step": 66193
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.370863260539064e-07,
+      "loss": 0.6909,
+      "step": 66194
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.370590253305976e-07,
+      "loss": 0.7188,
+      "step": 66195
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.370317247320446e-07,
+      "loss": 0.6763,
+      "step": 66196
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.370044242582693e-07,
+      "loss": 0.792,
+      "step": 66197
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36977123909292e-07,
+      "loss": 0.5811,
+      "step": 66198
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36949823685134e-07,
+      "loss": 0.6299,
+      "step": 66199
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36922523585816e-07,
+      "loss": 0.7861,
+      "step": 66200
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.368952236113589e-07,
+      "loss": 0.5474,
+      "step": 66201
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.368679237617834e-07,
+      "loss": 0.6196,
+      "step": 66202
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36840624037111e-07,
+      "loss": 0.624,
+      "step": 66203
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.368133244373619e-07,
+      "loss": 0.5334,
+      "step": 66204
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36786024962558e-07,
+      "loss": 0.6543,
+      "step": 66205
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.367587256127188e-07,
+      "loss": 0.6738,
+      "step": 66206
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.367314263878664e-07,
+      "loss": 0.688,
+      "step": 66207
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.367041272880209e-07,
+      "loss": 0.5875,
+      "step": 66208
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.366768283132038e-07,
+      "loss": 0.6357,
+      "step": 66209
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.366495294634354e-07,
+      "loss": 0.4175,
+      "step": 66210
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.366222307387372e-07,
+      "loss": 0.5835,
+      "step": 66211
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.365949321391297e-07,
+      "loss": 0.8076,
+      "step": 66212
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.365676336646341e-07,
+      "loss": 0.6172,
+      "step": 66213
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.365403353152709e-07,
+      "loss": 0.6924,
+      "step": 66214
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.365130370910618e-07,
+      "loss": 0.6973,
+      "step": 66215
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.364857389920268e-07,
+      "loss": 0.6299,
+      "step": 66216
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36458441018187e-07,
+      "loss": 0.4888,
+      "step": 66217
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.364311431695638e-07,
+      "loss": 0.5034,
+      "step": 66218
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.364038454461774e-07,
+      "loss": 0.5669,
+      "step": 66219
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.363765478480493e-07,
+      "loss": 0.6279,
+      "step": 66220
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.363492503751999e-07,
+      "loss": 0.5352,
+      "step": 66221
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.363219530276505e-07,
+      "loss": 0.5723,
+      "step": 66222
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.362946558054218e-07,
+      "loss": 0.8125,
+      "step": 66223
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.362673587085348e-07,
+      "loss": 0.554,
+      "step": 66224
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.362400617370107e-07,
+      "loss": 0.7314,
+      "step": 66225
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.362127648908698e-07,
+      "loss": 0.624,
+      "step": 66226
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.361854681701331e-07,
+      "loss": 0.8184,
+      "step": 66227
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.361581715748218e-07,
+      "loss": 0.6016,
+      "step": 66228
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.361308751049565e-07,
+      "loss": 0.5256,
+      "step": 66229
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.361035787605584e-07,
+      "loss": 0.7012,
+      "step": 66230
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.360762825416481e-07,
+      "loss": 0.5896,
+      "step": 66231
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.360489864482467e-07,
+      "loss": 0.6606,
+      "step": 66232
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.36021690480375e-07,
+      "loss": 0.6348,
+      "step": 66233
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.35994394638054e-07,
+      "loss": 0.6406,
+      "step": 66234
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.35967098921305e-07,
+      "loss": 0.6343,
+      "step": 66235
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.359398033301479e-07,
+      "loss": 0.6494,
+      "step": 66236
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.359125078646044e-07,
+      "loss": 0.8096,
+      "step": 66237
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.358852125246949e-07,
+      "loss": 0.6108,
+      "step": 66238
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.358579173104409e-07,
+      "loss": 0.4963,
+      "step": 66239
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.358306222218625e-07,
+      "loss": 0.5952,
+      "step": 66240
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.358033272589814e-07,
+      "loss": 0.7305,
+      "step": 66241
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.357760324218182e-07,
+      "loss": 0.8154,
+      "step": 66242
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.357487377103934e-07,
+      "loss": 0.6494,
+      "step": 66243
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.357214431247285e-07,
+      "loss": 0.5806,
+      "step": 66244
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.356941486648444e-07,
+      "loss": 0.5718,
+      "step": 66245
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.356668543307611e-07,
+      "loss": 0.6252,
+      "step": 66246
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.356395601225006e-07,
+      "loss": 0.5527,
+      "step": 66247
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.35612266040083e-07,
+      "loss": 0.4946,
+      "step": 66248
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.355849720835297e-07,
+      "loss": 0.5352,
+      "step": 66249
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.355576782528612e-07,
+      "loss": 0.7646,
+      "step": 66250
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.35530384548099e-07,
+      "loss": 0.6499,
+      "step": 66251
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.355030909692633e-07,
+      "loss": 0.6445,
+      "step": 66252
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.354757975163755e-07,
+      "loss": 0.7334,
+      "step": 66253
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.354485041894562e-07,
+      "loss": 0.8311,
+      "step": 66254
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.354212109885268e-07,
+      "loss": 0.6309,
+      "step": 66255
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.353939179136076e-07,
+      "loss": 0.6069,
+      "step": 66256
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.353666249647194e-07,
+      "loss": 0.7393,
+      "step": 66257
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.353393321418838e-07,
+      "loss": 0.6104,
+      "step": 66258
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.353120394451209e-07,
+      "loss": 0.5737,
+      "step": 66259
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.352847468744524e-07,
+      "loss": 0.4868,
+      "step": 66260
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.352574544298987e-07,
+      "loss": 0.6143,
+      "step": 66261
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.352301621114805e-07,
+      "loss": 0.5557,
+      "step": 66262
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.352028699192193e-07,
+      "loss": 0.6235,
+      "step": 66263
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.351755778531355e-07,
+      "loss": 0.4688,
+      "step": 66264
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.351482859132506e-07,
+      "loss": 0.624,
+      "step": 66265
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.351209940995848e-07,
+      "loss": 0.6343,
+      "step": 66266
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.35093702412159e-07,
+      "loss": 0.6494,
+      "step": 66267
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.350664108509947e-07,
+      "loss": 0.6431,
+      "step": 66268
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.350391194161123e-07,
+      "loss": 0.6875,
+      "step": 66269
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.350118281075329e-07,
+      "loss": 0.7568,
+      "step": 66270
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.349845369252773e-07,
+      "loss": 0.6353,
+      "step": 66271
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.349572458693665e-07,
+      "loss": 0.6216,
+      "step": 66272
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.349299549398213e-07,
+      "loss": 0.6387,
+      "step": 66273
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.349026641366628e-07,
+      "loss": 0.7324,
+      "step": 66274
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.34875373459912e-07,
+      "loss": 0.5933,
+      "step": 66275
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.348480829095889e-07,
+      "loss": 0.5137,
+      "step": 66276
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.348207924857154e-07,
+      "loss": 0.7432,
+      "step": 66277
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.347935021883121e-07,
+      "loss": 0.4852,
+      "step": 66278
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.347662120173995e-07,
+      "loss": 0.6484,
+      "step": 66279
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.347389219729991e-07,
+      "loss": 0.6514,
+      "step": 66280
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.347116320551312e-07,
+      "loss": 0.5781,
+      "step": 66281
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.346843422638174e-07,
+      "loss": 0.5386,
+      "step": 66282
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.346570525990777e-07,
+      "loss": 0.6934,
+      "step": 66283
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.34629763060934e-07,
+      "loss": 0.7178,
+      "step": 66284
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.346024736494066e-07,
+      "loss": 0.6553,
+      "step": 66285
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.345751843645164e-07,
+      "loss": 0.5767,
+      "step": 66286
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.345478952062843e-07,
+      "loss": 0.7129,
+      "step": 66287
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.345206061747312e-07,
+      "loss": 0.6299,
+      "step": 66288
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.344933172698782e-07,
+      "loss": 0.6724,
+      "step": 66289
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.344660284917458e-07,
+      "loss": 0.5493,
+      "step": 66290
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.344387398403554e-07,
+      "loss": 0.79,
+      "step": 66291
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.344114513157275e-07,
+      "loss": 0.6855,
+      "step": 66292
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.343841629178834e-07,
+      "loss": 0.834,
+      "step": 66293
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.343568746468433e-07,
+      "loss": 0.7441,
+      "step": 66294
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.343295865026294e-07,
+      "loss": 0.7393,
+      "step": 66295
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.34302298485261e-07,
+      "loss": 0.6689,
+      "step": 66296
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.342750105947599e-07,
+      "loss": 0.6396,
+      "step": 66297
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.342477228311465e-07,
+      "loss": 0.6846,
+      "step": 66298
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.342204351944425e-07,
+      "loss": 0.7529,
+      "step": 66299
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.341931476846678e-07,
+      "loss": 0.5918,
+      "step": 66300
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.34165860301844e-07,
+      "loss": 0.6221,
+      "step": 66301
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.341385730459917e-07,
+      "loss": 0.6562,
+      "step": 66302
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.34111285917132e-07,
+      "loss": 0.4323,
+      "step": 66303
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.340839989152854e-07,
+      "loss": 0.6274,
+      "step": 66304
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.340567120404735e-07,
+      "loss": 0.6318,
+      "step": 66305
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.340294252927167e-07,
+      "loss": 0.7021,
+      "step": 66306
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.340021386720356e-07,
+      "loss": 0.6611,
+      "step": 66307
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.339748521784517e-07,
+      "loss": 0.6562,
+      "step": 66308
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.339475658119852e-07,
+      "loss": 0.5732,
+      "step": 66309
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.339202795726579e-07,
+      "loss": 0.7158,
+      "step": 66310
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.338929934604898e-07,
+      "loss": 0.6367,
+      "step": 66311
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.338657074755026e-07,
+      "loss": 0.6416,
+      "step": 66312
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.338384216177163e-07,
+      "loss": 0.7451,
+      "step": 66313
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.338111358871527e-07,
+      "loss": 0.6152,
+      "step": 66314
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.337838502838325e-07,
+      "loss": 0.5271,
+      "step": 66315
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.33756564807776e-07,
+      "loss": 0.6099,
+      "step": 66316
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.337292794590044e-07,
+      "loss": 0.6309,
+      "step": 66317
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.337019942375387e-07,
+      "loss": 0.615,
+      "step": 66318
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.336747091433996e-07,
+      "loss": 0.6055,
+      "step": 66319
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.336474241766085e-07,
+      "loss": 0.7173,
+      "step": 66320
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.336201393371855e-07,
+      "loss": 0.5469,
+      "step": 66321
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.335928546251522e-07,
+      "loss": 0.5815,
+      "step": 66322
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.335655700405289e-07,
+      "loss": 0.5942,
+      "step": 66323
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.335382855833371e-07,
+      "loss": 0.4219,
+      "step": 66324
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.335110012535976e-07,
+      "loss": 0.5732,
+      "step": 66325
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.334837170513306e-07,
+      "loss": 0.5283,
+      "step": 66326
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.334564329765577e-07,
+      "loss": 0.6157,
+      "step": 66327
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.334291490292993e-07,
+      "loss": 0.6704,
+      "step": 66328
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.334018652095767e-07,
+      "loss": 0.7344,
+      "step": 66329
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.333745815174106e-07,
+      "loss": 0.6963,
+      "step": 66330
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.33347297952822e-07,
+      "loss": 0.7842,
+      "step": 66331
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.333200145158314e-07,
+      "loss": 0.6211,
+      "step": 66332
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.332927312064604e-07,
+      "loss": 0.5425,
+      "step": 66333
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.332654480247294e-07,
+      "loss": 0.582,
+      "step": 66334
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.332381649706597e-07,
+      "loss": 0.6138,
+      "step": 66335
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.332108820442713e-07,
+      "loss": 0.5532,
+      "step": 66336
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.33183599245586e-07,
+      "loss": 0.6396,
+      "step": 66337
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.33156316574624e-07,
+      "loss": 0.6182,
+      "step": 66338
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.331290340314069e-07,
+      "loss": 0.6199,
+      "step": 66339
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.331017516159549e-07,
+      "loss": 0.5896,
+      "step": 66340
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.330744693282895e-07,
+      "loss": 0.7666,
+      "step": 66341
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.33047187168431e-07,
+      "loss": 0.5298,
+      "step": 66342
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.330199051364009e-07,
+      "loss": 0.6885,
+      "step": 66343
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.329926232322195e-07,
+      "loss": 0.5991,
+      "step": 66344
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.329653414559085e-07,
+      "loss": 0.6113,
+      "step": 66345
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.329380598074878e-07,
+      "loss": 0.6963,
+      "step": 66346
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.329107782869787e-07,
+      "loss": 0.6895,
+      "step": 66347
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.328834968944024e-07,
+      "loss": 0.751,
+      "step": 66348
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.328562156297793e-07,
+      "loss": 0.7109,
+      "step": 66349
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.328289344931306e-07,
+      "loss": 0.6396,
+      "step": 66350
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.328016534844773e-07,
+      "loss": 0.4336,
+      "step": 66351
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.327743726038396e-07,
+      "loss": 0.5508,
+      "step": 66352
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.327470918512393e-07,
+      "loss": 0.4797,
+      "step": 66353
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.327198112266967e-07,
+      "loss": 0.6216,
+      "step": 66354
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.32692530730233e-07,
+      "loss": 0.6318,
+      "step": 66355
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.326652503618689e-07,
+      "loss": 0.9189,
+      "step": 66356
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.326379701216251e-07,
+      "loss": 0.5366,
+      "step": 66357
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.326106900095229e-07,
+      "loss": 0.7793,
+      "step": 66358
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.325834100255829e-07,
+      "loss": 0.7358,
+      "step": 66359
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.325561301698262e-07,
+      "loss": 0.5986,
+      "step": 66360
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.325288504422733e-07,
+      "loss": 0.6514,
+      "step": 66361
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.325015708429455e-07,
+      "loss": 0.8203,
+      "step": 66362
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.324742913718637e-07,
+      "loss": 0.667,
+      "step": 66363
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.324470120290485e-07,
+      "loss": 0.4736,
+      "step": 66364
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.324197328145212e-07,
+      "loss": 0.387,
+      "step": 66365
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.32392453728302e-07,
+      "loss": 0.7422,
+      "step": 66366
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.323651747704124e-07,
+      "loss": 0.7949,
+      "step": 66367
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.323378959408731e-07,
+      "loss": 0.6748,
+      "step": 66368
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.323106172397047e-07,
+      "loss": 0.6816,
+      "step": 66369
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.322833386669288e-07,
+      "loss": 0.6084,
+      "step": 66370
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.322560602225653e-07,
+      "loss": 0.4545,
+      "step": 66371
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.322287819066361e-07,
+      "loss": 0.7266,
+      "step": 66372
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.322015037191613e-07,
+      "loss": 0.6416,
+      "step": 66373
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.321742256601622e-07,
+      "loss": 0.6245,
+      "step": 66374
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.321469477296599e-07,
+      "loss": 0.5356,
+      "step": 66375
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.321196699276744e-07,
+      "loss": 0.6157,
+      "step": 66376
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.320923922542275e-07,
+      "loss": 0.5261,
+      "step": 66377
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.320651147093395e-07,
+      "loss": 0.6143,
+      "step": 66378
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.320378372930317e-07,
+      "loss": 0.5,
+      "step": 66379
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.320105600053246e-07,
+      "loss": 0.5952,
+      "step": 66380
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.319832828462396e-07,
+      "loss": 0.5396,
+      "step": 66381
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.319560058157968e-07,
+      "loss": 0.5544,
+      "step": 66382
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.31928728914018e-07,
+      "loss": 0.6484,
+      "step": 66383
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.319014521409233e-07,
+      "loss": 0.668,
+      "step": 66384
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.318741754965347e-07,
+      "loss": 0.6416,
+      "step": 66385
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.318468989808717e-07,
+      "loss": 0.5928,
+      "step": 66386
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.318196225939559e-07,
+      "loss": 0.6772,
+      "step": 66387
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.317923463358079e-07,
+      "loss": 0.6274,
+      "step": 66388
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.317650702064489e-07,
+      "loss": 0.6953,
+      "step": 66389
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.317377942058996e-07,
+      "loss": 0.7598,
+      "step": 66390
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.31710518334181e-07,
+      "loss": 0.7148,
+      "step": 66391
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.316832425913138e-07,
+      "loss": 0.4584,
+      "step": 66392
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.316559669773192e-07,
+      "loss": 0.5938,
+      "step": 66393
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.316286914922176e-07,
+      "loss": 0.6943,
+      "step": 66394
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.316014161360307e-07,
+      "loss": 0.5215,
+      "step": 66395
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.315741409087785e-07,
+      "loss": 0.7598,
+      "step": 66396
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.315468658104822e-07,
+      "loss": 0.834,
+      "step": 66397
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.315195908411628e-07,
+      "loss": 0.5508,
+      "step": 66398
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.314923160008409e-07,
+      "loss": 0.5664,
+      "step": 66399
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.314650412895379e-07,
+      "loss": 0.5879,
+      "step": 66400
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.314377667072741e-07,
+      "loss": 0.6631,
+      "step": 66401
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.314104922540708e-07,
+      "loss": 0.6462,
+      "step": 66402
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.313832179299486e-07,
+      "loss": 0.6719,
+      "step": 66403
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.313559437349287e-07,
+      "loss": 0.626,
+      "step": 66404
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.313286696690321e-07,
+      "loss": 0.6172,
+      "step": 66405
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.31301395732279e-07,
+      "loss": 0.6035,
+      "step": 66406
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.312741219246904e-07,
+      "loss": 0.8047,
+      "step": 66407
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.312468482462878e-07,
+      "loss": 0.6348,
+      "step": 66408
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.312195746970915e-07,
+      "loss": 0.709,
+      "step": 66409
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.311923012771229e-07,
+      "loss": 0.4673,
+      "step": 66410
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.311650279864024e-07,
+      "loss": 0.6836,
+      "step": 66411
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.311377548249511e-07,
+      "loss": 0.6592,
+      "step": 66412
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.311104817927899e-07,
+      "loss": 0.7109,
+      "step": 66413
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.310832088899396e-07,
+      "loss": 0.7207,
+      "step": 66414
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.310559361164214e-07,
+      "loss": 0.5381,
+      "step": 66415
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.310286634722556e-07,
+      "loss": 0.5063,
+      "step": 66416
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.310013909574634e-07,
+      "loss": 0.6045,
+      "step": 66417
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.309741185720656e-07,
+      "loss": 0.7134,
+      "step": 66418
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.309468463160833e-07,
+      "loss": 0.7598,
+      "step": 66419
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.30919574189537e-07,
+      "loss": 0.4785,
+      "step": 66420
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.308923021924481e-07,
+      "loss": 0.448,
+      "step": 66421
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.308650303248369e-07,
+      "loss": 0.519,
+      "step": 66422
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.308377585867248e-07,
+      "loss": 0.7393,
+      "step": 66423
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.308104869781324e-07,
+      "loss": 0.5415,
+      "step": 66424
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.30783215499081e-07,
+      "loss": 0.5303,
+      "step": 66425
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.307559441495905e-07,
+      "loss": 0.6846,
+      "step": 66426
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.307286729296827e-07,
+      "loss": 0.4192,
+      "step": 66427
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.307014018393778e-07,
+      "loss": 0.6143,
+      "step": 66428
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.306741308786975e-07,
+      "loss": 0.5312,
+      "step": 66429
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.306468600476619e-07,
+      "loss": 0.6826,
+      "step": 66430
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.306195893462925e-07,
+      "loss": 0.6367,
+      "step": 66431
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.305923187746096e-07,
+      "loss": 0.5889,
+      "step": 66432
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.305650483326346e-07,
+      "loss": 0.689,
+      "step": 66433
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.30537778020388e-07,
+      "loss": 0.7246,
+      "step": 66434
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.305105078378913e-07,
+      "loss": 0.7607,
+      "step": 66435
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.304832377851645e-07,
+      "loss": 0.7002,
+      "step": 66436
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.304559678622288e-07,
+      "loss": 0.7349,
+      "step": 66437
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.304286980691054e-07,
+      "loss": 0.5635,
+      "step": 66438
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.304014284058148e-07,
+      "loss": 0.7656,
+      "step": 66439
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.303741588723781e-07,
+      "loss": 0.7656,
+      "step": 66440
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.303468894688162e-07,
+      "loss": 0.585,
+      "step": 66441
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.303196201951495e-07,
+      "loss": 0.6655,
+      "step": 66442
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.302923510513996e-07,
+      "loss": 0.5771,
+      "step": 66443
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.30265082037587e-07,
+      "loss": 0.7119,
+      "step": 66444
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.302378131537327e-07,
+      "loss": 0.6582,
+      "step": 66445
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.302105443998575e-07,
+      "loss": 0.4797,
+      "step": 66446
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.301832757759821e-07,
+      "loss": 0.5923,
+      "step": 66447
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.301560072821276e-07,
+      "loss": 0.5674,
+      "step": 66448
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.301287389183149e-07,
+      "loss": 0.6797,
+      "step": 66449
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.301014706845649e-07,
+      "loss": 0.6182,
+      "step": 66450
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.30074202580898e-07,
+      "loss": 0.4524,
+      "step": 66451
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.300469346073357e-07,
+      "loss": 0.5962,
+      "step": 66452
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.300196667638986e-07,
+      "loss": 0.665,
+      "step": 66453
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.299923990506079e-07,
+      "loss": 0.7139,
+      "step": 66454
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.299651314674843e-07,
+      "loss": 0.7793,
+      "step": 66455
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.299378640145481e-07,
+      "loss": 0.6309,
+      "step": 66456
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.299105966918209e-07,
+      "loss": 0.7725,
+      "step": 66457
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.298833294993235e-07,
+      "loss": 0.8096,
+      "step": 66458
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.298560624370762e-07,
+      "loss": 0.6924,
+      "step": 66459
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.298287955051005e-07,
+      "loss": 0.5662,
+      "step": 66460
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.29801528703417e-07,
+      "loss": 0.5518,
+      "step": 66461
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.297742620320468e-07,
+      "loss": 0.5962,
+      "step": 66462
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.297469954910104e-07,
+      "loss": 0.5391,
+      "step": 66463
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.297197290803291e-07,
+      "loss": 0.6069,
+      "step": 66464
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.296924628000237e-07,
+      "loss": 0.5216,
+      "step": 66465
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.296651966501147e-07,
+      "loss": 0.729,
+      "step": 66466
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.296379306306232e-07,
+      "loss": 0.4922,
+      "step": 66467
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.296106647415701e-07,
+      "loss": 0.626,
+      "step": 66468
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.295833989829764e-07,
+      "loss": 0.625,
+      "step": 66469
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.295561333548627e-07,
+      "loss": 0.5503,
+      "step": 66470
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.295288678572502e-07,
+      "loss": 0.7422,
+      "step": 66471
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.295016024901595e-07,
+      "loss": 0.522,
+      "step": 66472
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.294743372536118e-07,
+      "loss": 0.6387,
+      "step": 66473
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.294470721476273e-07,
+      "loss": 0.4585,
+      "step": 66474
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.294198071722282e-07,
+      "loss": 0.522,
+      "step": 66475
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.293925423274337e-07,
+      "loss": 0.5989,
+      "step": 66476
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.293652776132658e-07,
+      "loss": 0.7544,
+      "step": 66477
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.293380130297449e-07,
+      "loss": 0.7021,
+      "step": 66478
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.293107485768922e-07,
+      "loss": 0.488,
+      "step": 66479
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.292834842547282e-07,
+      "loss": 0.7559,
+      "step": 66480
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.292562200632742e-07,
+      "loss": 0.7275,
+      "step": 66481
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.292289560025506e-07,
+      "loss": 0.6504,
+      "step": 66482
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.292016920725788e-07,
+      "loss": 0.7041,
+      "step": 66483
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.291744282733792e-07,
+      "loss": 0.5798,
+      "step": 66484
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.291471646049733e-07,
+      "loss": 0.5444,
+      "step": 66485
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.291199010673813e-07,
+      "loss": 0.6514,
+      "step": 66486
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.290926376606242e-07,
+      "loss": 0.4392,
+      "step": 66487
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.290653743847231e-07,
+      "loss": 0.5269,
+      "step": 66488
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.290381112396987e-07,
+      "loss": 0.6562,
+      "step": 66489
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.290108482255722e-07,
+      "loss": 0.6709,
+      "step": 66490
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.289835853423639e-07,
+      "loss": 0.4834,
+      "step": 66491
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.289563225900953e-07,
+      "loss": 0.667,
+      "step": 66492
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.289290599687867e-07,
+      "loss": 0.7275,
+      "step": 66493
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.289017974784599e-07,
+      "loss": 0.7402,
+      "step": 66494
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.288745351191345e-07,
+      "loss": 0.5195,
+      "step": 66495
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.288472728908322e-07,
+      "loss": 0.7114,
+      "step": 66496
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.288200107935736e-07,
+      "loss": 0.5796,
+      "step": 66497
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.287927488273797e-07,
+      "loss": 0.5781,
+      "step": 66498
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.287654869922711e-07,
+      "loss": 0.6318,
+      "step": 66499
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.287382252882692e-07,
+      "loss": 0.7935,
+      "step": 66500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.287109637153942e-07,
+      "loss": 0.3816,
+      "step": 66501
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.286837022736678e-07,
+      "loss": 0.6562,
+      "step": 66502
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.286564409631101e-07,
+      "loss": 0.5605,
+      "step": 66503
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.286291797837426e-07,
+      "loss": 0.5269,
+      "step": 66504
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.286019187355857e-07,
+      "loss": 0.6279,
+      "step": 66505
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.285746578186602e-07,
+      "loss": 0.6162,
+      "step": 66506
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.285473970329875e-07,
+      "loss": 0.603,
+      "step": 66507
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.285201363785879e-07,
+      "loss": 0.7646,
+      "step": 66508
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.284928758554828e-07,
+      "loss": 0.563,
+      "step": 66509
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.284656154636925e-07,
+      "loss": 0.6841,
+      "step": 66510
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.284383552032386e-07,
+      "loss": 0.7178,
+      "step": 66511
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.284110950741411e-07,
+      "loss": 0.4133,
+      "step": 66512
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.283838350764217e-07,
+      "loss": 0.7461,
+      "step": 66513
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.283565752101011e-07,
+      "loss": 0.8242,
+      "step": 66514
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.283293154751998e-07,
+      "loss": 0.6777,
+      "step": 66515
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.283020558717386e-07,
+      "loss": 0.6211,
+      "step": 66516
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.282747963997387e-07,
+      "loss": 0.6533,
+      "step": 66517
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.282475370592209e-07,
+      "loss": 0.479,
+      "step": 66518
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.282202778502062e-07,
+      "loss": 0.5439,
+      "step": 66519
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.281930187727151e-07,
+      "loss": 0.5776,
+      "step": 66520
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.281657598267689e-07,
+      "loss": 0.5366,
+      "step": 66521
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 8.281385010123881e-07,
+      "loss": 0.6738,
+      "step": 66522
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.281112423295939e-07,
+      "loss": 0.541,
+      "step": 66523
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.280839837784074e-07,
+      "loss": 0.6006,
+      "step": 66524
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.280567253588485e-07,
+      "loss": 0.6113,
+      "step": 66525
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.280294670709389e-07,
+      "loss": 0.583,
+      "step": 66526
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.280022089146989e-07,
+      "loss": 0.5581,
+      "step": 66527
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.279749508901501e-07,
+      "loss": 0.582,
+      "step": 66528
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.279476929973128e-07,
+      "loss": 0.6934,
+      "step": 66529
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.279204352362081e-07,
+      "loss": 0.4775,
+      "step": 66530
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.278931776068569e-07,
+      "loss": 0.6709,
+      "step": 66531
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.278659201092797e-07,
+      "loss": 0.5078,
+      "step": 66532
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.278386627434979e-07,
+      "loss": 0.6162,
+      "step": 66533
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.278114055095324e-07,
+      "loss": 0.7607,
+      "step": 66534
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.277841484074033e-07,
+      "loss": 0.5205,
+      "step": 66535
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.277568914371322e-07,
+      "loss": 0.666,
+      "step": 66536
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.277296345987395e-07,
+      "loss": 0.8135,
+      "step": 66537
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.277023778922464e-07,
+      "loss": 0.6646,
+      "step": 66538
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.276751213176736e-07,
+      "loss": 0.4399,
+      "step": 66539
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.276478648750421e-07,
+      "loss": 0.4668,
+      "step": 66540
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.276206085643727e-07,
+      "loss": 0.7637,
+      "step": 66541
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.275933523856862e-07,
+      "loss": 0.7617,
+      "step": 66542
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.275660963390035e-07,
+      "loss": 0.6863,
+      "step": 66543
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.275388404243459e-07,
+      "loss": 0.7061,
+      "step": 66544
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.275115846417338e-07,
+      "loss": 0.7988,
+      "step": 66545
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.274843289911879e-07,
+      "loss": 0.6421,
+      "step": 66546
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.274570734727295e-07,
+      "loss": 0.7441,
+      "step": 66547
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.274298180863792e-07,
+      "loss": 0.6084,
+      "step": 66548
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.274025628321578e-07,
+      "loss": 0.6113,
+      "step": 66549
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.273753077100865e-07,
+      "loss": 0.5215,
+      "step": 66550
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.273480527201858e-07,
+      "loss": 0.709,
+      "step": 66551
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.27320797862477e-07,
+      "loss": 0.7383,
+      "step": 66552
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.272935431369806e-07,
+      "loss": 0.7168,
+      "step": 66553
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.272662885437179e-07,
+      "loss": 0.7568,
+      "step": 66554
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.272390340827091e-07,
+      "loss": 0.7119,
+      "step": 66555
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.272117797539755e-07,
+      "loss": 0.7026,
+      "step": 66556
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.271845255575379e-07,
+      "loss": 0.6758,
+      "step": 66557
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.27157271493417e-07,
+      "loss": 0.6641,
+      "step": 66558
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.271300175616341e-07,
+      "loss": 0.4893,
+      "step": 66559
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.271027637622095e-07,
+      "loss": 0.7344,
+      "step": 66560
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.270755100951648e-07,
+      "loss": 0.5344,
+      "step": 66561
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.270482565605199e-07,
+      "loss": 0.6572,
+      "step": 66562
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.270210031582966e-07,
+      "loss": 0.624,
+      "step": 66563
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.269937498885156e-07,
+      "loss": 0.6357,
+      "step": 66564
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.269664967511972e-07,
+      "loss": 0.75,
+      "step": 66565
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.269392437463625e-07,
+      "loss": 0.5015,
+      "step": 66566
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.269119908740325e-07,
+      "loss": 0.5134,
+      "step": 66567
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26884738134228e-07,
+      "loss": 0.5486,
+      "step": 66568
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.2685748552697e-07,
+      "loss": 0.7236,
+      "step": 66569
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.268302330522791e-07,
+      "loss": 0.6216,
+      "step": 66570
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.268029807101765e-07,
+      "loss": 0.5798,
+      "step": 66571
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.267757285006828e-07,
+      "loss": 0.7036,
+      "step": 66572
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26748476423819e-07,
+      "loss": 0.5229,
+      "step": 66573
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.267212244796062e-07,
+      "loss": 0.7134,
+      "step": 66574
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.266939726680646e-07,
+      "loss": 0.5781,
+      "step": 66575
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.266667209892157e-07,
+      "loss": 0.6748,
+      "step": 66576
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.266394694430798e-07,
+      "loss": 0.6724,
+      "step": 66577
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.266122180296784e-07,
+      "loss": 0.7158,
+      "step": 66578
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.265849667490319e-07,
+      "loss": 0.6975,
+      "step": 66579
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.265577156011614e-07,
+      "loss": 0.6836,
+      "step": 66580
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.265304645860874e-07,
+      "loss": 0.6187,
+      "step": 66581
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.265032137038315e-07,
+      "loss": 0.584,
+      "step": 66582
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.264759629544136e-07,
+      "loss": 0.5518,
+      "step": 66583
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26448712337856e-07,
+      "loss": 0.4969,
+      "step": 66584
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.264214618541778e-07,
+      "loss": 0.6348,
+      "step": 66585
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26394211503401e-07,
+      "loss": 0.7974,
+      "step": 66586
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26366961285546e-07,
+      "loss": 0.5469,
+      "step": 66587
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26339711200634e-07,
+      "loss": 0.7725,
+      "step": 66588
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.263124612486855e-07,
+      "loss": 0.6924,
+      "step": 66589
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.262852114297217e-07,
+      "loss": 0.7285,
+      "step": 66590
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.262579617437633e-07,
+      "loss": 0.6294,
+      "step": 66591
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.262307121908312e-07,
+      "loss": 0.6104,
+      "step": 66592
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.262034627709461e-07,
+      "loss": 0.5181,
+      "step": 66593
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.261762134841294e-07,
+      "loss": 0.4778,
+      "step": 66594
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.261489643304015e-07,
+      "loss": 0.5908,
+      "step": 66595
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.26121715309783e-07,
+      "loss": 0.646,
+      "step": 66596
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.260944664222955e-07,
+      "loss": 0.7471,
+      "step": 66597
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.260672176679591e-07,
+      "loss": 0.6362,
+      "step": 66598
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.260399690467953e-07,
+      "loss": 0.6685,
+      "step": 66599
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.260127205588245e-07,
+      "loss": 0.7036,
+      "step": 66600
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.259854722040679e-07,
+      "loss": 0.4072,
+      "step": 66601
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.25958223982546e-07,
+      "loss": 0.5415,
+      "step": 66602
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.259309758942802e-07,
+      "loss": 0.3787,
+      "step": 66603
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.259037279392912e-07,
+      "loss": 0.5298,
+      "step": 66604
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.258764801175994e-07,
+      "loss": 0.7158,
+      "step": 66605
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.25849232429226e-07,
+      "loss": 0.7261,
+      "step": 66606
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.258219848741918e-07,
+      "loss": 0.5869,
+      "step": 66607
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.257947374525177e-07,
+      "loss": 0.6973,
+      "step": 66608
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.257674901642247e-07,
+      "loss": 0.7041,
+      "step": 66609
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.257402430093332e-07,
+      "loss": 0.6133,
+      "step": 66610
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.257129959878647e-07,
+      "loss": 0.6406,
+      "step": 66611
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.256857490998394e-07,
+      "loss": 0.366,
+      "step": 66612
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.256585023452789e-07,
+      "loss": 0.6006,
+      "step": 66613
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.256312557242039e-07,
+      "loss": 0.4707,
+      "step": 66614
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.256040092366344e-07,
+      "loss": 0.4553,
+      "step": 66615
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.255767628825923e-07,
+      "loss": 0.5859,
+      "step": 66616
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.255495166620977e-07,
+      "loss": 0.8188,
+      "step": 66617
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.255222705751721e-07,
+      "loss": 0.5381,
+      "step": 66618
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.254950246218358e-07,
+      "loss": 0.4536,
+      "step": 66619
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.254677788021103e-07,
+      "loss": 0.5439,
+      "step": 66620
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.254405331160159e-07,
+      "loss": 0.5396,
+      "step": 66621
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.254132875635736e-07,
+      "loss": 0.6279,
+      "step": 66622
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.253860421448045e-07,
+      "loss": 0.6914,
+      "step": 66623
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.253587968597294e-07,
+      "loss": 0.7085,
+      "step": 66624
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.253315517083686e-07,
+      "loss": 0.8008,
+      "step": 66625
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.253043066907438e-07,
+      "loss": 0.7219,
+      "step": 66626
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.252770618068751e-07,
+      "loss": 0.6328,
+      "step": 66627
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.25249817056784e-07,
+      "loss": 0.6553,
+      "step": 66628
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.252225724404908e-07,
+      "loss": 0.6802,
+      "step": 66629
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.251953279580169e-07,
+      "loss": 0.5757,
+      "step": 66630
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.251680836093826e-07,
+      "loss": 0.5845,
+      "step": 66631
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.251408393946093e-07,
+      "loss": 0.6719,
+      "step": 66632
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.251135953137175e-07,
+      "loss": 0.5365,
+      "step": 66633
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.250863513667285e-07,
+      "loss": 0.6514,
+      "step": 66634
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.250591075536628e-07,
+      "loss": 0.6523,
+      "step": 66635
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.25031863874541e-07,
+      "loss": 0.7061,
+      "step": 66636
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.250046203293843e-07,
+      "loss": 0.4697,
+      "step": 66637
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.249773769182137e-07,
+      "loss": 0.6494,
+      "step": 66638
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.249501336410496e-07,
+      "loss": 0.7119,
+      "step": 66639
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.249228904979133e-07,
+      "loss": 0.5352,
+      "step": 66640
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.248956474888254e-07,
+      "loss": 0.7998,
+      "step": 66641
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.248684046138071e-07,
+      "loss": 0.3916,
+      "step": 66642
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.248411618728785e-07,
+      "loss": 0.6357,
+      "step": 66643
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.248139192660617e-07,
+      "loss": 0.71,
+      "step": 66644
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.247866767933765e-07,
+      "loss": 0.5171,
+      "step": 66645
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.247594344548438e-07,
+      "loss": 0.7236,
+      "step": 66646
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.24732192250485e-07,
+      "loss": 0.627,
+      "step": 66647
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.247049501803206e-07,
+      "loss": 0.6299,
+      "step": 66648
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.246777082443717e-07,
+      "loss": 0.6289,
+      "step": 66649
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.246504664426588e-07,
+      "loss": 0.5405,
+      "step": 66650
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.246232247752032e-07,
+      "loss": 0.7432,
+      "step": 66651
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.245959832420253e-07,
+      "loss": 0.7686,
+      "step": 66652
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.245687418431464e-07,
+      "loss": 0.7275,
+      "step": 66653
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.245415005785874e-07,
+      "loss": 0.8096,
+      "step": 66654
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.245142594483686e-07,
+      "loss": 0.6074,
+      "step": 66655
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.24487018452511e-07,
+      "loss": 0.6777,
+      "step": 66656
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.244597775910358e-07,
+      "loss": 0.5847,
+      "step": 66657
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.244325368639635e-07,
+      "loss": 0.71,
+      "step": 66658
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.244052962713154e-07,
+      "loss": 0.7314,
+      "step": 66659
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.243780558131118e-07,
+      "loss": 0.7354,
+      "step": 66660
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.243508154893741e-07,
+      "loss": 0.6992,
+      "step": 66661
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.243235753001228e-07,
+      "loss": 0.7275,
+      "step": 66662
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.242963352453789e-07,
+      "loss": 0.4755,
+      "step": 66663
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.242690953251635e-07,
+      "loss": 0.7808,
+      "step": 66664
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.242418555394967e-07,
+      "loss": 0.7324,
+      "step": 66665
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.242146158884e-07,
+      "loss": 0.6245,
+      "step": 66666
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.241873763718941e-07,
+      "loss": 0.5977,
+      "step": 66667
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.241601369899997e-07,
+      "loss": 0.5747,
+      "step": 66668
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.24132897742738e-07,
+      "loss": 0.5684,
+      "step": 66669
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.241056586301296e-07,
+      "loss": 0.5166,
+      "step": 66670
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.240784196521953e-07,
+      "loss": 0.5947,
+      "step": 66671
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.240511808089561e-07,
+      "loss": 0.5654,
+      "step": 66672
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.240239421004328e-07,
+      "loss": 0.668,
+      "step": 66673
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.239967035266468e-07,
+      "loss": 0.7041,
+      "step": 66674
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.239694650876179e-07,
+      "loss": 0.6797,
+      "step": 66675
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.239422267833677e-07,
+      "loss": 0.6836,
+      "step": 66676
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.239149886139165e-07,
+      "loss": 0.6313,
+      "step": 66677
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.238877505792858e-07,
+      "loss": 0.5322,
+      "step": 66678
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.238605126794958e-07,
+      "loss": 0.6426,
+      "step": 66679
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.238332749145681e-07,
+      "loss": 0.6582,
+      "step": 66680
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.23806037284523e-07,
+      "loss": 0.5493,
+      "step": 66681
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.237787997893815e-07,
+      "loss": 0.5679,
+      "step": 66682
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.237515624291644e-07,
+      "loss": 0.5725,
+      "step": 66683
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.23724325203893e-07,
+      "loss": 0.7227,
+      "step": 66684
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.236970881135876e-07,
+      "loss": 0.7852,
+      "step": 66685
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.23669851158269e-07,
+      "loss": 0.5476,
+      "step": 66686
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.236426143379586e-07,
+      "loss": 0.5879,
+      "step": 66687
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.236153776526766e-07,
+      "loss": 0.5928,
+      "step": 66688
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.235881411024444e-07,
+      "loss": 0.4468,
+      "step": 66689
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.235609046872824e-07,
+      "loss": 0.6772,
+      "step": 66690
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.23533668407212e-07,
+      "loss": 0.603,
+      "step": 66691
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.235064322622535e-07,
+      "loss": 0.6982,
+      "step": 66692
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.234791962524284e-07,
+      "loss": 0.583,
+      "step": 66693
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.234519603777573e-07,
+      "loss": 0.6577,
+      "step": 66694
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.234247246382606e-07,
+      "loss": 0.6323,
+      "step": 66695
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.233974890339593e-07,
+      "loss": 0.6304,
+      "step": 66696
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.233702535648745e-07,
+      "loss": 0.5853,
+      "step": 66697
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.233430182310271e-07,
+      "loss": 0.6348,
+      "step": 66698
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.233157830324378e-07,
+      "loss": 0.6631,
+      "step": 66699
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.232885479691273e-07,
+      "loss": 0.5715,
+      "step": 66700
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.232613130411168e-07,
+      "loss": 0.532,
+      "step": 66701
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.232340782484268e-07,
+      "loss": 0.7266,
+      "step": 66702
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.232068435910786e-07,
+      "loss": 0.71,
+      "step": 66703
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.23179609069093e-07,
+      "loss": 0.707,
+      "step": 66704
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.231523746824902e-07,
+      "loss": 0.6699,
+      "step": 66705
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.231251404312916e-07,
+      "loss": 0.7466,
+      "step": 66706
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.230979063155179e-07,
+      "loss": 0.4719,
+      "step": 66707
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.230706723351901e-07,
+      "loss": 0.5874,
+      "step": 66708
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.230434384903289e-07,
+      "loss": 0.5615,
+      "step": 66709
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.230162047809553e-07,
+      "loss": 0.5078,
+      "step": 66710
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.2298897120709e-07,
+      "loss": 0.666,
+      "step": 66711
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.229617377687538e-07,
+      "loss": 0.4598,
+      "step": 66712
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.229345044659678e-07,
+      "loss": 0.5581,
+      "step": 66713
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.22907271298753e-07,
+      "loss": 0.3634,
+      "step": 66714
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.228800382671294e-07,
+      "loss": 0.707,
+      "step": 66715
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.228528053711187e-07,
+      "loss": 0.71,
+      "step": 66716
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.228255726107412e-07,
+      "loss": 0.6416,
+      "step": 66717
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.227983399860183e-07,
+      "loss": 0.7188,
+      "step": 66718
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.227711074969703e-07,
+      "loss": 0.6123,
+      "step": 66719
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.227438751436186e-07,
+      "loss": 0.5505,
+      "step": 66720
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.227166429259835e-07,
+      "loss": 0.5903,
+      "step": 66721
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.226894108440863e-07,
+      "loss": 0.6113,
+      "step": 66722
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.226621788979476e-07,
+      "loss": 0.4746,
+      "step": 66723
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.226349470875885e-07,
+      "loss": 0.8701,
+      "step": 66724
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.226077154130296e-07,
+      "loss": 0.6455,
+      "step": 66725
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.225804838742915e-07,
+      "loss": 0.7197,
+      "step": 66726
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.225532524713958e-07,
+      "loss": 0.6006,
+      "step": 66727
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.225260212043627e-07,
+      "loss": 0.7148,
+      "step": 66728
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.224987900732131e-07,
+      "loss": 0.5728,
+      "step": 66729
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.224715590779682e-07,
+      "loss": 0.6924,
+      "step": 66730
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.224443282186486e-07,
+      "loss": 0.5381,
+      "step": 66731
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.224170974952753e-07,
+      "loss": 0.6729,
+      "step": 66732
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.223898669078688e-07,
+      "loss": 0.7998,
+      "step": 66733
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.223626364564507e-07,
+      "loss": 0.5562,
+      "step": 66734
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.223354061410412e-07,
+      "loss": 0.499,
+      "step": 66735
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.22308175961661e-07,
+      "loss": 0.7344,
+      "step": 66736
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.222809459183316e-07,
+      "loss": 0.6338,
+      "step": 66737
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.222537160110732e-07,
+      "loss": 0.7178,
+      "step": 66738
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.222264862399072e-07,
+      "loss": 0.6689,
+      "step": 66739
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.22199256604854e-07,
+      "loss": 0.7822,
+      "step": 66740
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.221720271059349e-07,
+      "loss": 0.4966,
+      "step": 66741
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.221447977431702e-07,
+      "loss": 0.4877,
+      "step": 66742
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.221175685165813e-07,
+      "loss": 0.469,
+      "step": 66743
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.220903394261891e-07,
+      "loss": 0.8154,
+      "step": 66744
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.220631104720138e-07,
+      "loss": 0.6616,
+      "step": 66745
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.220358816540765e-07,
+      "loss": 0.6021,
+      "step": 66746
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.220086529723982e-07,
+      "loss": 0.8105,
+      "step": 66747
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.219814244269996e-07,
+      "loss": 0.7095,
+      "step": 66748
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.219541960179018e-07,
+      "loss": 0.6128,
+      "step": 66749
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.219269677451253e-07,
+      "loss": 0.6328,
+      "step": 66750
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.218997396086913e-07,
+      "loss": 0.6592,
+      "step": 66751
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.218725116086203e-07,
+      "loss": 0.5786,
+      "step": 66752
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.218452837449337e-07,
+      "loss": 0.6973,
+      "step": 66753
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.218180560176519e-07,
+      "loss": 0.6904,
+      "step": 66754
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.217908284267955e-07,
+      "loss": 0.4648,
+      "step": 66755
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.21763600972386e-07,
+      "loss": 0.6982,
+      "step": 66756
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.217363736544436e-07,
+      "loss": 0.7109,
+      "step": 66757
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.217091464729897e-07,
+      "loss": 0.7822,
+      "step": 66758
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.216819194280447e-07,
+      "loss": 0.7031,
+      "step": 66759
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.216546925196298e-07,
+      "loss": 0.7383,
+      "step": 66760
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.216274657477656e-07,
+      "loss": 0.5874,
+      "step": 66761
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.216002391124731e-07,
+      "loss": 0.7007,
+      "step": 66762
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.21573012613773e-07,
+      "loss": 0.7129,
+      "step": 66763
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.215457862516869e-07,
+      "loss": 0.5435,
+      "step": 66764
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.215185600262343e-07,
+      "loss": 0.6572,
+      "step": 66765
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.21491333937437e-07,
+      "loss": 0.5276,
+      "step": 66766
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.214641079853153e-07,
+      "loss": 0.5742,
+      "step": 66767
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.214368821698905e-07,
+      "loss": 0.5078,
+      "step": 66768
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.214096564911831e-07,
+      "loss": 0.7695,
+      "step": 66769
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.213824309492143e-07,
+      "loss": 0.4634,
+      "step": 66770
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.213552055440046e-07,
+      "loss": 0.7451,
+      "step": 66771
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.213279802755752e-07,
+      "loss": 0.7412,
+      "step": 66772
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.213007551439465e-07,
+      "loss": 0.6191,
+      "step": 66773
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.212735301491401e-07,
+      "loss": 0.6514,
+      "step": 66774
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.212463052911761e-07,
+      "loss": 0.5518,
+      "step": 66775
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.212190805700753e-07,
+      "loss": 0.6685,
+      "step": 66776
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.211918559858592e-07,
+      "loss": 0.667,
+      "step": 66777
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.211646315385479e-07,
+      "loss": 0.6689,
+      "step": 66778
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.21137407228163e-07,
+      "loss": 0.4761,
+      "step": 66779
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.211101830547245e-07,
+      "loss": 0.5122,
+      "step": 66780
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.210829590182541e-07,
+      "loss": 0.5859,
+      "step": 66781
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.21055735118772e-07,
+      "loss": 0.6602,
+      "step": 66782
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.210285113562995e-07,
+      "loss": 0.5464,
+      "step": 66783
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.210012877308574e-07,
+      "loss": 0.6587,
+      "step": 66784
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.209740642424662e-07,
+      "loss": 0.6777,
+      "step": 66785
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.209468408911469e-07,
+      "loss": 0.594,
+      "step": 66786
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.209196176769204e-07,
+      "loss": 0.6953,
+      "step": 66787
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.208923945998073e-07,
+      "loss": 0.6001,
+      "step": 66788
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.20865171659829e-07,
+      "loss": 0.6504,
+      "step": 66789
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.208379488570057e-07,
+      "loss": 0.5636,
+      "step": 66790
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.208107261913586e-07,
+      "loss": 0.7686,
+      "step": 66791
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.207835036629084e-07,
+      "loss": 0.6118,
+      "step": 66792
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.207562812716766e-07,
+      "loss": 0.645,
+      "step": 66793
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.207290590176831e-07,
+      "loss": 0.5962,
+      "step": 66794
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.207018369009488e-07,
+      "loss": 0.6396,
+      "step": 66795
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.206746149214954e-07,
+      "loss": 0.7129,
+      "step": 66796
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.206473930793427e-07,
+      "loss": 0.6167,
+      "step": 66797
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.206201713745125e-07,
+      "loss": 0.6934,
+      "step": 66798
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.205929498070248e-07,
+      "loss": 0.6001,
+      "step": 66799
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.20565728376901e-07,
+      "loss": 0.5522,
+      "step": 66800
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.205385070841619e-07,
+      "loss": 0.7451,
+      "step": 66801
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.205112859288279e-07,
+      "loss": 0.3633,
+      "step": 66802
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.204840649109206e-07,
+      "loss": 0.6943,
+      "step": 66803
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.204568440304602e-07,
+      "loss": 0.6396,
+      "step": 66804
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.204296232874675e-07,
+      "loss": 0.7578,
+      "step": 66805
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.20402402681964e-07,
+      "loss": 0.7041,
+      "step": 66806
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.203751822139696e-07,
+      "loss": 0.6597,
+      "step": 66807
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.203479618835061e-07,
+      "loss": 0.792,
+      "step": 66808
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.203207416905935e-07,
+      "loss": 0.6479,
+      "step": 66809
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.202935216352534e-07,
+      "loss": 0.5869,
+      "step": 66810
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.202663017175061e-07,
+      "loss": 0.5635,
+      "step": 66811
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.202390819373728e-07,
+      "loss": 0.5381,
+      "step": 66812
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.202118622948743e-07,
+      "loss": 0.6992,
+      "step": 66813
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.20184642790031e-07,
+      "loss": 0.8086,
+      "step": 66814
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.201574234228642e-07,
+      "loss": 0.6567,
+      "step": 66815
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.201302041933944e-07,
+      "loss": 0.7764,
+      "step": 66816
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.201029851016428e-07,
+      "loss": 0.4706,
+      "step": 66817
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.200757661476302e-07,
+      "loss": 0.6421,
+      "step": 66818
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.20048547331377e-07,
+      "loss": 0.5469,
+      "step": 66819
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.200213286529045e-07,
+      "loss": 0.792,
+      "step": 66820
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.199941101122332e-07,
+      "loss": 0.5647,
+      "step": 66821
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.199668917093845e-07,
+      "loss": 0.6602,
+      "step": 66822
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.19939673444379e-07,
+      "loss": 0.542,
+      "step": 66823
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.19912455317237e-07,
+      "loss": 0.7876,
+      "step": 66824
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.198852373279798e-07,
+      "loss": 0.6055,
+      "step": 66825
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.198580194766282e-07,
+      "loss": 0.6382,
+      "step": 66826
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.198308017632032e-07,
+      "loss": 0.6216,
+      "step": 66827
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.198035841877252e-07,
+      "loss": 0.749,
+      "step": 66828
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.197763667502156e-07,
+      "loss": 0.4946,
+      "step": 66829
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.197491494506946e-07,
+      "loss": 0.7461,
+      "step": 66830
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.197219322891837e-07,
+      "loss": 0.8701,
+      "step": 66831
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.196947152657032e-07,
+      "loss": 0.7314,
+      "step": 66832
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.196674983802746e-07,
+      "loss": 0.6895,
+      "step": 66833
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.196402816329181e-07,
+      "loss": 0.6357,
+      "step": 66834
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.196130650236547e-07,
+      "loss": 0.4448,
+      "step": 66835
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.195858485525052e-07,
+      "loss": 0.7188,
+      "step": 66836
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.195586322194905e-07,
+      "loss": 0.6436,
+      "step": 66837
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.195314160246314e-07,
+      "loss": 0.6362,
+      "step": 66838
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.19504199967949e-07,
+      "loss": 0.6875,
+      "step": 66839
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.194769840494636e-07,
+      "loss": 0.5757,
+      "step": 66840
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.194497682691968e-07,
+      "loss": 0.6255,
+      "step": 66841
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.194225526271686e-07,
+      "loss": 0.6172,
+      "step": 66842
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.193953371234008e-07,
+      "loss": 0.8545,
+      "step": 66843
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.193681217579134e-07,
+      "loss": 0.7012,
+      "step": 66844
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.193409065307273e-07,
+      "loss": 0.6621,
+      "step": 66845
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.193136914418639e-07,
+      "loss": 0.6309,
+      "step": 66846
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.192864764913434e-07,
+      "loss": 0.6179,
+      "step": 66847
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.192592616791871e-07,
+      "loss": 0.6113,
+      "step": 66848
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.192320470054155e-07,
+      "loss": 0.6509,
+      "step": 66849
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.192048324700499e-07,
+      "loss": 0.5776,
+      "step": 66850
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.191776180731104e-07,
+      "loss": 0.6123,
+      "step": 66851
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.191504038146188e-07,
+      "loss": 0.71,
+      "step": 66852
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.191231896945955e-07,
+      "loss": 0.7549,
+      "step": 66853
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.19095975713061e-07,
+      "loss": 0.7451,
+      "step": 66854
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.190687618700362e-07,
+      "loss": 0.6846,
+      "step": 66855
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.190415481655424e-07,
+      "loss": 0.6606,
+      "step": 66856
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.190143345995999e-07,
+      "loss": 0.5161,
+      "step": 66857
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.189871211722301e-07,
+      "loss": 0.7041,
+      "step": 66858
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.189599078834533e-07,
+      "loss": 0.7686,
+      "step": 66859
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.189326947332906e-07,
+      "loss": 0.6655,
+      "step": 66860
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.189054817217628e-07,
+      "loss": 0.5386,
+      "step": 66861
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.188782688488909e-07,
+      "loss": 0.6777,
+      "step": 66862
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.188510561146957e-07,
+      "loss": 0.5889,
+      "step": 66863
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.188238435191976e-07,
+      "loss": 0.7036,
+      "step": 66864
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.18796631062418e-07,
+      "loss": 0.5801,
+      "step": 66865
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.187694187443772e-07,
+      "loss": 0.6865,
+      "step": 66866
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.187422065650966e-07,
+      "loss": 0.75,
+      "step": 66867
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.187149945245965e-07,
+      "loss": 0.6636,
+      "step": 66868
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.186877826228982e-07,
+      "loss": 0.6953,
+      "step": 66869
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.186605708600221e-07,
+      "loss": 0.6709,
+      "step": 66870
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.186333592359895e-07,
+      "loss": 0.5591,
+      "step": 66871
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.186061477508209e-07,
+      "loss": 0.6816,
+      "step": 66872
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.185789364045377e-07,
+      "loss": 0.7373,
+      "step": 66873
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.185517251971596e-07,
+      "loss": 0.52,
+      "step": 66874
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.185245141287085e-07,
+      "loss": 0.6875,
+      "step": 66875
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.184973031992044e-07,
+      "loss": 0.7871,
+      "step": 66876
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.18470092408669e-07,
+      "loss": 0.6108,
+      "step": 66877
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.184428817571223e-07,
+      "loss": 0.7012,
+      "step": 66878
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.184156712445859e-07,
+      "loss": 0.6851,
+      "step": 66879
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.1838846087108e-07,
+      "loss": 0.6426,
+      "step": 66880
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.183612506366258e-07,
+      "loss": 0.709,
+      "step": 66881
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.18334040541244e-07,
+      "loss": 0.7803,
+      "step": 66882
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.183068305849558e-07,
+      "loss": 0.6021,
+      "step": 66883
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.182796207677815e-07,
+      "loss": 0.7754,
+      "step": 66884
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.182524110897421e-07,
+      "loss": 0.7734,
+      "step": 66885
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.182252015508585e-07,
+      "loss": 0.7158,
+      "step": 66886
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.181979921511514e-07,
+      "loss": 0.6064,
+      "step": 66887
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.181707828906419e-07,
+      "loss": 0.6279,
+      "step": 66888
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.181435737693504e-07,
+      "loss": 0.7734,
+      "step": 66889
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.181163647872982e-07,
+      "loss": 0.5959,
+      "step": 66890
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.18089155944506e-07,
+      "loss": 0.6538,
+      "step": 66891
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.180619472409944e-07,
+      "loss": 0.7266,
+      "step": 66892
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.180347386767848e-07,
+      "loss": 0.5791,
+      "step": 66893
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.180075302518974e-07,
+      "loss": 0.6318,
+      "step": 66894
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.17980321966353e-07,
+      "loss": 0.4624,
+      "step": 66895
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.17953113820173e-07,
+      "loss": 0.7148,
+      "step": 66896
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.179259058133776e-07,
+      "loss": 0.7002,
+      "step": 66897
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.178986979459883e-07,
+      "loss": 0.6514,
+      "step": 66898
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.178714902180254e-07,
+      "loss": 0.8896,
+      "step": 66899
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.1784428262951e-07,
+      "loss": 0.7251,
+      "step": 66900
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.178170751804627e-07,
+      "loss": 0.48,
+      "step": 66901
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.177898678709048e-07,
+      "loss": 0.7285,
+      "step": 66902
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.17762660700857e-07,
+      "loss": 0.5581,
+      "step": 66903
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.177354536703393e-07,
+      "loss": 0.71,
+      "step": 66904
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.177082467793736e-07,
+      "loss": 0.6519,
+      "step": 66905
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.176810400279801e-07,
+      "loss": 0.5188,
+      "step": 66906
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.176538334161801e-07,
+      "loss": 0.6338,
+      "step": 66907
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.176266269439941e-07,
+      "loss": 0.5786,
+      "step": 66908
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.175994206114428e-07,
+      "loss": 0.7295,
+      "step": 66909
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.175722144185475e-07,
+      "loss": 0.7656,
+      "step": 66910
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.175450083653285e-07,
+      "loss": 0.582,
+      "step": 66911
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 8.175178024518071e-07,
+      "loss": 0.7256,
+      "step": 66912
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.174905966780041e-07,
+      "loss": 0.4478,
+      "step": 66913
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.174633910439398e-07,
+      "loss": 0.7725,
+      "step": 66914
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.174361855496356e-07,
+      "loss": 0.7607,
+      "step": 66915
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.174089801951119e-07,
+      "loss": 0.8369,
+      "step": 66916
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.1738177498039e-07,
+      "loss": 0.7158,
+      "step": 66917
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.173545699054902e-07,
+      "loss": 0.5967,
+      "step": 66918
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.173273649704337e-07,
+      "loss": 0.5933,
+      "step": 66919
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.173001601752413e-07,
+      "loss": 0.4891,
+      "step": 66920
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.172729555199338e-07,
+      "loss": 0.564,
+      "step": 66921
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.172457510045319e-07,
+      "loss": 0.6831,
+      "step": 66922
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.172185466290569e-07,
+      "loss": 0.5967,
+      "step": 66923
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.171913423935289e-07,
+      "loss": 0.6447,
+      "step": 66924
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.171641382979691e-07,
+      "loss": 0.626,
+      "step": 66925
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.171369343423984e-07,
+      "loss": 0.6074,
+      "step": 66926
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.171097305268375e-07,
+      "loss": 0.6084,
+      "step": 66927
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.170825268513071e-07,
+      "loss": 0.7871,
+      "step": 66928
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.170553233158285e-07,
+      "loss": 0.5977,
+      "step": 66929
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.170281199204219e-07,
+      "loss": 0.5239,
+      "step": 66930
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.170009166651087e-07,
+      "loss": 0.6362,
+      "step": 66931
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.169737135499093e-07,
+      "loss": 0.6279,
+      "step": 66932
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.169465105748451e-07,
+      "loss": 0.7061,
+      "step": 66933
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.169193077399363e-07,
+      "loss": 0.6748,
+      "step": 66934
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.168921050452037e-07,
+      "loss": 0.71,
+      "step": 66935
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.168649024906688e-07,
+      "loss": 0.7656,
+      "step": 66936
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.168377000763518e-07,
+      "loss": 0.7495,
+      "step": 66937
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.168104978022738e-07,
+      "loss": 0.415,
+      "step": 66938
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.167832956684555e-07,
+      "loss": 0.7344,
+      "step": 66939
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.167560936749179e-07,
+      "loss": 0.4961,
+      "step": 66940
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.167288918216816e-07,
+      "loss": 0.8472,
+      "step": 66941
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.167016901087678e-07,
+      "loss": 0.7002,
+      "step": 66942
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.166744885361972e-07,
+      "loss": 0.5854,
+      "step": 66943
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.166472871039904e-07,
+      "loss": 0.7461,
+      "step": 66944
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.166200858121681e-07,
+      "loss": 0.7402,
+      "step": 66945
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.165928846607516e-07,
+      "loss": 0.6187,
+      "step": 66946
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.165656836497612e-07,
+      "loss": 0.6641,
+      "step": 66947
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.165384827792183e-07,
+      "loss": 0.6436,
+      "step": 66948
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.165112820491433e-07,
+      "loss": 0.623,
+      "step": 66949
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.164840814595573e-07,
+      "loss": 0.5371,
+      "step": 66950
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.164568810104807e-07,
+      "loss": 0.5464,
+      "step": 66951
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.16429680701935e-07,
+      "loss": 0.4688,
+      "step": 66952
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.164024805339408e-07,
+      "loss": 0.7891,
+      "step": 66953
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.163752805065182e-07,
+      "loss": 0.5732,
+      "step": 66954
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.16348080619689e-07,
+      "loss": 0.3974,
+      "step": 66955
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.163208808734735e-07,
+      "loss": 0.8438,
+      "step": 66956
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.162936812678928e-07,
+      "loss": 0.7188,
+      "step": 66957
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.162664818029673e-07,
+      "loss": 0.533,
+      "step": 66958
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.162392824787184e-07,
+      "loss": 0.5977,
+      "step": 66959
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.162120832951662e-07,
+      "loss": 0.7969,
+      "step": 66960
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.161848842523324e-07,
+      "loss": 0.6685,
+      "step": 66961
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.16157685350237e-07,
+      "loss": 0.7227,
+      "step": 66962
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.16130486588902e-07,
+      "loss": 0.6982,
+      "step": 66963
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.161032879683468e-07,
+      "loss": 0.6885,
+      "step": 66964
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.16076089488593e-07,
+      "loss": 0.48,
+      "step": 66965
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.160488911496611e-07,
+      "loss": 0.6953,
+      "step": 66966
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.160216929515723e-07,
+      "loss": 0.6963,
+      "step": 66967
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.15994494894347e-07,
+      "loss": 0.4879,
+      "step": 66968
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.159672969780065e-07,
+      "loss": 0.7334,
+      "step": 66969
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.159400992025711e-07,
+      "loss": 0.8311,
+      "step": 66970
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.159129015680622e-07,
+      "loss": 0.6057,
+      "step": 66971
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.158857040745001e-07,
+      "loss": 0.3887,
+      "step": 66972
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.158585067219063e-07,
+      "loss": 0.7178,
+      "step": 66973
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.158313095103009e-07,
+      "loss": 0.7305,
+      "step": 66974
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.158041124397048e-07,
+      "loss": 0.4568,
+      "step": 66975
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.157769155101392e-07,
+      "loss": 0.7168,
+      "step": 66976
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.157497187216245e-07,
+      "loss": 0.5933,
+      "step": 66977
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.157225220741821e-07,
+      "loss": 0.6592,
+      "step": 66978
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.156953255678323e-07,
+      "loss": 0.7012,
+      "step": 66979
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.156681292025962e-07,
+      "loss": 0.7656,
+      "step": 66980
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.156409329784946e-07,
+      "loss": 0.6743,
+      "step": 66981
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.156137368955481e-07,
+      "loss": 0.6055,
+      "step": 66982
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.15586540953778e-07,
+      "loss": 0.7324,
+      "step": 66983
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.155593451532046e-07,
+      "loss": 0.6289,
+      "step": 66984
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.155321494938488e-07,
+      "loss": 0.6885,
+      "step": 66985
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.155049539757317e-07,
+      "loss": 0.6558,
+      "step": 66986
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.154777585988737e-07,
+      "loss": 0.6963,
+      "step": 66987
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.154505633632963e-07,
+      "loss": 0.5708,
+      "step": 66988
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.154233682690196e-07,
+      "loss": 0.6284,
+      "step": 66989
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.15396173316065e-07,
+      "loss": 0.4614,
+      "step": 66990
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.153689785044528e-07,
+      "loss": 0.6768,
+      "step": 66991
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.153417838342043e-07,
+      "loss": 0.7744,
+      "step": 66992
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.153145893053403e-07,
+      "loss": 0.6621,
+      "step": 66993
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.15287394917881e-07,
+      "loss": 0.5576,
+      "step": 66994
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.152602006718479e-07,
+      "loss": 0.6406,
+      "step": 66995
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.152330065672613e-07,
+      "loss": 0.6592,
+      "step": 66996
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.152058126041426e-07,
+      "loss": 0.6162,
+      "step": 66997
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.151786187825121e-07,
+      "loss": 0.5449,
+      "step": 66998
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.151514251023908e-07,
+      "loss": 0.627,
+      "step": 66999
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.151242315637998e-07,
+      "loss": 0.5801,
+      "step": 67000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.150970381667595e-07,
+      "loss": 0.6587,
+      "step": 67001
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.150698449112909e-07,
+      "loss": 0.6768,
+      "step": 67002
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.150426517974152e-07,
+      "loss": 0.6714,
+      "step": 67003
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.150154588251523e-07,
+      "loss": 0.7109,
+      "step": 67004
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.149882659945238e-07,
+      "loss": 0.4888,
+      "step": 67005
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.1496107330555e-07,
+      "loss": 0.6089,
+      "step": 67006
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.149338807582525e-07,
+      "loss": 0.5518,
+      "step": 67007
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.149066883526511e-07,
+      "loss": 0.6416,
+      "step": 67008
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.148794960887675e-07,
+      "loss": 0.7129,
+      "step": 67009
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.148523039666218e-07,
+      "loss": 0.7559,
+      "step": 67010
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.148251119862355e-07,
+      "loss": 0.6641,
+      "step": 67011
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.147979201476289e-07,
+      "loss": 0.5396,
+      "step": 67012
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.147707284508234e-07,
+      "loss": 0.6597,
+      "step": 67013
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.147435368958392e-07,
+      "loss": 0.7441,
+      "step": 67014
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.147163454826972e-07,
+      "loss": 0.6611,
+      "step": 67015
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.146891542114186e-07,
+      "loss": 0.4202,
+      "step": 67016
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.146619630820239e-07,
+      "loss": 0.6299,
+      "step": 67017
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.146347720945339e-07,
+      "loss": 0.7334,
+      "step": 67018
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.146075812489696e-07,
+      "loss": 0.7734,
+      "step": 67019
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.145803905453518e-07,
+      "loss": 0.7227,
+      "step": 67020
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.145531999837012e-07,
+      "loss": 0.708,
+      "step": 67021
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.145260095640387e-07,
+      "loss": 0.4767,
+      "step": 67022
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.144988192863854e-07,
+      "loss": 0.5752,
+      "step": 67023
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.144716291507616e-07,
+      "loss": 0.5938,
+      "step": 67024
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.144444391571882e-07,
+      "loss": 0.6321,
+      "step": 67025
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.144172493056863e-07,
+      "loss": 0.6758,
+      "step": 67026
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.143900595962765e-07,
+      "loss": 0.5029,
+      "step": 67027
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.143628700289797e-07,
+      "loss": 0.5747,
+      "step": 67028
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.143356806038167e-07,
+      "loss": 0.5928,
+      "step": 67029
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.143084913208085e-07,
+      "loss": 0.6167,
+      "step": 67030
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.142813021799755e-07,
+      "loss": 0.6592,
+      "step": 67031
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.142541131813389e-07,
+      "loss": 0.7324,
+      "step": 67032
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.142269243249198e-07,
+      "loss": 0.6929,
+      "step": 67033
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.141997356107381e-07,
+      "loss": 0.6494,
+      "step": 67034
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.141725470388152e-07,
+      "loss": 0.572,
+      "step": 67035
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.141453586091719e-07,
+      "loss": 0.6003,
+      "step": 67036
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.141181703218287e-07,
+      "loss": 0.6431,
+      "step": 67037
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.140909821768071e-07,
+      "loss": 0.6323,
+      "step": 67038
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.14063794174127e-07,
+      "loss": 0.668,
+      "step": 67039
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.140366063138101e-07,
+      "loss": 0.4912,
+      "step": 67040
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.140094185958766e-07,
+      "loss": 0.585,
+      "step": 67041
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.139822310203477e-07,
+      "loss": 0.7305,
+      "step": 67042
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.139550435872442e-07,
+      "loss": 0.6367,
+      "step": 67043
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.139278562965865e-07,
+      "loss": 0.6094,
+      "step": 67044
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.139006691483958e-07,
+      "loss": 0.5125,
+      "step": 67045
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.138734821426924e-07,
+      "loss": 0.5737,
+      "step": 67046
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.13846295279498e-07,
+      "loss": 0.4315,
+      "step": 67047
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.138191085588326e-07,
+      "loss": 0.7988,
+      "step": 67048
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.137919219807177e-07,
+      "loss": 0.5742,
+      "step": 67049
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.137647355451733e-07,
+      "loss": 0.6196,
+      "step": 67050
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.137375492522212e-07,
+      "loss": 0.561,
+      "step": 67051
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.137103631018813e-07,
+      "loss": 0.7603,
+      "step": 67052
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.136831770941755e-07,
+      "loss": 0.7041,
+      "step": 67053
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.136559912291232e-07,
+      "loss": 0.5718,
+      "step": 67054
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.136288055067462e-07,
+      "loss": 0.3956,
+      "step": 67055
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.136016199270647e-07,
+      "loss": 0.7715,
+      "step": 67056
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.135744344901003e-07,
+      "loss": 0.5977,
+      "step": 67057
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.135472491958732e-07,
+      "loss": 0.6689,
+      "step": 67058
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.135200640444044e-07,
+      "loss": 0.6777,
+      "step": 67059
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.134928790357147e-07,
+      "loss": 0.6465,
+      "step": 67060
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.13465694169825e-07,
+      "loss": 0.6089,
+      "step": 67061
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.134385094467559e-07,
+      "loss": 0.5898,
+      "step": 67062
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.134113248665287e-07,
+      "loss": 0.605,
+      "step": 67063
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.133841404291638e-07,
+      "loss": 0.5358,
+      "step": 67064
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.133569561346818e-07,
+      "loss": 0.7559,
+      "step": 67065
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.13329771983104e-07,
+      "loss": 0.6797,
+      "step": 67066
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.133025879744507e-07,
+      "loss": 0.8164,
+      "step": 67067
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.132754041087434e-07,
+      "loss": 0.7295,
+      "step": 67068
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.132482203860022e-07,
+      "loss": 0.6143,
+      "step": 67069
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.132210368062485e-07,
+      "loss": 0.6729,
+      "step": 67070
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.131938533695029e-07,
+      "loss": 0.8203,
+      "step": 67071
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.13166670075786e-07,
+      "loss": 0.5679,
+      "step": 67072
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.131394869251191e-07,
+      "loss": 0.5535,
+      "step": 67073
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.131123039175227e-07,
+      "loss": 0.7002,
+      "step": 67074
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.130851210530172e-07,
+      "loss": 0.5103,
+      "step": 67075
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.130579383316241e-07,
+      "loss": 0.5376,
+      "step": 67076
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.130307557533637e-07,
+      "loss": 0.6587,
+      "step": 67077
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.130035733182572e-07,
+      "loss": 0.5244,
+      "step": 67078
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.129763910263252e-07,
+      "loss": 0.5835,
+      "step": 67079
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.129492088775887e-07,
+      "loss": 0.6157,
+      "step": 67080
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.129220268720682e-07,
+      "loss": 0.5898,
+      "step": 67081
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.128948450097853e-07,
+      "loss": 0.4208,
+      "step": 67082
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.128676632907597e-07,
+      "loss": 0.532,
+      "step": 67083
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.128404817150127e-07,
+      "loss": 0.3574,
+      "step": 67084
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.128133002825653e-07,
+      "loss": 0.52,
+      "step": 67085
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.127861189934378e-07,
+      "loss": 0.5186,
+      "step": 67086
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.127589378476518e-07,
+      "loss": 0.6592,
+      "step": 67087
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.127317568452274e-07,
+      "loss": 0.6299,
+      "step": 67088
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.127045759861858e-07,
+      "loss": 0.4712,
+      "step": 67089
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.126773952705477e-07,
+      "loss": 0.5652,
+      "step": 67090
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.126502146983339e-07,
+      "loss": 0.7114,
+      "step": 67091
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.126230342695655e-07,
+      "loss": 0.7051,
+      "step": 67092
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.125958539842628e-07,
+      "loss": 0.7383,
+      "step": 67093
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.125686738424465e-07,
+      "loss": 0.6914,
+      "step": 67094
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.125414938441381e-07,
+      "loss": 0.6523,
+      "step": 67095
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.125143139893579e-07,
+      "loss": 0.7256,
+      "step": 67096
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.124871342781269e-07,
+      "loss": 0.4692,
+      "step": 67097
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.124599547104659e-07,
+      "loss": 0.6392,
+      "step": 67098
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.124327752863957e-07,
+      "loss": 0.6436,
+      "step": 67099
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.124055960059371e-07,
+      "loss": 0.5273,
+      "step": 67100
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.12378416869111e-07,
+      "loss": 0.5059,
+      "step": 67101
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.123512378759382e-07,
+      "loss": 0.6367,
+      "step": 67102
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.123240590264391e-07,
+      "loss": 0.5151,
+      "step": 67103
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.122968803206352e-07,
+      "loss": 0.6592,
+      "step": 67104
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.122697017585466e-07,
+      "loss": 0.5312,
+      "step": 67105
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.122425233401947e-07,
+      "loss": 0.6201,
+      "step": 67106
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.122153450656001e-07,
+      "loss": 0.7061,
+      "step": 67107
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.121881669347832e-07,
+      "loss": 0.5344,
+      "step": 67108
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.121609889477657e-07,
+      "loss": 0.5271,
+      "step": 67109
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.121338111045673e-07,
+      "loss": 0.5928,
+      "step": 67110
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.121066334052099e-07,
+      "loss": 0.4517,
+      "step": 67111
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.12079455849714e-07,
+      "loss": 0.8369,
+      "step": 67112
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.120522784380997e-07,
+      "loss": 0.6338,
+      "step": 67113
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.120251011703884e-07,
+      "loss": 0.5581,
+      "step": 67114
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.119979240466009e-07,
+      "loss": 0.6792,
+      "step": 67115
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.11970747066758e-07,
+      "loss": 0.5381,
+      "step": 67116
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.119435702308803e-07,
+      "loss": 0.666,
+      "step": 67117
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.11916393538989e-07,
+      "loss": 0.6033,
+      "step": 67118
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.118892169911044e-07,
+      "loss": 0.5986,
+      "step": 67119
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.118620405872478e-07,
+      "loss": 0.519,
+      "step": 67120
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.118348643274396e-07,
+      "loss": 0.6484,
+      "step": 67121
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.118076882117013e-07,
+      "loss": 0.6226,
+      "step": 67122
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.117805122400529e-07,
+      "loss": 0.5857,
+      "step": 67123
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.117533364125155e-07,
+      "loss": 0.7031,
+      "step": 67124
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.117261607291097e-07,
+      "loss": 0.5903,
+      "step": 67125
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.116989851898567e-07,
+      "loss": 0.5437,
+      "step": 67126
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.116718097947769e-07,
+      "loss": 0.8486,
+      "step": 67127
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.116446345438916e-07,
+      "loss": 0.5854,
+      "step": 67128
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.116174594372212e-07,
+      "loss": 0.4541,
+      "step": 67129
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.115902844747867e-07,
+      "loss": 0.6865,
+      "step": 67130
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.115631096566088e-07,
+      "loss": 0.5015,
+      "step": 67131
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.115359349827087e-07,
+      "loss": 0.6367,
+      "step": 67132
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.115087604531067e-07,
+      "loss": 0.5146,
+      "step": 67133
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.114815860678235e-07,
+      "loss": 0.5054,
+      "step": 67134
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.114544118268804e-07,
+      "loss": 0.5898,
+      "step": 67135
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.114272377302978e-07,
+      "loss": 0.7959,
+      "step": 67136
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.11400063778097e-07,
+      "loss": 0.668,
+      "step": 67137
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.113728899702981e-07,
+      "loss": 0.6416,
+      "step": 67138
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.113457163069226e-07,
+      "loss": 0.6904,
+      "step": 67139
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.113185427879909e-07,
+      "loss": 0.4801,
+      "step": 67140
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.11291369413524e-07,
+      "loss": 0.7158,
+      "step": 67141
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.112641961835429e-07,
+      "loss": 0.6797,
+      "step": 67142
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.112370230980679e-07,
+      "loss": 0.5879,
+      "step": 67143
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.112098501571199e-07,
+      "loss": 0.5945,
+      "step": 67144
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.111826773607199e-07,
+      "loss": 0.7344,
+      "step": 67145
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.111555047088885e-07,
+      "loss": 0.6304,
+      "step": 67146
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.111283322016468e-07,
+      "loss": 0.7246,
+      "step": 67147
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.111011598390153e-07,
+      "loss": 0.5879,
+      "step": 67148
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.11073987621015e-07,
+      "loss": 0.7749,
+      "step": 67149
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.110468155476668e-07,
+      "loss": 0.666,
+      "step": 67150
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.110196436189913e-07,
+      "loss": 0.6152,
+      "step": 67151
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.109924718350097e-07,
+      "loss": 0.6138,
+      "step": 67152
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.109653001957421e-07,
+      "loss": 0.6602,
+      "step": 67153
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.109381287012097e-07,
+      "loss": 0.7603,
+      "step": 67154
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.109109573514332e-07,
+      "loss": 0.6465,
+      "step": 67155
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.108837861464337e-07,
+      "loss": 0.4375,
+      "step": 67156
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.108566150862314e-07,
+      "loss": 0.6895,
+      "step": 67157
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.108294441708479e-07,
+      "loss": 0.6135,
+      "step": 67158
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.108022734003033e-07,
+      "loss": 0.541,
+      "step": 67159
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.107751027746191e-07,
+      "loss": 0.8047,
+      "step": 67160
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.107479322938156e-07,
+      "loss": 0.6543,
+      "step": 67161
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.107207619579137e-07,
+      "loss": 0.7188,
+      "step": 67162
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.10693591766934e-07,
+      "loss": 0.5762,
+      "step": 67163
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.106664217208978e-07,
+      "loss": 0.6797,
+      "step": 67164
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.106392518198252e-07,
+      "loss": 0.6582,
+      "step": 67165
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.106120820637377e-07,
+      "loss": 0.7402,
+      "step": 67166
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.105849124526556e-07,
+      "loss": 0.5186,
+      "step": 67167
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.105577429866002e-07,
+      "loss": 0.7891,
+      "step": 67168
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.105305736655918e-07,
+      "loss": 0.5044,
+      "step": 67169
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.105034044896515e-07,
+      "loss": 0.6499,
+      "step": 67170
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.104762354587999e-07,
+      "loss": 0.6758,
+      "step": 67171
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.104490665730585e-07,
+      "loss": 0.6133,
+      "step": 67172
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.104218978324472e-07,
+      "loss": 0.4756,
+      "step": 67173
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.103947292369868e-07,
+      "loss": 0.667,
+      "step": 67174
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.103675607866989e-07,
+      "loss": 0.5161,
+      "step": 67175
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.103403924816034e-07,
+      "loss": 0.6909,
+      "step": 67176
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.103132243217219e-07,
+      "loss": 0.6938,
+      "step": 67177
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.102860563070746e-07,
+      "loss": 0.7148,
+      "step": 67178
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.102588884376827e-07,
+      "loss": 0.6709,
+      "step": 67179
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.10231720713567e-07,
+      "loss": 0.5718,
+      "step": 67180
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.102045531347478e-07,
+      "loss": 0.6445,
+      "step": 67181
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.101773857012466e-07,
+      "loss": 0.6328,
+      "step": 67182
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.101502184130837e-07,
+      "loss": 0.6343,
+      "step": 67183
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.101230512702799e-07,
+      "loss": 0.6958,
+      "step": 67184
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.100958842728564e-07,
+      "loss": 0.5537,
+      "step": 67185
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.100687174208334e-07,
+      "loss": 0.7271,
+      "step": 67186
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.100415507142323e-07,
+      "loss": 0.7402,
+      "step": 67187
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.100143841530735e-07,
+      "loss": 0.6299,
+      "step": 67188
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.09987217737378e-07,
+      "loss": 0.6289,
+      "step": 67189
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.099600514671666e-07,
+      "loss": 0.7529,
+      "step": 67190
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.099328853424601e-07,
+      "loss": 0.6714,
+      "step": 67191
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.099057193632795e-07,
+      "loss": 0.6523,
+      "step": 67192
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.098785535296448e-07,
+      "loss": 0.7676,
+      "step": 67193
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.098513878415778e-07,
+      "loss": 0.7251,
+      "step": 67194
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.098242222990985e-07,
+      "loss": 0.6372,
+      "step": 67195
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.097970569022283e-07,
+      "loss": 0.5283,
+      "step": 67196
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.097698916509876e-07,
+      "loss": 0.5391,
+      "step": 67197
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.097427265453974e-07,
+      "loss": 0.6289,
+      "step": 67198
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.097155615854785e-07,
+      "loss": 0.6538,
+      "step": 67199
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.096883967712513e-07,
+      "loss": 0.583,
+      "step": 67200
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.096612321027374e-07,
+      "loss": 0.667,
+      "step": 67201
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.096340675799572e-07,
+      "loss": 0.8018,
+      "step": 67202
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.096069032029311e-07,
+      "loss": 0.6558,
+      "step": 67203
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.095797389716804e-07,
+      "loss": 0.582,
+      "step": 67204
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.095525748862257e-07,
+      "loss": 0.55,
+      "step": 67205
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.095254109465878e-07,
+      "loss": 0.6768,
+      "step": 67206
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.094982471527873e-07,
+      "loss": 0.6543,
+      "step": 67207
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.094710835048457e-07,
+      "loss": 0.5894,
+      "step": 67208
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.094439200027829e-07,
+      "loss": 0.5983,
+      "step": 67209
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.094167566466204e-07,
+      "loss": 0.6875,
+      "step": 67210
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.093895934363785e-07,
+      "loss": 0.6426,
+      "step": 67211
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.093624303720788e-07,
+      "loss": 0.749,
+      "step": 67212
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.093352674537412e-07,
+      "loss": 0.3982,
+      "step": 67213
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.093081046813868e-07,
+      "loss": 0.8145,
+      "step": 67214
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.092809420550363e-07,
+      "loss": 0.6963,
+      "step": 67215
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.092537795747108e-07,
+      "loss": 0.605,
+      "step": 67216
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.092266172404305e-07,
+      "loss": 0.605,
+      "step": 67217
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.09199455052217e-07,
+      "loss": 0.5007,
+      "step": 67218
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.091722930100904e-07,
+      "loss": 0.554,
+      "step": 67219
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.091451311140721e-07,
+      "loss": 0.4814,
+      "step": 67220
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.091179693641823e-07,
+      "loss": 0.5942,
+      "step": 67221
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.090908077604426e-07,
+      "loss": 0.5781,
+      "step": 67222
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.09063646302873e-07,
+      "loss": 0.5117,
+      "step": 67223
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.090364849914945e-07,
+      "loss": 0.6523,
+      "step": 67224
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.090093238263281e-07,
+      "loss": 0.5786,
+      "step": 67225
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.089821628073943e-07,
+      "loss": 0.5698,
+      "step": 67226
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.089550019347142e-07,
+      "loss": 0.4849,
+      "step": 67227
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.089278412083084e-07,
+      "loss": 0.6943,
+      "step": 67228
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.08900680628198e-07,
+      "loss": 0.6348,
+      "step": 67229
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.088735201944031e-07,
+      "loss": 0.5815,
+      "step": 67230
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.088463599069454e-07,
+      "loss": 0.6934,
+      "step": 67231
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.088191997658453e-07,
+      "loss": 0.5996,
+      "step": 67232
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.087920397711236e-07,
+      "loss": 0.6885,
+      "step": 67233
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.087648799228006e-07,
+      "loss": 0.6309,
+      "step": 67234
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.087377202208978e-07,
+      "loss": 0.7129,
+      "step": 67235
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.087105606654355e-07,
+      "loss": 0.6807,
+      "step": 67236
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.08683401256435e-07,
+      "loss": 0.8564,
+      "step": 67237
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.086562419939166e-07,
+      "loss": 0.4568,
+      "step": 67238
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.086290828779015e-07,
+      "loss": 0.6782,
+      "step": 67239
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.086019239084101e-07,
+      "loss": 0.4985,
+      "step": 67240
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.085747650854636e-07,
+      "loss": 0.4873,
+      "step": 67241
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.085476064090828e-07,
+      "loss": 0.7422,
+      "step": 67242
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.08520447879288e-07,
+      "loss": 0.4656,
+      "step": 67243
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.084932894961004e-07,
+      "loss": 0.6011,
+      "step": 67244
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.084661312595405e-07,
+      "loss": 0.4176,
+      "step": 67245
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.084389731696295e-07,
+      "loss": 0.6631,
+      "step": 67246
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.084118152263877e-07,
+      "loss": 0.6807,
+      "step": 67247
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.083846574298363e-07,
+      "loss": 0.6885,
+      "step": 67248
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.083574997799958e-07,
+      "loss": 0.6421,
+      "step": 67249
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.083303422768875e-07,
+      "loss": 0.6226,
+      "step": 67250
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.083031849205315e-07,
+      "loss": 0.5991,
+      "step": 67251
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.082760277109497e-07,
+      "loss": 0.4849,
+      "step": 67252
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.082488706481613e-07,
+      "loss": 0.6357,
+      "step": 67253
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.082217137321882e-07,
+      "loss": 0.8027,
+      "step": 67254
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.081945569630508e-07,
+      "loss": 0.7461,
+      "step": 67255
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.081674003407702e-07,
+      "loss": 0.7144,
+      "step": 67256
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.081402438653668e-07,
+      "loss": 0.6914,
+      "step": 67257
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.081130875368618e-07,
+      "loss": 0.573,
+      "step": 67258
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.080859313552755e-07,
+      "loss": 0.7197,
+      "step": 67259
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.080587753206293e-07,
+      "loss": 0.3683,
+      "step": 67260
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.080316194329433e-07,
+      "loss": 0.5571,
+      "step": 67261
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.080044636922392e-07,
+      "loss": 0.6733,
+      "step": 67262
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.07977308098537e-07,
+      "loss": 0.6914,
+      "step": 67263
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.079501526518576e-07,
+      "loss": 0.5298,
+      "step": 67264
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.07922997352222e-07,
+      "loss": 0.6191,
+      "step": 67265
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.078958421996509e-07,
+      "loss": 0.478,
+      "step": 67266
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.078686871941652e-07,
+      "loss": 0.5864,
+      "step": 67267
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.078415323357855e-07,
+      "loss": 0.5752,
+      "step": 67268
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.078143776245327e-07,
+      "loss": 0.457,
+      "step": 67269
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.077872230604279e-07,
+      "loss": 0.6807,
+      "step": 67270
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.077600686434913e-07,
+      "loss": 0.6025,
+      "step": 67271
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.077329143737443e-07,
+      "loss": 0.6729,
+      "step": 67272
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.07705760251207e-07,
+      "loss": 0.7681,
+      "step": 67273
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.076786062759007e-07,
+      "loss": 0.4907,
+      "step": 67274
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.076514524478461e-07,
+      "loss": 0.3872,
+      "step": 67275
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.076242987670637e-07,
+      "loss": 0.7061,
+      "step": 67276
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.075971452335747e-07,
+      "loss": 0.6213,
+      "step": 67277
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.075699918473995e-07,
+      "loss": 0.5381,
+      "step": 67278
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.075428386085595e-07,
+      "loss": 0.8379,
+      "step": 67279
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.075156855170747e-07,
+      "loss": 0.5547,
+      "step": 67280
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.074885325729666e-07,
+      "loss": 0.6133,
+      "step": 67281
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.074613797762558e-07,
+      "loss": 0.5542,
+      "step": 67282
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.074342271269627e-07,
+      "loss": 0.5728,
+      "step": 67283
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.074070746251083e-07,
+      "loss": 0.6455,
+      "step": 67284
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.073799222707135e-07,
+      "loss": 0.5552,
+      "step": 67285
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.073527700637992e-07,
+      "loss": 0.5295,
+      "step": 67286
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.073256180043859e-07,
+      "loss": 0.7324,
+      "step": 67287
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.072984660924944e-07,
+      "loss": 0.3894,
+      "step": 67288
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.072713143281459e-07,
+      "loss": 0.6426,
+      "step": 67289
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.072441627113605e-07,
+      "loss": 0.709,
+      "step": 67290
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.072170112421597e-07,
+      "loss": 0.6465,
+      "step": 67291
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.071898599205642e-07,
+      "loss": 0.625,
+      "step": 67292
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.071627087465941e-07,
+      "loss": 0.606,
+      "step": 67293
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.071355577202709e-07,
+      "loss": 0.6973,
+      "step": 67294
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.071084068416149e-07,
+      "loss": 0.71,
+      "step": 67295
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.070812561106472e-07,
+      "loss": 0.584,
+      "step": 67296
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.070541055273884e-07,
+      "loss": 0.5654,
+      "step": 67297
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.070269550918597e-07,
+      "loss": 0.6147,
+      "step": 67298
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.069998048040813e-07,
+      "loss": 0.6528,
+      "step": 67299
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.069726546640745e-07,
+      "loss": 0.6543,
+      "step": 67300
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.069455046718596e-07,
+      "loss": 0.6719,
+      "step": 67301
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 8.069183548274581e-07,
+      "loss": 0.6328,
+      "step": 67302
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.068912051308901e-07,
+      "loss": 0.4861,
+      "step": 67303
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.068640555821768e-07,
+      "loss": 0.7256,
+      "step": 67304
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.068369061813384e-07,
+      "loss": 0.4995,
+      "step": 67305
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.068097569283964e-07,
+      "loss": 0.5869,
+      "step": 67306
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.06782607823371e-07,
+      "loss": 0.583,
+      "step": 67307
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.067554588662835e-07,
+      "loss": 0.7461,
+      "step": 67308
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.067283100571542e-07,
+      "loss": 0.6616,
+      "step": 67309
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.067011613960045e-07,
+      "loss": 0.6924,
+      "step": 67310
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.066740128828546e-07,
+      "loss": 0.5754,
+      "step": 67311
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.066468645177259e-07,
+      "loss": 0.6133,
+      "step": 67312
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.066197163006386e-07,
+      "loss": 0.7256,
+      "step": 67313
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.065925682316135e-07,
+      "loss": 0.4976,
+      "step": 67314
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.065654203106717e-07,
+      "loss": 0.6689,
+      "step": 67315
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.065382725378337e-07,
+      "loss": 0.7178,
+      "step": 67316
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.065111249131207e-07,
+      "loss": 0.4836,
+      "step": 67317
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.064839774365531e-07,
+      "loss": 0.7109,
+      "step": 67318
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.064568301081518e-07,
+      "loss": 0.6826,
+      "step": 67319
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.064296829279375e-07,
+      "loss": 0.6787,
+      "step": 67320
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.064025358959314e-07,
+      "loss": 0.5874,
+      "step": 67321
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.063753890121541e-07,
+      "loss": 0.6748,
+      "step": 67322
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.063482422766261e-07,
+      "loss": 0.5542,
+      "step": 67323
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.06321095689368e-07,
+      "loss": 0.7432,
+      "step": 67324
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.062939492504012e-07,
+      "loss": 0.71,
+      "step": 67325
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.062668029597461e-07,
+      "loss": 0.7021,
+      "step": 67326
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.062396568174238e-07,
+      "loss": 0.6309,
+      "step": 67327
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.062125108234547e-07,
+      "loss": 0.5254,
+      "step": 67328
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.061853649778598e-07,
+      "loss": 0.7178,
+      "step": 67329
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.061582192806599e-07,
+      "loss": 0.5605,
+      "step": 67330
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.061310737318757e-07,
+      "loss": 0.5708,
+      "step": 67331
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.061039283315284e-07,
+      "loss": 0.4834,
+      "step": 67332
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.060767830796379e-07,
+      "loss": 0.6562,
+      "step": 67333
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.060496379762257e-07,
+      "loss": 0.7812,
+      "step": 67334
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.060224930213121e-07,
+      "loss": 0.6865,
+      "step": 67335
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.059953482149186e-07,
+      "loss": 0.7568,
+      "step": 67336
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.059682035570651e-07,
+      "loss": 0.6802,
+      "step": 67337
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.059410590477732e-07,
+      "loss": 0.6934,
+      "step": 67338
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.05913914687063e-07,
+      "loss": 0.6948,
+      "step": 67339
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.058867704749559e-07,
+      "loss": 0.6367,
+      "step": 67340
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.058596264114719e-07,
+      "loss": 0.6057,
+      "step": 67341
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.058324824966332e-07,
+      "loss": 0.7129,
+      "step": 67342
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.058053387304588e-07,
+      "loss": 0.6555,
+      "step": 67343
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.057781951129708e-07,
+      "loss": 0.7451,
+      "step": 67344
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.05751051644189e-07,
+      "loss": 0.6758,
+      "step": 67345
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.05723908324135e-07,
+      "loss": 0.5957,
+      "step": 67346
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.056967651528292e-07,
+      "loss": 0.5088,
+      "step": 67347
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.056696221302924e-07,
+      "loss": 0.5767,
+      "step": 67348
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.056424792565454e-07,
+      "loss": 0.5483,
+      "step": 67349
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.056153365316092e-07,
+      "loss": 0.4844,
+      "step": 67350
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.055881939555041e-07,
+      "loss": 0.8154,
+      "step": 67351
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.055610515282517e-07,
+      "loss": 0.6553,
+      "step": 67352
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.055339092498721e-07,
+      "loss": 0.573,
+      "step": 67353
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.055067671203859e-07,
+      "loss": 0.6807,
+      "step": 67354
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.054796251398145e-07,
+      "loss": 0.626,
+      "step": 67355
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.054524833081782e-07,
+      "loss": 0.6348,
+      "step": 67356
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.054253416254982e-07,
+      "loss": 0.8086,
+      "step": 67357
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.053982000917948e-07,
+      "loss": 0.5753,
+      "step": 67358
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.053710587070894e-07,
+      "loss": 0.6294,
+      "step": 67359
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.053439174714023e-07,
+      "loss": 0.6084,
+      "step": 67360
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.053167763847543e-07,
+      "loss": 0.6323,
+      "step": 67361
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.052896354471666e-07,
+      "loss": 0.3367,
+      "step": 67362
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.052624946586595e-07,
+      "loss": 0.6353,
+      "step": 67363
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.052353540192538e-07,
+      "loss": 0.6611,
+      "step": 67364
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.052082135289706e-07,
+      "loss": 0.6333,
+      "step": 67365
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.051810731878304e-07,
+      "loss": 0.8057,
+      "step": 67366
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.051539329958541e-07,
+      "loss": 0.583,
+      "step": 67367
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.051267929530624e-07,
+      "loss": 0.7012,
+      "step": 67368
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.050996530594764e-07,
+      "loss": 0.8379,
+      "step": 67369
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.050725133151164e-07,
+      "loss": 0.4436,
+      "step": 67370
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.050453737200036e-07,
+      "loss": 0.4143,
+      "step": 67371
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.050182342741587e-07,
+      "loss": 0.7363,
+      "step": 67372
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.049910949776021e-07,
+      "loss": 0.6924,
+      "step": 67373
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.049639558303549e-07,
+      "loss": 0.707,
+      "step": 67374
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.049368168324378e-07,
+      "loss": 0.6904,
+      "step": 67375
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.049096779838717e-07,
+      "loss": 0.4886,
+      "step": 67376
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.048825392846774e-07,
+      "loss": 0.7627,
+      "step": 67377
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.048554007348753e-07,
+      "loss": 0.604,
+      "step": 67378
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.048282623344867e-07,
+      "loss": 0.4971,
+      "step": 67379
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.048011240835318e-07,
+      "loss": 0.6152,
+      "step": 67380
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.047739859820322e-07,
+      "loss": 0.5339,
+      "step": 67381
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.04746848030008e-07,
+      "loss": 0.4279,
+      "step": 67382
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.047197102274798e-07,
+      "loss": 0.6011,
+      "step": 67383
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.046925725744692e-07,
+      "loss": 0.7471,
+      "step": 67384
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.046654350709962e-07,
+      "loss": 0.9062,
+      "step": 67385
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.04638297717082e-07,
+      "loss": 0.623,
+      "step": 67386
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.046111605127471e-07,
+      "loss": 0.6226,
+      "step": 67387
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.045840234580128e-07,
+      "loss": 0.6816,
+      "step": 67388
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.045568865528991e-07,
+      "loss": 0.7783,
+      "step": 67389
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.045297497974276e-07,
+      "loss": 0.5479,
+      "step": 67390
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.045026131916188e-07,
+      "loss": 0.7842,
+      "step": 67391
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.04475476735493e-07,
+      "loss": 0.3367,
+      "step": 67392
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.044483404290715e-07,
+      "loss": 0.5977,
+      "step": 67393
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.044212042723749e-07,
+      "loss": 0.5278,
+      "step": 67394
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.043940682654238e-07,
+      "loss": 0.6599,
+      "step": 67395
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.043669324082394e-07,
+      "loss": 0.7969,
+      "step": 67396
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.043397967008421e-07,
+      "loss": 0.6118,
+      "step": 67397
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.043126611432529e-07,
+      "loss": 0.5859,
+      "step": 67398
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.042855257354923e-07,
+      "loss": 0.6763,
+      "step": 67399
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.042583904775814e-07,
+      "loss": 0.7725,
+      "step": 67400
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.042312553695414e-07,
+      "loss": 0.731,
+      "step": 67401
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.042041204113918e-07,
+      "loss": 0.5996,
+      "step": 67402
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.041769856031544e-07,
+      "loss": 0.5591,
+      "step": 67403
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.041498509448494e-07,
+      "loss": 0.6494,
+      "step": 67404
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.041227164364982e-07,
+      "loss": 0.6689,
+      "step": 67405
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.04095582078121e-07,
+      "loss": 0.5938,
+      "step": 67406
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.040684478697389e-07,
+      "loss": 0.4883,
+      "step": 67407
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.040413138113724e-07,
+      "loss": 0.6348,
+      "step": 67408
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.040141799030428e-07,
+      "loss": 0.5781,
+      "step": 67409
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.039870461447702e-07,
+      "loss": 0.5762,
+      "step": 67410
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.039599125365763e-07,
+      "loss": 0.5586,
+      "step": 67411
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.039327790784808e-07,
+      "loss": 0.7075,
+      "step": 67412
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.03905645770505e-07,
+      "loss": 0.6221,
+      "step": 67413
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.038785126126694e-07,
+      "loss": 0.6953,
+      "step": 67414
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.038513796049954e-07,
+      "loss": 0.6646,
+      "step": 67415
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.038242467475032e-07,
+      "loss": 0.7314,
+      "step": 67416
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.037971140402138e-07,
+      "loss": 0.7227,
+      "step": 67417
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.037699814831477e-07,
+      "loss": 0.7344,
+      "step": 67418
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.037428490763262e-07,
+      "loss": 0.4812,
+      "step": 67419
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.037157168197695e-07,
+      "loss": 0.6021,
+      "step": 67420
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.036885847134991e-07,
+      "loss": 0.4878,
+      "step": 67421
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.036614527575352e-07,
+      "loss": 0.6641,
+      "step": 67422
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.036343209518983e-07,
+      "loss": 0.5728,
+      "step": 67423
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.0360718929661e-07,
+      "loss": 0.6543,
+      "step": 67424
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.035800577916902e-07,
+      "loss": 0.5686,
+      "step": 67425
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.035529264371605e-07,
+      "loss": 0.8232,
+      "step": 67426
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.035257952330411e-07,
+      "loss": 0.8418,
+      "step": 67427
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.034986641793531e-07,
+      "loss": 0.6274,
+      "step": 67428
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.03471533276117e-07,
+      "loss": 0.4644,
+      "step": 67429
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.034444025233537e-07,
+      "loss": 0.6318,
+      "step": 67430
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.034172719210845e-07,
+      "loss": 0.7207,
+      "step": 67431
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.033901414693292e-07,
+      "loss": 0.571,
+      "step": 67432
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.033630111681089e-07,
+      "loss": 0.4241,
+      "step": 67433
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.033358810174448e-07,
+      "loss": 0.5487,
+      "step": 67434
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.033087510173571e-07,
+      "loss": 0.5913,
+      "step": 67435
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.03281621167867e-07,
+      "loss": 0.5078,
+      "step": 67436
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.032544914689949e-07,
+      "loss": 0.5947,
+      "step": 67437
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.032273619207621e-07,
+      "loss": 0.6436,
+      "step": 67438
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.032002325231888e-07,
+      "loss": 0.4116,
+      "step": 67439
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.031731032762963e-07,
+      "loss": 0.6934,
+      "step": 67440
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.031459741801052e-07,
+      "loss": 0.7266,
+      "step": 67441
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.031188452346358e-07,
+      "loss": 0.6289,
+      "step": 67442
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.030917164399094e-07,
+      "loss": 0.7231,
+      "step": 67443
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.030645877959465e-07,
+      "loss": 0.6055,
+      "step": 67444
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.030374593027682e-07,
+      "loss": 0.5977,
+      "step": 67445
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.030103309603948e-07,
+      "loss": 0.7568,
+      "step": 67446
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.029832027688477e-07,
+      "loss": 0.5901,
+      "step": 67447
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02956074728147e-07,
+      "loss": 0.7124,
+      "step": 67448
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02928946838314e-07,
+      "loss": 0.5063,
+      "step": 67449
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02901819099369e-07,
+      "loss": 0.54,
+      "step": 67450
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.028746915113336e-07,
+      "loss": 0.5132,
+      "step": 67451
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.028475640742273e-07,
+      "loss": 0.4878,
+      "step": 67452
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.028204367880719e-07,
+      "loss": 0.5352,
+      "step": 67453
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.027933096528875e-07,
+      "loss": 0.8701,
+      "step": 67454
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.027661826686955e-07,
+      "loss": 0.666,
+      "step": 67455
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02739055835516e-07,
+      "loss": 0.4695,
+      "step": 67456
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.027119291533706e-07,
+      "loss": 0.5449,
+      "step": 67457
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.026848026222792e-07,
+      "loss": 0.7236,
+      "step": 67458
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.026576762422632e-07,
+      "loss": 0.4275,
+      "step": 67459
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02630550013343e-07,
+      "loss": 0.5635,
+      "step": 67460
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.026034239355399e-07,
+      "loss": 0.6182,
+      "step": 67461
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.025762980088742e-07,
+      "loss": 0.5767,
+      "step": 67462
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.025491722333664e-07,
+      "loss": 0.686,
+      "step": 67463
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.025220466090378e-07,
+      "loss": 0.7314,
+      "step": 67464
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.024949211359089e-07,
+      "loss": 0.5415,
+      "step": 67465
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.024677958140007e-07,
+      "loss": 0.5811,
+      "step": 67466
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.024406706433338e-07,
+      "loss": 0.6416,
+      "step": 67467
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.024135456239289e-07,
+      "loss": 0.6714,
+      "step": 67468
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.023864207558069e-07,
+      "loss": 0.4604,
+      "step": 67469
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.023592960389884e-07,
+      "loss": 0.7656,
+      "step": 67470
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.023321714734948e-07,
+      "loss": 0.542,
+      "step": 67471
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02305047059346e-07,
+      "loss": 0.7188,
+      "step": 67472
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02277922796563e-07,
+      "loss": 0.5049,
+      "step": 67473
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.022507986851669e-07,
+      "loss": 0.562,
+      "step": 67474
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.022236747251781e-07,
+      "loss": 0.7773,
+      "step": 67475
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.021965509166178e-07,
+      "loss": 0.5591,
+      "step": 67476
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.021694272595061e-07,
+      "loss": 0.7456,
+      "step": 67477
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.021423037538646e-07,
+      "loss": 0.5835,
+      "step": 67478
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.021151803997133e-07,
+      "loss": 0.4683,
+      "step": 67479
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.020880571970735e-07,
+      "loss": 0.4727,
+      "step": 67480
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.020609341459661e-07,
+      "loss": 0.5549,
+      "step": 67481
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.02033811246411e-07,
+      "loss": 0.5894,
+      "step": 67482
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.020066884984298e-07,
+      "loss": 0.6357,
+      "step": 67483
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01979565902043e-07,
+      "loss": 0.4526,
+      "step": 67484
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.019524434572711e-07,
+      "loss": 0.7129,
+      "step": 67485
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.019253211641352e-07,
+      "loss": 0.7285,
+      "step": 67486
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.018981990226558e-07,
+      "loss": 0.4316,
+      "step": 67487
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.018710770328543e-07,
+      "loss": 0.7773,
+      "step": 67488
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.018439551947505e-07,
+      "loss": 0.6846,
+      "step": 67489
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.018168335083661e-07,
+      "loss": 0.6416,
+      "step": 67490
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.017897119737215e-07,
+      "loss": 0.5835,
+      "step": 67491
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01762590590837e-07,
+      "loss": 0.7207,
+      "step": 67492
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01735469359734e-07,
+      "loss": 0.748,
+      "step": 67493
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.017083482804329e-07,
+      "loss": 0.5449,
+      "step": 67494
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.016812273529548e-07,
+      "loss": 0.6038,
+      "step": 67495
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.016541065773199e-07,
+      "loss": 0.6987,
+      "step": 67496
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.016269859535496e-07,
+      "loss": 0.7598,
+      "step": 67497
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.015998654816642e-07,
+      "loss": 0.6709,
+      "step": 67498
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.015727451616849e-07,
+      "loss": 0.6611,
+      "step": 67499
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01545624993632e-07,
+      "loss": 0.5483,
+      "step": 67500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.015185049775271e-07,
+      "loss": 0.6421,
+      "step": 67501
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.014913851133897e-07,
+      "loss": 0.6069,
+      "step": 67502
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.014642654012416e-07,
+      "loss": 0.708,
+      "step": 67503
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.014371458411029e-07,
+      "loss": 0.5278,
+      "step": 67504
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.014100264329947e-07,
+      "loss": 0.6445,
+      "step": 67505
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.013829071769377e-07,
+      "loss": 0.7119,
+      "step": 67506
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.013557880729528e-07,
+      "loss": 0.6279,
+      "step": 67507
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.013286691210604e-07,
+      "loss": 0.606,
+      "step": 67508
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.013015503212817e-07,
+      "loss": 0.792,
+      "step": 67509
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.012744316736371e-07,
+      "loss": 0.5869,
+      "step": 67510
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01247313178148e-07,
+      "loss": 0.7354,
+      "step": 67511
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.012201948348343e-07,
+      "loss": 0.3353,
+      "step": 67512
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.011930766437171e-07,
+      "loss": 0.5918,
+      "step": 67513
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.011659586048174e-07,
+      "loss": 0.5688,
+      "step": 67514
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.011388407181556e-07,
+      "loss": 0.6455,
+      "step": 67515
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.011117229837529e-07,
+      "loss": 0.5164,
+      "step": 67516
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.010846054016295e-07,
+      "loss": 0.6621,
+      "step": 67517
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.010574879718067e-07,
+      "loss": 0.6221,
+      "step": 67518
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.01030370694305e-07,
+      "loss": 0.5723,
+      "step": 67519
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.010032535691451e-07,
+      "loss": 0.6338,
+      "step": 67520
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.009761365963482e-07,
+      "loss": 0.6167,
+      "step": 67521
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.009490197759346e-07,
+      "loss": 0.7109,
+      "step": 67522
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.009219031079249e-07,
+      "loss": 0.5815,
+      "step": 67523
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.008947865923404e-07,
+      "loss": 0.6147,
+      "step": 67524
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.008676702292013e-07,
+      "loss": 0.5303,
+      "step": 67525
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.008405540185288e-07,
+      "loss": 0.7676,
+      "step": 67526
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.008134379603435e-07,
+      "loss": 0.6509,
+      "step": 67527
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.007863220546663e-07,
+      "loss": 0.6758,
+      "step": 67528
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.007592063015177e-07,
+      "loss": 0.4158,
+      "step": 67529
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.007320907009189e-07,
+      "loss": 0.6396,
+      "step": 67530
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.007049752528905e-07,
+      "loss": 0.6245,
+      "step": 67531
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.006778599574526e-07,
+      "loss": 0.6536,
+      "step": 67532
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.006507448146269e-07,
+      "loss": 0.7998,
+      "step": 67533
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.006236298244334e-07,
+      "loss": 0.687,
+      "step": 67534
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.005965149868935e-07,
+      "loss": 0.6875,
+      "step": 67535
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.005694003020276e-07,
+      "loss": 0.6787,
+      "step": 67536
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.005422857698565e-07,
+      "loss": 0.7549,
+      "step": 67537
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.00515171390401e-07,
+      "loss": 0.6606,
+      "step": 67538
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.004880571636819e-07,
+      "loss": 0.5142,
+      "step": 67539
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.0046094308972e-07,
+      "loss": 0.627,
+      "step": 67540
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.004338291685362e-07,
+      "loss": 0.7715,
+      "step": 67541
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.004067154001505e-07,
+      "loss": 0.7861,
+      "step": 67542
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.003796017845847e-07,
+      "loss": 0.4829,
+      "step": 67543
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.003524883218587e-07,
+      "loss": 0.5298,
+      "step": 67544
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.003253750119938e-07,
+      "loss": 0.564,
+      "step": 67545
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.002982618550103e-07,
+      "loss": 0.7021,
+      "step": 67546
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.002711488509296e-07,
+      "loss": 0.6533,
+      "step": 67547
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.002440359997718e-07,
+      "loss": 0.5146,
+      "step": 67548
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.002169233015581e-07,
+      "loss": 0.5996,
+      "step": 67549
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.00189810756309e-07,
+      "loss": 0.7646,
+      "step": 67550
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.001626983640458e-07,
+      "loss": 0.7207,
+      "step": 67551
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.001355861247886e-07,
+      "loss": 0.6064,
+      "step": 67552
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.001084740385582e-07,
+      "loss": 0.6885,
+      "step": 67553
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.000813621053757e-07,
+      "loss": 0.6357,
+      "step": 67554
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.000542503252616e-07,
+      "loss": 0.7012,
+      "step": 67555
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.000271386982369e-07,
+      "loss": 0.6089,
+      "step": 67556
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 8.000000272243222e-07,
+      "loss": 0.6572,
+      "step": 67557
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.999729159035381e-07,
+      "loss": 0.6414,
+      "step": 67558
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.999458047359058e-07,
+      "loss": 0.5249,
+      "step": 67559
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.999186937214456e-07,
+      "loss": 0.6426,
+      "step": 67560
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.998915828601788e-07,
+      "loss": 0.5781,
+      "step": 67561
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.998644721521256e-07,
+      "loss": 0.665,
+      "step": 67562
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.998373615973068e-07,
+      "loss": 0.554,
+      "step": 67563
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.998102511957437e-07,
+      "loss": 0.4762,
+      "step": 67564
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.997831409474561e-07,
+      "loss": 0.6538,
+      "step": 67565
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.997560308524658e-07,
+      "loss": 0.5576,
+      "step": 67566
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.99728920910793e-07,
+      "loss": 0.5884,
+      "step": 67567
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.997018111224585e-07,
+      "loss": 0.6821,
+      "step": 67568
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.996747014874831e-07,
+      "loss": 0.4702,
+      "step": 67569
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.996475920058876e-07,
+      "loss": 0.5903,
+      "step": 67570
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.996204826776931e-07,
+      "loss": 0.4473,
+      "step": 67571
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.995933735029195e-07,
+      "loss": 0.6333,
+      "step": 67572
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.995662644815883e-07,
+      "loss": 0.5273,
+      "step": 67573
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.995391556137199e-07,
+      "loss": 0.8408,
+      "step": 67574
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.995120468993349e-07,
+      "loss": 0.5513,
+      "step": 67575
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.994849383384547e-07,
+      "loss": 0.8018,
+      "step": 67576
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.994578299310993e-07,
+      "loss": 0.6494,
+      "step": 67577
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.994307216772901e-07,
+      "loss": 0.7251,
+      "step": 67578
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.994036135770473e-07,
+      "loss": 0.6807,
+      "step": 67579
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.993765056303922e-07,
+      "loss": 0.7046,
+      "step": 67580
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.993493978373456e-07,
+      "loss": 0.5537,
+      "step": 67581
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.993222901979273e-07,
+      "loss": 0.4634,
+      "step": 67582
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.992951827121591e-07,
+      "loss": 0.666,
+      "step": 67583
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.992680753800609e-07,
+      "loss": 0.6602,
+      "step": 67584
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.992409682016544e-07,
+      "loss": 0.6543,
+      "step": 67585
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.992138611769595e-07,
+      "loss": 0.6309,
+      "step": 67586
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.991867543059975e-07,
+      "loss": 0.7441,
+      "step": 67587
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.991596475887888e-07,
+      "loss": 0.7139,
+      "step": 67588
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.991325410253545e-07,
+      "loss": 0.6343,
+      "step": 67589
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.99105434615715e-07,
+      "loss": 0.5298,
+      "step": 67590
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.990783283598916e-07,
+      "loss": 0.4927,
+      "step": 67591
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.990512222579046e-07,
+      "loss": 0.7363,
+      "step": 67592
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.990241163097747e-07,
+      "loss": 0.6017,
+      "step": 67593
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.989970105155227e-07,
+      "loss": 0.6602,
+      "step": 67594
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.989699048751696e-07,
+      "loss": 0.5469,
+      "step": 67595
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.989427993887357e-07,
+      "loss": 0.5591,
+      "step": 67596
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.989156940562425e-07,
+      "loss": 0.707,
+      "step": 67597
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.988885888777099e-07,
+      "loss": 0.604,
+      "step": 67598
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.988614838531594e-07,
+      "loss": 0.5923,
+      "step": 67599
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.988343789826111e-07,
+      "loss": 0.665,
+      "step": 67600
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.988072742660864e-07,
+      "loss": 0.6543,
+      "step": 67601
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.987801697036056e-07,
+      "loss": 0.6084,
+      "step": 67602
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.987530652951895e-07,
+      "loss": 0.8154,
+      "step": 67603
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98725961040859e-07,
+      "loss": 0.6318,
+      "step": 67604
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.986988569406346e-07,
+      "loss": 0.7764,
+      "step": 67605
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.986717529945374e-07,
+      "loss": 0.5825,
+      "step": 67606
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.986446492025878e-07,
+      "loss": 0.5791,
+      "step": 67607
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.986175455648068e-07,
+      "loss": 0.6134,
+      "step": 67608
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98590442081215e-07,
+      "loss": 0.7637,
+      "step": 67609
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.985633387518336e-07,
+      "loss": 0.7388,
+      "step": 67610
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98536235576683e-07,
+      "loss": 0.6099,
+      "step": 67611
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.985091325557838e-07,
+      "loss": 0.5703,
+      "step": 67612
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.984820296891567e-07,
+      "loss": 0.6006,
+      "step": 67613
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.984549269768228e-07,
+      "loss": 0.6484,
+      "step": 67614
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.984278244188025e-07,
+      "loss": 0.6328,
+      "step": 67615
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98400722015117e-07,
+      "loss": 0.5173,
+      "step": 67616
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.983736197657866e-07,
+      "loss": 0.6592,
+      "step": 67617
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.983465176708324e-07,
+      "loss": 0.7324,
+      "step": 67618
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.983194157302747e-07,
+      "loss": 0.7178,
+      "step": 67619
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.982923139441349e-07,
+      "loss": 0.5845,
+      "step": 67620
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.982652123124336e-07,
+      "loss": 0.5518,
+      "step": 67621
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.982381108351909e-07,
+      "loss": 0.6587,
+      "step": 67622
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.982110095124282e-07,
+      "loss": 0.7666,
+      "step": 67623
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.981839083441659e-07,
+      "loss": 0.502,
+      "step": 67624
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98156807330425e-07,
+      "loss": 0.6592,
+      "step": 67625
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.98129706471226e-07,
+      "loss": 0.6431,
+      "step": 67626
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.981026057665901e-07,
+      "loss": 0.6602,
+      "step": 67627
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.980755052165374e-07,
+      "loss": 0.8125,
+      "step": 67628
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.980484048210892e-07,
+      "loss": 0.5322,
+      "step": 67629
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.980213045802662e-07,
+      "loss": 0.5957,
+      "step": 67630
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.979942044940892e-07,
+      "loss": 0.6348,
+      "step": 67631
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.979671045625782e-07,
+      "loss": 0.5986,
+      "step": 67632
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.979400047857547e-07,
+      "loss": 0.6719,
+      "step": 67633
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.979129051636392e-07,
+      "loss": 0.4277,
+      "step": 67634
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.978858056962526e-07,
+      "loss": 0.5127,
+      "step": 67635
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.978587063836153e-07,
+      "loss": 0.562,
+      "step": 67636
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.978316072257486e-07,
+      "loss": 0.8027,
+      "step": 67637
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.978045082226727e-07,
+      "loss": 0.6255,
+      "step": 67638
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.977774093744087e-07,
+      "loss": 0.7217,
+      "step": 67639
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.977503106809772e-07,
+      "loss": 0.4811,
+      "step": 67640
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.977232121423994e-07,
+      "loss": 0.252,
+      "step": 67641
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.976961137586952e-07,
+      "loss": 0.5956,
+      "step": 67642
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.976690155298858e-07,
+      "loss": 0.7139,
+      "step": 67643
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.97641917455992e-07,
+      "loss": 0.6357,
+      "step": 67644
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.976148195370344e-07,
+      "loss": 0.6562,
+      "step": 67645
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.975877217730341e-07,
+      "loss": 0.7168,
+      "step": 67646
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.975606241640114e-07,
+      "loss": 0.6157,
+      "step": 67647
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.97533526709987e-07,
+      "loss": 0.7598,
+      "step": 67648
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.975064294109821e-07,
+      "loss": 0.5596,
+      "step": 67649
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.974793322670171e-07,
+      "loss": 0.6992,
+      "step": 67650
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.974522352781133e-07,
+      "loss": 0.6758,
+      "step": 67651
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.974251384442906e-07,
+      "loss": 0.675,
+      "step": 67652
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.973980417655702e-07,
+      "loss": 0.5532,
+      "step": 67653
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.973709452419728e-07,
+      "loss": 0.623,
+      "step": 67654
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.97343848873519e-07,
+      "loss": 0.7275,
+      "step": 67655
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.973167526602299e-07,
+      "loss": 0.7549,
+      "step": 67656
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.972896566021259e-07,
+      "loss": 0.4897,
+      "step": 67657
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.97262560699228e-07,
+      "loss": 0.6299,
+      "step": 67658
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.972354649515567e-07,
+      "loss": 0.4449,
+      "step": 67659
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.972083693591331e-07,
+      "loss": 0.5293,
+      "step": 67660
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.971812739219779e-07,
+      "loss": 0.6719,
+      "step": 67661
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.971541786401112e-07,
+      "loss": 0.71,
+      "step": 67662
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.971270835135546e-07,
+      "loss": 0.7344,
+      "step": 67663
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.970999885423284e-07,
+      "loss": 0.4392,
+      "step": 67664
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.970728937264531e-07,
+      "loss": 0.6655,
+      "step": 67665
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.9704579906595e-07,
+      "loss": 0.4204,
+      "step": 67666
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.970187045608394e-07,
+      "loss": 0.665,
+      "step": 67667
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.969916102111425e-07,
+      "loss": 0.6924,
+      "step": 67668
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.969645160168795e-07,
+      "loss": 0.8125,
+      "step": 67669
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.969374219780718e-07,
+      "loss": 0.5566,
+      "step": 67670
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.969103280947398e-07,
+      "loss": 0.6963,
+      "step": 67671
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.96883234366904e-07,
+      "loss": 0.4714,
+      "step": 67672
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.968561407945854e-07,
+      "loss": 0.6084,
+      "step": 67673
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.968290473778045e-07,
+      "loss": 0.6755,
+      "step": 67674
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.968019541165826e-07,
+      "loss": 0.8301,
+      "step": 67675
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.967748610109399e-07,
+      "loss": 0.7207,
+      "step": 67676
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.967477680608974e-07,
+      "loss": 0.519,
+      "step": 67677
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.967206752664757e-07,
+      "loss": 0.5576,
+      "step": 67678
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.966935826276957e-07,
+      "loss": 0.6885,
+      "step": 67679
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.966664901445784e-07,
+      "loss": 0.6592,
+      "step": 67680
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.966393978171438e-07,
+      "loss": 0.583,
+      "step": 67681
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.966123056454133e-07,
+      "loss": 0.7627,
+      "step": 67682
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.965852136294074e-07,
+      "loss": 0.5996,
+      "step": 67683
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.965581217691465e-07,
+      "loss": 0.6785,
+      "step": 67684
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.96531030064652e-07,
+      "loss": 0.6045,
+      "step": 67685
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.965039385159441e-07,
+      "loss": 0.5142,
+      "step": 67686
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.964768471230441e-07,
+      "loss": 0.5967,
+      "step": 67687
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.964497558859721e-07,
+      "loss": 0.6279,
+      "step": 67688
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.964226648047494e-07,
+      "loss": 0.7417,
+      "step": 67689
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.963955738793967e-07,
+      "loss": 0.7451,
+      "step": 67690
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.963684831099341e-07,
+      "loss": 0.5527,
+      "step": 67691
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 7.963413924963831e-07,
+      "loss": 0.7861,
+      "step": 67692
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.963143020387638e-07,
+      "loss": 0.6802,
+      "step": 67693
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.962872117370976e-07,
+      "loss": 0.4609,
+      "step": 67694
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.962601215914046e-07,
+      "loss": 0.8662,
+      "step": 67695
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.962330316017061e-07,
+      "loss": 0.7295,
+      "step": 67696
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.962059417680224e-07,
+      "loss": 0.6099,
+      "step": 67697
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.961788520903746e-07,
+      "loss": 0.415,
+      "step": 67698
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.961517625687832e-07,
+      "loss": 0.6934,
+      "step": 67699
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.961246732032696e-07,
+      "loss": 0.7012,
+      "step": 67700
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.960975839938533e-07,
+      "loss": 0.7109,
+      "step": 67701
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.960704949405559e-07,
+      "loss": 0.6567,
+      "step": 67702
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.960434060433977e-07,
+      "loss": 0.6465,
+      "step": 67703
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.960163173024e-07,
+      "loss": 0.6138,
+      "step": 67704
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.959892287175829e-07,
+      "loss": 0.8516,
+      "step": 67705
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.959621402889678e-07,
+      "loss": 0.6182,
+      "step": 67706
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.959350520165748e-07,
+      "loss": 0.5649,
+      "step": 67707
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.959079639004252e-07,
+      "loss": 0.6602,
+      "step": 67708
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.958808759405393e-07,
+      "loss": 0.6323,
+      "step": 67709
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.958537881369385e-07,
+      "loss": 0.4602,
+      "step": 67710
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.958267004896427e-07,
+      "loss": 0.7002,
+      "step": 67711
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.957996129986728e-07,
+      "loss": 0.5657,
+      "step": 67712
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.957725256640502e-07,
+      "loss": 0.6895,
+      "step": 67713
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.957454384857947e-07,
+      "loss": 0.5066,
+      "step": 67714
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.957183514639278e-07,
+      "loss": 0.537,
+      "step": 67715
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.956912645984699e-07,
+      "loss": 0.6533,
+      "step": 67716
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.956641778894419e-07,
+      "loss": 0.71,
+      "step": 67717
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.956370913368642e-07,
+      "loss": 0.4661,
+      "step": 67718
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.956100049407581e-07,
+      "loss": 0.5464,
+      "step": 67719
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.955829187011442e-07,
+      "loss": 0.7129,
+      "step": 67720
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.955558326180428e-07,
+      "loss": 0.634,
+      "step": 67721
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.955287466914747e-07,
+      "loss": 0.7236,
+      "step": 67722
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.955016609214611e-07,
+      "loss": 0.623,
+      "step": 67723
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.954745753080222e-07,
+      "loss": 0.533,
+      "step": 67724
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.954474898511792e-07,
+      "loss": 0.6846,
+      "step": 67725
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.954204045509525e-07,
+      "loss": 0.5844,
+      "step": 67726
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.953933194073632e-07,
+      "loss": 0.6738,
+      "step": 67727
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.953662344204316e-07,
+      "loss": 0.7104,
+      "step": 67728
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.953391495901788e-07,
+      "loss": 0.7021,
+      "step": 67729
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.953120649166259e-07,
+      "loss": 0.666,
+      "step": 67730
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.952849803997924e-07,
+      "loss": 0.5464,
+      "step": 67731
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.952578960397001e-07,
+      "loss": 0.6572,
+      "step": 67732
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.952308118363692e-07,
+      "loss": 0.6895,
+      "step": 67733
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.95203727789821e-07,
+      "loss": 0.6523,
+      "step": 67734
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.951766439000754e-07,
+      "loss": 0.5728,
+      "step": 67735
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.95149560167154e-07,
+      "loss": 0.447,
+      "step": 67736
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.951224765910771e-07,
+      "loss": 0.7002,
+      "step": 67737
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.950953931718654e-07,
+      "loss": 0.6855,
+      "step": 67738
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.950683099095399e-07,
+      "loss": 0.752,
+      "step": 67739
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.950412268041214e-07,
+      "loss": 0.5386,
+      "step": 67740
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.9501414385563e-07,
+      "loss": 0.7178,
+      "step": 67741
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.949870610640868e-07,
+      "loss": 0.5203,
+      "step": 67742
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.949599784295127e-07,
+      "loss": 0.6475,
+      "step": 67743
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.949328959519283e-07,
+      "loss": 0.5183,
+      "step": 67744
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.949058136313543e-07,
+      "loss": 0.4617,
+      "step": 67745
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.948787314678116e-07,
+      "loss": 0.4321,
+      "step": 67746
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.948516494613207e-07,
+      "loss": 0.6135,
+      "step": 67747
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.948245676119026e-07,
+      "loss": 0.8008,
+      "step": 67748
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.947974859195777e-07,
+      "loss": 0.624,
+      "step": 67749
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.947704043843674e-07,
+      "loss": 0.6335,
+      "step": 67750
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.947433230062918e-07,
+      "loss": 0.627,
+      "step": 67751
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.947162417853714e-07,
+      "loss": 0.6816,
+      "step": 67752
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.946891607216278e-07,
+      "loss": 0.8389,
+      "step": 67753
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.946620798150807e-07,
+      "loss": 0.6035,
+      "step": 67754
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.94634999065752e-07,
+      "loss": 0.3752,
+      "step": 67755
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.946079184736616e-07,
+      "loss": 0.6836,
+      "step": 67756
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.945808380388302e-07,
+      "loss": 0.6772,
+      "step": 67757
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.945537577612792e-07,
+      "loss": 0.5991,
+      "step": 67758
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.945266776410287e-07,
+      "loss": 0.5872,
+      "step": 67759
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.944995976781001e-07,
+      "loss": 0.5854,
+      "step": 67760
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.944725178725134e-07,
+      "loss": 0.7236,
+      "step": 67761
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.944454382242896e-07,
+      "loss": 0.4858,
+      "step": 67762
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.944183587334495e-07,
+      "loss": 0.4695,
+      "step": 67763
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.943912794000137e-07,
+      "loss": 0.6318,
+      "step": 67764
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.943642002240032e-07,
+      "loss": 0.6895,
+      "step": 67765
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.943371212054384e-07,
+      "loss": 0.6025,
+      "step": 67766
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.943100423443404e-07,
+      "loss": 0.6846,
+      "step": 67767
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.942829636407296e-07,
+      "loss": 0.6436,
+      "step": 67768
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.94255885094627e-07,
+      "loss": 0.7148,
+      "step": 67769
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.942288067060536e-07,
+      "loss": 0.6187,
+      "step": 67770
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.942017284750291e-07,
+      "loss": 0.7285,
+      "step": 67771
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.941746504015751e-07,
+      "loss": 0.6504,
+      "step": 67772
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.941475724857122e-07,
+      "loss": 0.7549,
+      "step": 67773
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.941204947274608e-07,
+      "loss": 0.573,
+      "step": 67774
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.940934171268421e-07,
+      "loss": 0.8271,
+      "step": 67775
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.940663396838764e-07,
+      "loss": 0.6255,
+      "step": 67776
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.940392623985848e-07,
+      "loss": 0.3284,
+      "step": 67777
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.940121852709875e-07,
+      "loss": 0.6875,
+      "step": 67778
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.939851083011062e-07,
+      "loss": 0.6416,
+      "step": 67779
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.939580314889609e-07,
+      "loss": 0.71,
+      "step": 67780
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.939309548345722e-07,
+      "loss": 0.5747,
+      "step": 67781
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.939038783379612e-07,
+      "loss": 0.6016,
+      "step": 67782
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.938768019991484e-07,
+      "loss": 0.7734,
+      "step": 67783
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.938497258181549e-07,
+      "loss": 0.6553,
+      "step": 67784
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.938226497950009e-07,
+      "loss": 0.708,
+      "step": 67785
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.937955739297077e-07,
+      "loss": 0.6504,
+      "step": 67786
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.937684982222954e-07,
+      "loss": 0.6934,
+      "step": 67787
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.937414226727854e-07,
+      "loss": 0.6406,
+      "step": 67788
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.937143472811978e-07,
+      "loss": 0.3868,
+      "step": 67789
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.936872720475545e-07,
+      "loss": 0.4309,
+      "step": 67790
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.936601969718744e-07,
+      "loss": 0.5439,
+      "step": 67791
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.936331220541796e-07,
+      "loss": 0.7354,
+      "step": 67792
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.936060472944903e-07,
+      "loss": 0.604,
+      "step": 67793
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.935789726928275e-07,
+      "loss": 0.6362,
+      "step": 67794
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.935518982492115e-07,
+      "loss": 0.5615,
+      "step": 67795
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.935248239636636e-07,
+      "loss": 0.6143,
+      "step": 67796
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.934977498362039e-07,
+      "loss": 0.375,
+      "step": 67797
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.934706758668539e-07,
+      "loss": 0.5376,
+      "step": 67798
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.934436020556336e-07,
+      "loss": 0.5312,
+      "step": 67799
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.934165284025645e-07,
+      "loss": 0.6016,
+      "step": 67800
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.933894549076666e-07,
+      "loss": 0.6572,
+      "step": 67801
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.933623815709607e-07,
+      "loss": 0.7334,
+      "step": 67802
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.93335308392468e-07,
+      "loss": 0.5815,
+      "step": 67803
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.933082353722086e-07,
+      "loss": 0.627,
+      "step": 67804
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.932811625102038e-07,
+      "loss": 0.5542,
+      "step": 67805
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.93254089806474e-07,
+      "loss": 0.5728,
+      "step": 67806
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.932270172610403e-07,
+      "loss": 0.7939,
+      "step": 67807
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.931999448739228e-07,
+      "loss": 0.6963,
+      "step": 67808
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.931728726451429e-07,
+      "loss": 0.6226,
+      "step": 67809
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.931458005747213e-07,
+      "loss": 0.5237,
+      "step": 67810
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.931187286626781e-07,
+      "loss": 0.623,
+      "step": 67811
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.930916569090342e-07,
+      "loss": 0.645,
+      "step": 67812
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.930645853138107e-07,
+      "loss": 0.6294,
+      "step": 67813
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.93037513877028e-07,
+      "loss": 0.6797,
+      "step": 67814
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.930104425987071e-07,
+      "loss": 0.6572,
+      "step": 67815
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.929833714788683e-07,
+      "loss": 0.606,
+      "step": 67816
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92956300517533e-07,
+      "loss": 0.4856,
+      "step": 67817
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.929292297147212e-07,
+      "loss": 0.6177,
+      "step": 67818
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.929021590704543e-07,
+      "loss": 0.5403,
+      "step": 67819
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.928750885847527e-07,
+      "loss": 0.6963,
+      "step": 67820
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.928480182576369e-07,
+      "loss": 0.8545,
+      "step": 67821
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.928209480891279e-07,
+      "loss": 0.6768,
+      "step": 67822
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.927938780792463e-07,
+      "loss": 0.5874,
+      "step": 67823
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92766808228013e-07,
+      "loss": 0.6172,
+      "step": 67824
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.927397385354485e-07,
+      "loss": 0.7207,
+      "step": 67825
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.927126690015738e-07,
+      "loss": 0.6353,
+      "step": 67826
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.926855996264096e-07,
+      "loss": 0.468,
+      "step": 67827
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.926585304099762e-07,
+      "loss": 0.7275,
+      "step": 67828
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.926314613522949e-07,
+      "loss": 0.6802,
+      "step": 67829
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.926043924533862e-07,
+      "loss": 0.6401,
+      "step": 67830
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.925773237132705e-07,
+      "loss": 0.7842,
+      "step": 67831
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92550255131969e-07,
+      "loss": 0.7559,
+      "step": 67832
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92523186709502e-07,
+      "loss": 0.6484,
+      "step": 67833
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.924961184458907e-07,
+      "loss": 0.6235,
+      "step": 67834
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.924690503411553e-07,
+      "loss": 0.4771,
+      "step": 67835
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.924419823953171e-07,
+      "loss": 0.8535,
+      "step": 67836
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.924149146083963e-07,
+      "loss": 0.7539,
+      "step": 67837
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92387846980414e-07,
+      "loss": 0.4558,
+      "step": 67838
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.923607795113907e-07,
+      "loss": 0.7139,
+      "step": 67839
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.923337122013475e-07,
+      "loss": 0.6636,
+      "step": 67840
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.923066450503046e-07,
+      "loss": 0.8125,
+      "step": 67841
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.922795780582828e-07,
+      "loss": 0.625,
+      "step": 67842
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.922525112253033e-07,
+      "loss": 0.8184,
+      "step": 67843
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92225444551386e-07,
+      "loss": 0.4097,
+      "step": 67844
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.921983780365526e-07,
+      "loss": 0.668,
+      "step": 67845
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.921713116808233e-07,
+      "loss": 0.6499,
+      "step": 67846
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.921442454842187e-07,
+      "loss": 0.6836,
+      "step": 67847
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.921171794467598e-07,
+      "loss": 0.6904,
+      "step": 67848
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.92090113568467e-07,
+      "loss": 0.6812,
+      "step": 67849
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.920630478493617e-07,
+      "loss": 0.6328,
+      "step": 67850
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.920359822894638e-07,
+      "loss": 0.7104,
+      "step": 67851
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.920089168887945e-07,
+      "loss": 0.7354,
+      "step": 67852
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.919818516473743e-07,
+      "loss": 0.5811,
+      "step": 67853
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.919547865652239e-07,
+      "loss": 0.6865,
+      "step": 67854
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.919277216423646e-07,
+      "loss": 1.0908,
+      "step": 67855
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.919006568788162e-07,
+      "loss": 0.7319,
+      "step": 67856
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.918735922746002e-07,
+      "loss": 0.5308,
+      "step": 67857
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.918465278297368e-07,
+      "loss": 0.6011,
+      "step": 67858
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.918194635442471e-07,
+      "loss": 0.7998,
+      "step": 67859
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.91792399418152e-07,
+      "loss": 0.4629,
+      "step": 67860
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.917653354514714e-07,
+      "loss": 0.6553,
+      "step": 67861
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.917382716442267e-07,
+      "loss": 0.7129,
+      "step": 67862
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.917112079964384e-07,
+      "loss": 0.5737,
+      "step": 67863
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.916841445081272e-07,
+      "loss": 0.6396,
+      "step": 67864
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.916570811793138e-07,
+      "loss": 0.7041,
+      "step": 67865
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.91630018010019e-07,
+      "loss": 0.6338,
+      "step": 67866
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.916029550002636e-07,
+      "loss": 0.6636,
+      "step": 67867
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.915758921500682e-07,
+      "loss": 0.4731,
+      "step": 67868
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.915488294594537e-07,
+      "loss": 0.5718,
+      "step": 67869
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.915217669284409e-07,
+      "loss": 0.6055,
+      "step": 67870
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.914947045570496e-07,
+      "loss": 0.6455,
+      "step": 67871
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.914676423453019e-07,
+      "loss": 0.6963,
+      "step": 67872
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.914405802932173e-07,
+      "loss": 0.6475,
+      "step": 67873
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.914135184008173e-07,
+      "loss": 0.6836,
+      "step": 67874
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.913864566681223e-07,
+      "loss": 0.7598,
+      "step": 67875
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.913593950951532e-07,
+      "loss": 0.79,
+      "step": 67876
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.913323336819304e-07,
+      "loss": 0.5923,
+      "step": 67877
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.913052724284751e-07,
+      "loss": 0.7446,
+      "step": 67878
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.912782113348076e-07,
+      "loss": 0.7402,
+      "step": 67879
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.912511504009493e-07,
+      "loss": 0.6152,
+      "step": 67880
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.912240896269198e-07,
+      "loss": 0.6602,
+      "step": 67881
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.911970290127405e-07,
+      "loss": 0.7363,
+      "step": 67882
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.911699685584321e-07,
+      "loss": 0.5771,
+      "step": 67883
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.911429082640153e-07,
+      "loss": 0.5977,
+      "step": 67884
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.911158481295106e-07,
+      "loss": 0.7061,
+      "step": 67885
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.910887881549391e-07,
+      "loss": 0.6387,
+      "step": 67886
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.910617283403209e-07,
+      "loss": 0.6885,
+      "step": 67887
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.910346686856776e-07,
+      "loss": 0.6484,
+      "step": 67888
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.910076091910291e-07,
+      "loss": 0.71,
+      "step": 67889
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.90980549856397e-07,
+      "loss": 0.6885,
+      "step": 67890
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.909534906818012e-07,
+      "loss": 0.7637,
+      "step": 67891
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.909264316672625e-07,
+      "loss": 0.427,
+      "step": 67892
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.90899372812802e-07,
+      "loss": 0.5176,
+      "step": 67893
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.9087231411844e-07,
+      "loss": 0.5967,
+      "step": 67894
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.908452555841978e-07,
+      "loss": 0.7861,
+      "step": 67895
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.908181972100954e-07,
+      "loss": 0.5781,
+      "step": 67896
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.907911389961541e-07,
+      "loss": 0.6318,
+      "step": 67897
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.907640809423943e-07,
+      "loss": 0.7344,
+      "step": 67898
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.90737023048837e-07,
+      "loss": 0.541,
+      "step": 67899
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.90709965315503e-07,
+      "loss": 0.5767,
+      "step": 67900
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.906829077424125e-07,
+      "loss": 0.6112,
+      "step": 67901
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.906558503295863e-07,
+      "loss": 0.7959,
+      "step": 67902
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.906287930770454e-07,
+      "loss": 0.6348,
+      "step": 67903
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.906017359848103e-07,
+      "loss": 0.8193,
+      "step": 67904
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.905746790529019e-07,
+      "loss": 0.6011,
+      "step": 67905
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.905476222813408e-07,
+      "loss": 0.5478,
+      "step": 67906
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.905205656701478e-07,
+      "loss": 0.7168,
+      "step": 67907
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.904935092193435e-07,
+      "loss": 0.7432,
+      "step": 67908
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.904664529289487e-07,
+      "loss": 0.5874,
+      "step": 67909
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.904393967989845e-07,
+      "loss": 0.6001,
+      "step": 67910
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.904123408294707e-07,
+      "loss": 0.6787,
+      "step": 67911
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.903852850204289e-07,
+      "loss": 0.6519,
+      "step": 67912
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.90358229371879e-07,
+      "loss": 0.6943,
+      "step": 67913
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.903311738838425e-07,
+      "loss": 0.6094,
+      "step": 67914
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.903041185563396e-07,
+      "loss": 0.6138,
+      "step": 67915
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.902770633893914e-07,
+      "loss": 0.6323,
+      "step": 67916
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.902500083830181e-07,
+      "loss": 0.7002,
+      "step": 67917
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.902229535372411e-07,
+      "loss": 0.6655,
+      "step": 67918
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.901958988520806e-07,
+      "loss": 0.6328,
+      "step": 67919
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.901688443275577e-07,
+      "loss": 0.7627,
+      "step": 67920
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.901417899636925e-07,
+      "loss": 0.4741,
+      "step": 67921
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.901147357605063e-07,
+      "loss": 0.5591,
+      "step": 67922
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.900876817180193e-07,
+      "loss": 0.5547,
+      "step": 67923
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.900606278362527e-07,
+      "loss": 0.6318,
+      "step": 67924
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.900335741152269e-07,
+      "loss": 0.7256,
+      "step": 67925
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.900065205549629e-07,
+      "loss": 0.7222,
+      "step": 67926
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.899794671554811e-07,
+      "loss": 0.6821,
+      "step": 67927
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.899524139168024e-07,
+      "loss": 0.6064,
+      "step": 67928
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.899253608389474e-07,
+      "loss": 0.5879,
+      "step": 67929
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.898983079219375e-07,
+      "loss": 0.4756,
+      "step": 67930
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.898712551657924e-07,
+      "loss": 0.6572,
+      "step": 67931
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.89844202570533e-07,
+      "loss": 0.5518,
+      "step": 67932
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.898171501361803e-07,
+      "loss": 0.6504,
+      "step": 67933
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.89790097862755e-07,
+      "loss": 0.7275,
+      "step": 67934
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.897630457502779e-07,
+      "loss": 0.5737,
+      "step": 67935
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.897359937987694e-07,
+      "loss": 0.6396,
+      "step": 67936
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.897089420082503e-07,
+      "loss": 0.7153,
+      "step": 67937
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.896818903787417e-07,
+      "loss": 0.6626,
+      "step": 67938
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.896548389102636e-07,
+      "loss": 0.6582,
+      "step": 67939
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.896277876028376e-07,
+      "loss": 0.5659,
+      "step": 67940
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.896007364564838e-07,
+      "loss": 0.5718,
+      "step": 67941
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.895736854712228e-07,
+      "loss": 0.521,
+      "step": 67942
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.895466346470758e-07,
+      "loss": 0.481,
+      "step": 67943
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.895195839840629e-07,
+      "loss": 0.5703,
+      "step": 67944
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.894925334822055e-07,
+      "loss": 0.6797,
+      "step": 67945
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.894654831415238e-07,
+      "loss": 0.7227,
+      "step": 67946
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.894384329620389e-07,
+      "loss": 0.5488,
+      "step": 67947
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.894113829437711e-07,
+      "loss": 0.8799,
+      "step": 67948
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.893843330867415e-07,
+      "loss": 0.7471,
+      "step": 67949
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.893572833909709e-07,
+      "loss": 0.667,
+      "step": 67950
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.893302338564793e-07,
+      "loss": 0.7812,
+      "step": 67951
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.893031844832879e-07,
+      "loss": 0.5312,
+      "step": 67952
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.892761352714176e-07,
+      "loss": 0.7744,
+      "step": 67953
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.892490862208885e-07,
+      "loss": 0.5073,
+      "step": 67954
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.89222037331722e-07,
+      "loss": 0.5698,
+      "step": 67955
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.891949886039382e-07,
+      "loss": 0.5693,
+      "step": 67956
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.891679400375584e-07,
+      "loss": 0.6553,
+      "step": 67957
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.891408916326026e-07,
+      "loss": 0.5183,
+      "step": 67958
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.891138433890923e-07,
+      "loss": 0.7471,
+      "step": 67959
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.890867953070481e-07,
+      "loss": 0.6328,
+      "step": 67960
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.8905974738649e-07,
+      "loss": 0.6597,
+      "step": 67961
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.890326996274394e-07,
+      "loss": 0.7549,
+      "step": 67962
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.890056520299164e-07,
+      "loss": 0.5337,
+      "step": 67963
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.889786045939423e-07,
+      "loss": 0.6572,
+      "step": 67964
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.889515573195375e-07,
+      "loss": 0.7441,
+      "step": 67965
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.889245102067229e-07,
+      "loss": 0.6416,
+      "step": 67966
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.888974632555189e-07,
+      "loss": 0.603,
+      "step": 67967
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.888704164659465e-07,
+      "loss": 0.6787,
+      "step": 67968
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.888433698380268e-07,
+      "loss": 0.5093,
+      "step": 67969
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.888163233717796e-07,
+      "loss": 0.6392,
+      "step": 67970
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.887892770672259e-07,
+      "loss": 0.7178,
+      "step": 67971
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.887622309243867e-07,
+      "loss": 0.6035,
+      "step": 67972
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.887351849432824e-07,
+      "loss": 0.6523,
+      "step": 67973
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.887081391239341e-07,
+      "loss": 0.8076,
+      "step": 67974
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.886810934663619e-07,
+      "loss": 0.644,
+      "step": 67975
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.886540479705872e-07,
+      "loss": 0.531,
+      "step": 67976
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.8862700263663e-07,
+      "loss": 0.6074,
+      "step": 67977
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.885999574645118e-07,
+      "loss": 0.6582,
+      "step": 67978
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.88572912454253e-07,
+      "loss": 0.4626,
+      "step": 67979
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.885458676058739e-07,
+      "loss": 0.6582,
+      "step": 67980
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.885188229193955e-07,
+      "loss": 0.5044,
+      "step": 67981
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.884917783948385e-07,
+      "loss": 0.5811,
+      "step": 67982
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.884647340322237e-07,
+      "loss": 0.5825,
+      "step": 67983
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.884376898315717e-07,
+      "loss": 0.668,
+      "step": 67984
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.884106457929032e-07,
+      "loss": 0.7217,
+      "step": 67985
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.883836019162389e-07,
+      "loss": 0.583,
+      "step": 67986
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.883565582015997e-07,
+      "loss": 0.71,
+      "step": 67987
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.883295146490059e-07,
+      "loss": 0.5361,
+      "step": 67988
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.883024712584792e-07,
+      "loss": 0.8213,
+      "step": 67989
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.882754280300388e-07,
+      "loss": 0.6499,
+      "step": 67990
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.882483849637065e-07,
+      "loss": 0.6914,
+      "step": 67991
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.882213420595023e-07,
+      "loss": 0.6045,
+      "step": 67992
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.881942993174477e-07,
+      "loss": 0.4005,
+      "step": 67993
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.881672567375627e-07,
+      "loss": 0.5811,
+      "step": 67994
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.881402143198684e-07,
+      "loss": 0.6016,
+      "step": 67995
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.881131720643853e-07,
+      "loss": 0.6465,
+      "step": 67996
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.880861299711343e-07,
+      "loss": 0.6206,
+      "step": 67997
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.880590880401359e-07,
+      "loss": 0.7178,
+      "step": 67998
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.880320462714113e-07,
+      "loss": 0.6631,
+      "step": 67999
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.880050046649806e-07,
+      "loss": 0.5186,
+      "step": 68000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.879779632208645e-07,
+      "loss": 0.6104,
+      "step": 68001
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.879509219390842e-07,
+      "loss": 0.7109,
+      "step": 68002
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.879238808196598e-07,
+      "loss": 0.5667,
+      "step": 68003
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.878968398626127e-07,
+      "loss": 0.6494,
+      "step": 68004
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.878697990679629e-07,
+      "loss": 0.5898,
+      "step": 68005
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.878427584357316e-07,
+      "loss": 0.5396,
+      "step": 68006
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.878157179659393e-07,
+      "loss": 0.6377,
+      "step": 68007
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.877886776586069e-07,
+      "loss": 0.7852,
+      "step": 68008
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.877616375137551e-07,
+      "loss": 0.626,
+      "step": 68009
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.877345975314044e-07,
+      "loss": 0.688,
+      "step": 68010
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.877075577115751e-07,
+      "loss": 0.604,
+      "step": 68011
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.876805180542888e-07,
+      "loss": 0.5244,
+      "step": 68012
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.876534785595654e-07,
+      "loss": 0.5737,
+      "step": 68013
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.876264392274263e-07,
+      "loss": 0.5537,
+      "step": 68014
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.875994000578916e-07,
+      "loss": 0.6567,
+      "step": 68015
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.875723610509825e-07,
+      "loss": 0.583,
+      "step": 68016
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.875453222067192e-07,
+      "loss": 0.6035,
+      "step": 68017
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.87518283525123e-07,
+      "loss": 0.5454,
+      "step": 68018
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.874912450062145e-07,
+      "loss": 0.7573,
+      "step": 68019
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.874642066500137e-07,
+      "loss": 0.7012,
+      "step": 68020
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.87437168456542e-07,
+      "loss": 0.6426,
+      "step": 68021
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.874101304258196e-07,
+      "loss": 0.6284,
+      "step": 68022
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.873830925578677e-07,
+      "loss": 0.6377,
+      "step": 68023
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.873560548527066e-07,
+      "loss": 0.5635,
+      "step": 68024
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.873290173103575e-07,
+      "loss": 0.8652,
+      "step": 68025
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.873019799308407e-07,
+      "loss": 0.7598,
+      "step": 68026
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.872749427141768e-07,
+      "loss": 0.6621,
+      "step": 68027
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.872479056603869e-07,
+      "loss": 0.7764,
+      "step": 68028
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.872208687694918e-07,
+      "loss": 0.5459,
+      "step": 68029
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.871938320415114e-07,
+      "loss": 0.345,
+      "step": 68030
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.871667954764671e-07,
+      "loss": 0.5024,
+      "step": 68031
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.871397590743792e-07,
+      "loss": 0.6045,
+      "step": 68032
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.871127228352689e-07,
+      "loss": 0.7168,
+      "step": 68033
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.870856867591563e-07,
+      "loss": 0.6377,
+      "step": 68034
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.870586508460627e-07,
+      "loss": 0.686,
+      "step": 68035
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.87031615096008e-07,
+      "loss": 0.6069,
+      "step": 68036
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.870045795090139e-07,
+      "loss": 0.6335,
+      "step": 68037
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.869775440851002e-07,
+      "loss": 0.5981,
+      "step": 68038
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.869505088242886e-07,
+      "loss": 0.5811,
+      "step": 68039
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.869234737265988e-07,
+      "loss": 0.5469,
+      "step": 68040
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.868964387920519e-07,
+      "loss": 0.665,
+      "step": 68041
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.868694040206687e-07,
+      "loss": 0.6914,
+      "step": 68042
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.868423694124698e-07,
+      "loss": 0.7334,
+      "step": 68043
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.868153349674757e-07,
+      "loss": 0.6328,
+      "step": 68044
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.867883006857074e-07,
+      "loss": 0.6191,
+      "step": 68045
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.867612665671855e-07,
+      "loss": 0.3959,
+      "step": 68046
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.867342326119307e-07,
+      "loss": 0.6914,
+      "step": 68047
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.867071988199637e-07,
+      "loss": 0.7129,
+      "step": 68048
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.866801651913053e-07,
+      "loss": 0.7109,
+      "step": 68049
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.866531317259761e-07,
+      "loss": 0.6885,
+      "step": 68050
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.866260984239966e-07,
+      "loss": 0.6396,
+      "step": 68051
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.865990652853878e-07,
+      "loss": 0.7266,
+      "step": 68052
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.8657203231017e-07,
+      "loss": 0.5115,
+      "step": 68053
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.865449994983645e-07,
+      "loss": 0.7041,
+      "step": 68054
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.865179668499915e-07,
+      "loss": 0.6846,
+      "step": 68055
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.864909343650721e-07,
+      "loss": 0.6919,
+      "step": 68056
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.864639020436265e-07,
+      "loss": 0.6562,
+      "step": 68057
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.864368698856758e-07,
+      "loss": 0.6621,
+      "step": 68058
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.86409837891241e-07,
+      "loss": 0.6099,
+      "step": 68059
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.863828060603419e-07,
+      "loss": 0.4417,
+      "step": 68060
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.863557743929997e-07,
+      "loss": 0.7803,
+      "step": 68061
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.863287428892351e-07,
+      "loss": 0.4756,
+      "step": 68062
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.863017115490685e-07,
+      "loss": 0.7402,
+      "step": 68063
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.862746803725212e-07,
+      "loss": 0.6719,
+      "step": 68064
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.862476493596133e-07,
+      "loss": 0.5491,
+      "step": 68065
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.86220618510366e-07,
+      "loss": 0.4438,
+      "step": 68066
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.861935878247994e-07,
+      "loss": 0.6963,
+      "step": 68067
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.861665573029348e-07,
+      "loss": 0.5903,
+      "step": 68068
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.86139526944793e-07,
+      "loss": 0.4945,
+      "step": 68069
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.861124967503937e-07,
+      "loss": 0.5811,
+      "step": 68070
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.860854667197585e-07,
+      "loss": 0.4714,
+      "step": 68071
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.860584368529077e-07,
+      "loss": 0.8369,
+      "step": 68072
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.860314071498622e-07,
+      "loss": 0.5137,
+      "step": 68073
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.860043776106425e-07,
+      "loss": 0.6396,
+      "step": 68074
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.859773482352696e-07,
+      "loss": 0.707,
+      "step": 68075
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.859503190237638e-07,
+      "loss": 0.8359,
+      "step": 68076
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.859232899761462e-07,
+      "loss": 0.5977,
+      "step": 68077
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.858962610924371e-07,
+      "loss": 0.6343,
+      "step": 68078
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.85869232372658e-07,
+      "loss": 0.751,
+      "step": 68079
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.858422038168283e-07,
+      "loss": 0.5752,
+      "step": 68080
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.858151754249697e-07,
+      "loss": 0.6309,
+      "step": 68081
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 7.857881471971022e-07,
+      "loss": 0.6885,
+      "step": 68082
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.857611191332471e-07,
+      "loss": 0.7949,
+      "step": 68083
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.857340912334249e-07,
+      "loss": 0.6909,
+      "step": 68084
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.857070634976562e-07,
+      "loss": 0.561,
+      "step": 68085
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.856800359259616e-07,
+      "loss": 0.5364,
+      "step": 68086
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.856530085183621e-07,
+      "loss": 0.7686,
+      "step": 68087
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.856259812748782e-07,
+      "loss": 0.5889,
+      "step": 68088
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.85598954195531e-07,
+      "loss": 0.6255,
+      "step": 68089
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.855719272803404e-07,
+      "loss": 0.6816,
+      "step": 68090
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.855449005293276e-07,
+      "loss": 0.5381,
+      "step": 68091
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.855178739425132e-07,
+      "loss": 0.5723,
+      "step": 68092
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.854908475199178e-07,
+      "loss": 0.7227,
+      "step": 68093
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.854638212615624e-07,
+      "loss": 0.75,
+      "step": 68094
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.854367951674673e-07,
+      "loss": 0.7041,
+      "step": 68095
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.854097692376535e-07,
+      "loss": 0.623,
+      "step": 68096
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.853827434721414e-07,
+      "loss": 0.6855,
+      "step": 68097
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.853557178709521e-07,
+      "loss": 0.689,
+      "step": 68098
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.853286924341063e-07,
+      "loss": 0.6841,
+      "step": 68099
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.853016671616242e-07,
+      "loss": 0.6802,
+      "step": 68100
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.852746420535265e-07,
+      "loss": 0.5791,
+      "step": 68101
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.852476171098343e-07,
+      "loss": 0.6465,
+      "step": 68102
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.85220592330568e-07,
+      "loss": 0.6733,
+      "step": 68103
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.851935677157486e-07,
+      "loss": 0.583,
+      "step": 68104
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.851665432653965e-07,
+      "loss": 0.7285,
+      "step": 68105
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.851395189795326e-07,
+      "loss": 0.6992,
+      "step": 68106
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.851124948581773e-07,
+      "loss": 0.4077,
+      "step": 68107
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.850854709013517e-07,
+      "loss": 0.5283,
+      "step": 68108
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.850584471090764e-07,
+      "loss": 0.4645,
+      "step": 68109
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.850314234813716e-07,
+      "loss": 0.709,
+      "step": 68110
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.850044000182586e-07,
+      "loss": 0.5969,
+      "step": 68111
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.849773767197577e-07,
+      "loss": 0.5864,
+      "step": 68112
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.849503535858898e-07,
+      "loss": 0.5864,
+      "step": 68113
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.849233306166754e-07,
+      "loss": 0.6587,
+      "step": 68114
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.848963078121354e-07,
+      "loss": 0.6829,
+      "step": 68115
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.848692851722906e-07,
+      "loss": 0.6641,
+      "step": 68116
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.848422626971611e-07,
+      "loss": 0.5522,
+      "step": 68117
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.848152403867684e-07,
+      "loss": 0.5552,
+      "step": 68118
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.847882182411328e-07,
+      "loss": 0.7305,
+      "step": 68119
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.847611962602747e-07,
+      "loss": 0.6162,
+      "step": 68120
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.847341744442152e-07,
+      "loss": 0.5938,
+      "step": 68121
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.847071527929747e-07,
+      "loss": 0.8057,
+      "step": 68122
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.846801313065741e-07,
+      "loss": 0.5981,
+      "step": 68123
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.846531099850338e-07,
+      "loss": 0.4714,
+      "step": 68124
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.84626088828375e-07,
+      "loss": 0.6084,
+      "step": 68125
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.845990678366179e-07,
+      "loss": 0.6523,
+      "step": 68126
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.845720470097837e-07,
+      "loss": 0.665,
+      "step": 68127
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.845450263478923e-07,
+      "loss": 0.6431,
+      "step": 68128
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.845180058509655e-07,
+      "loss": 0.7139,
+      "step": 68129
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.844909855190231e-07,
+      "loss": 0.6089,
+      "step": 68130
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.844639653520858e-07,
+      "loss": 0.5869,
+      "step": 68131
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.844369453501748e-07,
+      "loss": 0.492,
+      "step": 68132
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.844099255133105e-07,
+      "loss": 0.6753,
+      "step": 68133
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.843829058415134e-07,
+      "loss": 0.644,
+      "step": 68134
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.843558863348045e-07,
+      "loss": 0.6504,
+      "step": 68135
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.843288669932043e-07,
+      "loss": 0.6504,
+      "step": 68136
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.843018478167338e-07,
+      "loss": 0.6694,
+      "step": 68137
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.84274828805413e-07,
+      "loss": 0.6128,
+      "step": 68138
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.842478099592638e-07,
+      "loss": 0.3174,
+      "step": 68139
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.842207912783058e-07,
+      "loss": 0.5105,
+      "step": 68140
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.841937727625597e-07,
+      "loss": 0.6016,
+      "step": 68141
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.841667544120468e-07,
+      "loss": 0.5889,
+      "step": 68142
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.841397362267873e-07,
+      "loss": 0.5166,
+      "step": 68143
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.841127182068023e-07,
+      "loss": 0.4692,
+      "step": 68144
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.84085700352112e-07,
+      "loss": 0.5498,
+      "step": 68145
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.840586826627375e-07,
+      "loss": 0.6562,
+      "step": 68146
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.840316651386992e-07,
+      "loss": 0.6309,
+      "step": 68147
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.840046477800182e-07,
+      "loss": 0.7202,
+      "step": 68148
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.83977630586715e-07,
+      "loss": 0.8574,
+      "step": 68149
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.839506135588101e-07,
+      "loss": 0.707,
+      "step": 68150
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.83923596696324e-07,
+      "loss": 0.7227,
+      "step": 68151
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.838965799992779e-07,
+      "loss": 0.6016,
+      "step": 68152
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.83869563467692e-07,
+      "loss": 0.8447,
+      "step": 68153
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.838425471015875e-07,
+      "loss": 0.4248,
+      "step": 68154
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.838155309009846e-07,
+      "loss": 0.7051,
+      "step": 68155
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.837885148659045e-07,
+      "loss": 0.4003,
+      "step": 68156
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.837614989963672e-07,
+      "loss": 0.6201,
+      "step": 68157
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.837344832923941e-07,
+      "loss": 0.6118,
+      "step": 68158
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.837074677540057e-07,
+      "loss": 0.4536,
+      "step": 68159
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.836804523812223e-07,
+      "loss": 0.5183,
+      "step": 68160
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.836534371740648e-07,
+      "loss": 0.6338,
+      "step": 68161
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.836264221325539e-07,
+      "loss": 0.6328,
+      "step": 68162
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.835994072567105e-07,
+      "loss": 0.6807,
+      "step": 68163
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.835723925465547e-07,
+      "loss": 0.7041,
+      "step": 68164
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.835453780021079e-07,
+      "loss": 0.5674,
+      "step": 68165
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.835183636233903e-07,
+      "loss": 0.6104,
+      "step": 68166
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.834913494104228e-07,
+      "loss": 0.6709,
+      "step": 68167
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.834643353632259e-07,
+      "loss": 0.6816,
+      "step": 68168
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.834373214818211e-07,
+      "loss": 0.7207,
+      "step": 68169
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.834103077662276e-07,
+      "loss": 0.6968,
+      "step": 68170
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.833832942164672e-07,
+      "loss": 0.6348,
+      "step": 68171
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.8335628083256e-07,
+      "loss": 0.5542,
+      "step": 68172
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.833292676145271e-07,
+      "loss": 0.5869,
+      "step": 68173
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.833022545623889e-07,
+      "loss": 0.5481,
+      "step": 68174
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.832752416761663e-07,
+      "loss": 0.7207,
+      "step": 68175
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.832482289558798e-07,
+      "loss": 0.5957,
+      "step": 68176
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.832212164015501e-07,
+      "loss": 0.5664,
+      "step": 68177
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.831942040131981e-07,
+      "loss": 0.8555,
+      "step": 68178
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.831671917908445e-07,
+      "loss": 0.5308,
+      "step": 68179
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.831401797345096e-07,
+      "loss": 0.6807,
+      "step": 68180
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.831131678442142e-07,
+      "loss": 0.7178,
+      "step": 68181
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.830861561199793e-07,
+      "loss": 0.7119,
+      "step": 68182
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.830591445618251e-07,
+      "loss": 0.7598,
+      "step": 68183
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.830321331697727e-07,
+      "loss": 0.5439,
+      "step": 68184
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.830051219438425e-07,
+      "loss": 0.8105,
+      "step": 68185
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.829781108840554e-07,
+      "loss": 0.6992,
+      "step": 68186
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.829510999904318e-07,
+      "loss": 0.6885,
+      "step": 68187
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.829240892629929e-07,
+      "loss": 0.7681,
+      "step": 68188
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.828970787017591e-07,
+      "loss": 0.4771,
+      "step": 68189
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.82870068306751e-07,
+      "loss": 0.543,
+      "step": 68190
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.828430580779889e-07,
+      "loss": 0.6592,
+      "step": 68191
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.828160480154942e-07,
+      "loss": 0.6592,
+      "step": 68192
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.82789038119287e-07,
+      "loss": 0.583,
+      "step": 68193
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.827620283893885e-07,
+      "loss": 0.5742,
+      "step": 68194
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.827350188258188e-07,
+      "loss": 0.7373,
+      "step": 68195
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.827080094285993e-07,
+      "loss": 0.5669,
+      "step": 68196
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.826810001977499e-07,
+      "loss": 0.5103,
+      "step": 68197
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.82653991133292e-07,
+      "loss": 0.6616,
+      "step": 68198
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.826269822352461e-07,
+      "loss": 0.6396,
+      "step": 68199
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.825999735036324e-07,
+      "loss": 0.563,
+      "step": 68200
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.825729649384721e-07,
+      "loss": 0.5472,
+      "step": 68201
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.825459565397853e-07,
+      "loss": 0.687,
+      "step": 68202
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.825189483075934e-07,
+      "loss": 0.7207,
+      "step": 68203
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.824919402419165e-07,
+      "loss": 0.4476,
+      "step": 68204
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.824649323427758e-07,
+      "loss": 0.748,
+      "step": 68205
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.824379246101916e-07,
+      "loss": 0.5908,
+      "step": 68206
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.824109170441845e-07,
+      "loss": 0.5562,
+      "step": 68207
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.823839096447755e-07,
+      "loss": 0.5332,
+      "step": 68208
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.823569024119855e-07,
+      "loss": 0.6729,
+      "step": 68209
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.823298953458343e-07,
+      "loss": 0.6772,
+      "step": 68210
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.823028884463433e-07,
+      "loss": 0.6113,
+      "step": 68211
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.822758817135328e-07,
+      "loss": 0.7334,
+      "step": 68212
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.822488751474239e-07,
+      "loss": 0.7144,
+      "step": 68213
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.822218687480366e-07,
+      "loss": 0.6479,
+      "step": 68214
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.821948625153924e-07,
+      "loss": 0.7354,
+      "step": 68215
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.821678564495114e-07,
+      "loss": 0.7207,
+      "step": 68216
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.821408505504146e-07,
+      "loss": 0.6772,
+      "step": 68217
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.821138448181222e-07,
+      "loss": 0.6611,
+      "step": 68218
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.820868392526557e-07,
+      "loss": 0.8555,
+      "step": 68219
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.820598338540351e-07,
+      "loss": 0.6133,
+      "step": 68220
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.820328286222812e-07,
+      "loss": 0.7725,
+      "step": 68221
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.820058235574147e-07,
+      "loss": 0.5161,
+      "step": 68222
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.819788186594565e-07,
+      "loss": 0.6245,
+      "step": 68223
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.819518139284268e-07,
+      "loss": 0.6196,
+      "step": 68224
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.819248093643468e-07,
+      "loss": 0.6121,
+      "step": 68225
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.818978049672368e-07,
+      "loss": 0.707,
+      "step": 68226
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.818708007371178e-07,
+      "loss": 0.5996,
+      "step": 68227
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.8184379667401e-07,
+      "loss": 0.8438,
+      "step": 68228
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.818167927779349e-07,
+      "loss": 0.6211,
+      "step": 68229
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.817897890489123e-07,
+      "loss": 0.6709,
+      "step": 68230
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.817627854869631e-07,
+      "loss": 0.6362,
+      "step": 68231
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.817357820921083e-07,
+      "loss": 0.601,
+      "step": 68232
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.817087788643683e-07,
+      "loss": 0.7339,
+      "step": 68233
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.816817758037639e-07,
+      "loss": 0.6357,
+      "step": 68234
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.816547729103155e-07,
+      "loss": 0.4135,
+      "step": 68235
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.816277701840444e-07,
+      "loss": 0.6069,
+      "step": 68236
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.816007676249704e-07,
+      "loss": 0.7012,
+      "step": 68237
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.81573765233115e-07,
+      "loss": 0.6396,
+      "step": 68238
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.815467630084988e-07,
+      "loss": 0.7627,
+      "step": 68239
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.815197609511419e-07,
+      "loss": 0.4907,
+      "step": 68240
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.814927590610651e-07,
+      "loss": 0.502,
+      "step": 68241
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.814657573382894e-07,
+      "loss": 0.7959,
+      "step": 68242
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.814387557828352e-07,
+      "loss": 0.6279,
+      "step": 68243
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.814117543947234e-07,
+      "loss": 0.6895,
+      "step": 68244
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.813847531739744e-07,
+      "loss": 0.5688,
+      "step": 68245
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.813577521206092e-07,
+      "loss": 0.4875,
+      "step": 68246
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.813307512346483e-07,
+      "loss": 0.6187,
+      "step": 68247
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.813037505161124e-07,
+      "loss": 0.5996,
+      "step": 68248
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.812767499650225e-07,
+      "loss": 0.7197,
+      "step": 68249
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.812497495813984e-07,
+      "loss": 0.6675,
+      "step": 68250
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.812227493652616e-07,
+      "loss": 0.6201,
+      "step": 68251
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.811957493166322e-07,
+      "loss": 0.5942,
+      "step": 68252
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.811687494355315e-07,
+      "loss": 0.5859,
+      "step": 68253
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.811417497219794e-07,
+      "loss": 0.6558,
+      "step": 68254
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.811147501759974e-07,
+      "loss": 0.5127,
+      "step": 68255
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.810877507976055e-07,
+      "loss": 0.7168,
+      "step": 68256
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.810607515868248e-07,
+      "loss": 0.7295,
+      "step": 68257
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.810337525436758e-07,
+      "loss": 0.3813,
+      "step": 68258
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.810067536681796e-07,
+      "loss": 0.5516,
+      "step": 68259
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.80979754960356e-07,
+      "loss": 0.6631,
+      "step": 68260
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.809527564202261e-07,
+      "loss": 0.5303,
+      "step": 68261
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.809257580478106e-07,
+      "loss": 0.6118,
+      "step": 68262
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.808987598431302e-07,
+      "loss": 0.7227,
+      "step": 68263
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.808717618062055e-07,
+      "loss": 0.8096,
+      "step": 68264
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.808447639370575e-07,
+      "loss": 0.5527,
+      "step": 68265
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.808177662357062e-07,
+      "loss": 0.5879,
+      "step": 68266
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.80790768702173e-07,
+      "loss": 0.7275,
+      "step": 68267
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.807637713364783e-07,
+      "loss": 0.6421,
+      "step": 68268
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.807367741386423e-07,
+      "loss": 0.7393,
+      "step": 68269
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.807097771086864e-07,
+      "loss": 0.5476,
+      "step": 68270
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.806827802466306e-07,
+      "loss": 0.6787,
+      "step": 68271
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.806557835524962e-07,
+      "loss": 0.7422,
+      "step": 68272
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.806287870263034e-07,
+      "loss": 0.4905,
+      "step": 68273
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.806017906680732e-07,
+      "loss": 0.6553,
+      "step": 68274
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.805747944778258e-07,
+      "loss": 0.5889,
+      "step": 68275
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.805477984555826e-07,
+      "loss": 0.5552,
+      "step": 68276
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.805208026013636e-07,
+      "loss": 0.7104,
+      "step": 68277
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.804938069151903e-07,
+      "loss": 0.6284,
+      "step": 68278
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.804668113970822e-07,
+      "loss": 0.6123,
+      "step": 68279
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.804398160470608e-07,
+      "loss": 0.8096,
+      "step": 68280
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.804128208651462e-07,
+      "loss": 0.6104,
+      "step": 68281
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.803858258513598e-07,
+      "loss": 0.6494,
+      "step": 68282
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.803588310057216e-07,
+      "loss": 0.6484,
+      "step": 68283
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.803318363282527e-07,
+      "loss": 0.5962,
+      "step": 68284
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.803048418189734e-07,
+      "loss": 0.7168,
+      "step": 68285
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.802778474779049e-07,
+      "loss": 0.625,
+      "step": 68286
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.802508533050673e-07,
+      "loss": 0.7275,
+      "step": 68287
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.802238593004819e-07,
+      "loss": 0.6875,
+      "step": 68288
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.801968654641688e-07,
+      "loss": 0.562,
+      "step": 68289
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.801698717961485e-07,
+      "loss": 0.6885,
+      "step": 68290
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.801428782964424e-07,
+      "loss": 0.5723,
+      "step": 68291
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.801158849650706e-07,
+      "loss": 0.626,
+      "step": 68292
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.80088891802054e-07,
+      "loss": 0.8223,
+      "step": 68293
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.800618988074132e-07,
+      "loss": 0.6973,
+      "step": 68294
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.80034905981169e-07,
+      "loss": 0.7285,
+      "step": 68295
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.80007913323342e-07,
+      "loss": 0.5414,
+      "step": 68296
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.799809208339525e-07,
+      "loss": 0.6592,
+      "step": 68297
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.799539285130221e-07,
+      "loss": 0.7637,
+      "step": 68298
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.799269363605705e-07,
+      "loss": 0.7119,
+      "step": 68299
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.798999443766185e-07,
+      "loss": 0.5776,
+      "step": 68300
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.798729525611874e-07,
+      "loss": 0.561,
+      "step": 68301
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.798459609142971e-07,
+      "loss": 0.5859,
+      "step": 68302
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.798189694359688e-07,
+      "loss": 0.666,
+      "step": 68303
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.797919781262229e-07,
+      "loss": 0.5781,
+      "step": 68304
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.797649869850802e-07,
+      "loss": 0.7363,
+      "step": 68305
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.797379960125613e-07,
+      "loss": 0.5952,
+      "step": 68306
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.79711005208687e-07,
+      "loss": 0.667,
+      "step": 68307
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.796840145734781e-07,
+      "loss": 0.7256,
+      "step": 68308
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.796570241069545e-07,
+      "loss": 0.5245,
+      "step": 68309
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.796300338091378e-07,
+      "loss": 0.6509,
+      "step": 68310
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.796030436800479e-07,
+      "loss": 0.7451,
+      "step": 68311
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.795760537197059e-07,
+      "loss": 0.6636,
+      "step": 68312
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.795490639281325e-07,
+      "loss": 0.4402,
+      "step": 68313
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.795220743053481e-07,
+      "loss": 0.7646,
+      "step": 68314
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.794950848513737e-07,
+      "loss": 0.6086,
+      "step": 68315
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.794680955662295e-07,
+      "loss": 0.6392,
+      "step": 68316
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.794411064499366e-07,
+      "loss": 0.5845,
+      "step": 68317
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.794141175025159e-07,
+      "loss": 0.7002,
+      "step": 68318
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.793871287239871e-07,
+      "loss": 0.6851,
+      "step": 68319
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.793601401143717e-07,
+      "loss": 0.5508,
+      "step": 68320
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.793331516736899e-07,
+      "loss": 0.6938,
+      "step": 68321
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.793061634019628e-07,
+      "loss": 0.7471,
+      "step": 68322
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.792791752992105e-07,
+      "loss": 0.6738,
+      "step": 68323
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.792521873654543e-07,
+      "loss": 0.6689,
+      "step": 68324
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.792251996007142e-07,
+      "loss": 0.6855,
+      "step": 68325
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.791982120050115e-07,
+      "loss": 0.5479,
+      "step": 68326
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.791712245783664e-07,
+      "loss": 0.5898,
+      "step": 68327
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.791442373208001e-07,
+      "loss": 0.6953,
+      "step": 68328
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.791172502323329e-07,
+      "loss": 0.5791,
+      "step": 68329
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.790902633129853e-07,
+      "loss": 0.5342,
+      "step": 68330
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.790632765627779e-07,
+      "loss": 0.7295,
+      "step": 68331
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.790362899817317e-07,
+      "loss": 0.7012,
+      "step": 68332
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.790093035698672e-07,
+      "loss": 0.5608,
+      "step": 68333
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.789823173272053e-07,
+      "loss": 0.6606,
+      "step": 68334
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.789553312537662e-07,
+      "loss": 0.4453,
+      "step": 68335
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.789283453495712e-07,
+      "loss": 0.5164,
+      "step": 68336
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.789013596146402e-07,
+      "loss": 0.6108,
+      "step": 68337
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.788743740489948e-07,
+      "loss": 0.5283,
+      "step": 68338
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.788473886526549e-07,
+      "loss": 0.6841,
+      "step": 68339
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.788204034256411e-07,
+      "loss": 0.6611,
+      "step": 68340
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.787934183679746e-07,
+      "loss": 0.6162,
+      "step": 68341
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.787664334796757e-07,
+      "loss": 0.5381,
+      "step": 68342
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.787394487607653e-07,
+      "loss": 0.4832,
+      "step": 68343
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.787124642112637e-07,
+      "loss": 0.6152,
+      "step": 68344
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.786854798311921e-07,
+      "loss": 0.7998,
+      "step": 68345
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.786584956205704e-07,
+      "loss": 0.4814,
+      "step": 68346
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.786315115794202e-07,
+      "loss": 0.6318,
+      "step": 68347
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.786045277077617e-07,
+      "loss": 0.6514,
+      "step": 68348
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.785775440056154e-07,
+      "loss": 0.5554,
+      "step": 68349
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.785505604730018e-07,
+      "loss": 0.6201,
+      "step": 68350
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.785235771099422e-07,
+      "loss": 0.4951,
+      "step": 68351
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.784965939164566e-07,
+      "loss": 0.6858,
+      "step": 68352
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.784696108925663e-07,
+      "loss": 0.6284,
+      "step": 68353
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.784426280382913e-07,
+      "loss": 0.6982,
+      "step": 68354
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.784156453536528e-07,
+      "loss": 0.5186,
+      "step": 68355
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.783886628386711e-07,
+      "loss": 0.4883,
+      "step": 68356
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.783616804933673e-07,
+      "loss": 0.6211,
+      "step": 68357
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.783346983177619e-07,
+      "loss": 0.5693,
+      "step": 68358
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.78307716311875e-07,
+      "loss": 0.7383,
+      "step": 68359
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.782807344757278e-07,
+      "loss": 0.8438,
+      "step": 68360
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.782537528093407e-07,
+      "loss": 0.5786,
+      "step": 68361
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.782267713127349e-07,
+      "loss": 0.6323,
+      "step": 68362
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.781997899859301e-07,
+      "loss": 0.5576,
+      "step": 68363
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.781728088289481e-07,
+      "loss": 0.6763,
+      "step": 68364
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.781458278418085e-07,
+      "loss": 0.8428,
+      "step": 68365
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.781188470245328e-07,
+      "loss": 0.541,
+      "step": 68366
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.780918663771409e-07,
+      "loss": 0.6938,
+      "step": 68367
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.780648858996548e-07,
+      "loss": 0.5234,
+      "step": 68368
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.780379055920934e-07,
+      "loss": 0.6323,
+      "step": 68369
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.780109254544783e-07,
+      "loss": 0.52,
+      "step": 68370
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.7798394548683e-07,
+      "loss": 0.8574,
+      "step": 68371
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.779569656891693e-07,
+      "loss": 0.792,
+      "step": 68372
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.779299860615165e-07,
+      "loss": 0.6543,
+      "step": 68373
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.779030066038928e-07,
+      "loss": 0.7212,
+      "step": 68374
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.778760273163182e-07,
+      "loss": 0.8467,
+      "step": 68375
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.77849048198814e-07,
+      "loss": 0.6455,
+      "step": 68376
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.778220692514003e-07,
+      "loss": 0.7656,
+      "step": 68377
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.777950904740985e-07,
+      "loss": 0.6016,
+      "step": 68378
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.777681118669286e-07,
+      "loss": 0.6846,
+      "step": 68379
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.777411334299113e-07,
+      "loss": 0.6768,
+      "step": 68380
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.777141551630675e-07,
+      "loss": 0.7412,
+      "step": 68381
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.776871770664175e-07,
+      "loss": 0.7158,
+      "step": 68382
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.776601991399825e-07,
+      "loss": 0.665,
+      "step": 68383
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.776332213837826e-07,
+      "loss": 0.6797,
+      "step": 68384
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.776062437978389e-07,
+      "loss": 0.6904,
+      "step": 68385
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.77579266382172e-07,
+      "loss": 0.6182,
+      "step": 68386
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.775522891368021e-07,
+      "loss": 0.6357,
+      "step": 68387
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.775253120617506e-07,
+      "loss": 0.5457,
+      "step": 68388
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.774983351570376e-07,
+      "loss": 0.6377,
+      "step": 68389
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.774713584226838e-07,
+      "loss": 0.709,
+      "step": 68390
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.774443818587098e-07,
+      "loss": 0.7764,
+      "step": 68391
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.774174054651366e-07,
+      "loss": 0.7256,
+      "step": 68392
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.773904292419846e-07,
+      "loss": 0.5327,
+      "step": 68393
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.773634531892743e-07,
+      "loss": 0.4839,
+      "step": 68394
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.773364773070269e-07,
+      "loss": 0.5215,
+      "step": 68395
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.773095015952625e-07,
+      "loss": 0.7676,
+      "step": 68396
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.772825260540021e-07,
+      "loss": 0.4886,
+      "step": 68397
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.772555506832665e-07,
+      "loss": 0.5508,
+      "step": 68398
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.772285754830756e-07,
+      "loss": 0.7354,
+      "step": 68399
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.77201600453451e-07,
+      "loss": 0.4409,
+      "step": 68400
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.771746255944124e-07,
+      "loss": 0.5747,
+      "step": 68401
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.771476509059812e-07,
+      "loss": 0.4113,
+      "step": 68402
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.771206763881779e-07,
+      "loss": 0.5767,
+      "step": 68403
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.770937020410226e-07,
+      "loss": 0.5342,
+      "step": 68404
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.770667278645368e-07,
+      "loss": 0.6182,
+      "step": 68405
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.770397538587407e-07,
+      "loss": 0.6143,
+      "step": 68406
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.77012780023655e-07,
+      "loss": 0.4905,
+      "step": 68407
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.769858063593007e-07,
+      "loss": 0.5015,
+      "step": 68408
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.769588328656975e-07,
+      "loss": 0.7246,
+      "step": 68409
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.76931859542867e-07,
+      "loss": 0.6299,
+      "step": 68410
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.769048863908293e-07,
+      "loss": 0.5203,
+      "step": 68411
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.768779134096055e-07,
+      "loss": 0.6934,
+      "step": 68412
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.768509405992157e-07,
+      "loss": 0.6216,
+      "step": 68413
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.768239679596812e-07,
+      "loss": 0.6433,
+      "step": 68414
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.767969954910221e-07,
+      "loss": 0.6426,
+      "step": 68415
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.767700231932595e-07,
+      "loss": 0.6133,
+      "step": 68416
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.767430510664135e-07,
+      "loss": 0.4834,
+      "step": 68417
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.767160791105056e-07,
+      "loss": 0.52,
+      "step": 68418
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.766891073255558e-07,
+      "loss": 0.5486,
+      "step": 68419
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.766621357115847e-07,
+      "loss": 0.6865,
+      "step": 68420
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.766351642686131e-07,
+      "loss": 0.6743,
+      "step": 68421
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.766081929966619e-07,
+      "loss": 0.5977,
+      "step": 68422
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.765812218957513e-07,
+      "loss": 0.6167,
+      "step": 68423
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.765542509659022e-07,
+      "loss": 0.5986,
+      "step": 68424
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.765272802071352e-07,
+      "loss": 0.7686,
+      "step": 68425
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.765003096194711e-07,
+      "loss": 0.3273,
+      "step": 68426
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.764733392029303e-07,
+      "loss": 0.5986,
+      "step": 68427
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.764463689575341e-07,
+      "loss": 0.5876,
+      "step": 68428
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.764193988833022e-07,
+      "loss": 0.5791,
+      "step": 68429
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.763924289802557e-07,
+      "loss": 0.5312,
+      "step": 68430
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.763654592484152e-07,
+      "loss": 0.562,
+      "step": 68431
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.763384896878014e-07,
+      "loss": 0.6992,
+      "step": 68432
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.763115202984351e-07,
+      "loss": 0.5603,
+      "step": 68433
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.762845510803364e-07,
+      "loss": 0.6426,
+      "step": 68434
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.762575820335268e-07,
+      "loss": 0.4414,
+      "step": 68435
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.76230613158026e-07,
+      "loss": 0.5571,
+      "step": 68436
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.762036444538556e-07,
+      "loss": 0.6206,
+      "step": 68437
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.761766759210358e-07,
+      "loss": 0.6001,
+      "step": 68438
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.761497075595872e-07,
+      "loss": 0.6289,
+      "step": 68439
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.761227393695301e-07,
+      "loss": 0.6924,
+      "step": 68440
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.760957713508858e-07,
+      "loss": 0.7832,
+      "step": 68441
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.760688035036745e-07,
+      "loss": 0.6013,
+      "step": 68442
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.760418358279172e-07,
+      "loss": 0.6182,
+      "step": 68443
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.76014868323634e-07,
+      "loss": 0.6772,
+      "step": 68444
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.759879009908463e-07,
+      "loss": 0.5093,
+      "step": 68445
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.759609338295741e-07,
+      "loss": 0.625,
+      "step": 68446
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.759339668398386e-07,
+      "loss": 0.7437,
+      "step": 68447
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.759070000216604e-07,
+      "loss": 0.5723,
+      "step": 68448
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.758800333750594e-07,
+      "loss": 0.6255,
+      "step": 68449
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75853066900057e-07,
+      "loss": 0.5967,
+      "step": 68450
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.758261005966733e-07,
+      "loss": 0.7109,
+      "step": 68451
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.757991344649295e-07,
+      "loss": 0.6597,
+      "step": 68452
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.757721685048457e-07,
+      "loss": 0.5859,
+      "step": 68453
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.757452027164431e-07,
+      "loss": 0.4956,
+      "step": 68454
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.757182370997421e-07,
+      "loss": 0.6785,
+      "step": 68455
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.756912716547632e-07,
+      "loss": 0.522,
+      "step": 68456
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.756643063815272e-07,
+      "loss": 0.6562,
+      "step": 68457
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.756373412800553e-07,
+      "loss": 0.4312,
+      "step": 68458
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75610376350367e-07,
+      "loss": 0.7158,
+      "step": 68459
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.755834115924837e-07,
+      "loss": 0.7227,
+      "step": 68460
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.755564470064255e-07,
+      "loss": 0.7871,
+      "step": 68461
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.755294825922137e-07,
+      "loss": 0.8496,
+      "step": 68462
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.755025183498685e-07,
+      "loss": 0.5986,
+      "step": 68463
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.754755542794109e-07,
+      "loss": 0.6572,
+      "step": 68464
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75448590380861e-07,
+      "loss": 0.4089,
+      "step": 68465
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.754216266542401e-07,
+      "loss": 0.6069,
+      "step": 68466
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.753946630995684e-07,
+      "loss": 0.5342,
+      "step": 68467
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.753676997168671e-07,
+      "loss": 0.6133,
+      "step": 68468
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.753407365061562e-07,
+      "loss": 0.7705,
+      "step": 68469
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.753137734674563e-07,
+      "loss": 0.4678,
+      "step": 68470
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.752868106007885e-07,
+      "loss": 0.7144,
+      "step": 68471
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75259847906173e-07,
+      "loss": 0.5518,
+      "step": 68472
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.75232885383631e-07,
+      "loss": 0.6426,
+      "step": 68473
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.752059230331826e-07,
+      "loss": 0.8447,
+      "step": 68474
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.75178960854849e-07,
+      "loss": 0.533,
+      "step": 68475
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.751519988486504e-07,
+      "loss": 0.5693,
+      "step": 68476
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.751250370146074e-07,
+      "loss": 0.5815,
+      "step": 68477
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.750980753527413e-07,
+      "loss": 0.6733,
+      "step": 68478
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.750711138630721e-07,
+      "loss": 0.5713,
+      "step": 68479
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.750441525456203e-07,
+      "loss": 0.8271,
+      "step": 68480
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.75017191400407e-07,
+      "loss": 0.6704,
+      "step": 68481
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.749902304274526e-07,
+      "loss": 0.5469,
+      "step": 68482
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.74963269626778e-07,
+      "loss": 0.6138,
+      "step": 68483
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.749363089984033e-07,
+      "loss": 0.5859,
+      "step": 68484
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7490934854235e-07,
+      "loss": 0.5728,
+      "step": 68485
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.74882388258638e-07,
+      "loss": 0.5774,
+      "step": 68486
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.748554281472883e-07,
+      "loss": 0.6318,
+      "step": 68487
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.748284682083217e-07,
+      "loss": 0.7554,
+      "step": 68488
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.748015084417582e-07,
+      "loss": 0.7266,
+      "step": 68489
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.74774548847619e-07,
+      "loss": 0.6758,
+      "step": 68490
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.747475894259243e-07,
+      "loss": 0.6328,
+      "step": 68491
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.747206301766955e-07,
+      "loss": 0.4814,
+      "step": 68492
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.746936710999524e-07,
+      "loss": 0.5305,
+      "step": 68493
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.74666712195716e-07,
+      "loss": 0.5405,
+      "step": 68494
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.746397534640071e-07,
+      "loss": 0.6377,
+      "step": 68495
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.74612794904846e-07,
+      "loss": 0.5796,
+      "step": 68496
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.745858365182535e-07,
+      "loss": 0.6104,
+      "step": 68497
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.745588783042508e-07,
+      "loss": 0.5522,
+      "step": 68498
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.745319202628574e-07,
+      "loss": 0.6934,
+      "step": 68499
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.745049623940948e-07,
+      "loss": 0.626,
+      "step": 68500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.744780046979832e-07,
+      "loss": 0.6519,
+      "step": 68501
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.744510471745436e-07,
+      "loss": 0.5952,
+      "step": 68502
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.744240898237963e-07,
+      "loss": 0.7129,
+      "step": 68503
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.743971326457622e-07,
+      "loss": 0.6621,
+      "step": 68504
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.743701756404617e-07,
+      "loss": 0.5605,
+      "step": 68505
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.743432188079157e-07,
+      "loss": 0.5171,
+      "step": 68506
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.743162621481447e-07,
+      "loss": 0.6543,
+      "step": 68507
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.742893056611697e-07,
+      "loss": 0.5278,
+      "step": 68508
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.742623493470107e-07,
+      "loss": 0.5859,
+      "step": 68509
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.742353932056885e-07,
+      "loss": 0.4253,
+      "step": 68510
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.742084372372241e-07,
+      "loss": 0.7432,
+      "step": 68511
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.741814814416379e-07,
+      "loss": 0.4819,
+      "step": 68512
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.741545258189503e-07,
+      "loss": 0.6035,
+      "step": 68513
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.741275703691826e-07,
+      "loss": 0.4594,
+      "step": 68514
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.741006150923546e-07,
+      "loss": 0.6592,
+      "step": 68515
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.740736599884876e-07,
+      "loss": 0.6475,
+      "step": 68516
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.740467050576019e-07,
+      "loss": 0.7832,
+      "step": 68517
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.740197502997186e-07,
+      "loss": 0.6426,
+      "step": 68518
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.739927957148577e-07,
+      "loss": 0.6494,
+      "step": 68519
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.739658413030401e-07,
+      "loss": 0.5435,
+      "step": 68520
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.739388870642867e-07,
+      "loss": 0.6865,
+      "step": 68521
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.739119329986174e-07,
+      "loss": 0.3762,
+      "step": 68522
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.738849791060538e-07,
+      "loss": 0.458,
+      "step": 68523
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.738580253866158e-07,
+      "loss": 0.7446,
+      "step": 68524
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.738310718403245e-07,
+      "loss": 0.5679,
+      "step": 68525
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.738041184672002e-07,
+      "loss": 0.6758,
+      "step": 68526
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.737771652672637e-07,
+      "loss": 0.8613,
+      "step": 68527
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.737502122405362e-07,
+      "loss": 0.6943,
+      "step": 68528
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.737232593870372e-07,
+      "loss": 0.5479,
+      "step": 68529
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.736963067067878e-07,
+      "loss": 0.4258,
+      "step": 68530
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.736693541998089e-07,
+      "loss": 0.6787,
+      "step": 68531
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.73642401866121e-07,
+      "loss": 0.6311,
+      "step": 68532
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.736154497057447e-07,
+      "loss": 0.6299,
+      "step": 68533
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.735884977187003e-07,
+      "loss": 0.5732,
+      "step": 68534
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.735615459050092e-07,
+      "loss": 0.6084,
+      "step": 68535
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.735345942646913e-07,
+      "loss": 0.6846,
+      "step": 68536
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.735076427977679e-07,
+      "loss": 0.6455,
+      "step": 68537
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.734806915042594e-07,
+      "loss": 0.627,
+      "step": 68538
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.734537403841859e-07,
+      "loss": 0.6641,
+      "step": 68539
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.734267894375687e-07,
+      "loss": 0.6553,
+      "step": 68540
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.733998386644279e-07,
+      "loss": 0.5498,
+      "step": 68541
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.733728880647848e-07,
+      "loss": 0.5483,
+      "step": 68542
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.733459376386592e-07,
+      "loss": 0.6909,
+      "step": 68543
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.733189873860726e-07,
+      "loss": 0.5625,
+      "step": 68544
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.732920373070449e-07,
+      "loss": 0.6982,
+      "step": 68545
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.732650874015973e-07,
+      "loss": 0.8184,
+      "step": 68546
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7323813766975e-07,
+      "loss": 0.7666,
+      "step": 68547
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.732111881115245e-07,
+      "loss": 0.3745,
+      "step": 68548
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7318423872694e-07,
+      "loss": 0.4338,
+      "step": 68549
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.731572895160183e-07,
+      "loss": 0.5056,
+      "step": 68550
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.731303404787793e-07,
+      "loss": 0.5531,
+      "step": 68551
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.731033916152443e-07,
+      "loss": 0.707,
+      "step": 68552
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.730764429254333e-07,
+      "loss": 0.7012,
+      "step": 68553
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.730494944093674e-07,
+      "loss": 0.6631,
+      "step": 68554
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.73022546067067e-07,
+      "loss": 0.5859,
+      "step": 68555
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.729955978985528e-07,
+      "loss": 0.6499,
+      "step": 68556
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.729686499038454e-07,
+      "loss": 0.6406,
+      "step": 68557
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.729417020829658e-07,
+      "loss": 0.6455,
+      "step": 68558
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.729147544359341e-07,
+      "loss": 0.6377,
+      "step": 68559
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.728878069627709e-07,
+      "loss": 0.6992,
+      "step": 68560
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.728608596634974e-07,
+      "loss": 0.6191,
+      "step": 68561
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.728339125381336e-07,
+      "loss": 0.7705,
+      "step": 68562
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.728069655867005e-07,
+      "loss": 0.7715,
+      "step": 68563
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.727800188092186e-07,
+      "loss": 0.4717,
+      "step": 68564
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.727530722057087e-07,
+      "loss": 0.6255,
+      "step": 68565
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.727261257761915e-07,
+      "loss": 0.6592,
+      "step": 68566
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.726991795206877e-07,
+      "loss": 0.5859,
+      "step": 68567
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.726722334392171e-07,
+      "loss": 0.7017,
+      "step": 68568
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.72645287531801e-07,
+      "loss": 0.6543,
+      "step": 68569
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7261834179846e-07,
+      "loss": 0.3845,
+      "step": 68570
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.725913962392148e-07,
+      "loss": 0.6401,
+      "step": 68571
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.725644508540856e-07,
+      "loss": 0.6145,
+      "step": 68572
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.725375056430937e-07,
+      "loss": 0.6357,
+      "step": 68573
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.725105606062591e-07,
+      "loss": 0.4558,
+      "step": 68574
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.72483615743603e-07,
+      "loss": 0.667,
+      "step": 68575
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.724566710551453e-07,
+      "loss": 0.5669,
+      "step": 68576
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.724297265409077e-07,
+      "loss": 0.7012,
+      "step": 68577
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.724027822009099e-07,
+      "loss": 0.6074,
+      "step": 68578
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.723758380351725e-07,
+      "loss": 0.6035,
+      "step": 68579
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.72348894043717e-07,
+      "loss": 0.7583,
+      "step": 68580
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.72321950226563e-07,
+      "loss": 0.6841,
+      "step": 68581
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.722950065837319e-07,
+      "loss": 0.6516,
+      "step": 68582
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.722680631152441e-07,
+      "loss": 0.5718,
+      "step": 68583
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.722411198211198e-07,
+      "loss": 0.6387,
+      "step": 68584
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.722141767013803e-07,
+      "loss": 0.5488,
+      "step": 68585
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.721872337560458e-07,
+      "loss": 0.5635,
+      "step": 68586
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.721602909851375e-07,
+      "loss": 0.4854,
+      "step": 68587
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.721333483886752e-07,
+      "loss": 0.6875,
+      "step": 68588
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.721064059666798e-07,
+      "loss": 0.583,
+      "step": 68589
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.720794637191722e-07,
+      "loss": 0.5898,
+      "step": 68590
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.720525216461728e-07,
+      "loss": 0.6729,
+      "step": 68591
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.720255797477025e-07,
+      "loss": 0.6904,
+      "step": 68592
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.719986380237813e-07,
+      "loss": 0.7798,
+      "step": 68593
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.719716964744306e-07,
+      "loss": 0.7002,
+      "step": 68594
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.719447550996705e-07,
+      "loss": 0.666,
+      "step": 68595
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.71917813899522e-07,
+      "loss": 0.5898,
+      "step": 68596
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.718908728740059e-07,
+      "loss": 0.6836,
+      "step": 68597
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.718639320231419e-07,
+      "loss": 0.6758,
+      "step": 68598
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.718369913469514e-07,
+      "loss": 0.7051,
+      "step": 68599
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.718100508454546e-07,
+      "loss": 0.627,
+      "step": 68600
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.717831105186726e-07,
+      "loss": 0.5747,
+      "step": 68601
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.717561703666259e-07,
+      "loss": 0.7549,
+      "step": 68602
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.717292303893346e-07,
+      "loss": 0.6143,
+      "step": 68603
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7170229058682e-07,
+      "loss": 0.6514,
+      "step": 68604
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.716753509591023e-07,
+      "loss": 0.6816,
+      "step": 68605
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.716484115062026e-07,
+      "loss": 0.5532,
+      "step": 68606
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.716214722281412e-07,
+      "loss": 0.8662,
+      "step": 68607
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.715945331249383e-07,
+      "loss": 0.6709,
+      "step": 68608
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.715675941966153e-07,
+      "loss": 0.8389,
+      "step": 68609
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.715406554431923e-07,
+      "loss": 0.6787,
+      "step": 68610
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.715137168646903e-07,
+      "loss": 0.6777,
+      "step": 68611
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.714867784611295e-07,
+      "loss": 0.582,
+      "step": 68612
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.714598402325309e-07,
+      "loss": 0.7314,
+      "step": 68613
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.71432902178915e-07,
+      "loss": 0.5732,
+      "step": 68614
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.714059643003025e-07,
+      "loss": 0.6165,
+      "step": 68615
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.713790265967137e-07,
+      "loss": 0.5608,
+      "step": 68616
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.713520890681699e-07,
+      "loss": 0.4237,
+      "step": 68617
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.713251517146911e-07,
+      "loss": 0.8184,
+      "step": 68618
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.712982145362982e-07,
+      "loss": 0.6187,
+      "step": 68619
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.712712775330115e-07,
+      "loss": 0.5562,
+      "step": 68620
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.712443407048519e-07,
+      "loss": 0.4248,
+      "step": 68621
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.7121740405184e-07,
+      "loss": 0.6973,
+      "step": 68622
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.711904675739965e-07,
+      "loss": 0.6328,
+      "step": 68623
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.711635312713417e-07,
+      "loss": 0.6836,
+      "step": 68624
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.711365951438968e-07,
+      "loss": 0.6123,
+      "step": 68625
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.711096591916818e-07,
+      "loss": 0.5569,
+      "step": 68626
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.71082723414718e-07,
+      "loss": 0.6304,
+      "step": 68627
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.710557878130255e-07,
+      "loss": 0.4595,
+      "step": 68628
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.710288523866248e-07,
+      "loss": 0.7393,
+      "step": 68629
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.71001917135537e-07,
+      "loss": 0.5415,
+      "step": 68630
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.709749820597823e-07,
+      "loss": 0.5737,
+      "step": 68631
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.709480471593817e-07,
+      "loss": 0.6445,
+      "step": 68632
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.709211124343555e-07,
+      "loss": 0.7168,
+      "step": 68633
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.708941778847246e-07,
+      "loss": 0.6992,
+      "step": 68634
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.708672435105093e-07,
+      "loss": 0.5391,
+      "step": 68635
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.708403093117306e-07,
+      "loss": 0.7168,
+      "step": 68636
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.708133752884093e-07,
+      "loss": 0.7002,
+      "step": 68637
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.707864414405654e-07,
+      "loss": 0.6138,
+      "step": 68638
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.707595077682195e-07,
+      "loss": 0.7041,
+      "step": 68639
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.707325742713926e-07,
+      "loss": 0.7373,
+      "step": 68640
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.707056409501051e-07,
+      "loss": 0.6216,
+      "step": 68641
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.706787078043782e-07,
+      "loss": 0.6035,
+      "step": 68642
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.706517748342316e-07,
+      "loss": 0.9268,
+      "step": 68643
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.706248420396866e-07,
+      "loss": 0.4668,
+      "step": 68644
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.705979094207635e-07,
+      "loss": 0.6489,
+      "step": 68645
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.705709769774832e-07,
+      "loss": 0.7021,
+      "step": 68646
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.705440447098663e-07,
+      "loss": 0.7852,
+      "step": 68647
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.705171126179331e-07,
+      "loss": 0.6924,
+      "step": 68648
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.704901807017044e-07,
+      "loss": 0.6316,
+      "step": 68649
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.704632489612007e-07,
+      "loss": 0.6279,
+      "step": 68650
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.704363173964427e-07,
+      "loss": 0.6177,
+      "step": 68651
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.704093860074511e-07,
+      "loss": 0.5581,
+      "step": 68652
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.703824547942465e-07,
+      "loss": 0.6514,
+      "step": 68653
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.703555237568495e-07,
+      "loss": 0.6787,
+      "step": 68654
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.703285928952808e-07,
+      "loss": 0.7129,
+      "step": 68655
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.703016622095611e-07,
+      "loss": 0.8242,
+      "step": 68656
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.702747316997109e-07,
+      "loss": 0.7539,
+      "step": 68657
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.702478013657503e-07,
+      "loss": 0.4314,
+      "step": 68658
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.702208712077007e-07,
+      "loss": 0.5981,
+      "step": 68659
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.701939412255822e-07,
+      "loss": 0.4202,
+      "step": 68660
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.701670114194159e-07,
+      "loss": 0.7969,
+      "step": 68661
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.701400817892219e-07,
+      "loss": 0.7686,
+      "step": 68662
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.701131523350211e-07,
+      "loss": 0.5254,
+      "step": 68663
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.700862230568342e-07,
+      "loss": 0.8223,
+      "step": 68664
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.700592939546817e-07,
+      "loss": 0.5889,
+      "step": 68665
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.70032365028584e-07,
+      "loss": 0.6211,
+      "step": 68666
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.700054362785625e-07,
+      "loss": 0.512,
+      "step": 68667
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.69978507704637e-07,
+      "loss": 0.6206,
+      "step": 68668
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.699515793068281e-07,
+      "loss": 0.668,
+      "step": 68669
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.699246510851571e-07,
+      "loss": 0.6992,
+      "step": 68670
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.698977230396437e-07,
+      "loss": 0.6318,
+      "step": 68671
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.698707951703095e-07,
+      "loss": 0.5571,
+      "step": 68672
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.698438674771743e-07,
+      "loss": 0.6426,
+      "step": 68673
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.698169399602593e-07,
+      "loss": 0.5186,
+      "step": 68674
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.69790012619585e-07,
+      "loss": 0.5703,
+      "step": 68675
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.697630854551715e-07,
+      "loss": 0.6733,
+      "step": 68676
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.697361584670404e-07,
+      "loss": 0.8164,
+      "step": 68677
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.697092316552115e-07,
+      "loss": 0.7031,
+      "step": 68678
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.696823050197055e-07,
+      "loss": 0.7295,
+      "step": 68679
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.696553785605432e-07,
+      "loss": 0.6992,
+      "step": 68680
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.696284522777451e-07,
+      "loss": 0.6113,
+      "step": 68681
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.69601526171332e-07,
+      "loss": 0.7471,
+      "step": 68682
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.695746002413243e-07,
+      "loss": 0.6606,
+      "step": 68683
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.695476744877429e-07,
+      "loss": 0.6953,
+      "step": 68684
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.695207489106082e-07,
+      "loss": 0.6494,
+      "step": 68685
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.694938235099409e-07,
+      "loss": 0.7119,
+      "step": 68686
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.694668982857618e-07,
+      "loss": 0.4263,
+      "step": 68687
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.69439973238091e-07,
+      "loss": 0.7529,
+      "step": 68688
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.694130483669494e-07,
+      "loss": 0.5435,
+      "step": 68689
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.693861236723575e-07,
+      "loss": 0.6357,
+      "step": 68690
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.693591991543364e-07,
+      "loss": 0.7549,
+      "step": 68691
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.693322748129063e-07,
+      "loss": 0.5151,
+      "step": 68692
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.693053506480876e-07,
+      "loss": 0.6162,
+      "step": 68693
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.692784266599014e-07,
+      "loss": 0.6187,
+      "step": 68694
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.692515028483679e-07,
+      "loss": 0.4042,
+      "step": 68695
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.692245792135081e-07,
+      "loss": 0.791,
+      "step": 68696
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.691976557553426e-07,
+      "loss": 0.7119,
+      "step": 68697
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.691707324738915e-07,
+      "loss": 0.8232,
+      "step": 68698
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.69143809369176e-07,
+      "loss": 0.6064,
+      "step": 68699
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.691168864412162e-07,
+      "loss": 0.4612,
+      "step": 68700
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.690899636900333e-07,
+      "loss": 0.6689,
+      "step": 68701
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.690630411156473e-07,
+      "loss": 0.5688,
+      "step": 68702
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.690361187180792e-07,
+      "loss": 0.6357,
+      "step": 68703
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.690091964973494e-07,
+      "loss": 0.582,
+      "step": 68704
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.689822744534789e-07,
+      "loss": 0.5586,
+      "step": 68705
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.689553525864878e-07,
+      "loss": 0.5869,
+      "step": 68706
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.689284308963975e-07,
+      "loss": 0.9199,
+      "step": 68707
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.689015093832277e-07,
+      "loss": 0.749,
+      "step": 68708
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.688745880469994e-07,
+      "loss": 0.6611,
+      "step": 68709
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68847666887733e-07,
+      "loss": 0.7881,
+      "step": 68710
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.688207459054497e-07,
+      "loss": 0.5322,
+      "step": 68711
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.687938251001694e-07,
+      "loss": 0.6479,
+      "step": 68712
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.687669044719131e-07,
+      "loss": 0.5002,
+      "step": 68713
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.687399840207014e-07,
+      "loss": 0.4712,
+      "step": 68714
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.687130637465548e-07,
+      "loss": 0.5239,
+      "step": 68715
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68686143649494e-07,
+      "loss": 0.8369,
+      "step": 68716
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.686592237295399e-07,
+      "loss": 0.7666,
+      "step": 68717
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.686323039867125e-07,
+      "loss": 0.7271,
+      "step": 68718
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.686053844210326e-07,
+      "loss": 0.7056,
+      "step": 68719
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.685784650325211e-07,
+      "loss": 0.4459,
+      "step": 68720
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.685515458211983e-07,
+      "loss": 0.7568,
+      "step": 68721
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68524626787085e-07,
+      "loss": 0.6504,
+      "step": 68722
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.684977079302016e-07,
+      "loss": 0.7383,
+      "step": 68723
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.684707892505692e-07,
+      "loss": 0.6611,
+      "step": 68724
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.684438707482077e-07,
+      "loss": 0.4934,
+      "step": 68725
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.684169524231383e-07,
+      "loss": 0.6445,
+      "step": 68726
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.683900342753817e-07,
+      "loss": 0.5857,
+      "step": 68727
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68363116304958e-07,
+      "loss": 0.5432,
+      "step": 68728
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.683361985118877e-07,
+      "loss": 0.8389,
+      "step": 68729
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68309280896192e-07,
+      "loss": 0.5479,
+      "step": 68730
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.682823634578909e-07,
+      "loss": 0.6372,
+      "step": 68731
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.682554461970056e-07,
+      "loss": 0.6533,
+      "step": 68732
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.682285291135563e-07,
+      "loss": 0.3787,
+      "step": 68733
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.68201612207564e-07,
+      "loss": 0.6455,
+      "step": 68734
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.681746954790487e-07,
+      "loss": 0.7402,
+      "step": 68735
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.681477789280317e-07,
+      "loss": 0.5518,
+      "step": 68736
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.681208625545335e-07,
+      "loss": 0.6465,
+      "step": 68737
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.680939463585741e-07,
+      "loss": 0.7783,
+      "step": 68738
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.680670303401746e-07,
+      "loss": 0.7734,
+      "step": 68739
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.680401144993554e-07,
+      "loss": 0.7539,
+      "step": 68740
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.680131988361374e-07,
+      "loss": 0.6924,
+      "step": 68741
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.679862833505409e-07,
+      "loss": 0.54,
+      "step": 68742
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.679593680425867e-07,
+      "loss": 0.6738,
+      "step": 68743
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.679324529122951e-07,
+      "loss": 0.603,
+      "step": 68744
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.679055379596872e-07,
+      "loss": 0.7527,
+      "step": 68745
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.678786231847835e-07,
+      "loss": 0.7695,
+      "step": 68746
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.678517085876045e-07,
+      "loss": 0.6165,
+      "step": 68747
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.678247941681705e-07,
+      "loss": 0.6504,
+      "step": 68748
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.677978799265024e-07,
+      "loss": 0.6748,
+      "step": 68749
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.677709658626208e-07,
+      "loss": 0.6826,
+      "step": 68750
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.677440519765464e-07,
+      "loss": 0.707,
+      "step": 68751
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.677171382682995e-07,
+      "loss": 0.6851,
+      "step": 68752
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.676902247379011e-07,
+      "loss": 0.7148,
+      "step": 68753
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.676633113853715e-07,
+      "loss": 0.6592,
+      "step": 68754
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.676363982107315e-07,
+      "loss": 0.7148,
+      "step": 68755
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.676094852140014e-07,
+      "loss": 0.7158,
+      "step": 68756
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.675825723952026e-07,
+      "loss": 0.5674,
+      "step": 68757
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.675556597543548e-07,
+      "loss": 0.7266,
+      "step": 68758
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.675287472914787e-07,
+      "loss": 0.6758,
+      "step": 68759
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.675018350065954e-07,
+      "loss": 0.5114,
+      "step": 68760
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.67474922899725e-07,
+      "loss": 0.6973,
+      "step": 68761
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.674480109708887e-07,
+      "loss": 0.6914,
+      "step": 68762
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.674210992201064e-07,
+      "loss": 0.5151,
+      "step": 68763
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.673941876473995e-07,
+      "loss": 0.7041,
+      "step": 68764
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.67367276252788e-07,
+      "loss": 0.7578,
+      "step": 68765
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.673403650362927e-07,
+      "loss": 0.5702,
+      "step": 68766
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.673134539979342e-07,
+      "loss": 0.6641,
+      "step": 68767
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.672865431377332e-07,
+      "loss": 0.5894,
+      "step": 68768
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.672596324557099e-07,
+      "loss": 0.7158,
+      "step": 68769
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.672327219518854e-07,
+      "loss": 0.499,
+      "step": 68770
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.672058116262798e-07,
+      "loss": 0.7295,
+      "step": 68771
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.671789014789144e-07,
+      "loss": 0.4697,
+      "step": 68772
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.671519915098091e-07,
+      "loss": 0.6445,
+      "step": 68773
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.671250817189849e-07,
+      "loss": 0.6348,
+      "step": 68774
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.670981721064624e-07,
+      "loss": 0.6279,
+      "step": 68775
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.670712626722621e-07,
+      "loss": 0.5273,
+      "step": 68776
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.670443534164049e-07,
+      "loss": 0.6636,
+      "step": 68777
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.670174443389107e-07,
+      "loss": 0.5444,
+      "step": 68778
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.669905354398008e-07,
+      "loss": 0.6143,
+      "step": 68779
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.669636267190952e-07,
+      "loss": 0.6807,
+      "step": 68780
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.669367181768152e-07,
+      "loss": 0.4927,
+      "step": 68781
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.669098098129809e-07,
+      "loss": 0.6431,
+      "step": 68782
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.668829016276128e-07,
+      "loss": 0.7749,
+      "step": 68783
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.66855993620732e-07,
+      "loss": 0.4912,
+      "step": 68784
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.668290857923587e-07,
+      "loss": 0.8145,
+      "step": 68785
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.668021781425138e-07,
+      "loss": 0.5093,
+      "step": 68786
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.66775270671218e-07,
+      "loss": 0.7275,
+      "step": 68787
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.667483633784913e-07,
+      "loss": 0.5884,
+      "step": 68788
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.667214562643547e-07,
+      "loss": 0.5278,
+      "step": 68789
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.666945493288287e-07,
+      "loss": 0.8076,
+      "step": 68790
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.666676425719341e-07,
+      "loss": 0.6621,
+      "step": 68791
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.66640735993691e-07,
+      "loss": 0.6289,
+      "step": 68792
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.666138295941208e-07,
+      "loss": 0.5774,
+      "step": 68793
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.665869233732434e-07,
+      "loss": 0.7188,
+      "step": 68794
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.665600173310798e-07,
+      "loss": 0.5205,
+      "step": 68795
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.665331114676503e-07,
+      "loss": 0.3379,
+      "step": 68796
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.665062057829761e-07,
+      "loss": 0.7168,
+      "step": 68797
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.66479300277077e-07,
+      "loss": 0.6348,
+      "step": 68798
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.664523949499741e-07,
+      "loss": 0.6387,
+      "step": 68799
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.664254898016877e-07,
+      "loss": 0.6841,
+      "step": 68800
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.663985848322386e-07,
+      "loss": 0.6719,
+      "step": 68801
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.663716800416473e-07,
+      "loss": 0.6768,
+      "step": 68802
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.663447754299346e-07,
+      "loss": 0.4719,
+      "step": 68803
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.663178709971207e-07,
+      "loss": 0.5439,
+      "step": 68804
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.662909667432268e-07,
+      "loss": 0.5427,
+      "step": 68805
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.662640626682729e-07,
+      "loss": 0.4548,
+      "step": 68806
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.662371587722804e-07,
+      "loss": 0.5173,
+      "step": 68807
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.662102550552689e-07,
+      "loss": 0.6318,
+      "step": 68808
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.661833515172594e-07,
+      "loss": 0.6211,
+      "step": 68809
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.661564481582727e-07,
+      "loss": 0.6245,
+      "step": 68810
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.661295449783292e-07,
+      "loss": 0.4683,
+      "step": 68811
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.661026419774496e-07,
+      "loss": 0.6685,
+      "step": 68812
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.660757391556543e-07,
+      "loss": 0.7129,
+      "step": 68813
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.660488365129643e-07,
+      "loss": 0.5461,
+      "step": 68814
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.660219340493996e-07,
+      "loss": 0.6455,
+      "step": 68815
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.659950317649815e-07,
+      "loss": 0.7578,
+      "step": 68816
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.659681296597303e-07,
+      "loss": 0.6782,
+      "step": 68817
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.659412277336666e-07,
+      "loss": 0.6895,
+      "step": 68818
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.659143259868104e-07,
+      "loss": 0.541,
+      "step": 68819
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.658874244191833e-07,
+      "loss": 0.6299,
+      "step": 68820
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.658605230308051e-07,
+      "loss": 0.832,
+      "step": 68821
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.65833621821697e-07,
+      "loss": 0.8594,
+      "step": 68822
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.65806720791879e-07,
+      "loss": 0.6064,
+      "step": 68823
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.657798199413724e-07,
+      "loss": 0.6528,
+      "step": 68824
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.657529192701971e-07,
+      "loss": 0.5586,
+      "step": 68825
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.657260187783743e-07,
+      "loss": 0.6221,
+      "step": 68826
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.656991184659245e-07,
+      "loss": 0.6465,
+      "step": 68827
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.656722183328677e-07,
+      "loss": 0.4785,
+      "step": 68828
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.656453183792251e-07,
+      "loss": 0.6245,
+      "step": 68829
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.656184186050168e-07,
+      "loss": 0.7324,
+      "step": 68830
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.655915190102641e-07,
+      "loss": 0.6499,
+      "step": 68831
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.655646195949869e-07,
+      "loss": 0.6582,
+      "step": 68832
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.655377203592063e-07,
+      "loss": 0.6074,
+      "step": 68833
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.655108213029424e-07,
+      "loss": 0.585,
+      "step": 68834
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.654839224262164e-07,
+      "loss": 0.5488,
+      "step": 68835
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.654570237290483e-07,
+      "loss": 0.5938,
+      "step": 68836
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.654301252114597e-07,
+      "loss": 0.7793,
+      "step": 68837
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.654032268734697e-07,
+      "loss": 0.5283,
+      "step": 68838
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.653763287151e-07,
+      "loss": 0.5774,
+      "step": 68839
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.653494307363706e-07,
+      "loss": 0.7344,
+      "step": 68840
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.653225329373025e-07,
+      "loss": 0.6748,
+      "step": 68841
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.65295635317916e-07,
+      "loss": 0.5664,
+      "step": 68842
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.652687378782321e-07,
+      "loss": 0.542,
+      "step": 68843
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.652418406182708e-07,
+      "loss": 0.7061,
+      "step": 68844
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.652149435380533e-07,
+      "loss": 0.5107,
+      "step": 68845
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.651880466375998e-07,
+      "loss": 0.4658,
+      "step": 68846
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.651611499169313e-07,
+      "loss": 0.5398,
+      "step": 68847
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.651342533760679e-07,
+      "loss": 0.6709,
+      "step": 68848
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.651073570150303e-07,
+      "loss": 0.6006,
+      "step": 68849
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.650804608338393e-07,
+      "loss": 0.7051,
+      "step": 68850
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.650535648325151e-07,
+      "loss": 0.6865,
+      "step": 68851
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.65026669011079e-07,
+      "loss": 0.4832,
+      "step": 68852
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.64999773369551e-07,
+      "loss": 0.7261,
+      "step": 68853
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.649728779079518e-07,
+      "loss": 0.8027,
+      "step": 68854
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.649459826263023e-07,
+      "loss": 0.7578,
+      "step": 68855
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.64919087524623e-07,
+      "loss": 0.7227,
+      "step": 68856
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.64892192602934e-07,
+      "loss": 0.7773,
+      "step": 68857
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.648652978612562e-07,
+      "loss": 0.6963,
+      "step": 68858
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.648384032996102e-07,
+      "loss": 0.6245,
+      "step": 68859
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.648115089180168e-07,
+      "loss": 0.5286,
+      "step": 68860
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.647846147164962e-07,
+      "loss": 0.7422,
+      "step": 68861
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.647577206950694e-07,
+      "loss": 0.6042,
+      "step": 68862
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.647308268537567e-07,
+      "loss": 0.5801,
+      "step": 68863
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.647039331925789e-07,
+      "loss": 0.7646,
+      "step": 68864
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.646770397115563e-07,
+      "loss": 0.8799,
+      "step": 68865
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.6465014641071e-07,
+      "loss": 0.6572,
+      "step": 68866
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.646232532900601e-07,
+      "loss": 0.8604,
+      "step": 68867
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.645963603496271e-07,
+      "loss": 0.623,
+      "step": 68868
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.645694675894322e-07,
+      "loss": 0.5879,
+      "step": 68869
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.645425750094954e-07,
+      "loss": 0.4858,
+      "step": 68870
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.645156826098375e-07,
+      "loss": 0.6426,
+      "step": 68871
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.644887903904794e-07,
+      "loss": 0.6924,
+      "step": 68872
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.644618983514411e-07,
+      "loss": 0.6665,
+      "step": 68873
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.644350064927436e-07,
+      "loss": 0.791,
+      "step": 68874
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.644081148144073e-07,
+      "loss": 0.79,
+      "step": 68875
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.643812233164533e-07,
+      "loss": 0.5044,
+      "step": 68876
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.643543319989014e-07,
+      "loss": 0.5596,
+      "step": 68877
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.643274408617724e-07,
+      "loss": 0.6465,
+      "step": 68878
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.643005499050874e-07,
+      "loss": 0.7432,
+      "step": 68879
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.642736591288663e-07,
+      "loss": 0.7393,
+      "step": 68880
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.642467685331303e-07,
+      "loss": 0.522,
+      "step": 68881
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.642198781178995e-07,
+      "loss": 0.752,
+      "step": 68882
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.641929878831948e-07,
+      "loss": 0.7451,
+      "step": 68883
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.641660978290365e-07,
+      "loss": 0.7236,
+      "step": 68884
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.641392079554456e-07,
+      "loss": 0.4626,
+      "step": 68885
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.641123182624427e-07,
+      "loss": 0.6997,
+      "step": 68886
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.640854287500478e-07,
+      "loss": 0.7158,
+      "step": 68887
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.640585394182819e-07,
+      "loss": 0.6104,
+      "step": 68888
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.640316502671657e-07,
+      "loss": 0.4883,
+      "step": 68889
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.640047612967192e-07,
+      "loss": 0.6982,
+      "step": 68890
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.639778725069637e-07,
+      "loss": 0.6113,
+      "step": 68891
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.639509838979194e-07,
+      "loss": 0.5195,
+      "step": 68892
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63924095469607e-07,
+      "loss": 0.616,
+      "step": 68893
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63897207222047e-07,
+      "loss": 0.7734,
+      "step": 68894
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.638703191552602e-07,
+      "loss": 0.8438,
+      "step": 68895
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.638434312692672e-07,
+      "loss": 0.5371,
+      "step": 68896
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.638165435640881e-07,
+      "loss": 0.6689,
+      "step": 68897
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63789656039744e-07,
+      "loss": 0.6758,
+      "step": 68898
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.637627686962551e-07,
+      "loss": 0.6738,
+      "step": 68899
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.637358815336423e-07,
+      "loss": 0.8086,
+      "step": 68900
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.637089945519259e-07,
+      "loss": 0.8418,
+      "step": 68901
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63682107751127e-07,
+      "loss": 0.6445,
+      "step": 68902
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.636552211312654e-07,
+      "loss": 0.7236,
+      "step": 68903
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.636283346923626e-07,
+      "loss": 0.6177,
+      "step": 68904
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.636014484344383e-07,
+      "loss": 0.6099,
+      "step": 68905
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.635745623575143e-07,
+      "loss": 0.564,
+      "step": 68906
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.635476764616096e-07,
+      "loss": 0.6118,
+      "step": 68907
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.635207907467458e-07,
+      "loss": 0.8359,
+      "step": 68908
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.634939052129431e-07,
+      "loss": 0.6504,
+      "step": 68909
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.634670198602224e-07,
+      "loss": 0.6758,
+      "step": 68910
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63440134688604e-07,
+      "loss": 0.6992,
+      "step": 68911
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.634132496981088e-07,
+      "loss": 0.7998,
+      "step": 68912
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.63386364888757e-07,
+      "loss": 0.7539,
+      "step": 68913
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.633594802605696e-07,
+      "loss": 0.5776,
+      "step": 68914
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.633325958135667e-07,
+      "loss": 0.4922,
+      "step": 68915
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.633057115477696e-07,
+      "loss": 0.564,
+      "step": 68916
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.632788274631982e-07,
+      "loss": 0.7051,
+      "step": 68917
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.632519435598731e-07,
+      "loss": 0.7705,
+      "step": 68918
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.632250598378152e-07,
+      "loss": 0.583,
+      "step": 68919
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.631981762970449e-07,
+      "loss": 0.625,
+      "step": 68920
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.631712929375832e-07,
+      "loss": 0.6494,
+      "step": 68921
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.6314440975945e-07,
+      "loss": 0.728,
+      "step": 68922
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.631175267626663e-07,
+      "loss": 0.5249,
+      "step": 68923
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.630906439472527e-07,
+      "loss": 0.6162,
+      "step": 68924
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.630637613132298e-07,
+      "loss": 0.5815,
+      "step": 68925
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.630368788606183e-07,
+      "loss": 0.666,
+      "step": 68926
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.630099965894384e-07,
+      "loss": 0.7432,
+      "step": 68927
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.629831144997106e-07,
+      "loss": 0.729,
+      "step": 68928
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.629562325914558e-07,
+      "loss": 0.7363,
+      "step": 68929
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.629293508646943e-07,
+      "loss": 0.6455,
+      "step": 68930
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.629024693194474e-07,
+      "loss": 0.6086,
+      "step": 68931
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.628755879557347e-07,
+      "loss": 0.6709,
+      "step": 68932
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.628487067735776e-07,
+      "loss": 0.6953,
+      "step": 68933
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.62821825772996e-07,
+      "loss": 0.5273,
+      "step": 68934
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.627949449540112e-07,
+      "loss": 0.707,
+      "step": 68935
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.627680643166436e-07,
+      "loss": 0.584,
+      "step": 68936
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.62741183860913e-07,
+      "loss": 0.563,
+      "step": 68937
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.627143035868409e-07,
+      "loss": 0.4648,
+      "step": 68938
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.626874234944473e-07,
+      "loss": 0.6865,
+      "step": 68939
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.626605435837532e-07,
+      "loss": 0.5532,
+      "step": 68940
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.626336638547789e-07,
+      "loss": 0.6904,
+      "step": 68941
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.626067843075452e-07,
+      "loss": 0.4216,
+      "step": 68942
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.625799049420724e-07,
+      "loss": 0.709,
+      "step": 68943
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.625530257583814e-07,
+      "loss": 0.7285,
+      "step": 68944
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.625261467564927e-07,
+      "loss": 0.5896,
+      "step": 68945
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.62499267936427e-07,
+      "loss": 0.5911,
+      "step": 68946
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.624723892982043e-07,
+      "loss": 0.6406,
+      "step": 68947
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.624455108418458e-07,
+      "loss": 0.7891,
+      "step": 68948
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.624186325673716e-07,
+      "loss": 0.4708,
+      "step": 68949
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.623917544748028e-07,
+      "loss": 0.6343,
+      "step": 68950
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.623648765641594e-07,
+      "loss": 0.5139,
+      "step": 68951
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.623379988354626e-07,
+      "loss": 0.6182,
+      "step": 68952
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.623111212887323e-07,
+      "loss": 0.5928,
+      "step": 68953
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.622842439239898e-07,
+      "loss": 0.6685,
+      "step": 68954
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.622573667412551e-07,
+      "loss": 0.6221,
+      "step": 68955
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.622304897405494e-07,
+      "loss": 0.5781,
+      "step": 68956
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.622036129218926e-07,
+      "loss": 0.7134,
+      "step": 68957
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.621767362853055e-07,
+      "loss": 0.5174,
+      "step": 68958
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.621498598308087e-07,
+      "loss": 0.8086,
+      "step": 68959
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.621229835584229e-07,
+      "loss": 0.6885,
+      "step": 68960
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.620961074681686e-07,
+      "loss": 0.5757,
+      "step": 68961
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.620692315600664e-07,
+      "loss": 0.6699,
+      "step": 68962
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.620423558341366e-07,
+      "loss": 0.5781,
+      "step": 68963
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.620154802904004e-07,
+      "loss": 0.5349,
+      "step": 68964
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.619886049288777e-07,
+      "loss": 0.8184,
+      "step": 68965
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.619617297495898e-07,
+      "loss": 0.6484,
+      "step": 68966
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.619348547525567e-07,
+      "loss": 0.5598,
+      "step": 68967
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.619079799377989e-07,
+      "loss": 0.6758,
+      "step": 68968
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.618811053053373e-07,
+      "loss": 0.5195,
+      "step": 68969
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.618542308551923e-07,
+      "loss": 0.5928,
+      "step": 68970
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.618273565873847e-07,
+      "loss": 0.708,
+      "step": 68971
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.618004825019349e-07,
+      "loss": 0.5918,
+      "step": 68972
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.617736085988636e-07,
+      "loss": 0.5825,
+      "step": 68973
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.617467348781911e-07,
+      "loss": 0.375,
+      "step": 68974
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.617198613399383e-07,
+      "loss": 0.6704,
+      "step": 68975
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.61692987984126e-07,
+      "loss": 0.6323,
+      "step": 68976
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.61666114810774e-07,
+      "loss": 0.6157,
+      "step": 68977
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.616392418199033e-07,
+      "loss": 0.6582,
+      "step": 68978
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.616123690115347e-07,
+      "loss": 0.6313,
+      "step": 68979
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.615854963856882e-07,
+      "loss": 0.6118,
+      "step": 68980
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.61558623942385e-07,
+      "loss": 0.5625,
+      "step": 68981
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.615317516816452e-07,
+      "loss": 0.5947,
+      "step": 68982
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.615048796034897e-07,
+      "loss": 0.5654,
+      "step": 68983
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.614780077079388e-07,
+      "loss": 0.5198,
+      "step": 68984
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.614511359950134e-07,
+      "loss": 0.6035,
+      "step": 68985
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.614242644647342e-07,
+      "loss": 0.5048,
+      "step": 68986
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.613973931171209e-07,
+      "loss": 0.6846,
+      "step": 68987
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.61370521952195e-07,
+      "loss": 0.5288,
+      "step": 68988
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.613436509699765e-07,
+      "loss": 0.6719,
+      "step": 68989
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.613167801704862e-07,
+      "loss": 0.5337,
+      "step": 68990
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.612899095537448e-07,
+      "loss": 0.5757,
+      "step": 68991
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.612630391197727e-07,
+      "loss": 0.5913,
+      "step": 68992
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.612361688685903e-07,
+      "loss": 0.666,
+      "step": 68993
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.612092988002187e-07,
+      "loss": 0.5007,
+      "step": 68994
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.611824289146779e-07,
+      "loss": 0.6484,
+      "step": 68995
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.611555592119893e-07,
+      "loss": 0.5757,
+      "step": 68996
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.611286896921723e-07,
+      "loss": 0.7345,
+      "step": 68997
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.611018203552484e-07,
+      "loss": 0.6045,
+      "step": 68998
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.610749512012375e-07,
+      "loss": 0.5424,
+      "step": 68999
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.610480822301609e-07,
+      "loss": 0.5547,
+      "step": 69000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.610212134420385e-07,
+      "loss": 0.6401,
+      "step": 69001
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.609943448368914e-07,
+      "loss": 0.6836,
+      "step": 69002
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.609674764147397e-07,
+      "loss": 0.499,
+      "step": 69003
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.609406081756045e-07,
+      "loss": 0.4883,
+      "step": 69004
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.609137401195057e-07,
+      "loss": 0.7266,
+      "step": 69005
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.608868722464648e-07,
+      "loss": 0.6357,
+      "step": 69006
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.608600045565016e-07,
+      "loss": 0.6318,
+      "step": 69007
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.608331370496367e-07,
+      "loss": 0.7744,
+      "step": 69008
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.60806269725891e-07,
+      "loss": 0.6846,
+      "step": 69009
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.607794025852849e-07,
+      "loss": 0.7397,
+      "step": 69010
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.60752535627839e-07,
+      "loss": 0.7368,
+      "step": 69011
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.607256688535739e-07,
+      "loss": 0.6289,
+      "step": 69012
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.606988022625102e-07,
+      "loss": 0.6387,
+      "step": 69013
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.606719358546682e-07,
+      "loss": 0.6731,
+      "step": 69014
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.60645069630069e-07,
+      "loss": 0.7158,
+      "step": 69015
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.606182035887329e-07,
+      "loss": 0.6768,
+      "step": 69016
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.605913377306805e-07,
+      "loss": 0.7832,
+      "step": 69017
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.60564472055932e-07,
+      "loss": 0.6074,
+      "step": 69018
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.605376065645084e-07,
+      "loss": 0.5796,
+      "step": 69019
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.6051074125643e-07,
+      "loss": 0.3811,
+      "step": 69020
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.604838761317177e-07,
+      "loss": 0.5669,
+      "step": 69021
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.604570111903916e-07,
+      "loss": 0.6709,
+      "step": 69022
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.604301464324729e-07,
+      "loss": 0.7065,
+      "step": 69023
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.604032818579815e-07,
+      "loss": 0.6543,
+      "step": 69024
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.603764174669386e-07,
+      "loss": 0.7314,
+      "step": 69025
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.603495532593647e-07,
+      "loss": 0.75,
+      "step": 69026
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.603226892352796e-07,
+      "loss": 0.5715,
+      "step": 69027
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.602958253947045e-07,
+      "loss": 0.5273,
+      "step": 69028
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.602689617376598e-07,
+      "loss": 0.6094,
+      "step": 69029
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.602420982641664e-07,
+      "loss": 0.6064,
+      "step": 69030
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.602152349742443e-07,
+      "loss": 0.6514,
+      "step": 69031
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.601883718679146e-07,
+      "loss": 0.5957,
+      "step": 69032
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.601615089451974e-07,
+      "loss": 0.6152,
+      "step": 69033
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.601346462061138e-07,
+      "loss": 0.6123,
+      "step": 69034
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.601077836506839e-07,
+      "loss": 0.502,
+      "step": 69035
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.600809212789289e-07,
+      "loss": 0.6628,
+      "step": 69036
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.600540590908683e-07,
+      "loss": 0.5791,
+      "step": 69037
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.600271970865235e-07,
+      "loss": 0.6699,
+      "step": 69038
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.600003352659146e-07,
+      "loss": 0.4398,
+      "step": 69039
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.599734736290627e-07,
+      "loss": 0.6248,
+      "step": 69040
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.599466121759878e-07,
+      "loss": 0.6914,
+      "step": 69041
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.599197509067112e-07,
+      "loss": 0.689,
+      "step": 69042
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.598928898212524e-07,
+      "loss": 0.8286,
+      "step": 69043
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.59866028919633e-07,
+      "loss": 0.6436,
+      "step": 69044
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.598391682018729e-07,
+      "loss": 0.5264,
+      "step": 69045
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.598123076679935e-07,
+      "loss": 0.7334,
+      "step": 69046
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.597854473180143e-07,
+      "loss": 0.6519,
+      "step": 69047
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.597585871519563e-07,
+      "loss": 0.5552,
+      "step": 69048
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.597317271698402e-07,
+      "loss": 0.7871,
+      "step": 69049
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.597048673716863e-07,
+      "loss": 0.4691,
+      "step": 69050
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.596780077575156e-07,
+      "loss": 0.6787,
+      "step": 69051
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.596511483273482e-07,
+      "loss": 0.582,
+      "step": 69052
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.596242890812047e-07,
+      "loss": 0.6782,
+      "step": 69053
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.595974300191063e-07,
+      "loss": 0.6025,
+      "step": 69054
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.595705711410727e-07,
+      "loss": 0.7036,
+      "step": 69055
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.595437124471253e-07,
+      "loss": 0.582,
+      "step": 69056
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.59516853937284e-07,
+      "loss": 0.6797,
+      "step": 69057
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.594899956115693e-07,
+      "loss": 0.7451,
+      "step": 69058
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.594631374700025e-07,
+      "loss": 0.5752,
+      "step": 69059
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.594362795126032e-07,
+      "loss": 0.7178,
+      "step": 69060
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.594094217393929e-07,
+      "loss": 0.6333,
+      "step": 69061
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.593825641503915e-07,
+      "loss": 0.6626,
+      "step": 69062
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.5935570674562e-07,
+      "loss": 0.8037,
+      "step": 69063
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.593288495250985e-07,
+      "loss": 0.541,
+      "step": 69064
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.593019924888481e-07,
+      "loss": 0.5703,
+      "step": 69065
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.592751356368892e-07,
+      "loss": 0.6934,
+      "step": 69066
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.59248278969242e-07,
+      "loss": 0.4214,
+      "step": 69067
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.592214224859275e-07,
+      "loss": 0.7051,
+      "step": 69068
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.591945661869661e-07,
+      "loss": 0.5923,
+      "step": 69069
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.591677100723782e-07,
+      "loss": 0.6357,
+      "step": 69070
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.591408541421845e-07,
+      "loss": 0.8467,
+      "step": 69071
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.591139983964054e-07,
+      "loss": 0.6084,
+      "step": 69072
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.590871428350621e-07,
+      "loss": 0.644,
+      "step": 69073
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.590602874581742e-07,
+      "loss": 0.7842,
+      "step": 69074
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.590334322657631e-07,
+      "loss": 0.6982,
+      "step": 69075
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.590065772578492e-07,
+      "loss": 0.5361,
+      "step": 69076
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.589797224344525e-07,
+      "loss": 0.7383,
+      "step": 69077
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.589528677955941e-07,
+      "loss": 0.5981,
+      "step": 69078
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.589260133412942e-07,
+      "loss": 0.5029,
+      "step": 69079
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.588991590715738e-07,
+      "loss": 0.7021,
+      "step": 69080
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.58872304986453e-07,
+      "loss": 0.916,
+      "step": 69081
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.588454510859528e-07,
+      "loss": 0.7705,
+      "step": 69082
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.588185973700934e-07,
+      "loss": 0.7441,
+      "step": 69083
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.587917438388958e-07,
+      "loss": 0.5654,
+      "step": 69084
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.587648904923798e-07,
+      "loss": 0.6484,
+      "step": 69085
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.587380373305673e-07,
+      "loss": 0.5898,
+      "step": 69086
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.587111843534773e-07,
+      "loss": 0.5596,
+      "step": 69087
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.586843315611311e-07,
+      "loss": 0.6118,
+      "step": 69088
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.586574789535492e-07,
+      "loss": 0.6191,
+      "step": 69089
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.586306265307523e-07,
+      "loss": 0.6606,
+      "step": 69090
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.586037742927606e-07,
+      "loss": 0.5703,
+      "step": 69091
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.585769222395951e-07,
+      "loss": 0.5137,
+      "step": 69092
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.585500703712759e-07,
+      "loss": 0.6406,
+      "step": 69093
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.585232186878242e-07,
+      "loss": 0.5991,
+      "step": 69094
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.584963671892598e-07,
+      "loss": 0.6265,
+      "step": 69095
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.584695158756042e-07,
+      "loss": 0.5806,
+      "step": 69096
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.58442664746877e-07,
+      "loss": 0.561,
+      "step": 69097
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.584158138030989e-07,
+      "loss": 0.7441,
+      "step": 69098
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.583889630442911e-07,
+      "loss": 0.6895,
+      "step": 69099
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.583621124704734e-07,
+      "loss": 0.708,
+      "step": 69100
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.58335262081667e-07,
+      "loss": 0.6865,
+      "step": 69101
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.58308411877892e-07,
+      "loss": 0.501,
+      "step": 69102
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.582815618591694e-07,
+      "loss": 0.7549,
+      "step": 69103
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.582547120255191e-07,
+      "loss": 0.6279,
+      "step": 69104
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.582278623769623e-07,
+      "loss": 0.6611,
+      "step": 69105
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.582010129135197e-07,
+      "loss": 0.6621,
+      "step": 69106
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.581741636352111e-07,
+      "loss": 0.5605,
+      "step": 69107
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.581473145420571e-07,
+      "loss": 0.6729,
+      "step": 69108
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.58120465634079e-07,
+      "loss": 0.5723,
+      "step": 69109
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.580936169112967e-07,
+      "loss": 0.6028,
+      "step": 69110
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.580667683737312e-07,
+      "loss": 0.7568,
+      "step": 69111
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.580399200214026e-07,
+      "loss": 0.4736,
+      "step": 69112
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.580130718543318e-07,
+      "loss": 0.5376,
+      "step": 69113
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.579862238725393e-07,
+      "loss": 0.564,
+      "step": 69114
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.579593760760457e-07,
+      "loss": 0.6929,
+      "step": 69115
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.579325284648716e-07,
+      "loss": 0.6094,
+      "step": 69116
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.579056810390371e-07,
+      "loss": 0.7656,
+      "step": 69117
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.578788337985633e-07,
+      "loss": 0.6675,
+      "step": 69118
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.578519867434704e-07,
+      "loss": 0.604,
+      "step": 69119
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.578251398737793e-07,
+      "loss": 0.542,
+      "step": 69120
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.577982931895101e-07,
+      "loss": 0.6084,
+      "step": 69121
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.577714466906837e-07,
+      "loss": 0.6924,
+      "step": 69122
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.577446003773204e-07,
+      "loss": 0.6523,
+      "step": 69123
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.577177542494412e-07,
+      "loss": 0.6963,
+      "step": 69124
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.576909083070665e-07,
+      "loss": 0.8154,
+      "step": 69125
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.576640625502169e-07,
+      "loss": 0.7988,
+      "step": 69126
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.576372169789122e-07,
+      "loss": 0.6128,
+      "step": 69127
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.576103715931739e-07,
+      "loss": 0.7549,
+      "step": 69128
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.575835263930218e-07,
+      "loss": 0.4854,
+      "step": 69129
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.575566813784771e-07,
+      "loss": 0.4802,
+      "step": 69130
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.5752983654956e-07,
+      "loss": 0.6045,
+      "step": 69131
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.575029919062914e-07,
+      "loss": 0.4613,
+      "step": 69132
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.574761474486913e-07,
+      "loss": 0.6079,
+      "step": 69133
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.574493031767808e-07,
+      "loss": 0.5361,
+      "step": 69134
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.5742245909058e-07,
+      "loss": 0.6885,
+      "step": 69135
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.573956151901101e-07,
+      "loss": 0.7388,
+      "step": 69136
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.57368771475391e-07,
+      "loss": 0.4663,
+      "step": 69137
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.573419279464433e-07,
+      "loss": 0.6182,
+      "step": 69138
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.573150846032879e-07,
+      "loss": 0.5903,
+      "step": 69139
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.572882414459449e-07,
+      "loss": 0.4796,
+      "step": 69140
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.572613984744355e-07,
+      "loss": 0.7686,
+      "step": 69141
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.572345556887798e-07,
+      "loss": 0.8164,
+      "step": 69142
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.572077130889982e-07,
+      "loss": 0.6621,
+      "step": 69143
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.571808706751117e-07,
+      "loss": 0.499,
+      "step": 69144
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.571540284471405e-07,
+      "loss": 0.582,
+      "step": 69145
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.571271864051056e-07,
+      "loss": 0.4155,
+      "step": 69146
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.571003445490272e-07,
+      "loss": 0.7173,
+      "step": 69147
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.570735028789256e-07,
+      "loss": 0.7725,
+      "step": 69148
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.570466613948218e-07,
+      "loss": 0.4927,
+      "step": 69149
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.570198200967362e-07,
+      "loss": 0.54,
+      "step": 69150
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.569929789846893e-07,
+      "loss": 0.5552,
+      "step": 69151
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.569661380587016e-07,
+      "loss": 0.8271,
+      "step": 69152
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.56939297318794e-07,
+      "loss": 0.6982,
+      "step": 69153
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.569124567649865e-07,
+      "loss": 0.5654,
+      "step": 69154
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.568856163973005e-07,
+      "loss": 0.5459,
+      "step": 69155
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.568587762157557e-07,
+      "loss": 0.6846,
+      "step": 69156
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.568319362203728e-07,
+      "loss": 0.5898,
+      "step": 69157
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.568050964111727e-07,
+      "loss": 0.793,
+      "step": 69158
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.567782567881757e-07,
+      "loss": 0.4506,
+      "step": 69159
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.567514173514023e-07,
+      "loss": 0.4351,
+      "step": 69160
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.567245781008734e-07,
+      "loss": 0.729,
+      "step": 69161
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.566977390366088e-07,
+      "loss": 0.6372,
+      "step": 69162
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.5667090015863e-07,
+      "loss": 0.5887,
+      "step": 69163
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.566440614669569e-07,
+      "loss": 0.6436,
+      "step": 69164
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.566172229616106e-07,
+      "loss": 0.4547,
+      "step": 69165
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.565903846426111e-07,
+      "loss": 0.4746,
+      "step": 69166
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.56563546509979e-07,
+      "loss": 0.6543,
+      "step": 69167
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.565367085637352e-07,
+      "loss": 0.668,
+      "step": 69168
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.565098708038997e-07,
+      "loss": 0.6899,
+      "step": 69169
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.564830332304938e-07,
+      "loss": 0.8311,
+      "step": 69170
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.564561958435373e-07,
+      "loss": 0.7422,
+      "step": 69171
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.564293586430514e-07,
+      "loss": 0.6812,
+      "step": 69172
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.56402521629056e-07,
+      "loss": 0.623,
+      "step": 69173
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.563756848015723e-07,
+      "loss": 0.6704,
+      "step": 69174
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.563488481606209e-07,
+      "loss": 0.5903,
+      "step": 69175
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.563220117062212e-07,
+      "loss": 0.7266,
+      "step": 69176
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.562951754383951e-07,
+      "loss": 0.6372,
+      "step": 69177
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.562683393571622e-07,
+      "loss": 0.5547,
+      "step": 69178
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.562415034625436e-07,
+      "loss": 0.749,
+      "step": 69179
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.562146677545596e-07,
+      "loss": 0.6953,
+      "step": 69180
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.561878322332307e-07,
+      "loss": 0.5498,
+      "step": 69181
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.561609968985778e-07,
+      "loss": 0.645,
+      "step": 69182
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.561341617506211e-07,
+      "loss": 0.6582,
+      "step": 69183
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.561073267893813e-07,
+      "loss": 0.6514,
+      "step": 69184
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.560804920148792e-07,
+      "loss": 0.6885,
+      "step": 69185
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.560536574271345e-07,
+      "loss": 0.7515,
+      "step": 69186
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.560268230261687e-07,
+      "loss": 0.5889,
+      "step": 69187
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.559999888120017e-07,
+      "loss": 0.4897,
+      "step": 69188
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.559731547846544e-07,
+      "loss": 0.6128,
+      "step": 69189
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.559463209441471e-07,
+      "loss": 0.623,
+      "step": 69190
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.559194872905005e-07,
+      "loss": 0.6782,
+      "step": 69191
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.558926538237353e-07,
+      "loss": 0.6777,
+      "step": 69192
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.558658205438718e-07,
+      "loss": 0.4617,
+      "step": 69193
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.558389874509305e-07,
+      "loss": 0.5347,
+      "step": 69194
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.558121545449327e-07,
+      "loss": 0.4531,
+      "step": 69195
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.557853218258976e-07,
+      "loss": 0.6504,
+      "step": 69196
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.557584892938466e-07,
+      "loss": 0.7432,
+      "step": 69197
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.557316569488e-07,
+      "loss": 0.6211,
+      "step": 69198
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.557048247907787e-07,
+      "loss": 0.6326,
+      "step": 69199
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.556779928198027e-07,
+      "loss": 0.5432,
+      "step": 69200
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55651161035893e-07,
+      "loss": 0.7031,
+      "step": 69201
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.556243294390699e-07,
+      "loss": 0.6167,
+      "step": 69202
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55597498029354e-07,
+      "loss": 0.4595,
+      "step": 69203
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.555706668067658e-07,
+      "loss": 0.6367,
+      "step": 69204
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.555438357713262e-07,
+      "loss": 0.7324,
+      "step": 69205
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.555170049230554e-07,
+      "loss": 0.5688,
+      "step": 69206
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.554901742619737e-07,
+      "loss": 0.6958,
+      "step": 69207
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.554633437881021e-07,
+      "loss": 0.6328,
+      "step": 69208
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.554365135014609e-07,
+      "loss": 0.791,
+      "step": 69209
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.554096834020707e-07,
+      "loss": 0.6782,
+      "step": 69210
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55382853489952e-07,
+      "loss": 0.5142,
+      "step": 69211
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.553560237651256e-07,
+      "loss": 0.7373,
+      "step": 69212
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.553291942276116e-07,
+      "loss": 0.6875,
+      "step": 69213
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55302364877431e-07,
+      "loss": 0.552,
+      "step": 69214
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.552755357146044e-07,
+      "loss": 0.5933,
+      "step": 69215
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.552487067391518e-07,
+      "loss": 0.7246,
+      "step": 69216
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55221877951094e-07,
+      "loss": 0.8877,
+      "step": 69217
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.551950493504514e-07,
+      "loss": 0.4077,
+      "step": 69218
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.551682209372447e-07,
+      "loss": 0.644,
+      "step": 69219
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.551413927114946e-07,
+      "loss": 0.6133,
+      "step": 69220
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.551145646732213e-07,
+      "loss": 0.7183,
+      "step": 69221
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.550877368224458e-07,
+      "loss": 0.6895,
+      "step": 69222
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.550609091591879e-07,
+      "loss": 0.6416,
+      "step": 69223
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.55034081683469e-07,
+      "loss": 0.6885,
+      "step": 69224
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.550072543953095e-07,
+      "loss": 0.6719,
+      "step": 69225
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.549804272947292e-07,
+      "loss": 0.6058,
+      "step": 69226
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.549536003817493e-07,
+      "loss": 0.5225,
+      "step": 69227
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.549267736563899e-07,
+      "loss": 0.7031,
+      "step": 69228
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.548999471186722e-07,
+      "loss": 0.6216,
+      "step": 69229
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.548731207686159e-07,
+      "loss": 0.6543,
+      "step": 69230
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.548462946062423e-07,
+      "loss": 0.5562,
+      "step": 69231
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.548194686315715e-07,
+      "loss": 0.7451,
+      "step": 69232
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54792642844624e-07,
+      "loss": 0.8564,
+      "step": 69233
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.547658172454209e-07,
+      "loss": 0.5991,
+      "step": 69234
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.547389918339823e-07,
+      "loss": 0.5151,
+      "step": 69235
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.547121666103286e-07,
+      "loss": 0.5918,
+      "step": 69236
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.546853415744805e-07,
+      "loss": 0.7773,
+      "step": 69237
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.546585167264586e-07,
+      "loss": 0.6021,
+      "step": 69238
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.546316920662833e-07,
+      "loss": 0.6836,
+      "step": 69239
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.546048675939751e-07,
+      "loss": 0.4785,
+      "step": 69240
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54578043309555e-07,
+      "loss": 0.6699,
+      "step": 69241
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54551219213043e-07,
+      "loss": 0.7314,
+      "step": 69242
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.5452439530446e-07,
+      "loss": 0.5371,
+      "step": 69243
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54497571583826e-07,
+      "loss": 0.6566,
+      "step": 69244
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.544707480511626e-07,
+      "loss": 0.5835,
+      "step": 69245
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.544439247064893e-07,
+      "loss": 0.7744,
+      "step": 69246
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54417101549827e-07,
+      "loss": 0.6162,
+      "step": 69247
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.543902785811962e-07,
+      "loss": 0.6006,
+      "step": 69248
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.543634558006176e-07,
+      "loss": 0.6523,
+      "step": 69249
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.543366332081114e-07,
+      "loss": 0.7559,
+      "step": 69250
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.543098108036985e-07,
+      "loss": 0.5723,
+      "step": 69251
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.542829885873991e-07,
+      "loss": 0.478,
+      "step": 69252
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.54256166559234e-07,
+      "loss": 0.7534,
+      "step": 69253
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.542293447192236e-07,
+      "loss": 0.5527,
+      "step": 69254
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.542025230673888e-07,
+      "loss": 0.4822,
+      "step": 69255
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.541757016037497e-07,
+      "loss": 0.5962,
+      "step": 69256
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.541488803283267e-07,
+      "loss": 0.5624,
+      "step": 69257
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.541220592411408e-07,
+      "loss": 0.4399,
+      "step": 69258
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.540952383422122e-07,
+      "loss": 0.6147,
+      "step": 69259
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.540684176315617e-07,
+      "loss": 0.5352,
+      "step": 69260
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.540415971092096e-07,
+      "loss": 0.6631,
+      "step": 69261
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.540147767751766e-07,
+      "loss": 0.6431,
+      "step": 69262
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.539879566294832e-07,
+      "loss": 0.4443,
+      "step": 69263
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.539611366721499e-07,
+      "loss": 0.7104,
+      "step": 69264
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.539343169031975e-07,
+      "loss": 0.5862,
+      "step": 69265
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.539074973226459e-07,
+      "loss": 0.4834,
+      "step": 69266
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.538806779305162e-07,
+      "loss": 0.5742,
+      "step": 69267
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.538538587268287e-07,
+      "loss": 0.5747,
+      "step": 69268
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.538270397116038e-07,
+      "loss": 0.6401,
+      "step": 69269
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.538002208848625e-07,
+      "loss": 0.6348,
+      "step": 69270
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.537734022466248e-07,
+      "loss": 0.6758,
+      "step": 69271
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.537465837969117e-07,
+      "loss": 0.4912,
+      "step": 69272
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.537197655357433e-07,
+      "loss": 0.5435,
+      "step": 69273
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.536929474631405e-07,
+      "loss": 0.7734,
+      "step": 69274
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.53666129579124e-07,
+      "loss": 0.7441,
+      "step": 69275
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.536393118837136e-07,
+      "loss": 0.8574,
+      "step": 69276
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.536124943769303e-07,
+      "loss": 0.52,
+      "step": 69277
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.535856770587945e-07,
+      "loss": 0.6689,
+      "step": 69278
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.53558859929327e-07,
+      "loss": 0.6533,
+      "step": 69279
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.535320429885478e-07,
+      "loss": 0.4573,
+      "step": 69280
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.535052262364782e-07,
+      "loss": 0.5977,
+      "step": 69281
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.534784096731381e-07,
+      "loss": 0.7471,
+      "step": 69282
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.534515932985484e-07,
+      "loss": 0.5576,
+      "step": 69283
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.534247771127293e-07,
+      "loss": 0.7188,
+      "step": 69284
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.533979611157021e-07,
+      "loss": 0.6089,
+      "step": 69285
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.533711453074861e-07,
+      "loss": 0.6372,
+      "step": 69286
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.533443296881028e-07,
+      "loss": 0.7109,
+      "step": 69287
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.533175142575721e-07,
+      "loss": 0.6079,
+      "step": 69288
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.532906990159149e-07,
+      "loss": 0.7129,
+      "step": 69289
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.532638839631516e-07,
+      "loss": 0.6763,
+      "step": 69290
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.532370690993031e-07,
+      "loss": 0.5933,
+      "step": 69291
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.532102544243893e-07,
+      "loss": 0.6416,
+      "step": 69292
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.531834399384314e-07,
+      "loss": 0.6699,
+      "step": 69293
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.531566256414492e-07,
+      "loss": 0.5094,
+      "step": 69294
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.531298115334641e-07,
+      "loss": 0.499,
+      "step": 69295
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.53102997614496e-07,
+      "loss": 0.5391,
+      "step": 69296
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.530761838845653e-07,
+      "loss": 0.6602,
+      "step": 69297
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.530493703436931e-07,
+      "loss": 0.6953,
+      "step": 69298
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.530225569918993e-07,
+      "loss": 0.5165,
+      "step": 69299
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.529957438292052e-07,
+      "loss": 0.5049,
+      "step": 69300
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.529689308556305e-07,
+      "loss": 0.4843,
+      "step": 69301
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.529421180711964e-07,
+      "loss": 0.6182,
+      "step": 69302
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.52915305475923e-07,
+      "loss": 0.8311,
+      "step": 69303
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.528884930698312e-07,
+      "loss": 0.5791,
+      "step": 69304
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.528616808529415e-07,
+      "loss": 0.7334,
+      "step": 69305
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.528348688252739e-07,
+      "loss": 0.6223,
+      "step": 69306
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.528080569868492e-07,
+      "loss": 0.5408,
+      "step": 69307
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.527812453376883e-07,
+      "loss": 0.6777,
+      "step": 69308
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.527544338778112e-07,
+      "loss": 0.625,
+      "step": 69309
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.527276226072386e-07,
+      "loss": 0.8169,
+      "step": 69310
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.527008115259911e-07,
+      "loss": 0.6484,
+      "step": 69311
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.526740006340894e-07,
+      "loss": 0.5552,
+      "step": 69312
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.526471899315537e-07,
+      "loss": 0.5886,
+      "step": 69313
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.526203794184048e-07,
+      "loss": 0.4897,
+      "step": 69314
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.525935690946633e-07,
+      "loss": 0.792,
+      "step": 69315
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.525667589603491e-07,
+      "loss": 0.791,
+      "step": 69316
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.525399490154834e-07,
+      "loss": 0.7402,
+      "step": 69317
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.525131392600863e-07,
+      "loss": 0.5674,
+      "step": 69318
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.524863296941788e-07,
+      "loss": 0.791,
+      "step": 69319
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.524595203177808e-07,
+      "loss": 0.6455,
+      "step": 69320
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.524327111309135e-07,
+      "loss": 0.6926,
+      "step": 69321
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.524059021335969e-07,
+      "loss": 0.7578,
+      "step": 69322
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.523790933258516e-07,
+      "loss": 0.5894,
+      "step": 69323
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.523522847076984e-07,
+      "loss": 0.709,
+      "step": 69324
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.52325476279158e-07,
+      "loss": 0.5425,
+      "step": 69325
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.5229866804025e-07,
+      "loss": 0.6335,
+      "step": 69326
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.522718599909958e-07,
+      "loss": 0.5415,
+      "step": 69327
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.522450521314155e-07,
+      "loss": 0.5469,
+      "step": 69328
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.522182444615299e-07,
+      "loss": 0.6108,
+      "step": 69329
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.521914369813594e-07,
+      "loss": 0.5757,
+      "step": 69330
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.521646296909244e-07,
+      "loss": 0.6484,
+      "step": 69331
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.521378225902455e-07,
+      "loss": 0.5439,
+      "step": 69332
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.521110156793435e-07,
+      "loss": 0.4326,
+      "step": 69333
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.520842089582386e-07,
+      "loss": 0.7725,
+      "step": 69334
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.520574024269517e-07,
+      "loss": 0.4594,
+      "step": 69335
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.520305960855028e-07,
+      "loss": 0.5566,
+      "step": 69336
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.520037899339124e-07,
+      "loss": 0.6135,
+      "step": 69337
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.519769839722017e-07,
+      "loss": 0.6006,
+      "step": 69338
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.519501782003905e-07,
+      "loss": 0.7637,
+      "step": 69339
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.519233726184999e-07,
+      "loss": 0.6533,
+      "step": 69340
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.518965672265502e-07,
+      "loss": 0.5884,
+      "step": 69341
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.518697620245616e-07,
+      "loss": 0.6533,
+      "step": 69342
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.518429570125551e-07,
+      "loss": 0.6406,
+      "step": 69343
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.51816152190551e-07,
+      "loss": 0.7539,
+      "step": 69344
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.5178934755857e-07,
+      "loss": 0.5403,
+      "step": 69345
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.517625431166325e-07,
+      "loss": 0.5488,
+      "step": 69346
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.517357388647587e-07,
+      "loss": 0.6499,
+      "step": 69347
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.517089348029697e-07,
+      "loss": 0.7188,
+      "step": 69348
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.516821309312855e-07,
+      "loss": 0.6621,
+      "step": 69349
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.51655327249727e-07,
+      "loss": 0.6626,
+      "step": 69350
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.516285237583146e-07,
+      "loss": 0.644,
+      "step": 69351
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.516017204570689e-07,
+      "loss": 0.6919,
+      "step": 69352
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.515749173460101e-07,
+      "loss": 0.5918,
+      "step": 69353
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.515481144251592e-07,
+      "loss": 0.7178,
+      "step": 69354
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.515213116945367e-07,
+      "loss": 0.5212,
+      "step": 69355
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.514945091541624e-07,
+      "loss": 0.6382,
+      "step": 69356
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.514677068040578e-07,
+      "loss": 0.7197,
+      "step": 69357
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.514409046442427e-07,
+      "loss": 0.5444,
+      "step": 69358
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.514141026747378e-07,
+      "loss": 0.5752,
+      "step": 69359
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.513873008955638e-07,
+      "loss": 0.6548,
+      "step": 69360
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.513604993067409e-07,
+      "loss": 0.5918,
+      "step": 69361
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.513336979082901e-07,
+      "loss": 0.6338,
+      "step": 69362
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.513068967002314e-07,
+      "loss": 0.499,
+      "step": 69363
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.512800956825859e-07,
+      "loss": 0.5188,
+      "step": 69364
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.512532948553739e-07,
+      "loss": 0.689,
+      "step": 69365
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.512264942186154e-07,
+      "loss": 0.5815,
+      "step": 69366
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.511996937723316e-07,
+      "loss": 0.4506,
+      "step": 69367
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.511728935165424e-07,
+      "loss": 0.7266,
+      "step": 69368
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.51146093451269e-07,
+      "loss": 0.5508,
+      "step": 69369
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.511192935765314e-07,
+      "loss": 0.709,
+      "step": 69370
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.510924938923506e-07,
+      "loss": 0.5825,
+      "step": 69371
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.510656943987464e-07,
+      "loss": 0.5972,
+      "step": 69372
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.510388950957401e-07,
+      "loss": 0.7363,
+      "step": 69373
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.510120959833517e-07,
+      "loss": 0.5283,
+      "step": 69374
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.509852970616024e-07,
+      "loss": 0.5718,
+      "step": 69375
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.509584983305116e-07,
+      "loss": 0.6924,
+      "step": 69376
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.509316997901008e-07,
+      "loss": 0.604,
+      "step": 69377
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.509049014403897e-07,
+      "loss": 0.7085,
+      "step": 69378
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.508781032813996e-07,
+      "loss": 0.6475,
+      "step": 69379
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.508513053131505e-07,
+      "loss": 0.377,
+      "step": 69380
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.508245075356631e-07,
+      "loss": 0.624,
+      "step": 69381
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.507977099489579e-07,
+      "loss": 0.6675,
+      "step": 69382
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.507709125530557e-07,
+      "loss": 0.6582,
+      "step": 69383
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.507441153479764e-07,
+      "loss": 0.5747,
+      "step": 69384
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.507173183337415e-07,
+      "loss": 0.6729,
+      "step": 69385
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.506905215103704e-07,
+      "loss": 0.5215,
+      "step": 69386
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.50663724877884e-07,
+      "loss": 0.4514,
+      "step": 69387
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.506369284363032e-07,
+      "loss": 0.5298,
+      "step": 69388
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.50610132185648e-07,
+      "loss": 0.6348,
+      "step": 69389
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.505833361259393e-07,
+      "loss": 0.6602,
+      "step": 69390
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.505565402571973e-07,
+      "loss": 0.6816,
+      "step": 69391
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.505297445794428e-07,
+      "loss": 0.7412,
+      "step": 69392
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.505029490926962e-07,
+      "loss": 0.7227,
+      "step": 69393
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.50476153796978e-07,
+      "loss": 0.6079,
+      "step": 69394
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.504493586923091e-07,
+      "loss": 0.6592,
+      "step": 69395
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.504225637787094e-07,
+      "loss": 0.7051,
+      "step": 69396
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.503957690561995e-07,
+      "loss": 0.6592,
+      "step": 69397
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.503689745248003e-07,
+      "loss": 0.6919,
+      "step": 69398
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.503421801845318e-07,
+      "loss": 0.6523,
+      "step": 69399
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.50315386035415e-07,
+      "loss": 0.5913,
+      "step": 69400
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.502885920774702e-07,
+      "loss": 0.7422,
+      "step": 69401
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.50261798310718e-07,
+      "loss": 0.7202,
+      "step": 69402
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.502350047351786e-07,
+      "loss": 0.6016,
+      "step": 69403
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.502082113508732e-07,
+      "loss": 0.7979,
+      "step": 69404
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.501814181578219e-07,
+      "loss": 0.6836,
+      "step": 69405
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.501546251560449e-07,
+      "loss": 0.6318,
+      "step": 69406
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.501278323455631e-07,
+      "loss": 0.7832,
+      "step": 69407
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.501010397263968e-07,
+      "loss": 0.686,
+      "step": 69408
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.500742472985668e-07,
+      "loss": 0.689,
+      "step": 69409
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.500474550620933e-07,
+      "loss": 0.6182,
+      "step": 69410
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.500206630169972e-07,
+      "loss": 0.6128,
+      "step": 69411
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.499938711632987e-07,
+      "loss": 0.5088,
+      "step": 69412
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.499670795010183e-07,
+      "loss": 0.5361,
+      "step": 69413
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.499402880301769e-07,
+      "loss": 0.5708,
+      "step": 69414
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.499134967507947e-07,
+      "loss": 0.686,
+      "step": 69415
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.49886705662892e-07,
+      "loss": 0.4614,
+      "step": 69416
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.498599147664896e-07,
+      "loss": 0.6465,
+      "step": 69417
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.498331240616078e-07,
+      "loss": 0.6357,
+      "step": 69418
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.498063335482675e-07,
+      "loss": 0.6743,
+      "step": 69419
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.497795432264888e-07,
+      "loss": 0.4443,
+      "step": 69420
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.497527530962926e-07,
+      "loss": 0.4478,
+      "step": 69421
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.49725963157699e-07,
+      "loss": 0.3788,
+      "step": 69422
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.49699173410729e-07,
+      "loss": 0.5432,
+      "step": 69423
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.496723838554025e-07,
+      "loss": 0.6426,
+      "step": 69424
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.496455944917408e-07,
+      "loss": 0.5784,
+      "step": 69425
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.496188053197637e-07,
+      "loss": 0.5796,
+      "step": 69426
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.495920163394918e-07,
+      "loss": 0.5518,
+      "step": 69427
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.495652275509461e-07,
+      "loss": 0.6045,
+      "step": 69428
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.495384389541464e-07,
+      "loss": 0.5386,
+      "step": 69429
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.495116505491139e-07,
+      "loss": 0.707,
+      "step": 69430
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.494848623358688e-07,
+      "loss": 0.6523,
+      "step": 69431
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.494580743144314e-07,
+      "loss": 0.4976,
+      "step": 69432
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.494312864848227e-07,
+      "loss": 0.6377,
+      "step": 69433
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.494044988470627e-07,
+      "loss": 0.6855,
+      "step": 69434
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.493777114011726e-07,
+      "loss": 0.7578,
+      "step": 69435
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.493509241471722e-07,
+      "loss": 0.6128,
+      "step": 69436
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.49324137085082e-07,
+      "loss": 0.4553,
+      "step": 69437
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.492973502149232e-07,
+      "loss": 0.5942,
+      "step": 69438
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.492705635367156e-07,
+      "loss": 0.5952,
+      "step": 69439
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.4924377705048e-07,
+      "loss": 0.5647,
+      "step": 69440
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.492169907562369e-07,
+      "loss": 0.7783,
+      "step": 69441
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.491902046540071e-07,
+      "loss": 0.6314,
+      "step": 69442
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.491634187438105e-07,
+      "loss": 0.6611,
+      "step": 69443
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.491366330256681e-07,
+      "loss": 0.5591,
+      "step": 69444
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.491098474996005e-07,
+      "loss": 0.6494,
+      "step": 69445
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.490830621656277e-07,
+      "loss": 0.6035,
+      "step": 69446
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.490562770237703e-07,
+      "loss": 0.5444,
+      "step": 69447
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.490294920740493e-07,
+      "loss": 0.6426,
+      "step": 69448
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.490027073164847e-07,
+      "loss": 0.7026,
+      "step": 69449
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.489759227510971e-07,
+      "loss": 0.5305,
+      "step": 69450
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.489491383779071e-07,
+      "loss": 0.5166,
+      "step": 69451
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.489223541969353e-07,
+      "loss": 0.7754,
+      "step": 69452
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.488955702082019e-07,
+      "loss": 0.6807,
+      "step": 69453
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.488687864117281e-07,
+      "loss": 0.5241,
+      "step": 69454
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.488420028075337e-07,
+      "loss": 0.6387,
+      "step": 69455
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.488152193956392e-07,
+      "loss": 0.6709,
+      "step": 69456
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.487884361760657e-07,
+      "loss": 0.4626,
+      "step": 69457
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.487616531488329e-07,
+      "loss": 0.6162,
+      "step": 69458
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.48734870313962e-07,
+      "loss": 0.5728,
+      "step": 69459
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.487080876714731e-07,
+      "loss": 0.6704,
+      "step": 69460
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.486813052213872e-07,
+      "loss": 0.6924,
+      "step": 69461
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.48654522963724e-07,
+      "loss": 0.7783,
+      "step": 69462
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.486277408985049e-07,
+      "loss": 0.5234,
+      "step": 69463
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.4860095902575e-07,
+      "loss": 0.6235,
+      "step": 69464
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.485741773454798e-07,
+      "loss": 0.5244,
+      "step": 69465
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.485473958577144e-07,
+      "loss": 0.5369,
+      "step": 69466
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.485206145624749e-07,
+      "loss": 0.707,
+      "step": 69467
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.484938334597815e-07,
+      "loss": 0.6963,
+      "step": 69468
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.484670525496551e-07,
+      "loss": 0.6963,
+      "step": 69469
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.484402718321155e-07,
+      "loss": 0.5732,
+      "step": 69470
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.48413491307184e-07,
+      "loss": 0.5557,
+      "step": 69471
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.483867109748804e-07,
+      "loss": 0.5649,
+      "step": 69472
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.483599308352258e-07,
+      "loss": 0.5557,
+      "step": 69473
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.483331508882405e-07,
+      "loss": 0.7041,
+      "step": 69474
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.483063711339446e-07,
+      "loss": 0.4834,
+      "step": 69475
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.482795915723592e-07,
+      "loss": 0.4556,
+      "step": 69476
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.482528122035044e-07,
+      "loss": 0.6719,
+      "step": 69477
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.482260330274009e-07,
+      "loss": 0.7637,
+      "step": 69478
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.481992540440689e-07,
+      "loss": 0.6465,
+      "step": 69479
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.481724752535294e-07,
+      "loss": 0.4971,
+      "step": 69480
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.481456966558026e-07,
+      "loss": 0.6094,
+      "step": 69481
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.481189182509092e-07,
+      "loss": 0.5626,
+      "step": 69482
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.480921400388693e-07,
+      "loss": 0.7881,
+      "step": 69483
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.480653620197044e-07,
+      "loss": 0.5217,
+      "step": 69484
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.480385841934335e-07,
+      "loss": 0.3916,
+      "step": 69485
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.480118065600781e-07,
+      "loss": 0.5977,
+      "step": 69486
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.479850291196583e-07,
+      "loss": 0.5146,
+      "step": 69487
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.47958251872195e-07,
+      "loss": 0.7373,
+      "step": 69488
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.479314748177082e-07,
+      "loss": 0.7178,
+      "step": 69489
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.47904697956219e-07,
+      "loss": 0.6289,
+      "step": 69490
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.478779212877473e-07,
+      "loss": 0.6567,
+      "step": 69491
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.478511448123141e-07,
+      "loss": 0.5845,
+      "step": 69492
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.478243685299394e-07,
+      "loss": 0.6282,
+      "step": 69493
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.477975924406446e-07,
+      "loss": 0.7314,
+      "step": 69494
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.477708165444491e-07,
+      "loss": 0.5942,
+      "step": 69495
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.47744040841374e-07,
+      "loss": 0.6846,
+      "step": 69496
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.477172653314397e-07,
+      "loss": 0.5874,
+      "step": 69497
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.476904900146666e-07,
+      "loss": 0.7734,
+      "step": 69498
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.476637148910754e-07,
+      "loss": 0.5942,
+      "step": 69499
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.476369399606862e-07,
+      "loss": 0.6162,
+      "step": 69500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.4761016522352e-07,
+      "loss": 0.5073,
+      "step": 69501
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.475833906795972e-07,
+      "loss": 0.6299,
+      "step": 69502
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.475566163289381e-07,
+      "loss": 0.6255,
+      "step": 69503
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.475298421715635e-07,
+      "loss": 0.5303,
+      "step": 69504
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.475030682074937e-07,
+      "loss": 0.6074,
+      "step": 69505
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.474762944367487e-07,
+      "loss": 0.7451,
+      "step": 69506
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.474495208593499e-07,
+      "loss": 0.7568,
+      "step": 69507
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.474227474753172e-07,
+      "loss": 0.6338,
+      "step": 69508
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.473959742846714e-07,
+      "loss": 0.5903,
+      "step": 69509
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.473692012874326e-07,
+      "loss": 0.7812,
+      "step": 69510
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.47342428483622e-07,
+      "loss": 0.688,
+      "step": 69511
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.473156558732594e-07,
+      "loss": 0.665,
+      "step": 69512
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.472888834563657e-07,
+      "loss": 0.6572,
+      "step": 69513
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.472621112329616e-07,
+      "loss": 0.7217,
+      "step": 69514
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.472353392030668e-07,
+      "loss": 0.6289,
+      "step": 69515
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.472085673667026e-07,
+      "loss": 0.6489,
+      "step": 69516
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.471817957238888e-07,
+      "loss": 0.5703,
+      "step": 69517
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.471550242746465e-07,
+      "loss": 0.6914,
+      "step": 69518
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.471282530189959e-07,
+      "loss": 0.4622,
+      "step": 69519
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.471014819569577e-07,
+      "loss": 0.6704,
+      "step": 69520
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.470747110885523e-07,
+      "loss": 0.7432,
+      "step": 69521
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.470479404137998e-07,
+      "loss": 0.624,
+      "step": 69522
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.470211699327214e-07,
+      "loss": 0.4297,
+      "step": 69523
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.469943996453375e-07,
+      "loss": 0.5068,
+      "step": 69524
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.469676295516678e-07,
+      "loss": 0.6357,
+      "step": 69525
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.469408596517337e-07,
+      "loss": 0.5137,
+      "step": 69526
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.46914089945555e-07,
+      "loss": 0.7783,
+      "step": 69527
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.468873204331529e-07,
+      "loss": 0.5311,
+      "step": 69528
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.468605511145474e-07,
+      "loss": 0.6616,
+      "step": 69529
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.468337819897592e-07,
+      "loss": 0.6855,
+      "step": 69530
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.468070130588085e-07,
+      "loss": 0.6172,
+      "step": 69531
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.467802443217162e-07,
+      "loss": 0.5469,
+      "step": 69532
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.467534757785026e-07,
+      "loss": 0.5703,
+      "step": 69533
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.467267074291885e-07,
+      "loss": 0.7061,
+      "step": 69534
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.46699939273794e-07,
+      "loss": 0.7168,
+      "step": 69535
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.466731713123393e-07,
+      "loss": 0.7568,
+      "step": 69536
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.466464035448457e-07,
+      "loss": 0.7749,
+      "step": 69537
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.466196359713333e-07,
+      "loss": 0.7041,
+      "step": 69538
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.465928685918224e-07,
+      "loss": 0.623,
+      "step": 69539
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.465661014063338e-07,
+      "loss": 0.373,
+      "step": 69540
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.465393344148876e-07,
+      "loss": 0.6738,
+      "step": 69541
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.46512567617505e-07,
+      "loss": 0.7031,
+      "step": 69542
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.464858010142056e-07,
+      "loss": 0.666,
+      "step": 69543
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.46459034605011e-07,
+      "loss": 0.6445,
+      "step": 69544
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.464322683899408e-07,
+      "loss": 0.4788,
+      "step": 69545
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.464055023690155e-07,
+      "loss": 0.6016,
+      "step": 69546
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.463787365422562e-07,
+      "loss": 0.7979,
+      "step": 69547
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.463519709096826e-07,
+      "loss": 0.6431,
+      "step": 69548
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.46325205471316e-07,
+      "loss": 0.751,
+      "step": 69549
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.462984402271764e-07,
+      "loss": 0.6064,
+      "step": 69550
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.462716751772844e-07,
+      "loss": 0.405,
+      "step": 69551
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.462449103216604e-07,
+      "loss": 0.688,
+      "step": 69552
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.462181456603253e-07,
+      "loss": 0.7593,
+      "step": 69553
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.461913811932994e-07,
+      "loss": 0.7271,
+      "step": 69554
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.461646169206029e-07,
+      "loss": 0.625,
+      "step": 69555
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.461378528422562e-07,
+      "loss": 0.5596,
+      "step": 69556
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.461110889582805e-07,
+      "loss": 0.7065,
+      "step": 69557
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.460843252686955e-07,
+      "loss": 0.6504,
+      "step": 69558
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.460575617735223e-07,
+      "loss": 0.666,
+      "step": 69559
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.460307984727809e-07,
+      "loss": 0.6782,
+      "step": 69560
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.460040353664923e-07,
+      "loss": 0.3838,
+      "step": 69561
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.459772724546765e-07,
+      "loss": 0.5693,
+      "step": 69562
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.459505097373544e-07,
+      "loss": 0.5801,
+      "step": 69563
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.459237472145465e-07,
+      "loss": 0.7021,
+      "step": 69564
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.458969848862727e-07,
+      "loss": 0.6836,
+      "step": 69565
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.458702227525542e-07,
+      "loss": 0.75,
+      "step": 69566
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.458434608134108e-07,
+      "loss": 0.6738,
+      "step": 69567
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.458166990688638e-07,
+      "loss": 0.8369,
+      "step": 69568
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.45789937518933e-07,
+      "loss": 0.5017,
+      "step": 69569
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.457631761636393e-07,
+      "loss": 0.6016,
+      "step": 69570
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.457364150030028e-07,
+      "loss": 0.6631,
+      "step": 69571
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.457096540370444e-07,
+      "loss": 0.532,
+      "step": 69572
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.456828932657843e-07,
+      "loss": 0.6631,
+      "step": 69573
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.456561326892439e-07,
+      "loss": 0.6565,
+      "step": 69574
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.456293723074421e-07,
+      "loss": 0.6689,
+      "step": 69575
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.456026121204004e-07,
+      "loss": 0.6763,
+      "step": 69576
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.45575852128139e-07,
+      "loss": 0.6807,
+      "step": 69577
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.455490923306787e-07,
+      "loss": 0.6904,
+      "step": 69578
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.455223327280393e-07,
+      "loss": 0.7334,
+      "step": 69579
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.454955733202421e-07,
+      "loss": 0.7266,
+      "step": 69580
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.454688141073071e-07,
+      "loss": 0.6743,
+      "step": 69581
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.45442055089255e-07,
+      "loss": 0.6528,
+      "step": 69582
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.45415296266106e-07,
+      "loss": 0.7168,
+      "step": 69583
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.453885376378815e-07,
+      "loss": 0.6025,
+      "step": 69584
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.453617792046007e-07,
+      "loss": 0.6133,
+      "step": 69585
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.453350209662848e-07,
+      "loss": 0.6943,
+      "step": 69586
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.453082629229541e-07,
+      "loss": 0.6553,
+      "step": 69587
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.45281505074629e-07,
+      "loss": 0.5625,
+      "step": 69588
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.452547474213304e-07,
+      "loss": 0.6533,
+      "step": 69589
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.452279899630783e-07,
+      "loss": 0.7109,
+      "step": 69590
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.452012326998938e-07,
+      "loss": 0.5679,
+      "step": 69591
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.451744756317966e-07,
+      "loss": 0.7852,
+      "step": 69592
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.451477187588077e-07,
+      "loss": 0.4087,
+      "step": 69593
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.451209620809479e-07,
+      "loss": 0.6182,
+      "step": 69594
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.450942055982369e-07,
+      "loss": 0.7578,
+      "step": 69595
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.450674493106954e-07,
+      "loss": 0.5991,
+      "step": 69596
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.450406932183443e-07,
+      "loss": 0.5474,
+      "step": 69597
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.450139373212035e-07,
+      "loss": 0.6567,
+      "step": 69598
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.449871816192941e-07,
+      "loss": 0.6865,
+      "step": 69599
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.449604261126361e-07,
+      "loss": 0.5366,
+      "step": 69600
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.449336708012503e-07,
+      "loss": 0.6421,
+      "step": 69601
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.449069156851568e-07,
+      "loss": 0.6375,
+      "step": 69602
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.448801607643767e-07,
+      "loss": 0.7256,
+      "step": 69603
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.448534060389302e-07,
+      "loss": 0.6162,
+      "step": 69604
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.448266515088374e-07,
+      "loss": 0.6401,
+      "step": 69605
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.447998971741193e-07,
+      "loss": 0.6836,
+      "step": 69606
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.447731430347958e-07,
+      "loss": 0.6758,
+      "step": 69607
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.447463890908883e-07,
+      "loss": 0.8271,
+      "step": 69608
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.447196353424164e-07,
+      "loss": 0.6045,
+      "step": 69609
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.446928817894011e-07,
+      "loss": 0.5291,
+      "step": 69610
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.446661284318627e-07,
+      "loss": 0.4719,
+      "step": 69611
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.446393752698214e-07,
+      "loss": 0.7036,
+      "step": 69612
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.446126223032984e-07,
+      "loss": 0.6279,
+      "step": 69613
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.44585869532314e-07,
+      "loss": 0.8853,
+      "step": 69614
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.445591169568879e-07,
+      "loss": 0.5464,
+      "step": 69615
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.445323645770414e-07,
+      "loss": 0.5088,
+      "step": 69616
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.445056123927945e-07,
+      "loss": 0.5398,
+      "step": 69617
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.444788604041681e-07,
+      "loss": 0.5713,
+      "step": 69618
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.444521086111824e-07,
+      "loss": 0.6182,
+      "step": 69619
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.44425357013858e-07,
+      "loss": 0.4998,
+      "step": 69620
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.443986056122152e-07,
+      "loss": 0.4751,
+      "step": 69621
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.443718544062749e-07,
+      "loss": 0.3971,
+      "step": 69622
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.443451033960571e-07,
+      "loss": 0.6816,
+      "step": 69623
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.443183525815829e-07,
+      "loss": 0.709,
+      "step": 69624
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.442916019628723e-07,
+      "loss": 0.5659,
+      "step": 69625
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.442648515399455e-07,
+      "loss": 0.7266,
+      "step": 69626
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.442381013128234e-07,
+      "loss": 0.3699,
+      "step": 69627
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.442113512815266e-07,
+      "loss": 0.532,
+      "step": 69628
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.441846014460753e-07,
+      "loss": 0.7358,
+      "step": 69629
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.441578518064901e-07,
+      "loss": 0.7266,
+      "step": 69630
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.441311023627915e-07,
+      "loss": 0.7061,
+      "step": 69631
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.44104353115e-07,
+      "loss": 0.5933,
+      "step": 69632
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.440776040631358e-07,
+      "loss": 0.709,
+      "step": 69633
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.4405085520722e-07,
+      "loss": 0.5674,
+      "step": 69634
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.440241065472725e-07,
+      "loss": 0.5835,
+      "step": 69635
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.439973580833139e-07,
+      "loss": 0.667,
+      "step": 69636
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.439706098153648e-07,
+      "loss": 0.7607,
+      "step": 69637
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.439438617434454e-07,
+      "loss": 0.478,
+      "step": 69638
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.439171138675768e-07,
+      "loss": 0.6069,
+      "step": 69639
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.438903661877788e-07,
+      "loss": 0.6113,
+      "step": 69640
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.438636187040724e-07,
+      "loss": 0.5547,
+      "step": 69641
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.438368714164775e-07,
+      "loss": 0.7256,
+      "step": 69642
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 7.438101243250153e-07,
+      "loss": 0.5911,
+      "step": 69643
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.437833774297061e-07,
+      "loss": 0.5815,
+      "step": 69644
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.4375663073057e-07,
+      "loss": 0.4531,
+      "step": 69645
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.437298842276274e-07,
+      "loss": 0.7778,
+      "step": 69646
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.437031379208992e-07,
+      "loss": 0.7021,
+      "step": 69647
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.436763918104056e-07,
+      "loss": 0.6475,
+      "step": 69648
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.436496458961674e-07,
+      "loss": 0.6621,
+      "step": 69649
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.436229001782046e-07,
+      "loss": 0.5505,
+      "step": 69650
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.435961546565383e-07,
+      "loss": 0.6938,
+      "step": 69651
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.435694093311882e-07,
+      "loss": 0.6333,
+      "step": 69652
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.435426642021756e-07,
+      "loss": 0.7217,
+      "step": 69653
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.435159192695208e-07,
+      "loss": 0.5347,
+      "step": 69654
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.434891745332436e-07,
+      "loss": 0.8574,
+      "step": 69655
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.434624299933651e-07,
+      "loss": 0.7578,
+      "step": 69656
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.434356856499055e-07,
+      "loss": 0.6455,
+      "step": 69657
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.434089415028857e-07,
+      "loss": 0.7109,
+      "step": 69658
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.433821975523255e-07,
+      "loss": 0.6514,
+      "step": 69659
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.43355453798246e-07,
+      "loss": 0.5278,
+      "step": 69660
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.433287102406672e-07,
+      "loss": 0.5117,
+      "step": 69661
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.433019668796101e-07,
+      "loss": 0.8408,
+      "step": 69662
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.432752237150945e-07,
+      "loss": 0.6821,
+      "step": 69663
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.432484807471421e-07,
+      "loss": 0.7324,
+      "step": 69664
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.432217379757717e-07,
+      "loss": 0.5972,
+      "step": 69665
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.431949954010051e-07,
+      "loss": 0.5532,
+      "step": 69666
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.431682530228618e-07,
+      "loss": 0.7715,
+      "step": 69667
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.43141510841363e-07,
+      "loss": 0.7295,
+      "step": 69668
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.431147688565288e-07,
+      "loss": 0.8525,
+      "step": 69669
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.4308802706838e-07,
+      "loss": 0.71,
+      "step": 69670
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.430612854769367e-07,
+      "loss": 0.7744,
+      "step": 69671
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.430345440822198e-07,
+      "loss": 0.7031,
+      "step": 69672
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.430078028842492e-07,
+      "loss": 0.5488,
+      "step": 69673
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.429810618830462e-07,
+      "loss": 0.5269,
+      "step": 69674
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.429543210786306e-07,
+      "loss": 0.5859,
+      "step": 69675
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.429275804710229e-07,
+      "loss": 0.6079,
+      "step": 69676
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.429008400602438e-07,
+      "loss": 0.5975,
+      "step": 69677
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.428740998463136e-07,
+      "loss": 0.6494,
+      "step": 69678
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.42847359829253e-07,
+      "loss": 0.5933,
+      "step": 69679
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.428206200090822e-07,
+      "loss": 0.6082,
+      "step": 69680
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.427938803858221e-07,
+      "loss": 0.4421,
+      "step": 69681
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.427671409594926e-07,
+      "loss": 0.665,
+      "step": 69682
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.427404017301147e-07,
+      "loss": 0.6885,
+      "step": 69683
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.427136626977089e-07,
+      "loss": 0.6162,
+      "step": 69684
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.42686923862295e-07,
+      "loss": 0.7178,
+      "step": 69685
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.426601852238939e-07,
+      "loss": 0.5908,
+      "step": 69686
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.426334467825262e-07,
+      "loss": 0.6631,
+      "step": 69687
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.42606708538212e-07,
+      "loss": 0.6426,
+      "step": 69688
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.425799704909723e-07,
+      "loss": 0.4897,
+      "step": 69689
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.425532326408271e-07,
+      "loss": 0.8359,
+      "step": 69690
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.425264949877972e-07,
+      "loss": 0.7012,
+      "step": 69691
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.424997575319027e-07,
+      "loss": 0.6172,
+      "step": 69692
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.424730202731644e-07,
+      "loss": 0.7188,
+      "step": 69693
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.42446283211603e-07,
+      "loss": 0.7666,
+      "step": 69694
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.424195463472382e-07,
+      "loss": 0.5156,
+      "step": 69695
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.423928096800911e-07,
+      "loss": 0.7158,
+      "step": 69696
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.423660732101818e-07,
+      "loss": 0.7012,
+      "step": 69697
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.423393369375313e-07,
+      "loss": 0.6162,
+      "step": 69698
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.423126008621593e-07,
+      "loss": 0.7588,
+      "step": 69699
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.422858649840869e-07,
+      "loss": 0.5928,
+      "step": 69700
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.422591293033344e-07,
+      "loss": 0.5461,
+      "step": 69701
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.422323938199222e-07,
+      "loss": 0.79,
+      "step": 69702
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.422056585338708e-07,
+      "loss": 0.6067,
+      "step": 69703
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.42178923445201e-07,
+      "loss": 0.4937,
+      "step": 69704
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.421521885539325e-07,
+      "loss": 0.54,
+      "step": 69705
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.421254538600864e-07,
+      "loss": 0.7275,
+      "step": 69706
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.420987193636828e-07,
+      "loss": 0.6592,
+      "step": 69707
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.420719850647425e-07,
+      "loss": 0.7002,
+      "step": 69708
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.420452509632857e-07,
+      "loss": 0.4304,
+      "step": 69709
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.420185170593332e-07,
+      "loss": 0.7007,
+      "step": 69710
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.419917833529049e-07,
+      "loss": 0.7266,
+      "step": 69711
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.41965049844022e-07,
+      "loss": 0.5181,
+      "step": 69712
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.419383165327042e-07,
+      "loss": 0.5452,
+      "step": 69713
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.41911583418973e-07,
+      "loss": 0.4404,
+      "step": 69714
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.418848505028479e-07,
+      "loss": 0.5908,
+      "step": 69715
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.418581177843495e-07,
+      "loss": 0.6367,
+      "step": 69716
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.418313852634988e-07,
+      "loss": 0.6533,
+      "step": 69717
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.418046529403157e-07,
+      "loss": 0.8564,
+      "step": 69718
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.417779208148209e-07,
+      "loss": 0.6802,
+      "step": 69719
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.41751188887035e-07,
+      "loss": 0.6504,
+      "step": 69720
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.417244571569781e-07,
+      "loss": 0.6553,
+      "step": 69721
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.416977256246713e-07,
+      "loss": 0.6978,
+      "step": 69722
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.416709942901343e-07,
+      "loss": 0.5967,
+      "step": 69723
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.416442631533883e-07,
+      "loss": 0.6729,
+      "step": 69724
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.416175322144534e-07,
+      "loss": 0.5386,
+      "step": 69725
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.415908014733497e-07,
+      "loss": 0.6494,
+      "step": 69726
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.415640709300983e-07,
+      "loss": 0.6289,
+      "step": 69727
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.415373405847192e-07,
+      "loss": 0.5381,
+      "step": 69728
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.415106104372333e-07,
+      "loss": 0.689,
+      "step": 69729
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.414838804876607e-07,
+      "loss": 0.6255,
+      "step": 69730
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.41457150736022e-07,
+      "loss": 0.7168,
+      "step": 69731
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.414304211823376e-07,
+      "loss": 0.6895,
+      "step": 69732
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.414036918266282e-07,
+      "loss": 0.6426,
+      "step": 69733
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.413769626689144e-07,
+      "loss": 0.7695,
+      "step": 69734
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.413502337092161e-07,
+      "loss": 0.6396,
+      "step": 69735
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.413235049475539e-07,
+      "loss": 0.7314,
+      "step": 69736
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.412967763839484e-07,
+      "loss": 0.5522,
+      "step": 69737
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.412700480184201e-07,
+      "loss": 0.6172,
+      "step": 69738
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.412433198509896e-07,
+      "loss": 0.7891,
+      "step": 69739
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.412165918816768e-07,
+      "loss": 0.7207,
+      "step": 69740
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.41189864110503e-07,
+      "loss": 0.5327,
+      "step": 69741
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.411631365374877e-07,
+      "loss": 0.552,
+      "step": 69742
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.411364091626526e-07,
+      "loss": 0.8086,
+      "step": 69743
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.411096819860172e-07,
+      "loss": 0.493,
+      "step": 69744
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.410829550076019e-07,
+      "loss": 0.5551,
+      "step": 69745
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.410562282274277e-07,
+      "loss": 0.7891,
+      "step": 69746
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.410295016455145e-07,
+      "loss": 0.561,
+      "step": 69747
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.410027752618837e-07,
+      "loss": 0.5801,
+      "step": 69748
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.409760490765546e-07,
+      "loss": 0.6592,
+      "step": 69749
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.409493230895486e-07,
+      "loss": 0.5654,
+      "step": 69750
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.409225973008855e-07,
+      "loss": 0.7432,
+      "step": 69751
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.408958717105864e-07,
+      "loss": 0.4244,
+      "step": 69752
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.408691463186714e-07,
+      "loss": 0.6069,
+      "step": 69753
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40842421125161e-07,
+      "loss": 0.4849,
+      "step": 69754
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.408156961300753e-07,
+      "loss": 0.4719,
+      "step": 69755
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.407889713334355e-07,
+      "loss": 0.5176,
+      "step": 69756
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.407622467352613e-07,
+      "loss": 0.4352,
+      "step": 69757
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.407355223355736e-07,
+      "loss": 0.7598,
+      "step": 69758
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.407087981343927e-07,
+      "loss": 0.6421,
+      "step": 69759
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.406820741317394e-07,
+      "loss": 0.7598,
+      "step": 69760
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.406553503276337e-07,
+      "loss": 0.75,
+      "step": 69761
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.406286267220965e-07,
+      "loss": 0.6899,
+      "step": 69762
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.406019033151483e-07,
+      "loss": 0.6191,
+      "step": 69763
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.405751801068088e-07,
+      "loss": 0.4777,
+      "step": 69764
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40548457097099e-07,
+      "loss": 0.5122,
+      "step": 69765
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.405217342860392e-07,
+      "loss": 0.6494,
+      "step": 69766
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.404950116736503e-07,
+      "loss": 0.5342,
+      "step": 69767
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.404682892599523e-07,
+      "loss": 0.644,
+      "step": 69768
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.404415670449657e-07,
+      "loss": 0.5957,
+      "step": 69769
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.404148450287111e-07,
+      "loss": 0.6006,
+      "step": 69770
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40388123211209e-07,
+      "loss": 0.7017,
+      "step": 69771
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.403614015924797e-07,
+      "loss": 0.3773,
+      "step": 69772
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.403346801725443e-07,
+      "loss": 0.8125,
+      "step": 69773
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40307958951422e-07,
+      "loss": 0.4778,
+      "step": 69774
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.402812379291343e-07,
+      "loss": 0.6162,
+      "step": 69775
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40254517105701e-07,
+      "loss": 0.4275,
+      "step": 69776
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40227796481143e-07,
+      "loss": 0.6631,
+      "step": 69777
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.402010760554807e-07,
+      "loss": 0.708,
+      "step": 69778
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.401743558287345e-07,
+      "loss": 0.4824,
+      "step": 69779
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.401476358009246e-07,
+      "loss": 0.5769,
+      "step": 69780
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.40120915972072e-07,
+      "loss": 0.561,
+      "step": 69781
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.400941963421967e-07,
+      "loss": 0.5656,
+      "step": 69782
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.400674769113197e-07,
+      "loss": 0.4346,
+      "step": 69783
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.400407576794608e-07,
+      "loss": 0.4524,
+      "step": 69784
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.400140386466406e-07,
+      "loss": 0.7256,
+      "step": 69785
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.399873198128799e-07,
+      "loss": 0.6191,
+      "step": 69786
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.399606011781987e-07,
+      "loss": 0.5381,
+      "step": 69787
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.399338827426179e-07,
+      "loss": 0.7261,
+      "step": 69788
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.399071645061576e-07,
+      "loss": 0.8467,
+      "step": 69789
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.398804464688387e-07,
+      "loss": 0.3276,
+      "step": 69790
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.398537286306814e-07,
+      "loss": 0.7275,
+      "step": 69791
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.398270109917058e-07,
+      "loss": 0.6235,
+      "step": 69792
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.398002935519332e-07,
+      "loss": 0.6582,
+      "step": 69793
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.397735763113832e-07,
+      "loss": 0.6602,
+      "step": 69794
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.397468592700766e-07,
+      "loss": 0.6973,
+      "step": 69795
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.397201424280339e-07,
+      "loss": 0.7295,
+      "step": 69796
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.396934257852754e-07,
+      "loss": 0.6426,
+      "step": 69797
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.396667093418219e-07,
+      "loss": 0.6855,
+      "step": 69798
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.396399930976934e-07,
+      "loss": 0.7256,
+      "step": 69799
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.396132770529107e-07,
+      "loss": 0.6265,
+      "step": 69800
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.395865612074941e-07,
+      "loss": 0.6938,
+      "step": 69801
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.395598455614643e-07,
+      "loss": 0.7441,
+      "step": 69802
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.395331301148417e-07,
+      "loss": 0.8184,
+      "step": 69803
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.395064148676461e-07,
+      "loss": 0.6021,
+      "step": 69804
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.394796998198988e-07,
+      "loss": 0.6758,
+      "step": 69805
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.394529849716195e-07,
+      "loss": 0.8242,
+      "step": 69806
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.394262703228295e-07,
+      "loss": 0.4727,
+      "step": 69807
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.393995558735488e-07,
+      "loss": 0.4482,
+      "step": 69808
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.393728416237976e-07,
+      "loss": 0.6069,
+      "step": 69809
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.393461275735968e-07,
+      "loss": 0.7393,
+      "step": 69810
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.393194137229666e-07,
+      "loss": 0.6709,
+      "step": 69811
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.392927000719277e-07,
+      "loss": 0.5752,
+      "step": 69812
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.392659866205006e-07,
+      "loss": 0.7627,
+      "step": 69813
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.392392733687051e-07,
+      "loss": 0.5415,
+      "step": 69814
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.392125603165622e-07,
+      "loss": 0.6152,
+      "step": 69815
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.391858474640922e-07,
+      "loss": 0.6484,
+      "step": 69816
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.391591348113157e-07,
+      "loss": 0.4443,
+      "step": 69817
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.39132422358253e-07,
+      "loss": 0.667,
+      "step": 69818
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.391057101049247e-07,
+      "loss": 0.561,
+      "step": 69819
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.390789980513509e-07,
+      "loss": 0.7451,
+      "step": 69820
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.390522861975527e-07,
+      "loss": 0.6553,
+      "step": 69821
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.390255745435499e-07,
+      "loss": 0.562,
+      "step": 69822
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.389988630893636e-07,
+      "loss": 0.6484,
+      "step": 69823
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.389721518350137e-07,
+      "loss": 0.5525,
+      "step": 69824
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.389454407805208e-07,
+      "loss": 0.8984,
+      "step": 69825
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.389187299259052e-07,
+      "loss": 0.6094,
+      "step": 69826
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.388920192711876e-07,
+      "loss": 0.6592,
+      "step": 69827
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.388653088163883e-07,
+      "loss": 0.6924,
+      "step": 69828
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.38838598561528e-07,
+      "loss": 0.667,
+      "step": 69829
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.388118885066268e-07,
+      "loss": 0.5031,
+      "step": 69830
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.387851786517054e-07,
+      "loss": 0.6353,
+      "step": 69831
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.387584689967841e-07,
+      "loss": 0.6191,
+      "step": 69832
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.387317595418838e-07,
+      "loss": 0.7021,
+      "step": 69833
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.387050502870243e-07,
+      "loss": 0.5724,
+      "step": 69834
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.386783412322262e-07,
+      "loss": 0.6133,
+      "step": 69835
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.386516323775103e-07,
+      "loss": 0.7822,
+      "step": 69836
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.386249237228967e-07,
+      "loss": 0.7373,
+      "step": 69837
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.38598215268406e-07,
+      "loss": 0.8203,
+      "step": 69838
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.385715070140586e-07,
+      "loss": 0.7129,
+      "step": 69839
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.38544798959875e-07,
+      "loss": 0.6855,
+      "step": 69840
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.385180911058755e-07,
+      "loss": 0.4854,
+      "step": 69841
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.384913834520809e-07,
+      "loss": 0.4663,
+      "step": 69842
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.384646759985116e-07,
+      "loss": 0.6328,
+      "step": 69843
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.384379687451877e-07,
+      "loss": 0.7485,
+      "step": 69844
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.384112616921297e-07,
+      "loss": 0.7041,
+      "step": 69845
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.383845548393582e-07,
+      "loss": 0.7734,
+      "step": 69846
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.383578481868935e-07,
+      "loss": 0.7148,
+      "step": 69847
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.383311417347563e-07,
+      "loss": 0.6543,
+      "step": 69848
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.383044354829667e-07,
+      "loss": 0.7334,
+      "step": 69849
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.382777294315457e-07,
+      "loss": 0.6494,
+      "step": 69850
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.382510235805131e-07,
+      "loss": 0.5938,
+      "step": 69851
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.382243179298899e-07,
+      "loss": 0.6533,
+      "step": 69852
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.381976124796965e-07,
+      "loss": 0.6104,
+      "step": 69853
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.381709072299527e-07,
+      "loss": 0.748,
+      "step": 69854
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.381442021806797e-07,
+      "loss": 0.5662,
+      "step": 69855
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.381174973318973e-07,
+      "loss": 0.5635,
+      "step": 69856
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.380907926836267e-07,
+      "loss": 0.7158,
+      "step": 69857
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.380640882358875e-07,
+      "loss": 0.5088,
+      "step": 69858
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.380373839887008e-07,
+      "loss": 0.5605,
+      "step": 69859
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.380106799420868e-07,
+      "loss": 0.6606,
+      "step": 69860
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.379839760960662e-07,
+      "loss": 0.7949,
+      "step": 69861
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.379572724506589e-07,
+      "loss": 0.6631,
+      "step": 69862
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.379305690058863e-07,
+      "loss": 0.5972,
+      "step": 69863
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.379038657617675e-07,
+      "loss": 0.6392,
+      "step": 69864
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.37877162718324e-07,
+      "loss": 0.9053,
+      "step": 69865
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.378504598755757e-07,
+      "loss": 0.6089,
+      "step": 69866
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.378237572335434e-07,
+      "loss": 0.55,
+      "step": 69867
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.377970547922472e-07,
+      "loss": 0.6279,
+      "step": 69868
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.377703525517079e-07,
+      "loss": 0.6587,
+      "step": 69869
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.377436505119456e-07,
+      "loss": 0.4351,
+      "step": 69870
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.377169486729812e-07,
+      "loss": 0.6973,
+      "step": 69871
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.376902470348346e-07,
+      "loss": 0.5112,
+      "step": 69872
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.376635455975271e-07,
+      "loss": 0.6836,
+      "step": 69873
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.37636844361078e-07,
+      "loss": 0.769,
+      "step": 69874
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.376101433255084e-07,
+      "loss": 0.5957,
+      "step": 69875
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.375834424908388e-07,
+      "loss": 0.7334,
+      "step": 69876
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.375567418570892e-07,
+      "loss": 0.6528,
+      "step": 69877
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.375300414242807e-07,
+      "loss": 0.6592,
+      "step": 69878
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.375033411924329e-07,
+      "loss": 0.667,
+      "step": 69879
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.374766411615672e-07,
+      "loss": 0.7363,
+      "step": 69880
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.374499413317034e-07,
+      "loss": 0.6812,
+      "step": 69881
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.37423241702862e-07,
+      "loss": 0.708,
+      "step": 69882
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.373965422750639e-07,
+      "loss": 0.5437,
+      "step": 69883
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.373698430483291e-07,
+      "loss": 0.7402,
+      "step": 69884
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.373431440226778e-07,
+      "loss": 0.7256,
+      "step": 69885
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.373164451981309e-07,
+      "loss": 0.7617,
+      "step": 69886
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.372897465747087e-07,
+      "loss": 0.5239,
+      "step": 69887
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.372630481524317e-07,
+      "loss": 0.6284,
+      "step": 69888
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.372363499313203e-07,
+      "loss": 0.5547,
+      "step": 69889
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.37209651911395e-07,
+      "loss": 0.6821,
+      "step": 69890
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.371829540926759e-07,
+      "loss": 0.7959,
+      "step": 69891
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.371562564751841e-07,
+      "loss": 0.6626,
+      "step": 69892
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.371295590589399e-07,
+      "loss": 0.5125,
+      "step": 69893
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.37102861843963e-07,
+      "loss": 0.6602,
+      "step": 69894
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.370761648302745e-07,
+      "loss": 0.6816,
+      "step": 69895
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.370494680178947e-07,
+      "loss": 0.7295,
+      "step": 69896
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.370227714068441e-07,
+      "loss": 0.627,
+      "step": 69897
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.369960749971431e-07,
+      "loss": 0.6011,
+      "step": 69898
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.369693787888118e-07,
+      "loss": 0.626,
+      "step": 69899
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.369426827818714e-07,
+      "loss": 0.6611,
+      "step": 69900
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.369159869763415e-07,
+      "loss": 0.6802,
+      "step": 69901
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.368892913722431e-07,
+      "loss": 0.4232,
+      "step": 69902
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.368625959695967e-07,
+      "loss": 0.543,
+      "step": 69903
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.368359007684221e-07,
+      "loss": 0.4802,
+      "step": 69904
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.368092057687405e-07,
+      "loss": 0.5532,
+      "step": 69905
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.367825109705716e-07,
+      "loss": 0.6274,
+      "step": 69906
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.367558163739367e-07,
+      "loss": 0.752,
+      "step": 69907
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.367291219788552e-07,
+      "loss": 0.6064,
+      "step": 69908
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.367024277853486e-07,
+      "loss": 0.5906,
+      "step": 69909
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.366757337934365e-07,
+      "loss": 0.4452,
+      "step": 69910
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.3664904000314e-07,
+      "loss": 0.5547,
+      "step": 69911
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.36622346414479e-07,
+      "loss": 0.707,
+      "step": 69912
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.365956530274744e-07,
+      "loss": 0.6582,
+      "step": 69913
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.365689598421463e-07,
+      "loss": 0.6963,
+      "step": 69914
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.365422668585153e-07,
+      "loss": 0.6094,
+      "step": 69915
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.365155740766013e-07,
+      "loss": 0.6387,
+      "step": 69916
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.364888814964258e-07,
+      "loss": 0.6719,
+      "step": 69917
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.364621891180081e-07,
+      "loss": 0.7627,
+      "step": 69918
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.364354969413696e-07,
+      "loss": 0.6172,
+      "step": 69919
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.3640880496653e-07,
+      "loss": 0.6426,
+      "step": 69920
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.363821131935104e-07,
+      "loss": 0.6123,
+      "step": 69921
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.363554216223307e-07,
+      "loss": 0.5762,
+      "step": 69922
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.363287302530118e-07,
+      "loss": 0.5249,
+      "step": 69923
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.363020390855737e-07,
+      "loss": 0.5908,
+      "step": 69924
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.362753481200367e-07,
+      "loss": 0.6592,
+      "step": 69925
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.36248657356422e-07,
+      "loss": 0.6094,
+      "step": 69926
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.362219667947491e-07,
+      "loss": 0.7314,
+      "step": 69927
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.361952764350394e-07,
+      "loss": 0.5293,
+      "step": 69928
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.361685862773124e-07,
+      "loss": 0.4436,
+      "step": 69929
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.361418963215893e-07,
+      "loss": 0.6777,
+      "step": 69930
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.3611520656789e-07,
+      "loss": 0.4849,
+      "step": 69931
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.360885170162353e-07,
+      "loss": 0.5398,
+      "step": 69932
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.360618276666457e-07,
+      "loss": 0.7544,
+      "step": 69933
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.360351385191412e-07,
+      "loss": 0.5942,
+      "step": 69934
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.360084495737424e-07,
+      "loss": 0.4827,
+      "step": 69935
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.359817608304699e-07,
+      "loss": 0.5898,
+      "step": 69936
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.359550722893438e-07,
+      "loss": 0.5254,
+      "step": 69937
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.35928383950385e-07,
+      "loss": 0.604,
+      "step": 69938
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.359016958136135e-07,
+      "loss": 0.584,
+      "step": 69939
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.358750078790501e-07,
+      "loss": 0.5845,
+      "step": 69940
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.35848320146715e-07,
+      "loss": 0.6846,
+      "step": 69941
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.358216326166287e-07,
+      "loss": 0.6392,
+      "step": 69942
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.35794945288812e-07,
+      "loss": 0.6641,
+      "step": 69943
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.357682581632843e-07,
+      "loss": 0.667,
+      "step": 69944
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.35741571240067e-07,
+      "loss": 0.6162,
+      "step": 69945
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.357148845191801e-07,
+      "loss": 0.7163,
+      "step": 69946
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.356881980006445e-07,
+      "loss": 0.6777,
+      "step": 69947
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.356615116844797e-07,
+      "loss": 0.7197,
+      "step": 69948
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.356348255707072e-07,
+      "loss": 0.4685,
+      "step": 69949
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.356081396593467e-07,
+      "loss": 0.6147,
+      "step": 69950
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.355814539504192e-07,
+      "loss": 0.6284,
+      "step": 69951
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.355547684439445e-07,
+      "loss": 0.5752,
+      "step": 69952
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.355280831399439e-07,
+      "loss": 0.5601,
+      "step": 69953
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.355013980384369e-07,
+      "loss": 0.686,
+      "step": 69954
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.354747131394443e-07,
+      "loss": 0.7734,
+      "step": 69955
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.354480284429865e-07,
+      "loss": 0.4375,
+      "step": 69956
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.354213439490842e-07,
+      "loss": 0.75,
+      "step": 69957
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.353946596577572e-07,
+      "loss": 0.731,
+      "step": 69958
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.353679755690267e-07,
+      "loss": 0.7246,
+      "step": 69959
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.353412916829125e-07,
+      "loss": 0.6719,
+      "step": 69960
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.353146079994356e-07,
+      "loss": 0.6816,
+      "step": 69961
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.352879245186158e-07,
+      "loss": 0.6934,
+      "step": 69962
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.352612412404744e-07,
+      "loss": 0.751,
+      "step": 69963
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.35234558165031e-07,
+      "loss": 0.5215,
+      "step": 69964
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.352078752923063e-07,
+      "loss": 0.6748,
+      "step": 69965
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.351811926223209e-07,
+      "loss": 0.751,
+      "step": 69966
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.351545101550947e-07,
+      "loss": 0.5537,
+      "step": 69967
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.351278278906488e-07,
+      "loss": 0.3777,
+      "step": 69968
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.351011458290031e-07,
+      "loss": 0.6592,
+      "step": 69969
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.350744639701785e-07,
+      "loss": 0.4027,
+      "step": 69970
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.350477823141953e-07,
+      "loss": 0.7666,
+      "step": 69971
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.350211008610735e-07,
+      "loss": 0.6973,
+      "step": 69972
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.349944196108345e-07,
+      "loss": 0.4917,
+      "step": 69973
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.349677385634976e-07,
+      "loss": 0.4526,
+      "step": 69974
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.349410577190835e-07,
+      "loss": 0.6162,
+      "step": 69975
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.349143770776133e-07,
+      "loss": 0.5894,
+      "step": 69976
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.348876966391067e-07,
+      "loss": 0.6846,
+      "step": 69977
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.348610164035844e-07,
+      "loss": 0.7676,
+      "step": 69978
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.348343363710667e-07,
+      "loss": 0.5903,
+      "step": 69979
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.348076565415746e-07,
+      "loss": 0.6362,
+      "step": 69980
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.347809769151277e-07,
+      "loss": 0.4045,
+      "step": 69981
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34754297491747e-07,
+      "loss": 0.5771,
+      "step": 69982
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34727618271453e-07,
+      "loss": 0.5889,
+      "step": 69983
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.347009392542654e-07,
+      "loss": 0.6045,
+      "step": 69984
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.346742604402053e-07,
+      "loss": 0.7139,
+      "step": 69985
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.346475818292927e-07,
+      "loss": 0.5654,
+      "step": 69986
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.346209034215484e-07,
+      "loss": 0.5762,
+      "step": 69987
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.345942252169928e-07,
+      "loss": 0.5522,
+      "step": 69988
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.345675472156461e-07,
+      "loss": 0.6738,
+      "step": 69989
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.345408694175288e-07,
+      "loss": 0.6204,
+      "step": 69990
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.345141918226612e-07,
+      "loss": 0.5371,
+      "step": 69991
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34487514431064e-07,
+      "loss": 0.5898,
+      "step": 69992
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.344608372427579e-07,
+      "loss": 0.6377,
+      "step": 69993
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.344341602577624e-07,
+      "loss": 0.5933,
+      "step": 69994
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.344074834760987e-07,
+      "loss": 0.6528,
+      "step": 69995
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.343808068977868e-07,
+      "loss": 0.4229,
+      "step": 69996
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.343541305228475e-07,
+      "loss": 0.5498,
+      "step": 69997
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.343274543513007e-07,
+      "loss": 0.6475,
+      "step": 69998
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.343007783831675e-07,
+      "loss": 0.635,
+      "step": 69999
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.342741026184676e-07,
+      "loss": 0.6748,
+      "step": 70000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.342474270572222e-07,
+      "loss": 0.4531,
+      "step": 70001
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.342207516994511e-07,
+      "loss": 0.6675,
+      "step": 70002
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.341940765451753e-07,
+      "loss": 0.541,
+      "step": 70003
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.341674015944146e-07,
+      "loss": 0.71,
+      "step": 70004
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.341407268471898e-07,
+      "loss": 0.6279,
+      "step": 70005
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34114052303521e-07,
+      "loss": 0.542,
+      "step": 70006
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34087377963429e-07,
+      "loss": 0.6123,
+      "step": 70007
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.34060703826934e-07,
+      "loss": 0.6152,
+      "step": 70008
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.340340298940565e-07,
+      "loss": 0.6611,
+      "step": 70009
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.340073561648168e-07,
+      "loss": 0.7607,
+      "step": 70010
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.339806826392356e-07,
+      "loss": 0.583,
+      "step": 70011
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.339540093173331e-07,
+      "loss": 0.436,
+      "step": 70012
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.339273361991301e-07,
+      "loss": 0.4478,
+      "step": 70013
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.339006632846465e-07,
+      "loss": 0.6743,
+      "step": 70014
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.338739905739027e-07,
+      "loss": 0.5164,
+      "step": 70015
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.338473180669196e-07,
+      "loss": 0.5039,
+      "step": 70016
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.338206457637171e-07,
+      "loss": 0.5361,
+      "step": 70017
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.337939736643161e-07,
+      "loss": 0.6411,
+      "step": 70018
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.337673017687365e-07,
+      "loss": 0.6079,
+      "step": 70019
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.337406300769995e-07,
+      "loss": 0.6465,
+      "step": 70020
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.337139585891247e-07,
+      "loss": 0.6445,
+      "step": 70021
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.336872873051331e-07,
+      "loss": 0.6377,
+      "step": 70022
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.336606162250451e-07,
+      "loss": 0.6406,
+      "step": 70023
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.336339453488808e-07,
+      "loss": 0.7344,
+      "step": 70024
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.336072746766604e-07,
+      "loss": 0.5073,
+      "step": 70025
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.335806042084049e-07,
+      "loss": 0.688,
+      "step": 70026
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.335539339441343e-07,
+      "loss": 0.7432,
+      "step": 70027
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.335272638838694e-07,
+      "loss": 0.4583,
+      "step": 70028
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.335005940276303e-07,
+      "loss": 0.6543,
+      "step": 70029
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.334739243754376e-07,
+      "loss": 0.7793,
+      "step": 70030
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.334472549273116e-07,
+      "loss": 0.4722,
+      "step": 70031
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.33420585683273e-07,
+      "loss": 0.7451,
+      "step": 70032
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 7.333939166433421e-07,
+      "loss": 0.5708,
+      "step": 70033
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.333672478075388e-07,
+      "loss": 0.6597,
+      "step": 70034
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.333405791758842e-07,
+      "loss": 0.6445,
+      "step": 70035
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.333139107483981e-07,
+      "loss": 0.7129,
+      "step": 70036
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.332872425251017e-07,
+      "loss": 0.6885,
+      "step": 70037
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.332605745060147e-07,
+      "loss": 0.3818,
+      "step": 70038
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.332339066911579e-07,
+      "loss": 0.5791,
+      "step": 70039
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.332072390805516e-07,
+      "loss": 0.5337,
+      "step": 70040
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.331805716742165e-07,
+      "loss": 0.6875,
+      "step": 70041
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.331539044721728e-07,
+      "loss": 0.6577,
+      "step": 70042
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.331272374744408e-07,
+      "loss": 0.5935,
+      "step": 70043
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.331005706810407e-07,
+      "loss": 0.7007,
+      "step": 70044
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.330739040919934e-07,
+      "loss": 0.7422,
+      "step": 70045
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.330472377073191e-07,
+      "loss": 0.5459,
+      "step": 70046
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.330205715270383e-07,
+      "loss": 0.4849,
+      "step": 70047
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.329939055511713e-07,
+      "loss": 0.6523,
+      "step": 70048
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.329672397797386e-07,
+      "loss": 0.5918,
+      "step": 70049
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.329405742127606e-07,
+      "loss": 0.6934,
+      "step": 70050
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.329139088502579e-07,
+      "loss": 0.6406,
+      "step": 70051
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.328872436922508e-07,
+      "loss": 0.623,
+      "step": 70052
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.328605787387594e-07,
+      "loss": 0.6045,
+      "step": 70053
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.328339139898046e-07,
+      "loss": 0.709,
+      "step": 70054
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.328072494454063e-07,
+      "loss": 0.5098,
+      "step": 70055
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.327805851055853e-07,
+      "loss": 0.6123,
+      "step": 70056
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.327539209703619e-07,
+      "loss": 0.6519,
+      "step": 70057
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.327272570397567e-07,
+      "loss": 0.6616,
+      "step": 70058
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.327005933137896e-07,
+      "loss": 0.4763,
+      "step": 70059
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.326739297924818e-07,
+      "loss": 0.7236,
+      "step": 70060
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.326472664758531e-07,
+      "loss": 0.707,
+      "step": 70061
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.326206033639244e-07,
+      "loss": 0.5859,
+      "step": 70062
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.325939404567153e-07,
+      "loss": 0.708,
+      "step": 70063
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.32567277754247e-07,
+      "loss": 0.6196,
+      "step": 70064
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.325406152565394e-07,
+      "loss": 0.7051,
+      "step": 70065
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.325139529636133e-07,
+      "loss": 0.6069,
+      "step": 70066
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.324872908754888e-07,
+      "loss": 0.6455,
+      "step": 70067
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.324606289921868e-07,
+      "loss": 0.5229,
+      "step": 70068
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.324339673137272e-07,
+      "loss": 0.709,
+      "step": 70069
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.324073058401308e-07,
+      "loss": 0.6465,
+      "step": 70070
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.323806445714175e-07,
+      "loss": 0.6216,
+      "step": 70071
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.323539835076084e-07,
+      "loss": 0.6953,
+      "step": 70072
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.323273226487234e-07,
+      "loss": 0.752,
+      "step": 70073
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.323006619947829e-07,
+      "loss": 0.7725,
+      "step": 70074
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.322740015458078e-07,
+      "loss": 0.571,
+      "step": 70075
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.322473413018178e-07,
+      "loss": 0.6758,
+      "step": 70076
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.32220681262834e-07,
+      "loss": 0.6331,
+      "step": 70077
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.321940214288765e-07,
+      "loss": 0.6987,
+      "step": 70078
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.321673617999654e-07,
+      "loss": 0.6191,
+      "step": 70079
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.321407023761217e-07,
+      "loss": 0.6099,
+      "step": 70080
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.321140431573653e-07,
+      "loss": 0.5723,
+      "step": 70081
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.320873841437174e-07,
+      "loss": 0.5615,
+      "step": 70082
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.320607253351976e-07,
+      "loss": 0.6973,
+      "step": 70083
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.320340667318265e-07,
+      "loss": 0.5393,
+      "step": 70084
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.320074083336246e-07,
+      "loss": 0.666,
+      "step": 70085
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.319807501406122e-07,
+      "loss": 0.6758,
+      "step": 70086
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.3195409215281e-07,
+      "loss": 0.4636,
+      "step": 70087
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.319274343702379e-07,
+      "loss": 0.4233,
+      "step": 70088
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.319007767929171e-07,
+      "loss": 0.4948,
+      "step": 70089
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.318741194208672e-07,
+      "loss": 0.6157,
+      "step": 70090
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.318474622541091e-07,
+      "loss": 0.564,
+      "step": 70091
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.318208052926634e-07,
+      "loss": 0.668,
+      "step": 70092
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.317941485365497e-07,
+      "loss": 0.5815,
+      "step": 70093
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.317674919857892e-07,
+      "loss": 0.4631,
+      "step": 70094
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.317408356404016e-07,
+      "loss": 0.7568,
+      "step": 70095
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.31714179500408e-07,
+      "loss": 0.7891,
+      "step": 70096
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.316875235658287e-07,
+      "loss": 0.6538,
+      "step": 70097
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.316608678366833e-07,
+      "loss": 0.7285,
+      "step": 70098
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.316342123129935e-07,
+      "loss": 0.6025,
+      "step": 70099
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.316075569947786e-07,
+      "loss": 0.6172,
+      "step": 70100
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.315809018820595e-07,
+      "loss": 0.6816,
+      "step": 70101
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.315542469748571e-07,
+      "loss": 0.5107,
+      "step": 70102
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.315275922731907e-07,
+      "loss": 0.6191,
+      "step": 70103
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.315009377770816e-07,
+      "loss": 0.5371,
+      "step": 70104
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.314742834865495e-07,
+      "loss": 0.5374,
+      "step": 70105
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.314476294016155e-07,
+      "loss": 0.5898,
+      "step": 70106
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.314209755222993e-07,
+      "loss": 0.4438,
+      "step": 70107
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.313943218486221e-07,
+      "loss": 0.5928,
+      "step": 70108
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.313676683806038e-07,
+      "loss": 0.5581,
+      "step": 70109
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.313410151182651e-07,
+      "loss": 0.5879,
+      "step": 70110
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.313143620616258e-07,
+      "loss": 0.5122,
+      "step": 70111
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.312877092107074e-07,
+      "loss": 0.6567,
+      "step": 70112
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.312610565655292e-07,
+      "loss": 0.7842,
+      "step": 70113
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.312344041261123e-07,
+      "loss": 0.5498,
+      "step": 70114
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.312077518924765e-07,
+      "loss": 0.5986,
+      "step": 70115
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.311810998646428e-07,
+      "loss": 0.4902,
+      "step": 70116
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.311544480426313e-07,
+      "loss": 0.6377,
+      "step": 70117
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.311277964264625e-07,
+      "loss": 0.52,
+      "step": 70118
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.311011450161566e-07,
+      "loss": 0.6006,
+      "step": 70119
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.310744938117344e-07,
+      "loss": 0.6357,
+      "step": 70120
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.310478428132161e-07,
+      "loss": 0.7207,
+      "step": 70121
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.310211920206222e-07,
+      "loss": 0.6675,
+      "step": 70122
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30994541433973e-07,
+      "loss": 0.7026,
+      "step": 70123
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.309678910532886e-07,
+      "loss": 0.6353,
+      "step": 70124
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.3094124087859e-07,
+      "loss": 0.4524,
+      "step": 70125
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30914590909897e-07,
+      "loss": 0.4702,
+      "step": 70126
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.308879411472306e-07,
+      "loss": 0.5322,
+      "step": 70127
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.308612915906108e-07,
+      "loss": 0.6562,
+      "step": 70128
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.308346422400583e-07,
+      "loss": 0.6167,
+      "step": 70129
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30807993095593e-07,
+      "loss": 0.6836,
+      "step": 70130
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30781344157236e-07,
+      "loss": 0.6016,
+      "step": 70131
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.307546954250076e-07,
+      "loss": 0.6963,
+      "step": 70132
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.307280468989276e-07,
+      "loss": 0.6343,
+      "step": 70133
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.307013985790166e-07,
+      "loss": 0.5552,
+      "step": 70134
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.306747504652954e-07,
+      "loss": 0.6499,
+      "step": 70135
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.306481025577838e-07,
+      "loss": 0.5735,
+      "step": 70136
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.306214548565029e-07,
+      "loss": 0.627,
+      "step": 70137
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.305948073614725e-07,
+      "loss": 0.5781,
+      "step": 70138
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.305681600727136e-07,
+      "loss": 0.6196,
+      "step": 70139
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30541512990246e-07,
+      "loss": 0.6167,
+      "step": 70140
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.305148661140905e-07,
+      "loss": 0.6914,
+      "step": 70141
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.304882194442677e-07,
+      "loss": 0.5254,
+      "step": 70142
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.304615729807971e-07,
+      "loss": 0.626,
+      "step": 70143
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.304349267237e-07,
+      "loss": 0.6792,
+      "step": 70144
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.304082806729964e-07,
+      "loss": 0.6865,
+      "step": 70145
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.303816348287069e-07,
+      "loss": 0.6855,
+      "step": 70146
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.303549891908514e-07,
+      "loss": 0.748,
+      "step": 70147
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.303283437594509e-07,
+      "loss": 0.5391,
+      "step": 70148
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.303016985345256e-07,
+      "loss": 0.7549,
+      "step": 70149
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.30275053516096e-07,
+      "loss": 0.5991,
+      "step": 70150
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.302484087041825e-07,
+      "loss": 0.7549,
+      "step": 70151
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.302217640988054e-07,
+      "loss": 0.6514,
+      "step": 70152
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.301951196999848e-07,
+      "loss": 0.59,
+      "step": 70153
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.301684755077415e-07,
+      "loss": 0.6099,
+      "step": 70154
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.301418315220958e-07,
+      "loss": 0.5889,
+      "step": 70155
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.301151877430681e-07,
+      "loss": 0.6113,
+      "step": 70156
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.300885441706786e-07,
+      "loss": 0.6514,
+      "step": 70157
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.300619008049481e-07,
+      "loss": 0.6758,
+      "step": 70158
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.300352576458966e-07,
+      "loss": 0.6841,
+      "step": 70159
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.300086146935449e-07,
+      "loss": 0.6143,
+      "step": 70160
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29981971947913e-07,
+      "loss": 0.6387,
+      "step": 70161
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.299553294090219e-07,
+      "loss": 0.791,
+      "step": 70162
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.299286870768914e-07,
+      "loss": 0.6538,
+      "step": 70163
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.299020449515419e-07,
+      "loss": 0.6523,
+      "step": 70164
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.298754030329941e-07,
+      "loss": 0.6045,
+      "step": 70165
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29848761321268e-07,
+      "loss": 0.7822,
+      "step": 70166
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.298221198163847e-07,
+      "loss": 0.7207,
+      "step": 70167
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.297954785183641e-07,
+      "loss": 0.7002,
+      "step": 70168
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.297688374272264e-07,
+      "loss": 0.6963,
+      "step": 70169
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.297421965429926e-07,
+      "loss": 0.5259,
+      "step": 70170
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.297155558656825e-07,
+      "loss": 0.5396,
+      "step": 70171
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.296889153953172e-07,
+      "loss": 0.4915,
+      "step": 70172
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.296622751319165e-07,
+      "loss": 0.7324,
+      "step": 70173
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.296356350755007e-07,
+      "loss": 0.7402,
+      "step": 70174
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.296089952260907e-07,
+      "loss": 0.5757,
+      "step": 70175
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.295823555837064e-07,
+      "loss": 0.8008,
+      "step": 70176
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.295557161483688e-07,
+      "loss": 0.7578,
+      "step": 70177
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.295290769200975e-07,
+      "loss": 0.4276,
+      "step": 70178
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.295024378989137e-07,
+      "loss": 0.6318,
+      "step": 70179
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.294757990848373e-07,
+      "loss": 0.7021,
+      "step": 70180
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29449160477889e-07,
+      "loss": 0.6729,
+      "step": 70181
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.294225220780892e-07,
+      "loss": 0.6865,
+      "step": 70182
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.293958838854577e-07,
+      "loss": 0.5201,
+      "step": 70183
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.293692459000156e-07,
+      "loss": 0.4905,
+      "step": 70184
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.293426081217828e-07,
+      "loss": 0.5747,
+      "step": 70185
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.293159705507801e-07,
+      "loss": 0.6484,
+      "step": 70186
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.292893331870277e-07,
+      "loss": 0.6294,
+      "step": 70187
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29262696030546e-07,
+      "loss": 0.6499,
+      "step": 70188
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.292360590813555e-07,
+      "loss": 0.7168,
+      "step": 70189
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.292094223394762e-07,
+      "loss": 0.6816,
+      "step": 70190
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.291827858049291e-07,
+      "loss": 0.6104,
+      "step": 70191
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.291561494777343e-07,
+      "loss": 0.5103,
+      "step": 70192
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29129513357912e-07,
+      "loss": 0.6895,
+      "step": 70193
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.291028774454828e-07,
+      "loss": 0.7119,
+      "step": 70194
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.29076241740467e-07,
+      "loss": 0.7666,
+      "step": 70195
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.290496062428853e-07,
+      "loss": 0.5776,
+      "step": 70196
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.290229709527575e-07,
+      "loss": 0.6172,
+      "step": 70197
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.289963358701048e-07,
+      "loss": 0.6426,
+      "step": 70198
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.289697009949467e-07,
+      "loss": 0.7344,
+      "step": 70199
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.289430663273044e-07,
+      "loss": 0.7959,
+      "step": 70200
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.289164318671976e-07,
+      "loss": 0.6289,
+      "step": 70201
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.288897976146475e-07,
+      "loss": 0.6836,
+      "step": 70202
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.288631635696738e-07,
+      "loss": 0.6665,
+      "step": 70203
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.288365297322971e-07,
+      "loss": 0.6279,
+      "step": 70204
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.288098961025375e-07,
+      "loss": 0.7188,
+      "step": 70205
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.28783262680416e-07,
+      "loss": 0.6235,
+      "step": 70206
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.287566294659526e-07,
+      "loss": 0.5723,
+      "step": 70207
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.287299964591677e-07,
+      "loss": 0.5908,
+      "step": 70208
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.287033636600819e-07,
+      "loss": 0.6636,
+      "step": 70209
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.286767310687154e-07,
+      "loss": 0.603,
+      "step": 70210
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.286500986850885e-07,
+      "loss": 0.6865,
+      "step": 70211
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.286234665092223e-07,
+      "loss": 0.5771,
+      "step": 70212
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.285968345411363e-07,
+      "loss": 0.543,
+      "step": 70213
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.28570202780851e-07,
+      "loss": 0.4961,
+      "step": 70214
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.285435712283873e-07,
+      "loss": 0.7549,
+      "step": 70215
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.285169398837651e-07,
+      "loss": 0.4469,
+      "step": 70216
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.284903087470051e-07,
+      "loss": 0.5469,
+      "step": 70217
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.284636778181275e-07,
+      "loss": 0.5281,
+      "step": 70218
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.284370470971528e-07,
+      "loss": 0.7139,
+      "step": 70219
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.284104165841014e-07,
+      "loss": 0.5718,
+      "step": 70220
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.283837862789936e-07,
+      "loss": 0.6592,
+      "step": 70221
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.283571561818502e-07,
+      "loss": 0.6045,
+      "step": 70222
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.283305262926911e-07,
+      "loss": 0.5923,
+      "step": 70223
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.283038966115364e-07,
+      "loss": 0.6909,
+      "step": 70224
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.282772671384074e-07,
+      "loss": 0.6992,
+      "step": 70225
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.282506378733236e-07,
+      "loss": 0.6367,
+      "step": 70226
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.282240088163061e-07,
+      "loss": 0.6008,
+      "step": 70227
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.281973799673747e-07,
+      "loss": 0.7676,
+      "step": 70228
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.281707513265503e-07,
+      "loss": 0.606,
+      "step": 70229
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.281441228938529e-07,
+      "loss": 0.8135,
+      "step": 70230
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.281174946693032e-07,
+      "loss": 0.7036,
+      "step": 70231
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.280908666529217e-07,
+      "loss": 0.6211,
+      "step": 70232
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.28064238844728e-07,
+      "loss": 0.7749,
+      "step": 70233
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.280376112447433e-07,
+      "loss": 0.595,
+      "step": 70234
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.280109838529875e-07,
+      "loss": 0.4766,
+      "step": 70235
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.279843566694812e-07,
+      "loss": 0.6162,
+      "step": 70236
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.279577296942448e-07,
+      "loss": 0.5864,
+      "step": 70237
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.279311029272988e-07,
+      "loss": 0.5989,
+      "step": 70238
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.279044763686631e-07,
+      "loss": 0.7075,
+      "step": 70239
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.278778500183589e-07,
+      "loss": 0.5679,
+      "step": 70240
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.278512238764059e-07,
+      "loss": 0.6963,
+      "step": 70241
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27824597942825e-07,
+      "loss": 0.6748,
+      "step": 70242
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.277979722176358e-07,
+      "loss": 0.5254,
+      "step": 70243
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.277713467008595e-07,
+      "loss": 0.6211,
+      "step": 70244
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.277447213925159e-07,
+      "loss": 0.6699,
+      "step": 70245
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.277180962926258e-07,
+      "loss": 0.6846,
+      "step": 70246
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.276914714012092e-07,
+      "loss": 0.71,
+      "step": 70247
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.276648467182871e-07,
+      "loss": 0.6641,
+      "step": 70248
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.276382222438792e-07,
+      "loss": 0.5249,
+      "step": 70249
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.276115979780064e-07,
+      "loss": 0.5112,
+      "step": 70250
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.275849739206887e-07,
+      "loss": 0.3789,
+      "step": 70251
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27558350071947e-07,
+      "loss": 0.5059,
+      "step": 70252
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.275317264318012e-07,
+      "loss": 0.5933,
+      "step": 70253
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.275051030002716e-07,
+      "loss": 0.6338,
+      "step": 70254
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27478479777379e-07,
+      "loss": 0.585,
+      "step": 70255
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.274518567631433e-07,
+      "loss": 0.5864,
+      "step": 70256
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.274252339575856e-07,
+      "loss": 0.5962,
+      "step": 70257
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.273986113607256e-07,
+      "loss": 0.7073,
+      "step": 70258
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27371988972584e-07,
+      "loss": 0.4501,
+      "step": 70259
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.273453667931813e-07,
+      "loss": 0.6743,
+      "step": 70260
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.273187448225374e-07,
+      "loss": 0.5342,
+      "step": 70261
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.272921230606734e-07,
+      "loss": 0.6689,
+      "step": 70262
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.272655015076093e-07,
+      "loss": 0.4846,
+      "step": 70263
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.272388801633651e-07,
+      "loss": 0.5918,
+      "step": 70264
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.272122590279617e-07,
+      "loss": 0.4517,
+      "step": 70265
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.271856381014192e-07,
+      "loss": 0.667,
+      "step": 70266
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.271590173837583e-07,
+      "loss": 0.6421,
+      "step": 70267
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27132396874999e-07,
+      "loss": 0.7549,
+      "step": 70268
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.27105776575162e-07,
+      "loss": 0.4805,
+      "step": 70269
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.270791564842673e-07,
+      "loss": 0.5781,
+      "step": 70270
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.270525366023359e-07,
+      "loss": 0.6289,
+      "step": 70271
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.270259169293879e-07,
+      "loss": 0.8623,
+      "step": 70272
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.269992974654434e-07,
+      "loss": 0.4548,
+      "step": 70273
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.269726782105231e-07,
+      "loss": 0.6641,
+      "step": 70274
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.26946059164647e-07,
+      "loss": 0.5771,
+      "step": 70275
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.26919440327836e-07,
+      "loss": 0.6367,
+      "step": 70276
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.268928217001102e-07,
+      "loss": 0.6089,
+      "step": 70277
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.268662032814899e-07,
+      "loss": 0.667,
+      "step": 70278
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.268395850719957e-07,
+      "loss": 0.6602,
+      "step": 70279
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.268129670716477e-07,
+      "loss": 0.5547,
+      "step": 70280
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.267863492804666e-07,
+      "loss": 0.543,
+      "step": 70281
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.267597316984728e-07,
+      "loss": 0.5923,
+      "step": 70282
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.267331143256863e-07,
+      "loss": 0.6094,
+      "step": 70283
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.267064971621277e-07,
+      "loss": 0.7383,
+      "step": 70284
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.266798802078172e-07,
+      "loss": 0.7783,
+      "step": 70285
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.266532634627755e-07,
+      "loss": 0.6787,
+      "step": 70286
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.266266469270228e-07,
+      "loss": 0.6448,
+      "step": 70287
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.266000306005796e-07,
+      "loss": 0.4209,
+      "step": 70288
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.265734144834659e-07,
+      "loss": 0.7607,
+      "step": 70289
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.265467985757028e-07,
+      "loss": 0.6533,
+      "step": 70290
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.265201828773097e-07,
+      "loss": 0.5762,
+      "step": 70291
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.264935673883084e-07,
+      "loss": 0.6152,
+      "step": 70292
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.264669521087178e-07,
+      "loss": 0.7725,
+      "step": 70293
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.26440337038559e-07,
+      "loss": 0.4767,
+      "step": 70294
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.264137221778521e-07,
+      "loss": 0.875,
+      "step": 70295
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.263871075266177e-07,
+      "loss": 0.5073,
+      "step": 70296
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.26360493084876e-07,
+      "loss": 0.6982,
+      "step": 70297
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.263338788526477e-07,
+      "loss": 0.6212,
+      "step": 70298
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.263072648299528e-07,
+      "loss": 0.3146,
+      "step": 70299
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.262806510168121e-07,
+      "loss": 0.559,
+      "step": 70300
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.262540374132453e-07,
+      "loss": 0.5425,
+      "step": 70301
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.262274240192737e-07,
+      "loss": 0.6084,
+      "step": 70302
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.26200810834917e-07,
+      "loss": 0.6602,
+      "step": 70303
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.261741978601955e-07,
+      "loss": 0.5552,
+      "step": 70304
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.261475850951301e-07,
+      "loss": 0.6367,
+      "step": 70305
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.261209725397406e-07,
+      "loss": 0.5327,
+      "step": 70306
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.260943601940479e-07,
+      "loss": 0.4844,
+      "step": 70307
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.260677480580722e-07,
+      "loss": 0.4421,
+      "step": 70308
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.260411361318337e-07,
+      "loss": 0.6816,
+      "step": 70309
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.260145244153528e-07,
+      "loss": 0.6064,
+      "step": 70310
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.259879129086503e-07,
+      "loss": 0.6353,
+      "step": 70311
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.259613016117463e-07,
+      "loss": 0.6538,
+      "step": 70312
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.259346905246611e-07,
+      "loss": 0.6807,
+      "step": 70313
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.259080796474146e-07,
+      "loss": 0.8154,
+      "step": 70314
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.258814689800281e-07,
+      "loss": 0.6421,
+      "step": 70315
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.258548585225214e-07,
+      "loss": 0.5225,
+      "step": 70316
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.258282482749152e-07,
+      "loss": 0.4937,
+      "step": 70317
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.258016382372294e-07,
+      "loss": 0.5334,
+      "step": 70318
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.25775028409485e-07,
+      "loss": 0.5312,
+      "step": 70319
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.257484187917018e-07,
+      "loss": 0.6597,
+      "step": 70320
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.257218093839005e-07,
+      "loss": 0.7148,
+      "step": 70321
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.256952001861019e-07,
+      "loss": 0.6748,
+      "step": 70322
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.256685911983252e-07,
+      "loss": 0.8379,
+      "step": 70323
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.256419824205917e-07,
+      "loss": 0.5608,
+      "step": 70324
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.256153738529214e-07,
+      "loss": 0.6357,
+      "step": 70325
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.25588765495335e-07,
+      "loss": 0.4661,
+      "step": 70326
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.255621573478523e-07,
+      "loss": 0.4341,
+      "step": 70327
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.255355494104945e-07,
+      "loss": 0.4458,
+      "step": 70328
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.255089416832812e-07,
+      "loss": 0.7734,
+      "step": 70329
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.25482334166233e-07,
+      "loss": 0.4089,
+      "step": 70330
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.254557268593707e-07,
+      "loss": 0.6294,
+      "step": 70331
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.254291197627145e-07,
+      "loss": 0.463,
+      "step": 70332
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.25402512876284e-07,
+      "loss": 0.6572,
+      "step": 70333
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.253759062001005e-07,
+      "loss": 0.6943,
+      "step": 70334
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.253492997341839e-07,
+      "loss": 0.7803,
+      "step": 70335
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.253226934785548e-07,
+      "loss": 0.5537,
+      "step": 70336
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.252960874332335e-07,
+      "loss": 0.5028,
+      "step": 70337
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.252694815982403e-07,
+      "loss": 0.4043,
+      "step": 70338
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.252428759735955e-07,
+      "loss": 0.6758,
+      "step": 70339
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.252162705593198e-07,
+      "loss": 0.553,
+      "step": 70340
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.251896653554336e-07,
+      "loss": 0.6562,
+      "step": 70341
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.251630603619566e-07,
+      "loss": 0.5542,
+      "step": 70342
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.251364555789098e-07,
+      "loss": 0.5442,
+      "step": 70343
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.251098510063131e-07,
+      "loss": 0.7505,
+      "step": 70344
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.250832466441876e-07,
+      "loss": 0.5962,
+      "step": 70345
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.250566424925528e-07,
+      "loss": 0.5608,
+      "step": 70346
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.250300385514296e-07,
+      "loss": 0.5674,
+      "step": 70347
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.250034348208383e-07,
+      "loss": 0.5693,
+      "step": 70348
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.249768313007994e-07,
+      "loss": 0.7178,
+      "step": 70349
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.249502279913329e-07,
+      "loss": 0.5615,
+      "step": 70350
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.249236248924597e-07,
+      "loss": 0.6091,
+      "step": 70351
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.248970220041994e-07,
+      "loss": 0.6646,
+      "step": 70352
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.248704193265731e-07,
+      "loss": 0.666,
+      "step": 70353
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.248438168596005e-07,
+      "loss": 0.6147,
+      "step": 70354
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.248172146033027e-07,
+      "loss": 0.6807,
+      "step": 70355
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.247906125576995e-07,
+      "loss": 0.7109,
+      "step": 70356
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.247640107228115e-07,
+      "loss": 0.751,
+      "step": 70357
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.247374090986589e-07,
+      "loss": 0.6543,
+      "step": 70358
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.247108076852625e-07,
+      "loss": 0.6782,
+      "step": 70359
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.246842064826421e-07,
+      "loss": 0.7378,
+      "step": 70360
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.246576054908188e-07,
+      "loss": 0.5308,
+      "step": 70361
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.246310047098123e-07,
+      "loss": 0.4829,
+      "step": 70362
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.246044041396429e-07,
+      "loss": 0.4998,
+      "step": 70363
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.245778037803316e-07,
+      "loss": 0.5698,
+      "step": 70364
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.245512036318982e-07,
+      "loss": 0.6128,
+      "step": 70365
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.245246036943633e-07,
+      "loss": 0.8086,
+      "step": 70366
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.244980039677473e-07,
+      "loss": 0.6162,
+      "step": 70367
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.244714044520703e-07,
+      "loss": 0.752,
+      "step": 70368
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.244448051473531e-07,
+      "loss": 0.6997,
+      "step": 70369
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.244182060536157e-07,
+      "loss": 0.7466,
+      "step": 70370
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.243916071708791e-07,
+      "loss": 0.5898,
+      "step": 70371
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.243650084991628e-07,
+      "loss": 0.5559,
+      "step": 70372
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.243384100384873e-07,
+      "loss": 0.8076,
+      "step": 70373
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.243118117888736e-07,
+      "loss": 0.5127,
+      "step": 70374
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.242852137503412e-07,
+      "loss": 0.6934,
+      "step": 70375
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.242586159229113e-07,
+      "loss": 0.6992,
+      "step": 70376
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.242320183066038e-07,
+      "loss": 0.6455,
+      "step": 70377
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.242054209014392e-07,
+      "loss": 0.7036,
+      "step": 70378
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.241788237074377e-07,
+      "loss": 0.7773,
+      "step": 70379
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.241522267246201e-07,
+      "loss": 0.572,
+      "step": 70380
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.241256299530065e-07,
+      "loss": 0.3828,
+      "step": 70381
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.240990333926168e-07,
+      "loss": 0.7095,
+      "step": 70382
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.240724370434721e-07,
+      "loss": 0.4741,
+      "step": 70383
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.240458409055923e-07,
+      "loss": 0.8047,
+      "step": 70384
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.240192449789979e-07,
+      "loss": 0.7432,
+      "step": 70385
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.239926492637094e-07,
+      "loss": 0.6934,
+      "step": 70386
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.239660537597467e-07,
+      "loss": 0.8896,
+      "step": 70387
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.239394584671308e-07,
+      "loss": 0.4351,
+      "step": 70388
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.239128633858816e-07,
+      "loss": 0.3593,
+      "step": 70389
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.238862685160199e-07,
+      "loss": 0.6172,
+      "step": 70390
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.238596738575657e-07,
+      "loss": 0.7861,
+      "step": 70391
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.238330794105394e-07,
+      "loss": 0.6689,
+      "step": 70392
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.238064851749614e-07,
+      "loss": 0.7305,
+      "step": 70393
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.237798911508519e-07,
+      "loss": 0.4209,
+      "step": 70394
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.237532973382316e-07,
+      "loss": 0.4961,
+      "step": 70395
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.237267037371206e-07,
+      "loss": 0.7505,
+      "step": 70396
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.237001103475396e-07,
+      "loss": 0.6904,
+      "step": 70397
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.236735171695084e-07,
+      "loss": 0.4932,
+      "step": 70398
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.236469242030479e-07,
+      "loss": 0.6211,
+      "step": 70399
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.23620331448178e-07,
+      "loss": 0.6465,
+      "step": 70400
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.2359373890492e-07,
+      "loss": 0.7695,
+      "step": 70401
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.23567146573293e-07,
+      "loss": 0.6396,
+      "step": 70402
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.235405544533179e-07,
+      "loss": 0.6211,
+      "step": 70403
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.235139625450152e-07,
+      "loss": 0.6514,
+      "step": 70404
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.234873708484052e-07,
+      "loss": 0.708,
+      "step": 70405
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.23460779363508e-07,
+      "loss": 0.6523,
+      "step": 70406
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.234341880903443e-07,
+      "loss": 0.5488,
+      "step": 70407
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.234075970289342e-07,
+      "loss": 0.5967,
+      "step": 70408
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.233810061792984e-07,
+      "loss": 0.5234,
+      "step": 70409
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.233544155414569e-07,
+      "loss": 0.45,
+      "step": 70410
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.233278251154304e-07,
+      "loss": 0.6699,
+      "step": 70411
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.23301234901239e-07,
+      "loss": 0.54,
+      "step": 70412
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.232746448989029e-07,
+      "loss": 0.6787,
+      "step": 70413
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.232480551084428e-07,
+      "loss": 0.6768,
+      "step": 70414
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.232214655298789e-07,
+      "loss": 0.7402,
+      "step": 70415
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.231948761632317e-07,
+      "loss": 0.533,
+      "step": 70416
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.231682870085212e-07,
+      "loss": 0.604,
+      "step": 70417
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.231416980657683e-07,
+      "loss": 0.5747,
+      "step": 70418
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.231151093349928e-07,
+      "loss": 0.6001,
+      "step": 70419
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.230885208162156e-07,
+      "loss": 0.6592,
+      "step": 70420
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.230619325094572e-07,
+      "loss": 0.5146,
+      "step": 70421
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.23035344414737e-07,
+      "loss": 0.5527,
+      "step": 70422
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 7.230087565320758e-07,
+      "loss": 0.3786,
+      "step": 70423
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.229821688614943e-07,
+      "loss": 0.5723,
+      "step": 70424
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.229555814030124e-07,
+      "loss": 0.5703,
+      "step": 70425
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.229289941566508e-07,
+      "loss": 0.623,
+      "step": 70426
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.229024071224297e-07,
+      "loss": 0.3494,
+      "step": 70427
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.228758203003695e-07,
+      "loss": 0.709,
+      "step": 70428
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.228492336904904e-07,
+      "loss": 0.8096,
+      "step": 70429
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.228226472928131e-07,
+      "loss": 0.5181,
+      "step": 70430
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22796061107358e-07,
+      "loss": 0.6621,
+      "step": 70431
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.227694751341448e-07,
+      "loss": 0.7588,
+      "step": 70432
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.227428893731945e-07,
+      "loss": 0.5168,
+      "step": 70433
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22716303824527e-07,
+      "loss": 0.665,
+      "step": 70434
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22689718488163e-07,
+      "loss": 0.5645,
+      "step": 70435
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.226631333641227e-07,
+      "loss": 0.582,
+      "step": 70436
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.226365484524264e-07,
+      "loss": 0.6562,
+      "step": 70437
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.226099637530946e-07,
+      "loss": 0.645,
+      "step": 70438
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.225833792661476e-07,
+      "loss": 0.7383,
+      "step": 70439
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22556794991606e-07,
+      "loss": 0.6606,
+      "step": 70440
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.2253021092949e-07,
+      "loss": 0.6611,
+      "step": 70441
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.225036270798193e-07,
+      "loss": 0.5408,
+      "step": 70442
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.224770434426152e-07,
+      "loss": 0.6846,
+      "step": 70443
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.224504600178973e-07,
+      "loss": 0.5349,
+      "step": 70444
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.224238768056867e-07,
+      "loss": 0.7139,
+      "step": 70445
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22397293806003e-07,
+      "loss": 0.6641,
+      "step": 70446
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.223707110188672e-07,
+      "loss": 0.5217,
+      "step": 70447
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.223441284442992e-07,
+      "loss": 0.5918,
+      "step": 70448
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.223175460823199e-07,
+      "loss": 0.54,
+      "step": 70449
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.222909639329489e-07,
+      "loss": 0.5923,
+      "step": 70450
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.222643819962073e-07,
+      "loss": 0.6367,
+      "step": 70451
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22237800272115e-07,
+      "loss": 0.6089,
+      "step": 70452
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.222112187606921e-07,
+      "loss": 0.6787,
+      "step": 70453
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.221846374619597e-07,
+      "loss": 0.6494,
+      "step": 70454
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.221580563759374e-07,
+      "loss": 0.5967,
+      "step": 70455
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.221314755026462e-07,
+      "loss": 0.5151,
+      "step": 70456
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.22104894842106e-07,
+      "loss": 0.5628,
+      "step": 70457
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.220783143943372e-07,
+      "loss": 0.6602,
+      "step": 70458
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.220517341593605e-07,
+      "loss": 0.6621,
+      "step": 70459
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.220251541371957e-07,
+      "loss": 0.6675,
+      "step": 70460
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.219985743278641e-07,
+      "loss": 0.7412,
+      "step": 70461
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.21971994731385e-07,
+      "loss": 0.5845,
+      "step": 70462
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.219454153477788e-07,
+      "loss": 0.6296,
+      "step": 70463
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.219188361770667e-07,
+      "loss": 0.6436,
+      "step": 70464
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.218922572192682e-07,
+      "loss": 0.6118,
+      "step": 70465
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.218656784744043e-07,
+      "loss": 0.7305,
+      "step": 70466
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.218390999424948e-07,
+      "loss": 0.6426,
+      "step": 70467
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.218125216235607e-07,
+      "loss": 0.5225,
+      "step": 70468
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.217859435176215e-07,
+      "loss": 0.8086,
+      "step": 70469
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.217593656246983e-07,
+      "loss": 0.5449,
+      "step": 70470
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.217327879448114e-07,
+      "loss": 0.6104,
+      "step": 70471
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.217062104779806e-07,
+      "loss": 0.5986,
+      "step": 70472
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.216796332242265e-07,
+      "loss": 0.79,
+      "step": 70473
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.216530561835696e-07,
+      "loss": 0.6865,
+      "step": 70474
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.216264793560301e-07,
+      "loss": 0.6431,
+      "step": 70475
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.215999027416286e-07,
+      "loss": 0.5957,
+      "step": 70476
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.21573326340385e-07,
+      "loss": 0.623,
+      "step": 70477
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.2154675015232e-07,
+      "loss": 0.5835,
+      "step": 70478
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.215201741774537e-07,
+      "loss": 0.5369,
+      "step": 70479
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.214935984158068e-07,
+      "loss": 0.7373,
+      "step": 70480
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.214670228673995e-07,
+      "loss": 0.8545,
+      "step": 70481
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.21440447532252e-07,
+      "loss": 0.6079,
+      "step": 70482
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.214138724103848e-07,
+      "loss": 0.5986,
+      "step": 70483
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.213872975018181e-07,
+      "loss": 0.6094,
+      "step": 70484
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.213607228065723e-07,
+      "loss": 0.6655,
+      "step": 70485
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.213341483246677e-07,
+      "loss": 0.6572,
+      "step": 70486
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.21307574056125e-07,
+      "loss": 0.5273,
+      "step": 70487
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.212810000009639e-07,
+      "loss": 0.6572,
+      "step": 70488
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.212544261592055e-07,
+      "loss": 0.6543,
+      "step": 70489
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.212278525308695e-07,
+      "loss": 0.5771,
+      "step": 70490
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.212012791159771e-07,
+      "loss": 0.7344,
+      "step": 70491
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.211747059145474e-07,
+      "loss": 0.6709,
+      "step": 70492
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.211481329266017e-07,
+      "loss": 0.6099,
+      "step": 70493
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.211215601521599e-07,
+      "loss": 0.7002,
+      "step": 70494
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.210949875912426e-07,
+      "loss": 0.793,
+      "step": 70495
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.210684152438698e-07,
+      "loss": 0.5693,
+      "step": 70496
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.210418431100624e-07,
+      "loss": 0.5571,
+      "step": 70497
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.210152711898402e-07,
+      "loss": 0.6016,
+      "step": 70498
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.20988699483224e-07,
+      "loss": 0.4774,
+      "step": 70499
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.209621279902336e-07,
+      "loss": 0.5492,
+      "step": 70500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.209355567108903e-07,
+      "loss": 0.5952,
+      "step": 70501
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.209089856452134e-07,
+      "loss": 0.5596,
+      "step": 70502
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.208824147932235e-07,
+      "loss": 0.5254,
+      "step": 70503
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.208558441549414e-07,
+      "loss": 0.7637,
+      "step": 70504
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.208292737303869e-07,
+      "loss": 0.7012,
+      "step": 70505
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.208027035195807e-07,
+      "loss": 0.7646,
+      "step": 70506
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.20776133522543e-07,
+      "loss": 0.3984,
+      "step": 70507
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.207495637392942e-07,
+      "loss": 0.3967,
+      "step": 70508
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.207229941698545e-07,
+      "loss": 0.3989,
+      "step": 70509
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.206964248142446e-07,
+      "loss": 0.7075,
+      "step": 70510
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.206698556724847e-07,
+      "loss": 0.6719,
+      "step": 70511
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.206432867445949e-07,
+      "loss": 0.5649,
+      "step": 70512
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.206167180305956e-07,
+      "loss": 0.4865,
+      "step": 70513
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.205901495305073e-07,
+      "loss": 0.6094,
+      "step": 70514
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.2056358124435e-07,
+      "loss": 0.2463,
+      "step": 70515
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.205370131721447e-07,
+      "loss": 0.7607,
+      "step": 70516
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.205104453139111e-07,
+      "loss": 0.562,
+      "step": 70517
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.2048387766967e-07,
+      "loss": 0.7314,
+      "step": 70518
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.204573102394414e-07,
+      "loss": 0.6956,
+      "step": 70519
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.204307430232458e-07,
+      "loss": 0.4785,
+      "step": 70520
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.20404176021104e-07,
+      "loss": 0.5996,
+      "step": 70521
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.203776092330352e-07,
+      "loss": 0.5874,
+      "step": 70522
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.203510426590606e-07,
+      "loss": 0.6206,
+      "step": 70523
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.203244762992004e-07,
+      "loss": 0.5649,
+      "step": 70524
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.20297910153475e-07,
+      "loss": 0.5977,
+      "step": 70525
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.202713442219042e-07,
+      "loss": 0.6475,
+      "step": 70526
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.202447785045091e-07,
+      "loss": 0.6572,
+      "step": 70527
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.202182130013097e-07,
+      "loss": 0.5679,
+      "step": 70528
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.201916477123264e-07,
+      "loss": 0.4874,
+      "step": 70529
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.201650826375795e-07,
+      "loss": 0.7183,
+      "step": 70530
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.201385177770896e-07,
+      "loss": 0.707,
+      "step": 70531
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.201119531308762e-07,
+      "loss": 0.4204,
+      "step": 70532
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.200853886989604e-07,
+      "loss": 0.5405,
+      "step": 70533
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.200588244813624e-07,
+      "loss": 0.6436,
+      "step": 70534
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.200322604781024e-07,
+      "loss": 0.6831,
+      "step": 70535
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.200056966892007e-07,
+      "loss": 0.4932,
+      "step": 70536
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.199791331146781e-07,
+      "loss": 0.6406,
+      "step": 70537
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.199525697545543e-07,
+      "loss": 0.7803,
+      "step": 70538
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.199260066088501e-07,
+      "loss": 0.7109,
+      "step": 70539
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.198994436775856e-07,
+      "loss": 0.5991,
+      "step": 70540
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.198728809607817e-07,
+      "loss": 0.5745,
+      "step": 70541
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.198463184584578e-07,
+      "loss": 0.7451,
+      "step": 70542
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.198197561706346e-07,
+      "loss": 0.5581,
+      "step": 70543
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.197931940973327e-07,
+      "loss": 0.5562,
+      "step": 70544
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.197666322385722e-07,
+      "loss": 0.4834,
+      "step": 70545
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.197400705943736e-07,
+      "loss": 0.5317,
+      "step": 70546
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.197135091647572e-07,
+      "loss": 0.8203,
+      "step": 70547
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.196869479497429e-07,
+      "loss": 0.5415,
+      "step": 70548
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.196603869493518e-07,
+      "loss": 0.5586,
+      "step": 70549
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.196338261636036e-07,
+      "loss": 0.5227,
+      "step": 70550
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.196072655925193e-07,
+      "loss": 0.6348,
+      "step": 70551
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.195807052361185e-07,
+      "loss": 0.6677,
+      "step": 70552
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.195541450944218e-07,
+      "loss": 0.6875,
+      "step": 70553
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.195275851674498e-07,
+      "loss": 0.5217,
+      "step": 70554
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.195010254552224e-07,
+      "loss": 0.6172,
+      "step": 70555
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.194744659577602e-07,
+      "loss": 0.7598,
+      "step": 70556
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.194479066750836e-07,
+      "loss": 0.4719,
+      "step": 70557
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.194213476072128e-07,
+      "loss": 0.5989,
+      "step": 70558
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.193947887541682e-07,
+      "loss": 0.5718,
+      "step": 70559
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.193682301159701e-07,
+      "loss": 0.749,
+      "step": 70560
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.193416716926392e-07,
+      "loss": 0.644,
+      "step": 70561
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.193151134841951e-07,
+      "loss": 0.6567,
+      "step": 70562
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.192885554906585e-07,
+      "loss": 0.6367,
+      "step": 70563
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.192619977120497e-07,
+      "loss": 0.6543,
+      "step": 70564
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.192354401483891e-07,
+      "loss": 0.6729,
+      "step": 70565
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.192088827996971e-07,
+      "loss": 0.7451,
+      "step": 70566
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.191823256659939e-07,
+      "loss": 0.665,
+      "step": 70567
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.191557687473e-07,
+      "loss": 0.5188,
+      "step": 70568
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.191292120436354e-07,
+      "loss": 0.6421,
+      "step": 70569
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.191026555550208e-07,
+      "loss": 0.5767,
+      "step": 70570
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.190760992814766e-07,
+      "loss": 0.6885,
+      "step": 70571
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.190495432230226e-07,
+      "loss": 0.8438,
+      "step": 70572
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.190229873796796e-07,
+      "loss": 0.5117,
+      "step": 70573
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.189964317514675e-07,
+      "loss": 0.8115,
+      "step": 70574
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.189698763384071e-07,
+      "loss": 0.6777,
+      "step": 70575
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.189433211405185e-07,
+      "loss": 0.5713,
+      "step": 70576
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.189167661578222e-07,
+      "loss": 0.624,
+      "step": 70577
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.188902113903382e-07,
+      "loss": 0.6719,
+      "step": 70578
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.188636568380872e-07,
+      "loss": 0.6738,
+      "step": 70579
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.188371025010891e-07,
+      "loss": 0.7695,
+      "step": 70580
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.188105483793653e-07,
+      "loss": 0.8086,
+      "step": 70581
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.187839944729347e-07,
+      "loss": 0.7324,
+      "step": 70582
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.187574407818184e-07,
+      "loss": 0.7412,
+      "step": 70583
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.187308873060364e-07,
+      "loss": 0.5605,
+      "step": 70584
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.187043340456094e-07,
+      "loss": 0.6533,
+      "step": 70585
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.186777810005573e-07,
+      "loss": 0.5824,
+      "step": 70586
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.186512281709011e-07,
+      "loss": 0.5659,
+      "step": 70587
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.186246755566602e-07,
+      "loss": 0.6924,
+      "step": 70588
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.18598123157856e-07,
+      "loss": 0.8564,
+      "step": 70589
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.185715709745079e-07,
+      "loss": 0.7275,
+      "step": 70590
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.185450190066372e-07,
+      "loss": 0.5757,
+      "step": 70591
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.185184672542631e-07,
+      "loss": 0.5713,
+      "step": 70592
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.184919157174064e-07,
+      "loss": 0.6313,
+      "step": 70593
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.184653643960878e-07,
+      "loss": 0.4771,
+      "step": 70594
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.18438813290327e-07,
+      "loss": 0.5166,
+      "step": 70595
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.184122624001449e-07,
+      "loss": 0.618,
+      "step": 70596
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.183857117255614e-07,
+      "loss": 0.6738,
+      "step": 70597
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.183591612665971e-07,
+      "loss": 0.7383,
+      "step": 70598
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.183326110232722e-07,
+      "loss": 0.793,
+      "step": 70599
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.183060609956073e-07,
+      "loss": 0.657,
+      "step": 70600
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.182795111836226e-07,
+      "loss": 0.7158,
+      "step": 70601
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.182529615873381e-07,
+      "loss": 0.4766,
+      "step": 70602
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.182264122067741e-07,
+      "loss": 0.4106,
+      "step": 70603
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.181998630419514e-07,
+      "loss": 0.7598,
+      "step": 70604
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.1817331409289e-07,
+      "loss": 0.6401,
+      "step": 70605
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.181467653596106e-07,
+      "loss": 0.4761,
+      "step": 70606
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.18120216842133e-07,
+      "loss": 0.6328,
+      "step": 70607
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.180936685404779e-07,
+      "loss": 0.6641,
+      "step": 70608
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.180671204546654e-07,
+      "loss": 0.7612,
+      "step": 70609
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.180405725847161e-07,
+      "loss": 0.6108,
+      "step": 70610
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.180140249306503e-07,
+      "loss": 0.7271,
+      "step": 70611
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.179874774924879e-07,
+      "loss": 0.6807,
+      "step": 70612
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.179609302702498e-07,
+      "loss": 0.6846,
+      "step": 70613
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.179343832639557e-07,
+      "loss": 0.6914,
+      "step": 70614
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.179078364736265e-07,
+      "loss": 0.5314,
+      "step": 70615
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.178812898992821e-07,
+      "loss": 0.6055,
+      "step": 70616
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.178547435409432e-07,
+      "loss": 0.6309,
+      "step": 70617
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.178281973986297e-07,
+      "loss": 0.7041,
+      "step": 70618
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.178016514723626e-07,
+      "loss": 0.8301,
+      "step": 70619
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.177751057621617e-07,
+      "loss": 0.7988,
+      "step": 70620
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.177485602680475e-07,
+      "loss": 0.6367,
+      "step": 70621
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.1772201499004e-07,
+      "loss": 0.7466,
+      "step": 70622
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.176954699281599e-07,
+      "loss": 0.4572,
+      "step": 70623
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.176689250824273e-07,
+      "loss": 0.5693,
+      "step": 70624
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.176423804528626e-07,
+      "loss": 0.5482,
+      "step": 70625
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.176158360394862e-07,
+      "loss": 0.4497,
+      "step": 70626
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.175892918423183e-07,
+      "loss": 0.5579,
+      "step": 70627
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.175627478613792e-07,
+      "loss": 0.6685,
+      "step": 70628
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.175362040966897e-07,
+      "loss": 0.5928,
+      "step": 70629
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.175096605482698e-07,
+      "loss": 0.6992,
+      "step": 70630
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.174831172161393e-07,
+      "loss": 0.6104,
+      "step": 70631
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.174565741003193e-07,
+      "loss": 0.6553,
+      "step": 70632
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.174300312008295e-07,
+      "loss": 0.499,
+      "step": 70633
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.174034885176908e-07,
+      "loss": 0.5078,
+      "step": 70634
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.173769460509231e-07,
+      "loss": 0.5366,
+      "step": 70635
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.17350403800547e-07,
+      "loss": 0.6265,
+      "step": 70636
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.173238617665827e-07,
+      "loss": 0.5913,
+      "step": 70637
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.172973199490505e-07,
+      "loss": 0.5591,
+      "step": 70638
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.172707783479705e-07,
+      "loss": 0.5303,
+      "step": 70639
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.172442369633639e-07,
+      "loss": 0.6128,
+      "step": 70640
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.172176957952498e-07,
+      "loss": 0.8516,
+      "step": 70641
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.171911548436494e-07,
+      "loss": 0.5216,
+      "step": 70642
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.171646141085824e-07,
+      "loss": 0.7432,
+      "step": 70643
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.171380735900699e-07,
+      "loss": 0.7373,
+      "step": 70644
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.171115332881313e-07,
+      "loss": 0.5112,
+      "step": 70645
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.170849932027876e-07,
+      "loss": 0.6221,
+      "step": 70646
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.170584533340589e-07,
+      "loss": 0.5972,
+      "step": 70647
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.170319136819656e-07,
+      "loss": 0.437,
+      "step": 70648
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.170053742465278e-07,
+      "loss": 0.6465,
+      "step": 70649
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.169788350277663e-07,
+      "loss": 0.6616,
+      "step": 70650
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.169522960257009e-07,
+      "loss": 0.606,
+      "step": 70651
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.16925757240352e-07,
+      "loss": 0.6338,
+      "step": 70652
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.168992186717402e-07,
+      "loss": 0.4312,
+      "step": 70653
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.168726803198857e-07,
+      "loss": 0.4712,
+      "step": 70654
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.168461421848085e-07,
+      "loss": 0.8076,
+      "step": 70655
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.168196042665293e-07,
+      "loss": 0.6987,
+      "step": 70656
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.167930665650683e-07,
+      "loss": 0.7366,
+      "step": 70657
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.167665290804459e-07,
+      "loss": 0.6875,
+      "step": 70658
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.167399918126822e-07,
+      "loss": 0.512,
+      "step": 70659
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.16713454761798e-07,
+      "loss": 0.5155,
+      "step": 70660
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.166869179278132e-07,
+      "loss": 0.6035,
+      "step": 70661
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.166603813107479e-07,
+      "loss": 0.6382,
+      "step": 70662
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.16633844910623e-07,
+      "loss": 0.6064,
+      "step": 70663
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.166073087274582e-07,
+      "loss": 0.6274,
+      "step": 70664
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.165807727612745e-07,
+      "loss": 0.5503,
+      "step": 70665
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.165542370120915e-07,
+      "loss": 0.5629,
+      "step": 70666
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.165277014799303e-07,
+      "loss": 0.485,
+      "step": 70667
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.165011661648104e-07,
+      "loss": 0.71,
+      "step": 70668
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.164746310667529e-07,
+      "loss": 0.5366,
+      "step": 70669
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.164480961857777e-07,
+      "loss": 0.6562,
+      "step": 70670
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.164215615219052e-07,
+      "loss": 0.5649,
+      "step": 70671
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.163950270751553e-07,
+      "loss": 0.5957,
+      "step": 70672
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.163684928455491e-07,
+      "loss": 0.7271,
+      "step": 70673
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.163419588331061e-07,
+      "loss": 0.5615,
+      "step": 70674
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.163154250378474e-07,
+      "loss": 0.5247,
+      "step": 70675
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.162888914597925e-07,
+      "loss": 0.5479,
+      "step": 70676
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.162623580989625e-07,
+      "loss": 0.6807,
+      "step": 70677
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.162358249553772e-07,
+      "loss": 0.7432,
+      "step": 70678
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.162092920290572e-07,
+      "loss": 0.6113,
+      "step": 70679
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.16182759320023e-07,
+      "loss": 0.5703,
+      "step": 70680
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.16156226828294e-07,
+      "loss": 0.7041,
+      "step": 70681
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.161296945538915e-07,
+      "loss": 0.5625,
+      "step": 70682
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.161031624968353e-07,
+      "loss": 0.5852,
+      "step": 70683
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.160766306571458e-07,
+      "loss": 0.6187,
+      "step": 70684
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.160500990348433e-07,
+      "loss": 0.6123,
+      "step": 70685
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.160235676299484e-07,
+      "loss": 0.6729,
+      "step": 70686
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.15997036442481e-07,
+      "loss": 0.5742,
+      "step": 70687
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.159705054724618e-07,
+      "loss": 0.7432,
+      "step": 70688
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.159439747199107e-07,
+      "loss": 0.7515,
+      "step": 70689
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.15917444184849e-07,
+      "loss": 0.752,
+      "step": 70690
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.158909138672955e-07,
+      "loss": 0.6226,
+      "step": 70691
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.158643837672714e-07,
+      "loss": 0.5681,
+      "step": 70692
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.158378538847968e-07,
+      "loss": 0.5896,
+      "step": 70693
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.158113242198922e-07,
+      "loss": 0.665,
+      "step": 70694
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.157847947725777e-07,
+      "loss": 0.7793,
+      "step": 70695
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.157582655428739e-07,
+      "loss": 0.5145,
+      "step": 70696
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.157317365308006e-07,
+      "loss": 0.7061,
+      "step": 70697
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.157052077363788e-07,
+      "loss": 0.5823,
+      "step": 70698
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.156786791596281e-07,
+      "loss": 0.7344,
+      "step": 70699
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.156521508005697e-07,
+      "loss": 0.6543,
+      "step": 70700
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.156256226592231e-07,
+      "loss": 0.6162,
+      "step": 70701
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.155990947356088e-07,
+      "loss": 0.52,
+      "step": 70702
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.155725670297473e-07,
+      "loss": 0.5972,
+      "step": 70703
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.155460395416585e-07,
+      "loss": 0.4755,
+      "step": 70704
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.155195122713633e-07,
+      "loss": 0.6172,
+      "step": 70705
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.154929852188816e-07,
+      "loss": 0.6797,
+      "step": 70706
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.15466458384234e-07,
+      "loss": 0.7285,
+      "step": 70707
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.154399317674404e-07,
+      "loss": 0.4541,
+      "step": 70708
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.154134053685216e-07,
+      "loss": 0.8516,
+      "step": 70709
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.15386879187498e-07,
+      "loss": 0.5508,
+      "step": 70710
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.153603532243891e-07,
+      "loss": 0.5864,
+      "step": 70711
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.153338274792157e-07,
+      "loss": 0.6196,
+      "step": 70712
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.153073019519982e-07,
+      "loss": 0.7402,
+      "step": 70713
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.152807766427567e-07,
+      "loss": 0.5526,
+      "step": 70714
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.152542515515116e-07,
+      "loss": 0.5986,
+      "step": 70715
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.152277266782834e-07,
+      "loss": 0.5894,
+      "step": 70716
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.152012020230921e-07,
+      "loss": 0.584,
+      "step": 70717
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.151746775859581e-07,
+      "loss": 0.5752,
+      "step": 70718
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.15148153366902e-07,
+      "loss": 0.6257,
+      "step": 70719
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.151216293659439e-07,
+      "loss": 0.7622,
+      "step": 70720
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.150951055831038e-07,
+      "loss": 0.6367,
+      "step": 70721
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.150685820184023e-07,
+      "loss": 0.8203,
+      "step": 70722
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.150420586718598e-07,
+      "loss": 0.6846,
+      "step": 70723
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.150155355434963e-07,
+      "loss": 0.5625,
+      "step": 70724
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.149890126333325e-07,
+      "loss": 0.4924,
+      "step": 70725
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.149624899413884e-07,
+      "loss": 0.6411,
+      "step": 70726
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.149359674676844e-07,
+      "loss": 0.5146,
+      "step": 70727
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.149094452122408e-07,
+      "loss": 0.7349,
+      "step": 70728
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.148829231750781e-07,
+      "loss": 0.7109,
+      "step": 70729
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.148564013562167e-07,
+      "loss": 0.6719,
+      "step": 70730
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.148298797556762e-07,
+      "loss": 0.4082,
+      "step": 70731
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.148033583734775e-07,
+      "loss": 0.623,
+      "step": 70732
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.147768372096405e-07,
+      "loss": 0.792,
+      "step": 70733
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.14750316264186e-07,
+      "loss": 0.7607,
+      "step": 70734
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.147237955371339e-07,
+      "loss": 0.5776,
+      "step": 70735
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.14697275028505e-07,
+      "loss": 0.6108,
+      "step": 70736
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.146707547383188e-07,
+      "loss": 0.8252,
+      "step": 70737
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.146442346665964e-07,
+      "loss": 0.4653,
+      "step": 70738
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.146177148133577e-07,
+      "loss": 0.6914,
+      "step": 70739
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.145911951786234e-07,
+      "loss": 0.626,
+      "step": 70740
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.145646757624133e-07,
+      "loss": 0.5996,
+      "step": 70741
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.145381565647477e-07,
+      "loss": 0.5571,
+      "step": 70742
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.145116375856472e-07,
+      "loss": 0.7783,
+      "step": 70743
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.144851188251322e-07,
+      "loss": 0.563,
+      "step": 70744
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.144586002832226e-07,
+      "loss": 0.8096,
+      "step": 70745
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.14432081959939e-07,
+      "loss": 0.6074,
+      "step": 70746
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.144055638553015e-07,
+      "loss": 0.5811,
+      "step": 70747
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.143790459693307e-07,
+      "loss": 0.7061,
+      "step": 70748
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.143525283020466e-07,
+      "loss": 0.6309,
+      "step": 70749
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.1432601085347e-07,
+      "loss": 0.6831,
+      "step": 70750
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.142994936236206e-07,
+      "loss": 0.7637,
+      "step": 70751
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.142729766125189e-07,
+      "loss": 0.6587,
+      "step": 70752
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.142464598201851e-07,
+      "loss": 0.6343,
+      "step": 70753
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.142199432466397e-07,
+      "loss": 0.4646,
+      "step": 70754
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.141934268919031e-07,
+      "loss": 0.4778,
+      "step": 70755
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.141669107559953e-07,
+      "loss": 0.6172,
+      "step": 70756
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.141403948389369e-07,
+      "loss": 0.459,
+      "step": 70757
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.141138791407478e-07,
+      "loss": 0.6152,
+      "step": 70758
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.140873636614488e-07,
+      "loss": 0.4519,
+      "step": 70759
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.140608484010602e-07,
+      "loss": 0.4971,
+      "step": 70760
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.140343333596017e-07,
+      "loss": 0.7871,
+      "step": 70761
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.140078185370941e-07,
+      "loss": 0.6279,
+      "step": 70762
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.139813039335576e-07,
+      "loss": 0.709,
+      "step": 70763
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.139547895490121e-07,
+      "loss": 0.7207,
+      "step": 70764
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.139282753834786e-07,
+      "loss": 0.6475,
+      "step": 70765
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.139017614369769e-07,
+      "loss": 0.6445,
+      "step": 70766
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.138752477095276e-07,
+      "loss": 0.5112,
+      "step": 70767
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.138487342011508e-07,
+      "loss": 0.5991,
+      "step": 70768
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.138222209118668e-07,
+      "loss": 0.3817,
+      "step": 70769
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.137957078416964e-07,
+      "loss": 0.7305,
+      "step": 70770
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.13769194990659e-07,
+      "loss": 0.6655,
+      "step": 70771
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.137426823587756e-07,
+      "loss": 0.6924,
+      "step": 70772
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.137161699460659e-07,
+      "loss": 0.6504,
+      "step": 70773
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.13689657752551e-07,
+      "loss": 0.4985,
+      "step": 70774
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.136631457782504e-07,
+      "loss": 0.6846,
+      "step": 70775
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.136366340231849e-07,
+      "loss": 0.5376,
+      "step": 70776
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.136101224873747e-07,
+      "loss": 0.6211,
+      "step": 70777
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.135836111708401e-07,
+      "loss": 0.8584,
+      "step": 70778
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.135571000736012e-07,
+      "loss": 0.6553,
+      "step": 70779
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.135305891956791e-07,
+      "loss": 0.5039,
+      "step": 70780
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.135040785370927e-07,
+      "loss": 0.5532,
+      "step": 70781
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.134775680978634e-07,
+      "loss": 0.5005,
+      "step": 70782
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.13451057878011e-07,
+      "loss": 0.7588,
+      "step": 70783
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.13424547877556e-07,
+      "loss": 0.5093,
+      "step": 70784
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.133980380965186e-07,
+      "loss": 0.6367,
+      "step": 70785
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.133715285349191e-07,
+      "loss": 0.6865,
+      "step": 70786
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.133450191927779e-07,
+      "loss": 0.5664,
+      "step": 70787
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.133185100701153e-07,
+      "loss": 0.7275,
+      "step": 70788
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.132920011669514e-07,
+      "loss": 0.6479,
+      "step": 70789
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.13265492483307e-07,
+      "loss": 0.5806,
+      "step": 70790
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.132389840192018e-07,
+      "loss": 0.7842,
+      "step": 70791
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.132124757746561e-07,
+      "loss": 0.6221,
+      "step": 70792
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.131859677496908e-07,
+      "loss": 0.8545,
+      "step": 70793
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.131594599443254e-07,
+      "loss": 0.6768,
+      "step": 70794
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.131329523585809e-07,
+      "loss": 0.6348,
+      "step": 70795
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.131064449924771e-07,
+      "loss": 0.6958,
+      "step": 70796
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.130799378460347e-07,
+      "loss": 0.5645,
+      "step": 70797
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.130534309192736e-07,
+      "loss": 0.7705,
+      "step": 70798
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.130269242122144e-07,
+      "loss": 0.5776,
+      "step": 70799
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.130004177248777e-07,
+      "loss": 0.6855,
+      "step": 70800
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.12973911457283e-07,
+      "loss": 0.708,
+      "step": 70801
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.129474054094509e-07,
+      "loss": 0.5439,
+      "step": 70802
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.12920899581402e-07,
+      "loss": 0.7412,
+      "step": 70803
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.12894393973156e-07,
+      "loss": 0.6216,
+      "step": 70804
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.128678885847339e-07,
+      "loss": 0.5615,
+      "step": 70805
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.128413834161554e-07,
+      "loss": 0.5615,
+      "step": 70806
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.128148784674412e-07,
+      "loss": 0.6562,
+      "step": 70807
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.127883737386113e-07,
+      "loss": 0.6367,
+      "step": 70808
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.127618692296864e-07,
+      "loss": 0.5718,
+      "step": 70809
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.127353649406867e-07,
+      "loss": 0.6272,
+      "step": 70810
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.127088608716319e-07,
+      "loss": 0.5342,
+      "step": 70811
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.126823570225429e-07,
+      "loss": 0.6709,
+      "step": 70812
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.126558533934396e-07,
+      "loss": 0.7363,
+      "step": 70813
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 7.126293499843427e-07,
+      "loss": 0.5837,
+      "step": 70814
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.12602846795272e-07,
+      "loss": 0.5674,
+      "step": 70815
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.125763438262485e-07,
+      "loss": 0.6128,
+      "step": 70816
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.12549841077292e-07,
+      "loss": 0.5645,
+      "step": 70817
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.125233385484226e-07,
+      "loss": 0.6973,
+      "step": 70818
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.124968362396611e-07,
+      "loss": 0.5332,
+      "step": 70819
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.124703341510278e-07,
+      "loss": 0.4814,
+      "step": 70820
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.124438322825423e-07,
+      "loss": 0.7275,
+      "step": 70821
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.124173306342255e-07,
+      "loss": 0.6099,
+      "step": 70822
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.123908292060973e-07,
+      "loss": 0.6846,
+      "step": 70823
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.123643279981786e-07,
+      "loss": 0.5781,
+      "step": 70824
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.12337827010489e-07,
+      "loss": 0.5066,
+      "step": 70825
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.123113262430492e-07,
+      "loss": 0.394,
+      "step": 70826
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.122848256958792e-07,
+      "loss": 0.6299,
+      "step": 70827
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.122583253689998e-07,
+      "loss": 0.7314,
+      "step": 70828
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.122318252624307e-07,
+      "loss": 0.5608,
+      "step": 70829
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.122053253761928e-07,
+      "loss": 0.6807,
+      "step": 70830
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.12178825710306e-07,
+      "loss": 0.6611,
+      "step": 70831
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.121523262647903e-07,
+      "loss": 0.6147,
+      "step": 70832
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.121258270396666e-07,
+      "loss": 0.8848,
+      "step": 70833
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.12099328034955e-07,
+      "loss": 0.4941,
+      "step": 70834
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.120728292506754e-07,
+      "loss": 0.7656,
+      "step": 70835
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.120463306868487e-07,
+      "loss": 0.7188,
+      "step": 70836
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.120198323434945e-07,
+      "loss": 0.4399,
+      "step": 70837
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.119933342206337e-07,
+      "loss": 0.7988,
+      "step": 70838
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.119668363182864e-07,
+      "loss": 0.6455,
+      "step": 70839
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.119403386364731e-07,
+      "loss": 0.5947,
+      "step": 70840
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.119138411752136e-07,
+      "loss": 0.5371,
+      "step": 70841
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.118873439345282e-07,
+      "loss": 0.5659,
+      "step": 70842
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.118608469144376e-07,
+      "loss": 0.5664,
+      "step": 70843
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.118343501149617e-07,
+      "loss": 0.6011,
+      "step": 70844
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.118078535361214e-07,
+      "loss": 0.6211,
+      "step": 70845
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.117813571779361e-07,
+      "loss": 0.5654,
+      "step": 70846
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.117548610404269e-07,
+      "loss": 0.5596,
+      "step": 70847
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.117283651236135e-07,
+      "loss": 0.7764,
+      "step": 70848
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.117018694275168e-07,
+      "loss": 0.5928,
+      "step": 70849
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.116753739521569e-07,
+      "loss": 0.6509,
+      "step": 70850
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.116488786975532e-07,
+      "loss": 0.7217,
+      "step": 70851
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.116223836637272e-07,
+      "loss": 0.707,
+      "step": 70852
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.115958888506986e-07,
+      "loss": 0.4626,
+      "step": 70853
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.115693942584875e-07,
+      "loss": 0.5261,
+      "step": 70854
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.115428998871147e-07,
+      "loss": 0.6343,
+      "step": 70855
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.115164057366e-07,
+      "loss": 0.5747,
+      "step": 70856
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.114899118069642e-07,
+      "loss": 0.6406,
+      "step": 70857
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.114634180982271e-07,
+      "loss": 0.5913,
+      "step": 70858
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.114369246104093e-07,
+      "loss": 0.6963,
+      "step": 70859
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.114104313435312e-07,
+      "loss": 0.3267,
+      "step": 70860
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.113839382976126e-07,
+      "loss": 0.6514,
+      "step": 70861
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.113574454726741e-07,
+      "loss": 0.7871,
+      "step": 70862
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.113309528687356e-07,
+      "loss": 0.6846,
+      "step": 70863
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.113044604858182e-07,
+      "loss": 0.6479,
+      "step": 70864
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.112779683239414e-07,
+      "loss": 0.7148,
+      "step": 70865
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.112514763831259e-07,
+      "loss": 0.7627,
+      "step": 70866
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.112249846633917e-07,
+      "loss": 0.6934,
+      "step": 70867
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.111984931647597e-07,
+      "loss": 0.751,
+      "step": 70868
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.111720018872493e-07,
+      "loss": 0.7432,
+      "step": 70869
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.111455108308818e-07,
+      "loss": 0.603,
+      "step": 70870
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.111190199956763e-07,
+      "loss": 0.6953,
+      "step": 70871
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.11092529381654e-07,
+      "loss": 0.5986,
+      "step": 70872
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.110660389888345e-07,
+      "loss": 0.7129,
+      "step": 70873
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.110395488172388e-07,
+      "loss": 0.6304,
+      "step": 70874
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.110130588668865e-07,
+      "loss": 0.5239,
+      "step": 70875
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.109865691377985e-07,
+      "loss": 0.502,
+      "step": 70876
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.109600796299946e-07,
+      "loss": 0.584,
+      "step": 70877
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.109335903434955e-07,
+      "loss": 0.6667,
+      "step": 70878
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.10907101278321e-07,
+      "loss": 0.7725,
+      "step": 70879
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.108806124344921e-07,
+      "loss": 0.624,
+      "step": 70880
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.108541238120284e-07,
+      "loss": 0.5322,
+      "step": 70881
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.108276354109501e-07,
+      "loss": 0.6968,
+      "step": 70882
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.108011472312782e-07,
+      "loss": 0.5303,
+      "step": 70883
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.107746592730322e-07,
+      "loss": 0.6567,
+      "step": 70884
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.107481715362329e-07,
+      "loss": 0.4507,
+      "step": 70885
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.107216840209003e-07,
+      "loss": 0.5701,
+      "step": 70886
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.106951967270551e-07,
+      "loss": 0.7744,
+      "step": 70887
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.106687096547169e-07,
+      "loss": 0.8223,
+      "step": 70888
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.106422228039068e-07,
+      "loss": 0.6826,
+      "step": 70889
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.106157361746446e-07,
+      "loss": 0.46,
+      "step": 70890
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.105892497669505e-07,
+      "loss": 0.6992,
+      "step": 70891
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.105627635808447e-07,
+      "loss": 0.4951,
+      "step": 70892
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.10536277616348e-07,
+      "loss": 0.5474,
+      "step": 70893
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.1050979187348e-07,
+      "loss": 0.6167,
+      "step": 70894
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.104833063522618e-07,
+      "loss": 0.6411,
+      "step": 70895
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.104568210527126e-07,
+      "loss": 0.625,
+      "step": 70896
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.104303359748538e-07,
+      "loss": 0.7783,
+      "step": 70897
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.104038511187049e-07,
+      "loss": 0.6287,
+      "step": 70898
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.103773664842867e-07,
+      "loss": 0.7227,
+      "step": 70899
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.103508820716194e-07,
+      "loss": 0.4171,
+      "step": 70900
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.103243978807227e-07,
+      "loss": 0.5425,
+      "step": 70901
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.102979139116176e-07,
+      "loss": 0.6458,
+      "step": 70902
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.102714301643237e-07,
+      "loss": 0.5986,
+      "step": 70903
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.102449466388619e-07,
+      "loss": 0.6064,
+      "step": 70904
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.10218463335252e-07,
+      "loss": 0.5923,
+      "step": 70905
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.101919802535148e-07,
+      "loss": 0.6934,
+      "step": 70906
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.101654973936702e-07,
+      "loss": 0.6714,
+      "step": 70907
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.101390147557382e-07,
+      "loss": 0.6807,
+      "step": 70908
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.101125323397399e-07,
+      "loss": 0.6699,
+      "step": 70909
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.100860501456953e-07,
+      "loss": 0.7881,
+      "step": 70910
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.100595681736239e-07,
+      "loss": 0.644,
+      "step": 70911
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.100330864235469e-07,
+      "loss": 0.627,
+      "step": 70912
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.100066048954838e-07,
+      "loss": 0.6401,
+      "step": 70913
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.099801235894557e-07,
+      "loss": 0.71,
+      "step": 70914
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.099536425054822e-07,
+      "loss": 0.6152,
+      "step": 70915
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.099271616435842e-07,
+      "loss": 0.5488,
+      "step": 70916
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.099006810037813e-07,
+      "loss": 0.6611,
+      "step": 70917
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.098742005860944e-07,
+      "loss": 0.6846,
+      "step": 70918
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.098477203905432e-07,
+      "loss": 0.7617,
+      "step": 70919
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.098212404171486e-07,
+      "loss": 0.5098,
+      "step": 70920
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.097947606659304e-07,
+      "loss": 0.6226,
+      "step": 70921
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.09768281136909e-07,
+      "loss": 0.7178,
+      "step": 70922
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.097418018301046e-07,
+      "loss": 0.6772,
+      "step": 70923
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.097153227455378e-07,
+      "loss": 0.5513,
+      "step": 70924
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.096888438832283e-07,
+      "loss": 0.3379,
+      "step": 70925
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.096623652431969e-07,
+      "loss": 0.5793,
+      "step": 70926
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.096358868254635e-07,
+      "loss": 0.6094,
+      "step": 70927
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.096094086300487e-07,
+      "loss": 0.6934,
+      "step": 70928
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.095829306569729e-07,
+      "loss": 0.5127,
+      "step": 70929
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.095564529062556e-07,
+      "loss": 0.5933,
+      "step": 70930
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.095299753779178e-07,
+      "loss": 0.8057,
+      "step": 70931
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.095034980719794e-07,
+      "loss": 0.7075,
+      "step": 70932
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.094770209884611e-07,
+      "loss": 0.4583,
+      "step": 70933
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.094505441273824e-07,
+      "loss": 0.7422,
+      "step": 70934
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.094240674887645e-07,
+      "loss": 0.6484,
+      "step": 70935
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.093975910726269e-07,
+      "loss": 0.7568,
+      "step": 70936
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.093711148789905e-07,
+      "loss": 0.519,
+      "step": 70937
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.09344638907875e-07,
+      "loss": 0.76,
+      "step": 70938
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.093181631593014e-07,
+      "loss": 0.6406,
+      "step": 70939
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.092916876332893e-07,
+      "loss": 0.6016,
+      "step": 70940
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.09265212329859e-07,
+      "loss": 0.5776,
+      "step": 70941
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.092387372490311e-07,
+      "loss": 0.5269,
+      "step": 70942
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.092122623908259e-07,
+      "loss": 0.7168,
+      "step": 70943
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.091857877552631e-07,
+      "loss": 0.6289,
+      "step": 70944
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.091593133423637e-07,
+      "loss": 0.7012,
+      "step": 70945
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.091328391521473e-07,
+      "loss": 0.5293,
+      "step": 70946
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.091063651846348e-07,
+      "loss": 0.665,
+      "step": 70947
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.090798914398459e-07,
+      "loss": 0.6611,
+      "step": 70948
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.090534179178017e-07,
+      "loss": 0.6865,
+      "step": 70949
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.090269446185217e-07,
+      "loss": 0.7217,
+      "step": 70950
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.090004715420261e-07,
+      "loss": 0.6299,
+      "step": 70951
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.089739986883356e-07,
+      "loss": 0.5767,
+      "step": 70952
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.089475260574703e-07,
+      "loss": 0.5454,
+      "step": 70953
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.089210536494506e-07,
+      "loss": 0.5552,
+      "step": 70954
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.088945814642964e-07,
+      "loss": 0.5594,
+      "step": 70955
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.088681095020285e-07,
+      "loss": 0.7158,
+      "step": 70956
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.088416377626667e-07,
+      "loss": 0.749,
+      "step": 70957
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.088151662462318e-07,
+      "loss": 0.6584,
+      "step": 70958
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.087886949527438e-07,
+      "loss": 0.5508,
+      "step": 70959
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.087622238822227e-07,
+      "loss": 0.5625,
+      "step": 70960
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.087357530346888e-07,
+      "loss": 0.4946,
+      "step": 70961
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.087092824101628e-07,
+      "loss": 0.5615,
+      "step": 70962
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.086828120086644e-07,
+      "loss": 0.627,
+      "step": 70963
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.086563418302145e-07,
+      "loss": 0.6475,
+      "step": 70964
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.086298718748327e-07,
+      "loss": 0.364,
+      "step": 70965
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.086034021425399e-07,
+      "loss": 0.7109,
+      "step": 70966
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.085769326333558e-07,
+      "loss": 0.7559,
+      "step": 70967
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.085504633473012e-07,
+      "loss": 0.7979,
+      "step": 70968
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.085239942843964e-07,
+      "loss": 0.5225,
+      "step": 70969
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.084975254446608e-07,
+      "loss": 0.6562,
+      "step": 70970
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.084710568281155e-07,
+      "loss": 0.6602,
+      "step": 70971
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.084445884347803e-07,
+      "loss": 0.5144,
+      "step": 70972
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.084181202646759e-07,
+      "loss": 0.7314,
+      "step": 70973
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.083916523178222e-07,
+      "loss": 0.8613,
+      "step": 70974
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.083651845942398e-07,
+      "loss": 0.5098,
+      "step": 70975
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.083387170939486e-07,
+      "loss": 0.3721,
+      "step": 70976
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.08312249816969e-07,
+      "loss": 0.5786,
+      "step": 70977
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.082857827633213e-07,
+      "loss": 0.6841,
+      "step": 70978
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.082593159330263e-07,
+      "loss": 0.5962,
+      "step": 70979
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.082328493261031e-07,
+      "loss": 0.6206,
+      "step": 70980
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.082063829425729e-07,
+      "loss": 0.5293,
+      "step": 70981
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.081799167824554e-07,
+      "loss": 0.5693,
+      "step": 70982
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.081534508457713e-07,
+      "loss": 0.7314,
+      "step": 70983
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.081269851325404e-07,
+      "loss": 0.6436,
+      "step": 70984
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.081005196427836e-07,
+      "loss": 0.7598,
+      "step": 70985
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.080740543765205e-07,
+      "loss": 0.5874,
+      "step": 70986
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.080475893337718e-07,
+      "loss": 0.7979,
+      "step": 70987
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.080211245145576e-07,
+      "loss": 0.6577,
+      "step": 70988
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.079946599188987e-07,
+      "loss": 0.5664,
+      "step": 70989
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.079681955468144e-07,
+      "loss": 0.5437,
+      "step": 70990
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.079417313983254e-07,
+      "loss": 0.6738,
+      "step": 70991
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.079152674734521e-07,
+      "loss": 0.54,
+      "step": 70992
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.078888037722144e-07,
+      "loss": 0.6196,
+      "step": 70993
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.078623402946331e-07,
+      "loss": 0.7324,
+      "step": 70994
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.07835877040728e-07,
+      "loss": 0.6143,
+      "step": 70995
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.078094140105197e-07,
+      "loss": 0.7314,
+      "step": 70996
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.077829512040283e-07,
+      "loss": 0.7852,
+      "step": 70997
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.077564886212739e-07,
+      "loss": 0.6748,
+      "step": 70998
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.077300262622772e-07,
+      "loss": 0.7285,
+      "step": 70999
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.077035641270581e-07,
+      "loss": 0.6963,
+      "step": 71000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.076771022156367e-07,
+      "loss": 0.5269,
+      "step": 71001
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.076506405280336e-07,
+      "loss": 0.6504,
+      "step": 71002
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.076241790642689e-07,
+      "loss": 0.7471,
+      "step": 71003
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.07597717824363e-07,
+      "loss": 0.6431,
+      "step": 71004
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.075712568083361e-07,
+      "loss": 0.5513,
+      "step": 71005
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.075447960162084e-07,
+      "loss": 0.4443,
+      "step": 71006
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.075183354480001e-07,
+      "loss": 0.5825,
+      "step": 71007
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.074918751037317e-07,
+      "loss": 0.6621,
+      "step": 71008
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.074654149834236e-07,
+      "loss": 0.6587,
+      "step": 71009
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.074389550870954e-07,
+      "loss": 0.6797,
+      "step": 71010
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.07412495414768e-07,
+      "loss": 0.7002,
+      "step": 71011
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.073860359664609e-07,
+      "loss": 0.7383,
+      "step": 71012
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.073595767421954e-07,
+      "loss": 0.748,
+      "step": 71013
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.073331177419907e-07,
+      "loss": 0.5693,
+      "step": 71014
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.073066589658681e-07,
+      "loss": 0.5459,
+      "step": 71015
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.072802004138472e-07,
+      "loss": 0.7676,
+      "step": 71016
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.072537420859481e-07,
+      "loss": 0.4183,
+      "step": 71017
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.072272839821916e-07,
+      "loss": 0.5259,
+      "step": 71018
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.072008261025981e-07,
+      "loss": 0.6558,
+      "step": 71019
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.071743684471868e-07,
+      "loss": 0.6689,
+      "step": 71020
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.071479110159789e-07,
+      "loss": 0.793,
+      "step": 71021
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.071214538089943e-07,
+      "loss": 0.7275,
+      "step": 71022
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.070949968262534e-07,
+      "loss": 0.7363,
+      "step": 71023
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.070685400677762e-07,
+      "loss": 0.7334,
+      "step": 71024
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.070420835335833e-07,
+      "loss": 0.3972,
+      "step": 71025
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.070156272236946e-07,
+      "loss": 0.6777,
+      "step": 71026
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.06989171138131e-07,
+      "loss": 0.5884,
+      "step": 71027
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.069627152769119e-07,
+      "loss": 0.6211,
+      "step": 71028
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.069362596400585e-07,
+      "loss": 0.7085,
+      "step": 71029
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.069098042275903e-07,
+      "loss": 0.5977,
+      "step": 71030
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.068833490395275e-07,
+      "loss": 0.6279,
+      "step": 71031
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.068568940758908e-07,
+      "loss": 0.427,
+      "step": 71032
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.068304393367005e-07,
+      "loss": 0.4592,
+      "step": 71033
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.068039848219763e-07,
+      "loss": 0.519,
+      "step": 71034
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.067775305317391e-07,
+      "loss": 0.4492,
+      "step": 71035
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.067510764660086e-07,
+      "loss": 0.8262,
+      "step": 71036
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.067246226248056e-07,
+      "loss": 0.5742,
+      "step": 71037
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.066981690081497e-07,
+      "loss": 0.5483,
+      "step": 71038
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.06671715616062e-07,
+      "loss": 0.6812,
+      "step": 71039
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.066452624485622e-07,
+      "loss": 0.6104,
+      "step": 71040
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.066188095056702e-07,
+      "loss": 0.4399,
+      "step": 71041
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.065923567874072e-07,
+      "loss": 0.7383,
+      "step": 71042
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.065659042937924e-07,
+      "loss": 0.5161,
+      "step": 71043
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.065394520248471e-07,
+      "loss": 0.6177,
+      "step": 71044
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.065129999805906e-07,
+      "loss": 0.5647,
+      "step": 71045
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.06486548161044e-07,
+      "loss": 0.5659,
+      "step": 71046
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.064600965662268e-07,
+      "loss": 0.562,
+      "step": 71047
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.064336451961599e-07,
+      "loss": 0.7627,
+      "step": 71048
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.064071940508635e-07,
+      "loss": 0.6211,
+      "step": 71049
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.063807431303573e-07,
+      "loss": 0.7744,
+      "step": 71050
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.063542924346618e-07,
+      "loss": 0.5652,
+      "step": 71051
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.063278419637973e-07,
+      "loss": 0.7773,
+      "step": 71052
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.063013917177841e-07,
+      "loss": 0.8066,
+      "step": 71053
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.062749416966425e-07,
+      "loss": 0.7256,
+      "step": 71054
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.062484919003924e-07,
+      "loss": 0.5879,
+      "step": 71055
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.062220423290545e-07,
+      "loss": 0.9297,
+      "step": 71056
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.061955929826489e-07,
+      "loss": 0.5815,
+      "step": 71057
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.061691438611959e-07,
+      "loss": 0.5452,
+      "step": 71058
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.061426949647159e-07,
+      "loss": 0.6836,
+      "step": 71059
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.061162462932284e-07,
+      "loss": 0.7568,
+      "step": 71060
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.060897978467546e-07,
+      "loss": 0.7285,
+      "step": 71061
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.06063349625314e-07,
+      "loss": 0.5532,
+      "step": 71062
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.060369016289274e-07,
+      "loss": 0.8428,
+      "step": 71063
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.060104538576146e-07,
+      "loss": 0.6924,
+      "step": 71064
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.059840063113962e-07,
+      "loss": 0.6538,
+      "step": 71065
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.059575589902923e-07,
+      "loss": 0.5947,
+      "step": 71066
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.059311118943233e-07,
+      "loss": 0.6592,
+      "step": 71067
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.05904665023509e-07,
+      "loss": 0.4854,
+      "step": 71068
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.058782183778709e-07,
+      "loss": 0.5002,
+      "step": 71069
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.058517719574275e-07,
+      "loss": 0.7178,
+      "step": 71070
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.058253257622e-07,
+      "loss": 0.6514,
+      "step": 71071
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.057988797922086e-07,
+      "loss": 0.5767,
+      "step": 71072
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.057724340474734e-07,
+      "loss": 0.6982,
+      "step": 71073
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.057459885280145e-07,
+      "loss": 0.6846,
+      "step": 71074
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.057195432338527e-07,
+      "loss": 0.6992,
+      "step": 71075
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.056930981650078e-07,
+      "loss": 0.6372,
+      "step": 71076
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.056666533215001e-07,
+      "loss": 0.5558,
+      "step": 71077
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.056402087033499e-07,
+      "loss": 0.7529,
+      "step": 71078
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.056137643105779e-07,
+      "loss": 0.5688,
+      "step": 71079
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.055873201432035e-07,
+      "loss": 0.6514,
+      "step": 71080
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.055608762012472e-07,
+      "loss": 0.5566,
+      "step": 71081
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.055344324847299e-07,
+      "loss": 0.5811,
+      "step": 71082
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.055079889936708e-07,
+      "loss": 0.603,
+      "step": 71083
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.05481545728091e-07,
+      "loss": 0.7275,
+      "step": 71084
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.054551026880103e-07,
+      "loss": 0.5603,
+      "step": 71085
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.054286598734491e-07,
+      "loss": 0.6987,
+      "step": 71086
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.054022172844278e-07,
+      "loss": 0.771,
+      "step": 71087
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.053757749209664e-07,
+      "loss": 0.6416,
+      "step": 71088
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.053493327830853e-07,
+      "loss": 0.5653,
+      "step": 71089
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.053228908708047e-07,
+      "loss": 0.666,
+      "step": 71090
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.052964491841445e-07,
+      "loss": 0.5934,
+      "step": 71091
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.052700077231255e-07,
+      "loss": 0.646,
+      "step": 71092
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.052435664877675e-07,
+      "loss": 0.6304,
+      "step": 71093
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.052171254780912e-07,
+      "loss": 0.6313,
+      "step": 71094
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.051906846941162e-07,
+      "loss": 0.7402,
+      "step": 71095
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.051642441358636e-07,
+      "loss": 0.6162,
+      "step": 71096
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.051378038033529e-07,
+      "loss": 0.7891,
+      "step": 71097
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.051113636966047e-07,
+      "loss": 0.7422,
+      "step": 71098
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.050849238156396e-07,
+      "loss": 0.5635,
+      "step": 71099
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.050584841604767e-07,
+      "loss": 0.6445,
+      "step": 71100
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.050320447311374e-07,
+      "loss": 0.6738,
+      "step": 71101
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.050056055276412e-07,
+      "loss": 0.5249,
+      "step": 71102
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.049791665500089e-07,
+      "loss": 0.5193,
+      "step": 71103
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.049527277982602e-07,
+      "loss": 0.7119,
+      "step": 71104
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.049262892724159e-07,
+      "loss": 0.6855,
+      "step": 71105
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.048998509724959e-07,
+      "loss": 0.6113,
+      "step": 71106
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.048734128985204e-07,
+      "loss": 0.6543,
+      "step": 71107
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.048469750505099e-07,
+      "loss": 0.6455,
+      "step": 71108
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.048205374284848e-07,
+      "loss": 0.6235,
+      "step": 71109
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.047941000324645e-07,
+      "loss": 0.4824,
+      "step": 71110
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.047676628624701e-07,
+      "loss": 0.6865,
+      "step": 71111
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.047412259185213e-07,
+      "loss": 0.7637,
+      "step": 71112
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.047147892006388e-07,
+      "loss": 0.7041,
+      "step": 71113
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.046883527088423e-07,
+      "loss": 0.7246,
+      "step": 71114
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.046619164431526e-07,
+      "loss": 0.5676,
+      "step": 71115
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.046354804035894e-07,
+      "loss": 0.6221,
+      "step": 71116
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.046090445901736e-07,
+      "loss": 0.6602,
+      "step": 71117
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.045826090029247e-07,
+      "loss": 0.5415,
+      "step": 71118
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.045561736418639e-07,
+      "loss": 0.7061,
+      "step": 71119
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.045297385070106e-07,
+      "loss": 0.6602,
+      "step": 71120
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.045033035983851e-07,
+      "loss": 0.5928,
+      "step": 71121
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.044768689160078e-07,
+      "loss": 0.6958,
+      "step": 71122
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.044504344598994e-07,
+      "loss": 0.6128,
+      "step": 71123
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.044240002300791e-07,
+      "loss": 0.5767,
+      "step": 71124
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.043975662265682e-07,
+      "loss": 0.7065,
+      "step": 71125
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.043711324493862e-07,
+      "loss": 0.6787,
+      "step": 71126
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.043446988985538e-07,
+      "loss": 0.5342,
+      "step": 71127
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.043182655740909e-07,
+      "loss": 0.5244,
+      "step": 71128
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.042918324760184e-07,
+      "loss": 0.5557,
+      "step": 71129
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.042653996043558e-07,
+      "loss": 0.8008,
+      "step": 71130
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.042389669591233e-07,
+      "loss": 0.6162,
+      "step": 71131
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.042125345403418e-07,
+      "loss": 0.4187,
+      "step": 71132
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.041861023480308e-07,
+      "loss": 0.5366,
+      "step": 71133
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.041596703822111e-07,
+      "loss": 0.8281,
+      "step": 71134
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.041332386429026e-07,
+      "loss": 0.6267,
+      "step": 71135
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.04106807130126e-07,
+      "loss": 0.7578,
+      "step": 71136
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.040803758439006e-07,
+      "loss": 0.6167,
+      "step": 71137
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.040539447842479e-07,
+      "loss": 0.6235,
+      "step": 71138
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.040275139511875e-07,
+      "loss": 0.6035,
+      "step": 71139
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.040010833447394e-07,
+      "loss": 0.5181,
+      "step": 71140
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.039746529649239e-07,
+      "loss": 0.6064,
+      "step": 71141
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.039482228117616e-07,
+      "loss": 0.7578,
+      "step": 71142
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.039217928852724e-07,
+      "loss": 0.6523,
+      "step": 71143
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.038953631854767e-07,
+      "loss": 0.615,
+      "step": 71144
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.038689337123946e-07,
+      "loss": 0.6079,
+      "step": 71145
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.038425044660465e-07,
+      "loss": 0.615,
+      "step": 71146
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.038160754464527e-07,
+      "loss": 0.6133,
+      "step": 71147
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.037896466536333e-07,
+      "loss": 0.5625,
+      "step": 71148
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.037632180876087e-07,
+      "loss": 0.5095,
+      "step": 71149
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.037367897483987e-07,
+      "loss": 0.6079,
+      "step": 71150
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.037103616360239e-07,
+      "loss": 0.7539,
+      "step": 71151
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.036839337505043e-07,
+      "loss": 0.6875,
+      "step": 71152
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.036575060918606e-07,
+      "loss": 0.7871,
+      "step": 71153
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.036310786601124e-07,
+      "loss": 0.6895,
+      "step": 71154
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.036046514552805e-07,
+      "loss": 0.7031,
+      "step": 71155
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.035782244773847e-07,
+      "loss": 0.6406,
+      "step": 71156
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.035517977264456e-07,
+      "loss": 0.5669,
+      "step": 71157
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.035253712024831e-07,
+      "loss": 0.4261,
+      "step": 71158
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.034989449055181e-07,
+      "loss": 0.5454,
+      "step": 71159
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.034725188355697e-07,
+      "loss": 0.7539,
+      "step": 71160
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.034460929926591e-07,
+      "loss": 0.5728,
+      "step": 71161
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.034196673768058e-07,
+      "loss": 0.6318,
+      "step": 71162
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.033932419880309e-07,
+      "loss": 0.7021,
+      "step": 71163
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.033668168263538e-07,
+      "loss": 0.8242,
+      "step": 71164
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.033403918917952e-07,
+      "loss": 0.6323,
+      "step": 71165
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.033139671843751e-07,
+      "loss": 0.6445,
+      "step": 71166
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.032875427041141e-07,
+      "loss": 0.5654,
+      "step": 71167
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.03261118451032e-07,
+      "loss": 0.6582,
+      "step": 71168
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.032346944251495e-07,
+      "loss": 0.5601,
+      "step": 71169
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.032082706264864e-07,
+      "loss": 0.7246,
+      "step": 71170
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.031818470550629e-07,
+      "loss": 0.6426,
+      "step": 71171
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.031554237108997e-07,
+      "loss": 0.6077,
+      "step": 71172
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.031290005940163e-07,
+      "loss": 0.5864,
+      "step": 71173
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.031025777044338e-07,
+      "loss": 0.6113,
+      "step": 71174
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.030761550421718e-07,
+      "loss": 0.7988,
+      "step": 71175
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.030497326072509e-07,
+      "loss": 0.6655,
+      "step": 71176
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.030233103996908e-07,
+      "loss": 0.6755,
+      "step": 71177
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.029968884195125e-07,
+      "loss": 0.645,
+      "step": 71178
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.029704666667359e-07,
+      "loss": 0.6143,
+      "step": 71179
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.029440451413811e-07,
+      "loss": 0.6416,
+      "step": 71180
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.02917623843468e-07,
+      "loss": 0.6367,
+      "step": 71181
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.028912027730175e-07,
+      "loss": 0.7593,
+      "step": 71182
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.028647819300493e-07,
+      "loss": 0.6943,
+      "step": 71183
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.028383613145842e-07,
+      "loss": 0.6084,
+      "step": 71184
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.028119409266417e-07,
+      "loss": 0.7842,
+      "step": 71185
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.027855207662429e-07,
+      "loss": 0.5693,
+      "step": 71186
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.027591008334072e-07,
+      "loss": 0.6641,
+      "step": 71187
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.027326811281554e-07,
+      "loss": 0.6719,
+      "step": 71188
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.027062616505078e-07,
+      "loss": 0.6245,
+      "step": 71189
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.026798424004838e-07,
+      "loss": 0.8369,
+      "step": 71190
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.026534233781045e-07,
+      "loss": 0.7568,
+      "step": 71191
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.026270045833896e-07,
+      "loss": 0.4941,
+      "step": 71192
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.026005860163596e-07,
+      "loss": 0.5601,
+      "step": 71193
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.025741676770344e-07,
+      "loss": 0.5894,
+      "step": 71194
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.02547749565435e-07,
+      "loss": 0.6636,
+      "step": 71195
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.025213316815808e-07,
+      "loss": 0.8184,
+      "step": 71196
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.024949140254923e-07,
+      "loss": 0.5801,
+      "step": 71197
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.024684965971899e-07,
+      "loss": 0.752,
+      "step": 71198
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.02442079396694e-07,
+      "loss": 0.6963,
+      "step": 71199
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.024156624240241e-07,
+      "loss": 0.6963,
+      "step": 71200
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.023892456792009e-07,
+      "loss": 0.5708,
+      "step": 71201
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.023628291622445e-07,
+      "loss": 0.6816,
+      "step": 71202
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.023364128731752e-07,
+      "loss": 0.667,
+      "step": 71203
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 7.023099968120133e-07,
+      "loss": 0.5264,
+      "step": 71204
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.02283580978779e-07,
+      "loss": 0.6436,
+      "step": 71205
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.022571653734924e-07,
+      "loss": 0.595,
+      "step": 71206
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.022307499961737e-07,
+      "loss": 0.5079,
+      "step": 71207
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.022043348468433e-07,
+      "loss": 0.584,
+      "step": 71208
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.021779199255216e-07,
+      "loss": 0.5234,
+      "step": 71209
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.021515052322285e-07,
+      "loss": 0.6904,
+      "step": 71210
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.02125090766984e-07,
+      "loss": 0.6484,
+      "step": 71211
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.020986765298089e-07,
+      "loss": 0.5308,
+      "step": 71212
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.020722625207231e-07,
+      "loss": 0.707,
+      "step": 71213
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.020458487397467e-07,
+      "loss": 0.6035,
+      "step": 71214
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.020194351869003e-07,
+      "loss": 0.6562,
+      "step": 71215
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.019930218622035e-07,
+      "loss": 0.4878,
+      "step": 71216
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.019666087656773e-07,
+      "loss": 0.6597,
+      "step": 71217
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.019401958973414e-07,
+      "loss": 0.604,
+      "step": 71218
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.019137832572166e-07,
+      "loss": 0.7021,
+      "step": 71219
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.018873708453226e-07,
+      "loss": 0.6201,
+      "step": 71220
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.018609586616793e-07,
+      "loss": 0.6802,
+      "step": 71221
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.018345467063077e-07,
+      "loss": 0.5801,
+      "step": 71222
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.018081349792274e-07,
+      "loss": 0.5596,
+      "step": 71223
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.017817234804592e-07,
+      "loss": 0.5957,
+      "step": 71224
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.017553122100227e-07,
+      "loss": 0.4199,
+      "step": 71225
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.017289011679388e-07,
+      "loss": 0.54,
+      "step": 71226
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.01702490354227e-07,
+      "loss": 0.678,
+      "step": 71227
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.016760797689084e-07,
+      "loss": 0.6709,
+      "step": 71228
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.016496694120024e-07,
+      "loss": 0.835,
+      "step": 71229
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.016232592835296e-07,
+      "loss": 0.7686,
+      "step": 71230
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.0159684938351e-07,
+      "loss": 0.5913,
+      "step": 71231
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.015704397119641e-07,
+      "loss": 0.5073,
+      "step": 71232
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.015440302689118e-07,
+      "loss": 0.7109,
+      "step": 71233
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.015176210543738e-07,
+      "loss": 0.5098,
+      "step": 71234
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.014912120683697e-07,
+      "loss": 0.6318,
+      "step": 71235
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.014648033109203e-07,
+      "loss": 0.6445,
+      "step": 71236
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.014383947820455e-07,
+      "loss": 0.6875,
+      "step": 71237
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.014119864817658e-07,
+      "loss": 0.6113,
+      "step": 71238
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.01385578410101e-07,
+      "loss": 0.7627,
+      "step": 71239
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.013591705670715e-07,
+      "loss": 0.6509,
+      "step": 71240
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.013327629526977e-07,
+      "loss": 0.4802,
+      "step": 71241
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.013063555669993e-07,
+      "loss": 0.6729,
+      "step": 71242
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.012799484099974e-07,
+      "loss": 0.7773,
+      "step": 71243
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.012535414817113e-07,
+      "loss": 0.5537,
+      "step": 71244
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.012271347821619e-07,
+      "loss": 0.5781,
+      "step": 71245
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.012007283113689e-07,
+      "loss": 0.6357,
+      "step": 71246
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.011743220693529e-07,
+      "loss": 0.4514,
+      "step": 71247
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.011479160561344e-07,
+      "loss": 0.5918,
+      "step": 71248
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.011215102717328e-07,
+      "loss": 0.6157,
+      "step": 71249
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.010951047161687e-07,
+      "loss": 0.5571,
+      "step": 71250
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.010686993894624e-07,
+      "loss": 0.4269,
+      "step": 71251
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.010422942916336e-07,
+      "loss": 0.6584,
+      "step": 71252
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.010158894227036e-07,
+      "loss": 0.6445,
+      "step": 71253
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.009894847826916e-07,
+      "loss": 0.6445,
+      "step": 71254
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.009630803716184e-07,
+      "loss": 0.4875,
+      "step": 71255
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.009366761895038e-07,
+      "loss": 0.7017,
+      "step": 71256
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.009102722363685e-07,
+      "loss": 0.834,
+      "step": 71257
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.008838685122327e-07,
+      "loss": 0.8193,
+      "step": 71258
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.008574650171157e-07,
+      "loss": 0.6616,
+      "step": 71259
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.008310617510389e-07,
+      "loss": 0.7002,
+      "step": 71260
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.008046587140218e-07,
+      "loss": 0.6553,
+      "step": 71261
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.007782559060847e-07,
+      "loss": 0.6807,
+      "step": 71262
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.00751853327248e-07,
+      "loss": 0.7285,
+      "step": 71263
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.00725450977532e-07,
+      "loss": 0.7959,
+      "step": 71264
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.006990488569564e-07,
+      "loss": 0.6416,
+      "step": 71265
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.006726469655422e-07,
+      "loss": 0.6182,
+      "step": 71266
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.006462453033089e-07,
+      "loss": 0.6475,
+      "step": 71267
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.006198438702775e-07,
+      "loss": 0.5654,
+      "step": 71268
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.005934426664673e-07,
+      "loss": 0.6953,
+      "step": 71269
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.005670416918991e-07,
+      "loss": 0.5669,
+      "step": 71270
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.005406409465926e-07,
+      "loss": 0.8037,
+      "step": 71271
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.005142404305688e-07,
+      "loss": 0.4171,
+      "step": 71272
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.00487840143847e-07,
+      "loss": 0.6104,
+      "step": 71273
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.004614400864483e-07,
+      "loss": 0.6309,
+      "step": 71274
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.004350402583922e-07,
+      "loss": 0.5942,
+      "step": 71275
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.004086406596994e-07,
+      "loss": 0.5837,
+      "step": 71276
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.003822412903898e-07,
+      "loss": 0.6572,
+      "step": 71277
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.003558421504842e-07,
+      "loss": 0.5996,
+      "step": 71278
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.00329443240002e-07,
+      "loss": 0.6602,
+      "step": 71279
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.003030445589637e-07,
+      "loss": 0.6133,
+      "step": 71280
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.002766461073897e-07,
+      "loss": 0.6924,
+      "step": 71281
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.002502478852998e-07,
+      "loss": 0.7236,
+      "step": 71282
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.002238498927149e-07,
+      "loss": 0.4683,
+      "step": 71283
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.001974521296545e-07,
+      "loss": 0.5046,
+      "step": 71284
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.001710545961394e-07,
+      "loss": 0.4714,
+      "step": 71285
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.001446572921894e-07,
+      "loss": 0.5669,
+      "step": 71286
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.001182602178247e-07,
+      "loss": 0.7656,
+      "step": 71287
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.000918633730661e-07,
+      "loss": 0.6416,
+      "step": 71288
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.000654667579332e-07,
+      "loss": 0.6064,
+      "step": 71289
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.000390703724461e-07,
+      "loss": 0.582,
+      "step": 71290
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 7.000126742166256e-07,
+      "loss": 0.6621,
+      "step": 71291
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.999862782904913e-07,
+      "loss": 0.7568,
+      "step": 71292
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.99959882594064e-07,
+      "loss": 0.627,
+      "step": 71293
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.999334871273634e-07,
+      "loss": 0.7607,
+      "step": 71294
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.9990709189041e-07,
+      "loss": 0.5265,
+      "step": 71295
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.998806968832238e-07,
+      "loss": 0.7207,
+      "step": 71296
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.998543021058254e-07,
+      "loss": 0.6299,
+      "step": 71297
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.998279075582349e-07,
+      "loss": 0.6455,
+      "step": 71298
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.998015132404721e-07,
+      "loss": 0.6445,
+      "step": 71299
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.997751191525576e-07,
+      "loss": 0.416,
+      "step": 71300
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.997487252945113e-07,
+      "loss": 0.6816,
+      "step": 71301
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.997223316663539e-07,
+      "loss": 0.6963,
+      "step": 71302
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.996959382681051e-07,
+      "loss": 0.5554,
+      "step": 71303
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.996695450997851e-07,
+      "loss": 0.5681,
+      "step": 71304
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.996431521614147e-07,
+      "loss": 0.604,
+      "step": 71305
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.996167594530134e-07,
+      "loss": 0.6201,
+      "step": 71306
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.99590366974602e-07,
+      "loss": 0.7451,
+      "step": 71307
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.995639747262005e-07,
+      "loss": 0.708,
+      "step": 71308
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.995375827078289e-07,
+      "loss": 0.7134,
+      "step": 71309
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.995111909195076e-07,
+      "loss": 0.5872,
+      "step": 71310
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.994847993612566e-07,
+      "loss": 0.5498,
+      "step": 71311
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.994584080330965e-07,
+      "loss": 0.7568,
+      "step": 71312
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.99432016935047e-07,
+      "loss": 0.6177,
+      "step": 71313
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.994056260671288e-07,
+      "loss": 0.6758,
+      "step": 71314
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.993792354293617e-07,
+      "loss": 0.5571,
+      "step": 71315
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.993528450217663e-07,
+      "loss": 0.7422,
+      "step": 71316
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.993264548443625e-07,
+      "loss": 0.6172,
+      "step": 71317
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.99300064897171e-07,
+      "loss": 0.5469,
+      "step": 71318
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.992736751802113e-07,
+      "loss": 0.5737,
+      "step": 71319
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.992472856935039e-07,
+      "loss": 0.5448,
+      "step": 71320
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.992208964370689e-07,
+      "loss": 0.6592,
+      "step": 71321
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.991945074109268e-07,
+      "loss": 0.6187,
+      "step": 71322
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.991681186150974e-07,
+      "loss": 0.5752,
+      "step": 71323
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.991417300496014e-07,
+      "loss": 0.6841,
+      "step": 71324
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.991153417144585e-07,
+      "loss": 0.5391,
+      "step": 71325
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.990889536096893e-07,
+      "loss": 0.6753,
+      "step": 71326
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.990625657353137e-07,
+      "loss": 0.6924,
+      "step": 71327
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.990361780913527e-07,
+      "loss": 0.5813,
+      "step": 71328
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.990097906778253e-07,
+      "loss": 0.6006,
+      "step": 71329
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.989834034947523e-07,
+      "loss": 0.4529,
+      "step": 71330
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.989570165421538e-07,
+      "loss": 0.7461,
+      "step": 71331
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.989306298200501e-07,
+      "loss": 0.6885,
+      "step": 71332
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.989042433284615e-07,
+      "loss": 0.7422,
+      "step": 71333
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.988778570674079e-07,
+      "loss": 0.5369,
+      "step": 71334
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.9885147103691e-07,
+      "loss": 0.7139,
+      "step": 71335
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.988250852369872e-07,
+      "loss": 0.7725,
+      "step": 71336
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.987986996676606e-07,
+      "loss": 0.5674,
+      "step": 71337
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.987723143289501e-07,
+      "loss": 0.5283,
+      "step": 71338
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.987459292208757e-07,
+      "loss": 0.7275,
+      "step": 71339
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.987195443434574e-07,
+      "loss": 0.6641,
+      "step": 71340
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.986931596967159e-07,
+      "loss": 0.5332,
+      "step": 71341
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.98666775280671e-07,
+      "loss": 0.6548,
+      "step": 71342
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.986403910953431e-07,
+      "loss": 0.6982,
+      "step": 71343
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.986140071407525e-07,
+      "loss": 0.6797,
+      "step": 71344
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.985876234169193e-07,
+      "loss": 0.6777,
+      "step": 71345
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.985612399238636e-07,
+      "loss": 0.5759,
+      "step": 71346
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.985348566616059e-07,
+      "loss": 0.4485,
+      "step": 71347
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.985084736301662e-07,
+      "loss": 0.6455,
+      "step": 71348
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.984820908295646e-07,
+      "loss": 0.5151,
+      "step": 71349
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.984557082598214e-07,
+      "loss": 0.7109,
+      "step": 71350
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.984293259209567e-07,
+      "loss": 0.5723,
+      "step": 71351
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.98402943812991e-07,
+      "loss": 0.5381,
+      "step": 71352
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.98376561935944e-07,
+      "loss": 0.6064,
+      "step": 71353
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.983501802898365e-07,
+      "loss": 0.6016,
+      "step": 71354
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.983237988746881e-07,
+      "loss": 0.6953,
+      "step": 71355
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.982974176905197e-07,
+      "loss": 0.7036,
+      "step": 71356
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.982710367373506e-07,
+      "loss": 0.4917,
+      "step": 71357
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.982446560152022e-07,
+      "loss": 0.6615,
+      "step": 71358
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.982182755240935e-07,
+      "loss": 0.6558,
+      "step": 71359
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.981918952640453e-07,
+      "loss": 0.5249,
+      "step": 71360
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.981655152350776e-07,
+      "loss": 0.6631,
+      "step": 71361
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.981391354372107e-07,
+      "loss": 0.8154,
+      "step": 71362
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.981127558704647e-07,
+      "loss": 0.7227,
+      "step": 71363
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.980863765348602e-07,
+      "loss": 0.5018,
+      "step": 71364
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.980599974304167e-07,
+      "loss": 0.6514,
+      "step": 71365
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.980336185571549e-07,
+      "loss": 0.6519,
+      "step": 71366
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.980072399150948e-07,
+      "loss": 0.6514,
+      "step": 71367
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.979808615042571e-07,
+      "loss": 0.7549,
+      "step": 71368
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.979544833246614e-07,
+      "loss": 0.8271,
+      "step": 71369
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.979281053763278e-07,
+      "loss": 0.6226,
+      "step": 71370
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.979017276592768e-07,
+      "loss": 0.5044,
+      "step": 71371
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.978753501735285e-07,
+      "loss": 0.5356,
+      "step": 71372
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.978489729191034e-07,
+      "loss": 0.6592,
+      "step": 71373
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.978225958960211e-07,
+      "loss": 0.7285,
+      "step": 71374
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.977962191043025e-07,
+      "loss": 0.6094,
+      "step": 71375
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.977698425439674e-07,
+      "loss": 0.6201,
+      "step": 71376
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.977434662150359e-07,
+      "loss": 0.7158,
+      "step": 71377
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.977170901175286e-07,
+      "loss": 0.5798,
+      "step": 71378
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.976907142514652e-07,
+      "loss": 0.7598,
+      "step": 71379
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.976643386168662e-07,
+      "loss": 0.5527,
+      "step": 71380
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.976379632137516e-07,
+      "loss": 0.7705,
+      "step": 71381
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.976115880421416e-07,
+      "loss": 0.5798,
+      "step": 71382
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.975852131020567e-07,
+      "loss": 0.6597,
+      "step": 71383
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.975588383935167e-07,
+      "loss": 0.7412,
+      "step": 71384
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.975324639165423e-07,
+      "loss": 0.6724,
+      "step": 71385
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.975060896711531e-07,
+      "loss": 0.6416,
+      "step": 71386
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.974797156573698e-07,
+      "loss": 0.5464,
+      "step": 71387
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.974533418752124e-07,
+      "loss": 0.6602,
+      "step": 71388
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.97426968324701e-07,
+      "loss": 0.4913,
+      "step": 71389
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.974005950058557e-07,
+      "loss": 0.8086,
+      "step": 71390
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.973742219186969e-07,
+      "loss": 0.8945,
+      "step": 71391
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.973478490632447e-07,
+      "loss": 0.5349,
+      "step": 71392
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.973214764395196e-07,
+      "loss": 0.4771,
+      "step": 71393
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.972951040475412e-07,
+      "loss": 0.5471,
+      "step": 71394
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.972687318873303e-07,
+      "loss": 0.6045,
+      "step": 71395
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.972423599589066e-07,
+      "loss": 0.6426,
+      "step": 71396
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.972159882622906e-07,
+      "loss": 0.4121,
+      "step": 71397
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.971896167975027e-07,
+      "loss": 0.5854,
+      "step": 71398
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.971632455645623e-07,
+      "loss": 0.7295,
+      "step": 71399
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.971368745634903e-07,
+      "loss": 0.4087,
+      "step": 71400
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.971105037943065e-07,
+      "loss": 0.7363,
+      "step": 71401
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.970841332570314e-07,
+      "loss": 0.7383,
+      "step": 71402
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.97057762951685e-07,
+      "loss": 0.6533,
+      "step": 71403
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.970313928782876e-07,
+      "loss": 0.5723,
+      "step": 71404
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.970050230368591e-07,
+      "loss": 0.6885,
+      "step": 71405
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.969786534274203e-07,
+      "loss": 0.7246,
+      "step": 71406
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.969522840499906e-07,
+      "loss": 0.7588,
+      "step": 71407
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.969259149045912e-07,
+      "loss": 0.7896,
+      "step": 71408
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.968995459912415e-07,
+      "loss": 0.707,
+      "step": 71409
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.968731773099616e-07,
+      "loss": 0.5547,
+      "step": 71410
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.968468088607719e-07,
+      "loss": 0.5688,
+      "step": 71411
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.96820440643693e-07,
+      "loss": 0.6531,
+      "step": 71412
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.967940726587444e-07,
+      "loss": 0.5796,
+      "step": 71413
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.967677049059469e-07,
+      "loss": 0.6699,
+      "step": 71414
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.967413373853202e-07,
+      "loss": 0.5527,
+      "step": 71415
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.96714970096885e-07,
+      "loss": 0.6243,
+      "step": 71416
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.966886030406609e-07,
+      "loss": 0.5439,
+      "step": 71417
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.966622362166689e-07,
+      "loss": 0.6553,
+      "step": 71418
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.966358696249283e-07,
+      "loss": 0.8652,
+      "step": 71419
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.966095032654597e-07,
+      "loss": 0.5864,
+      "step": 71420
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.965831371382833e-07,
+      "loss": 0.7119,
+      "step": 71421
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.965567712434191e-07,
+      "loss": 0.707,
+      "step": 71422
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.965304055808876e-07,
+      "loss": 0.6807,
+      "step": 71423
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.965040401507085e-07,
+      "loss": 0.541,
+      "step": 71424
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.964776749529027e-07,
+      "loss": 0.7363,
+      "step": 71425
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.964513099874895e-07,
+      "loss": 0.509,
+      "step": 71426
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.964249452544902e-07,
+      "loss": 0.7207,
+      "step": 71427
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.963985807539242e-07,
+      "loss": 0.6855,
+      "step": 71428
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.963722164858117e-07,
+      "loss": 0.5127,
+      "step": 71429
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.963458524501731e-07,
+      "loss": 0.525,
+      "step": 71430
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.963194886470285e-07,
+      "loss": 0.5015,
+      "step": 71431
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.962931250763979e-07,
+      "loss": 0.3411,
+      "step": 71432
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.962667617383019e-07,
+      "loss": 0.7471,
+      "step": 71433
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.962403986327603e-07,
+      "loss": 0.6895,
+      "step": 71434
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.962140357597936e-07,
+      "loss": 0.6211,
+      "step": 71435
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.961876731194218e-07,
+      "loss": 0.5928,
+      "step": 71436
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.961613107116652e-07,
+      "loss": 0.5063,
+      "step": 71437
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.96134948536544e-07,
+      "loss": 0.5654,
+      "step": 71438
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.961085865940781e-07,
+      "loss": 0.7637,
+      "step": 71439
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.960822248842881e-07,
+      "loss": 0.6328,
+      "step": 71440
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.960558634071936e-07,
+      "loss": 0.6074,
+      "step": 71441
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.960295021628156e-07,
+      "loss": 0.4888,
+      "step": 71442
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.960031411511733e-07,
+      "loss": 0.6089,
+      "step": 71443
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.959767803722878e-07,
+      "loss": 0.5073,
+      "step": 71444
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.959504198261787e-07,
+      "loss": 0.6553,
+      "step": 71445
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.959240595128666e-07,
+      "loss": 0.6067,
+      "step": 71446
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.958976994323712e-07,
+      "loss": 0.7148,
+      "step": 71447
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.958713395847136e-07,
+      "loss": 0.5781,
+      "step": 71448
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.958449799699126e-07,
+      "loss": 0.5049,
+      "step": 71449
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.958186205879896e-07,
+      "loss": 0.5991,
+      "step": 71450
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.957922614389639e-07,
+      "loss": 0.7754,
+      "step": 71451
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.957659025228562e-07,
+      "loss": 0.5334,
+      "step": 71452
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.957395438396865e-07,
+      "loss": 0.6377,
+      "step": 71453
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.957131853894752e-07,
+      "loss": 0.8252,
+      "step": 71454
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.956868271722422e-07,
+      "loss": 0.5801,
+      "step": 71455
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.956604691880079e-07,
+      "loss": 0.623,
+      "step": 71456
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.956341114367922e-07,
+      "loss": 0.5877,
+      "step": 71457
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.956077539186158e-07,
+      "loss": 0.6865,
+      "step": 71458
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.955813966334985e-07,
+      "loss": 0.4045,
+      "step": 71459
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.955550395814602e-07,
+      "loss": 0.624,
+      "step": 71460
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.955286827625217e-07,
+      "loss": 0.7861,
+      "step": 71461
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.955023261767026e-07,
+      "loss": 0.668,
+      "step": 71462
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.954759698240236e-07,
+      "loss": 0.7422,
+      "step": 71463
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.954496137045045e-07,
+      "loss": 0.5449,
+      "step": 71464
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.954232578181658e-07,
+      "loss": 0.7314,
+      "step": 71465
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.953969021650275e-07,
+      "loss": 0.8271,
+      "step": 71466
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.953705467451096e-07,
+      "loss": 0.5117,
+      "step": 71467
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.953441915584329e-07,
+      "loss": 0.6099,
+      "step": 71468
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.953178366050168e-07,
+      "loss": 0.6216,
+      "step": 71469
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.952914818848817e-07,
+      "loss": 0.6182,
+      "step": 71470
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.952651273980481e-07,
+      "loss": 0.5728,
+      "step": 71471
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.95238773144536e-07,
+      "loss": 0.7021,
+      "step": 71472
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.952124191243654e-07,
+      "loss": 0.7832,
+      "step": 71473
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.951860653375565e-07,
+      "loss": 0.6099,
+      "step": 71474
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.951597117841301e-07,
+      "loss": 0.4258,
+      "step": 71475
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.951333584641055e-07,
+      "loss": 0.6538,
+      "step": 71476
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.951070053775034e-07,
+      "loss": 0.5435,
+      "step": 71477
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.950806525243441e-07,
+      "loss": 0.5645,
+      "step": 71478
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.950542999046471e-07,
+      "loss": 0.4731,
+      "step": 71479
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.950279475184333e-07,
+      "loss": 0.5698,
+      "step": 71480
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.950015953657222e-07,
+      "loss": 0.6455,
+      "step": 71481
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.949752434465348e-07,
+      "loss": 0.6357,
+      "step": 71482
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.949488917608907e-07,
+      "loss": 0.5786,
+      "step": 71483
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.949225403088099e-07,
+      "loss": 0.5737,
+      "step": 71484
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.948961890903132e-07,
+      "loss": 0.6738,
+      "step": 71485
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.948698381054203e-07,
+      "loss": 0.5625,
+      "step": 71486
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.948434873541517e-07,
+      "loss": 0.5449,
+      "step": 71487
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.948171368365275e-07,
+      "loss": 0.6494,
+      "step": 71488
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.947907865525675e-07,
+      "loss": 0.6328,
+      "step": 71489
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.947644365022924e-07,
+      "loss": 0.5889,
+      "step": 71490
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.947380866857219e-07,
+      "loss": 0.4119,
+      "step": 71491
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.947117371028766e-07,
+      "loss": 0.7432,
+      "step": 71492
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.946853877537761e-07,
+      "loss": 0.6553,
+      "step": 71493
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.946590386384414e-07,
+      "loss": 0.5332,
+      "step": 71494
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.946326897568919e-07,
+      "loss": 0.7832,
+      "step": 71495
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.946063411091485e-07,
+      "loss": 0.6274,
+      "step": 71496
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.945799926952306e-07,
+      "loss": 0.5757,
+      "step": 71497
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.945536445151592e-07,
+      "loss": 0.5859,
+      "step": 71498
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.945272965689539e-07,
+      "loss": 0.5925,
+      "step": 71499
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.945009488566349e-07,
+      "loss": 0.6316,
+      "step": 71500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.944746013782223e-07,
+      "loss": 0.7383,
+      "step": 71501
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.944482541337366e-07,
+      "loss": 0.5508,
+      "step": 71502
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.944219071231978e-07,
+      "loss": 0.7725,
+      "step": 71503
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.943955603466263e-07,
+      "loss": 0.6885,
+      "step": 71504
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.943692138040416e-07,
+      "loss": 0.6519,
+      "step": 71505
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.943428674954648e-07,
+      "loss": 0.5635,
+      "step": 71506
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.943165214209152e-07,
+      "loss": 0.4011,
+      "step": 71507
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.942901755804139e-07,
+      "loss": 0.7314,
+      "step": 71508
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.942638299739804e-07,
+      "loss": 0.5334,
+      "step": 71509
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.942374846016348e-07,
+      "loss": 0.79,
+      "step": 71510
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.942111394633976e-07,
+      "loss": 0.7222,
+      "step": 71511
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.941847945592887e-07,
+      "loss": 0.7295,
+      "step": 71512
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.941584498893287e-07,
+      "loss": 0.5952,
+      "step": 71513
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.941321054535374e-07,
+      "loss": 0.4471,
+      "step": 71514
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.941057612519351e-07,
+      "loss": 0.7178,
+      "step": 71515
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.940794172845419e-07,
+      "loss": 0.5361,
+      "step": 71516
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.940530735513783e-07,
+      "loss": 0.7383,
+      "step": 71517
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.940267300524641e-07,
+      "loss": 0.606,
+      "step": 71518
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.940003867878195e-07,
+      "loss": 0.55,
+      "step": 71519
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.939740437574645e-07,
+      "loss": 0.6479,
+      "step": 71520
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.939477009614197e-07,
+      "loss": 0.707,
+      "step": 71521
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.93921358399705e-07,
+      "loss": 0.7188,
+      "step": 71522
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.938950160723407e-07,
+      "loss": 0.5732,
+      "step": 71523
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.938686739793468e-07,
+      "loss": 0.522,
+      "step": 71524
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.938423321207439e-07,
+      "loss": 0.5759,
+      "step": 71525
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.938159904965514e-07,
+      "loss": 0.709,
+      "step": 71526
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.937896491067904e-07,
+      "loss": 0.6816,
+      "step": 71527
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.937633079514804e-07,
+      "loss": 0.5195,
+      "step": 71528
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.937369670306416e-07,
+      "loss": 0.6499,
+      "step": 71529
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.937106263442945e-07,
+      "loss": 0.7803,
+      "step": 71530
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.936842858924588e-07,
+      "loss": 0.5146,
+      "step": 71531
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.936579456751552e-07,
+      "loss": 0.6982,
+      "step": 71532
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.936316056924034e-07,
+      "loss": 0.6245,
+      "step": 71533
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.936052659442242e-07,
+      "loss": 0.5984,
+      "step": 71534
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.935789264306369e-07,
+      "loss": 0.6416,
+      "step": 71535
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.935525871516625e-07,
+      "loss": 0.5127,
+      "step": 71536
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.93526248107321e-07,
+      "loss": 0.5322,
+      "step": 71537
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.934999092976319e-07,
+      "loss": 0.6709,
+      "step": 71538
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.934735707226159e-07,
+      "loss": 0.5894,
+      "step": 71539
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.934472323822932e-07,
+      "loss": 0.582,
+      "step": 71540
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.934208942766837e-07,
+      "loss": 0.6797,
+      "step": 71541
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.933945564058079e-07,
+      "loss": 0.6289,
+      "step": 71542
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.933682187696855e-07,
+      "loss": 0.6289,
+      "step": 71543
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.933418813683372e-07,
+      "loss": 0.8223,
+      "step": 71544
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.933155442017829e-07,
+      "loss": 0.6191,
+      "step": 71545
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.932892072700427e-07,
+      "loss": 0.5466,
+      "step": 71546
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.932628705731373e-07,
+      "loss": 0.5249,
+      "step": 71547
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.932365341110859e-07,
+      "loss": 0.7681,
+      "step": 71548
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.932101978839093e-07,
+      "loss": 0.7891,
+      "step": 71549
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.931838618916274e-07,
+      "loss": 0.5884,
+      "step": 71550
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.931575261342608e-07,
+      "loss": 0.7783,
+      "step": 71551
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.931311906118291e-07,
+      "loss": 0.3604,
+      "step": 71552
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.931048553243531e-07,
+      "loss": 0.8389,
+      "step": 71553
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.930785202718521e-07,
+      "loss": 0.6025,
+      "step": 71554
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.930521854543472e-07,
+      "loss": 0.5591,
+      "step": 71555
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.93025850871858e-07,
+      "loss": 0.6968,
+      "step": 71556
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.92999516524405e-07,
+      "loss": 0.6133,
+      "step": 71557
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.929731824120077e-07,
+      "loss": 0.5728,
+      "step": 71558
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.929468485346871e-07,
+      "loss": 0.4868,
+      "step": 71559
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.929205148924628e-07,
+      "loss": 0.5635,
+      "step": 71560
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.928941814853551e-07,
+      "loss": 0.5298,
+      "step": 71561
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.928678483133842e-07,
+      "loss": 0.6196,
+      "step": 71562
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.928415153765703e-07,
+      "loss": 0.4097,
+      "step": 71563
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.928151826749335e-07,
+      "loss": 0.6797,
+      "step": 71564
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.92788850208494e-07,
+      "loss": 0.5328,
+      "step": 71565
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.92762517977272e-07,
+      "loss": 0.6055,
+      "step": 71566
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.927361859812879e-07,
+      "loss": 0.6177,
+      "step": 71567
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.927098542205614e-07,
+      "loss": 0.7095,
+      "step": 71568
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.926835226951124e-07,
+      "loss": 0.6533,
+      "step": 71569
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.926571914049619e-07,
+      "loss": 0.7354,
+      "step": 71570
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.926308603501295e-07,
+      "loss": 0.6309,
+      "step": 71571
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.926045295306357e-07,
+      "loss": 0.4572,
+      "step": 71572
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.925781989465005e-07,
+      "loss": 0.7354,
+      "step": 71573
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.925518685977436e-07,
+      "loss": 0.624,
+      "step": 71574
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.92525538484386e-07,
+      "loss": 0.5889,
+      "step": 71575
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.924992086064472e-07,
+      "loss": 0.6377,
+      "step": 71576
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.924728789639481e-07,
+      "loss": 0.6904,
+      "step": 71577
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.924465495569082e-07,
+      "loss": 0.6396,
+      "step": 71578
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.924202203853475e-07,
+      "loss": 0.6216,
+      "step": 71579
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.923938914492868e-07,
+      "loss": 0.4771,
+      "step": 71580
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.923675627487457e-07,
+      "loss": 0.7432,
+      "step": 71581
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.923412342837448e-07,
+      "loss": 0.7227,
+      "step": 71582
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.923149060543039e-07,
+      "loss": 0.6963,
+      "step": 71583
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.922885780604435e-07,
+      "loss": 0.5908,
+      "step": 71584
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.922622503021833e-07,
+      "loss": 0.4656,
+      "step": 71585
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.922359227795441e-07,
+      "loss": 0.6636,
+      "step": 71586
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.922095954925458e-07,
+      "loss": 0.4858,
+      "step": 71587
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.921832684412081e-07,
+      "loss": 0.5476,
+      "step": 71588
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.921569416255518e-07,
+      "loss": 0.4795,
+      "step": 71589
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.921306150455967e-07,
+      "loss": 0.5212,
+      "step": 71590
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.921042887013629e-07,
+      "loss": 0.6777,
+      "step": 71591
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.920779625928708e-07,
+      "loss": 0.5881,
+      "step": 71592
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.920516367201402e-07,
+      "loss": 0.5957,
+      "step": 71593
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 6.920253110831917e-07,
+      "loss": 0.6675,
+      "step": 71594
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.919989856820451e-07,
+      "loss": 0.6709,
+      "step": 71595
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.91972660516721e-07,
+      "loss": 0.6777,
+      "step": 71596
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.919463355872395e-07,
+      "loss": 0.6738,
+      "step": 71597
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.9192001089362e-07,
+      "loss": 0.7354,
+      "step": 71598
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.918936864358835e-07,
+      "loss": 0.4158,
+      "step": 71599
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.918673622140495e-07,
+      "loss": 0.6177,
+      "step": 71600
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.918410382281388e-07,
+      "loss": 0.5645,
+      "step": 71601
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.91814714478171e-07,
+      "loss": 0.7275,
+      "step": 71602
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.917883909641666e-07,
+      "loss": 0.6572,
+      "step": 71603
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.917620676861455e-07,
+      "loss": 0.3787,
+      "step": 71604
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.917357446441284e-07,
+      "loss": 0.6123,
+      "step": 71605
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.917094218381347e-07,
+      "loss": 0.543,
+      "step": 71606
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.916830992681853e-07,
+      "loss": 0.4561,
+      "step": 71607
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.916567769342999e-07,
+      "loss": 0.6709,
+      "step": 71608
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.916304548364986e-07,
+      "loss": 0.5474,
+      "step": 71609
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.916041329748015e-07,
+      "loss": 0.5332,
+      "step": 71610
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.915778113492292e-07,
+      "loss": 0.5737,
+      "step": 71611
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.915514899598014e-07,
+      "loss": 0.6963,
+      "step": 71612
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.915251688065386e-07,
+      "loss": 0.4854,
+      "step": 71613
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.914988478894605e-07,
+      "loss": 0.5706,
+      "step": 71614
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.914725272085879e-07,
+      "loss": 0.5176,
+      "step": 71615
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.914462067639404e-07,
+      "loss": 0.6943,
+      "step": 71616
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.914198865555387e-07,
+      "loss": 0.6045,
+      "step": 71617
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.913935665834024e-07,
+      "loss": 0.605,
+      "step": 71618
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.913672468475517e-07,
+      "loss": 0.7373,
+      "step": 71619
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.913409273480071e-07,
+      "loss": 0.6484,
+      "step": 71620
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.913146080847882e-07,
+      "loss": 0.667,
+      "step": 71621
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.912882890579159e-07,
+      "loss": 0.5913,
+      "step": 71622
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.912619702674096e-07,
+      "loss": 0.834,
+      "step": 71623
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.912356517132902e-07,
+      "loss": 0.5532,
+      "step": 71624
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.912093333955771e-07,
+      "loss": 0.6729,
+      "step": 71625
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.911830153142911e-07,
+      "loss": 0.6821,
+      "step": 71626
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.911566974694522e-07,
+      "loss": 0.5322,
+      "step": 71627
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.911303798610804e-07,
+      "loss": 0.6489,
+      "step": 71628
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.911040624891955e-07,
+      "loss": 0.6426,
+      "step": 71629
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.910777453538183e-07,
+      "loss": 0.6514,
+      "step": 71630
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.910514284549683e-07,
+      "loss": 0.728,
+      "step": 71631
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.910251117926664e-07,
+      "loss": 0.4597,
+      "step": 71632
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.90998795366932e-07,
+      "loss": 0.6606,
+      "step": 71633
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.90972479177786e-07,
+      "loss": 0.4585,
+      "step": 71634
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.909461632252477e-07,
+      "loss": 0.7402,
+      "step": 71635
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.909198475093383e-07,
+      "loss": 0.6851,
+      "step": 71636
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.908935320300773e-07,
+      "loss": 0.5247,
+      "step": 71637
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.908672167874845e-07,
+      "loss": 0.6333,
+      "step": 71638
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.908409017815809e-07,
+      "loss": 0.339,
+      "step": 71639
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.908145870123857e-07,
+      "loss": 0.6499,
+      "step": 71640
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.907882724799199e-07,
+      "loss": 0.4492,
+      "step": 71641
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.907619581842031e-07,
+      "loss": 0.5601,
+      "step": 71642
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.907356441252558e-07,
+      "loss": 0.6611,
+      "step": 71643
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.907093303030978e-07,
+      "loss": 0.79,
+      "step": 71644
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.906830167177497e-07,
+      "loss": 0.7456,
+      "step": 71645
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.906567033692313e-07,
+      "loss": 0.5903,
+      "step": 71646
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.90630390257563e-07,
+      "loss": 0.625,
+      "step": 71647
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.906040773827646e-07,
+      "loss": 0.7148,
+      "step": 71648
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.905777647448565e-07,
+      "loss": 0.6328,
+      "step": 71649
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.905514523438586e-07,
+      "loss": 0.8652,
+      "step": 71650
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.905251401797913e-07,
+      "loss": 0.8994,
+      "step": 71651
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.904988282526747e-07,
+      "loss": 0.4536,
+      "step": 71652
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.904725165625289e-07,
+      "loss": 0.5425,
+      "step": 71653
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.90446205109374e-07,
+      "loss": 0.4907,
+      "step": 71654
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.904198938932304e-07,
+      "loss": 0.6255,
+      "step": 71655
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.903935829141179e-07,
+      "loss": 0.5182,
+      "step": 71656
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.90367272172057e-07,
+      "loss": 0.5249,
+      "step": 71657
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.903409616670677e-07,
+      "loss": 0.7881,
+      "step": 71658
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.903146513991697e-07,
+      "loss": 0.6792,
+      "step": 71659
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.902883413683838e-07,
+      "loss": 0.7617,
+      "step": 71660
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.902620315747296e-07,
+      "loss": 0.7598,
+      "step": 71661
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.902357220182278e-07,
+      "loss": 0.6108,
+      "step": 71662
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.902094126988982e-07,
+      "loss": 0.6924,
+      "step": 71663
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.901831036167609e-07,
+      "loss": 0.6133,
+      "step": 71664
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.901567947718363e-07,
+      "loss": 0.5557,
+      "step": 71665
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.901304861641442e-07,
+      "loss": 0.6592,
+      "step": 71666
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.901041777937053e-07,
+      "loss": 0.5552,
+      "step": 71667
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.900778696605392e-07,
+      "loss": 0.6465,
+      "step": 71668
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.900515617646661e-07,
+      "loss": 0.6431,
+      "step": 71669
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.900252541061064e-07,
+      "loss": 0.603,
+      "step": 71670
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.899989466848799e-07,
+      "loss": 0.8018,
+      "step": 71671
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.899726395010072e-07,
+      "loss": 0.6455,
+      "step": 71672
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.899463325545079e-07,
+      "loss": 0.5479,
+      "step": 71673
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.899200258454028e-07,
+      "loss": 0.6621,
+      "step": 71674
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.898937193737113e-07,
+      "loss": 0.7568,
+      "step": 71675
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.898674131394541e-07,
+      "loss": 0.6797,
+      "step": 71676
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.898411071426516e-07,
+      "loss": 0.6426,
+      "step": 71677
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.898148013833229e-07,
+      "loss": 0.5869,
+      "step": 71678
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.89788495861489e-07,
+      "loss": 0.6621,
+      "step": 71679
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.897621905771695e-07,
+      "loss": 0.7373,
+      "step": 71680
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.897358855303852e-07,
+      "loss": 0.5784,
+      "step": 71681
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.89709580721156e-07,
+      "loss": 0.4042,
+      "step": 71682
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.896832761495013e-07,
+      "loss": 0.7178,
+      "step": 71683
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.896569718154422e-07,
+      "loss": 0.7607,
+      "step": 71684
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.896306677189985e-07,
+      "loss": 0.6738,
+      "step": 71685
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.896043638601903e-07,
+      "loss": 0.4578,
+      "step": 71686
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.89578060239038e-07,
+      "loss": 0.7749,
+      "step": 71687
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.895517568555611e-07,
+      "loss": 0.6855,
+      "step": 71688
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.895254537097803e-07,
+      "loss": 0.6504,
+      "step": 71689
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.894991508017155e-07,
+      "loss": 0.6797,
+      "step": 71690
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.894728481313871e-07,
+      "loss": 0.4255,
+      "step": 71691
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.894465456988148e-07,
+      "loss": 0.3347,
+      "step": 71692
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.894202435040193e-07,
+      "loss": 0.6973,
+      "step": 71693
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.893939415470201e-07,
+      "loss": 0.5918,
+      "step": 71694
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.89367639827838e-07,
+      "loss": 0.6426,
+      "step": 71695
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.893413383464925e-07,
+      "loss": 0.6631,
+      "step": 71696
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.893150371030045e-07,
+      "loss": 0.709,
+      "step": 71697
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.892887360973936e-07,
+      "loss": 0.5439,
+      "step": 71698
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.892624353296798e-07,
+      "loss": 0.7061,
+      "step": 71699
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.892361347998835e-07,
+      "loss": 0.7051,
+      "step": 71700
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.892098345080249e-07,
+      "loss": 0.6356,
+      "step": 71701
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.891835344541239e-07,
+      "loss": 0.5403,
+      "step": 71702
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.891572346382009e-07,
+      "loss": 0.6104,
+      "step": 71703
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.891309350602758e-07,
+      "loss": 0.6172,
+      "step": 71704
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.891046357203689e-07,
+      "loss": 0.7451,
+      "step": 71705
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.890783366185003e-07,
+      "loss": 0.624,
+      "step": 71706
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.890520377546902e-07,
+      "loss": 0.7051,
+      "step": 71707
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.890257391289586e-07,
+      "loss": 0.5664,
+      "step": 71708
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.889994407413256e-07,
+      "loss": 0.6689,
+      "step": 71709
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.889731425918117e-07,
+      "loss": 0.5796,
+      "step": 71710
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.889468446804363e-07,
+      "loss": 0.6914,
+      "step": 71711
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.889205470072204e-07,
+      "loss": 0.8262,
+      "step": 71712
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.888942495721834e-07,
+      "loss": 0.5723,
+      "step": 71713
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.88867952375346e-07,
+      "loss": 0.5183,
+      "step": 71714
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.88841655416728e-07,
+      "loss": 0.604,
+      "step": 71715
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.888153586963498e-07,
+      "loss": 0.7666,
+      "step": 71716
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.887890622142316e-07,
+      "loss": 0.7568,
+      "step": 71717
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.887627659703929e-07,
+      "loss": 0.6924,
+      "step": 71718
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.887364699648542e-07,
+      "loss": 0.7236,
+      "step": 71719
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.88710174197636e-07,
+      "loss": 0.7207,
+      "step": 71720
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.886838786687576e-07,
+      "loss": 0.6719,
+      "step": 71721
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.886575833782401e-07,
+      "loss": 0.666,
+      "step": 71722
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.886312883261029e-07,
+      "loss": 0.6343,
+      "step": 71723
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.886049935123666e-07,
+      "loss": 0.7246,
+      "step": 71724
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.885786989370508e-07,
+      "loss": 0.542,
+      "step": 71725
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.885524046001763e-07,
+      "loss": 0.5798,
+      "step": 71726
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.885261105017633e-07,
+      "loss": 0.5125,
+      "step": 71727
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.884998166418308e-07,
+      "loss": 0.5576,
+      "step": 71728
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.884735230204e-07,
+      "loss": 0.6162,
+      "step": 71729
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.884472296374905e-07,
+      "loss": 0.5359,
+      "step": 71730
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.884209364931229e-07,
+      "loss": 0.5879,
+      "step": 71731
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.883946435873167e-07,
+      "loss": 0.7051,
+      "step": 71732
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.883683509200926e-07,
+      "loss": 0.7031,
+      "step": 71733
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.883420584914706e-07,
+      "loss": 0.6992,
+      "step": 71734
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.883157663014707e-07,
+      "loss": 0.6592,
+      "step": 71735
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.882894743501131e-07,
+      "loss": 0.665,
+      "step": 71736
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.882631826374181e-07,
+      "loss": 0.6309,
+      "step": 71737
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.882368911634053e-07,
+      "loss": 0.5439,
+      "step": 71738
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.882105999280954e-07,
+      "loss": 0.7998,
+      "step": 71739
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.881843089315081e-07,
+      "loss": 0.605,
+      "step": 71740
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.881580181736639e-07,
+      "loss": 0.5684,
+      "step": 71741
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.881317276545826e-07,
+      "loss": 0.5967,
+      "step": 71742
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.881054373742848e-07,
+      "loss": 0.5195,
+      "step": 71743
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.8807914733279e-07,
+      "loss": 0.6963,
+      "step": 71744
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.880528575301188e-07,
+      "loss": 0.7266,
+      "step": 71745
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.880265679662911e-07,
+      "loss": 0.7031,
+      "step": 71746
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.880002786413275e-07,
+      "loss": 0.6426,
+      "step": 71747
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.879739895552475e-07,
+      "loss": 0.6758,
+      "step": 71748
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.879477007080713e-07,
+      "loss": 0.5811,
+      "step": 71749
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.879214120998193e-07,
+      "loss": 0.7842,
+      "step": 71750
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.878951237305115e-07,
+      "loss": 0.5957,
+      "step": 71751
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.878688356001681e-07,
+      "loss": 0.4551,
+      "step": 71752
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.878425477088092e-07,
+      "loss": 0.5715,
+      "step": 71753
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.878162600564548e-07,
+      "loss": 0.7168,
+      "step": 71754
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.877899726431253e-07,
+      "loss": 0.6431,
+      "step": 71755
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.877636854688405e-07,
+      "loss": 0.7354,
+      "step": 71756
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.87737398533621e-07,
+      "loss": 0.668,
+      "step": 71757
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.877111118374863e-07,
+      "loss": 0.6309,
+      "step": 71758
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.876848253804568e-07,
+      "loss": 0.6963,
+      "step": 71759
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.876585391625529e-07,
+      "loss": 0.5991,
+      "step": 71760
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.876322531837943e-07,
+      "loss": 0.5459,
+      "step": 71761
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.876059674442014e-07,
+      "loss": 0.6279,
+      "step": 71762
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.875796819437942e-07,
+      "loss": 0.6992,
+      "step": 71763
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.87553396682593e-07,
+      "loss": 0.6851,
+      "step": 71764
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.875271116606175e-07,
+      "loss": 0.6328,
+      "step": 71765
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.875008268778884e-07,
+      "loss": 0.835,
+      "step": 71766
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.874745423344258e-07,
+      "loss": 0.7256,
+      "step": 71767
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.874482580302491e-07,
+      "loss": 0.6729,
+      "step": 71768
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.874219739653792e-07,
+      "loss": 0.6211,
+      "step": 71769
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.873956901398356e-07,
+      "loss": 0.667,
+      "step": 71770
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.873694065536391e-07,
+      "loss": 0.6479,
+      "step": 71771
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.873431232068092e-07,
+      "loss": 0.6641,
+      "step": 71772
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.873168400993663e-07,
+      "loss": 0.5938,
+      "step": 71773
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.872905572313306e-07,
+      "loss": 0.4219,
+      "step": 71774
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.872642746027221e-07,
+      "loss": 0.6416,
+      "step": 71775
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.872379922135612e-07,
+      "loss": 0.561,
+      "step": 71776
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.872117100638679e-07,
+      "loss": 0.5591,
+      "step": 71777
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.871854281536617e-07,
+      "loss": 0.5962,
+      "step": 71778
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.871591464829635e-07,
+      "loss": 0.5605,
+      "step": 71779
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.871328650517929e-07,
+      "loss": 0.5327,
+      "step": 71780
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.871065838601706e-07,
+      "loss": 0.5542,
+      "step": 71781
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.870803029081162e-07,
+      "loss": 0.7129,
+      "step": 71782
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.870540221956501e-07,
+      "loss": 0.6646,
+      "step": 71783
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.870277417227924e-07,
+      "loss": 0.6392,
+      "step": 71784
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.870014614895631e-07,
+      "loss": 0.5381,
+      "step": 71785
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.869751814959822e-07,
+      "loss": 0.8242,
+      "step": 71786
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.869489017420704e-07,
+      "loss": 0.4761,
+      "step": 71787
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.869226222278475e-07,
+      "loss": 0.5552,
+      "step": 71788
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.868963429533334e-07,
+      "loss": 0.5254,
+      "step": 71789
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.86870063918548e-07,
+      "loss": 0.8135,
+      "step": 71790
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.868437851235123e-07,
+      "loss": 0.7017,
+      "step": 71791
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.868175065682454e-07,
+      "loss": 0.707,
+      "step": 71792
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.867912282527684e-07,
+      "loss": 0.5645,
+      "step": 71793
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.867649501771006e-07,
+      "loss": 0.6216,
+      "step": 71794
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.867386723412627e-07,
+      "loss": 0.4746,
+      "step": 71795
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.867123947452744e-07,
+      "loss": 0.9297,
+      "step": 71796
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.866861173891565e-07,
+      "loss": 0.5928,
+      "step": 71797
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.866598402729283e-07,
+      "loss": 0.4189,
+      "step": 71798
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.866335633966101e-07,
+      "loss": 0.5659,
+      "step": 71799
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.866072867602224e-07,
+      "loss": 0.6392,
+      "step": 71800
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.865810103637849e-07,
+      "loss": 0.6245,
+      "step": 71801
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.865547342073181e-07,
+      "loss": 0.4849,
+      "step": 71802
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.865284582908417e-07,
+      "loss": 0.7383,
+      "step": 71803
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.865021826143762e-07,
+      "loss": 0.4551,
+      "step": 71804
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.864759071779415e-07,
+      "loss": 0.6392,
+      "step": 71805
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.864496319815578e-07,
+      "loss": 0.5684,
+      "step": 71806
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.864233570252455e-07,
+      "loss": 0.8037,
+      "step": 71807
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.863970823090244e-07,
+      "loss": 0.6836,
+      "step": 71808
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.863708078329143e-07,
+      "loss": 0.6992,
+      "step": 71809
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.863445335969356e-07,
+      "loss": 0.3436,
+      "step": 71810
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.863182596011085e-07,
+      "loss": 0.5715,
+      "step": 71811
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.862919858454533e-07,
+      "loss": 0.4187,
+      "step": 71812
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.862657123299897e-07,
+      "loss": 0.6157,
+      "step": 71813
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.862394390547381e-07,
+      "loss": 0.6318,
+      "step": 71814
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.862131660197185e-07,
+      "loss": 0.6514,
+      "step": 71815
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.861868932249513e-07,
+      "loss": 0.8379,
+      "step": 71816
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.861606206704562e-07,
+      "loss": 0.6348,
+      "step": 71817
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.861343483562534e-07,
+      "loss": 0.7617,
+      "step": 71818
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.861080762823633e-07,
+      "loss": 0.8174,
+      "step": 71819
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.860818044488053e-07,
+      "loss": 0.6582,
+      "step": 71820
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.860555328556006e-07,
+      "loss": 0.542,
+      "step": 71821
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.860292615027683e-07,
+      "loss": 0.4926,
+      "step": 71822
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.860029903903293e-07,
+      "loss": 0.6709,
+      "step": 71823
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.859767195183031e-07,
+      "loss": 0.5469,
+      "step": 71824
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.859504488867103e-07,
+      "loss": 0.5522,
+      "step": 71825
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.85924178495571e-07,
+      "loss": 0.4966,
+      "step": 71826
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.85897908344905e-07,
+      "loss": 0.7861,
+      "step": 71827
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.858716384347322e-07,
+      "loss": 0.6689,
+      "step": 71828
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.858453687650734e-07,
+      "loss": 0.6641,
+      "step": 71829
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.85819099335948e-07,
+      "loss": 0.6377,
+      "step": 71830
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.857928301473767e-07,
+      "loss": 0.3616,
+      "step": 71831
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.857665611993793e-07,
+      "loss": 0.584,
+      "step": 71832
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.85740292491976e-07,
+      "loss": 0.7939,
+      "step": 71833
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.857140240251868e-07,
+      "loss": 0.7236,
+      "step": 71834
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.856877557990321e-07,
+      "loss": 0.6367,
+      "step": 71835
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.856614878135321e-07,
+      "loss": 0.5615,
+      "step": 71836
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.856352200687062e-07,
+      "loss": 0.6855,
+      "step": 71837
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.856089525645752e-07,
+      "loss": 0.6094,
+      "step": 71838
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.855826853011587e-07,
+      "loss": 0.563,
+      "step": 71839
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.855564182784774e-07,
+      "loss": 0.6943,
+      "step": 71840
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.855301514965508e-07,
+      "loss": 0.6416,
+      "step": 71841
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.855038849553994e-07,
+      "loss": 0.5991,
+      "step": 71842
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.854776186550431e-07,
+      "loss": 0.6465,
+      "step": 71843
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.854513525955024e-07,
+      "loss": 0.6274,
+      "step": 71844
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.854250867767971e-07,
+      "loss": 0.6528,
+      "step": 71845
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.853988211989475e-07,
+      "loss": 0.5483,
+      "step": 71846
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.853725558619731e-07,
+      "loss": 0.6943,
+      "step": 71847
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.853462907658949e-07,
+      "loss": 0.6348,
+      "step": 71848
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.853200259107321e-07,
+      "loss": 0.7007,
+      "step": 71849
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.852937612965057e-07,
+      "loss": 0.6895,
+      "step": 71850
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.852674969232352e-07,
+      "loss": 0.5518,
+      "step": 71851
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.852412327909409e-07,
+      "loss": 0.5562,
+      "step": 71852
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.852149688996429e-07,
+      "loss": 0.5789,
+      "step": 71853
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.851887052493614e-07,
+      "loss": 0.667,
+      "step": 71854
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.851624418401163e-07,
+      "loss": 0.7812,
+      "step": 71855
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.851361786719283e-07,
+      "loss": 0.6826,
+      "step": 71856
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.851099157448168e-07,
+      "loss": 0.4934,
+      "step": 71857
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.850836530588019e-07,
+      "loss": 0.6172,
+      "step": 71858
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.850573906139042e-07,
+      "loss": 0.397,
+      "step": 71859
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.850311284101434e-07,
+      "loss": 0.7549,
+      "step": 71860
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.8500486644754e-07,
+      "loss": 0.6099,
+      "step": 71861
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.849786047261139e-07,
+      "loss": 0.5889,
+      "step": 71862
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.849523432458848e-07,
+      "loss": 0.5996,
+      "step": 71863
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.849260820068736e-07,
+      "loss": 0.4919,
+      "step": 71864
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.848998210090999e-07,
+      "loss": 0.7305,
+      "step": 71865
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.848735602525842e-07,
+      "loss": 0.6108,
+      "step": 71866
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.84847299737346e-07,
+      "loss": 0.7129,
+      "step": 71867
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.848210394634057e-07,
+      "loss": 0.7119,
+      "step": 71868
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.847947794307836e-07,
+      "loss": 0.7422,
+      "step": 71869
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.847685196394994e-07,
+      "loss": 0.6191,
+      "step": 71870
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.847422600895736e-07,
+      "loss": 0.4531,
+      "step": 71871
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.847160007810261e-07,
+      "loss": 0.5967,
+      "step": 71872
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.846897417138772e-07,
+      "loss": 0.5004,
+      "step": 71873
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.846634828881465e-07,
+      "loss": 0.5166,
+      "step": 71874
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.84637224303855e-07,
+      "loss": 0.5312,
+      "step": 71875
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.846109659610223e-07,
+      "loss": 0.6895,
+      "step": 71876
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.84584707859668e-07,
+      "loss": 0.6572,
+      "step": 71877
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.84558449999813e-07,
+      "loss": 0.624,
+      "step": 71878
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.845321923814771e-07,
+      "loss": 0.7363,
+      "step": 71879
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.845059350046803e-07,
+      "loss": 0.4333,
+      "step": 71880
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.844796778694427e-07,
+      "loss": 0.584,
+      "step": 71881
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.844534209757844e-07,
+      "loss": 0.4863,
+      "step": 71882
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.84427164323726e-07,
+      "loss": 0.8262,
+      "step": 71883
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.844009079132869e-07,
+      "loss": 0.5225,
+      "step": 71884
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.843746517444876e-07,
+      "loss": 0.7969,
+      "step": 71885
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.843483958173484e-07,
+      "loss": 0.6367,
+      "step": 71886
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.843221401318889e-07,
+      "loss": 0.7344,
+      "step": 71887
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.842958846881293e-07,
+      "loss": 0.6118,
+      "step": 71888
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.842696294860898e-07,
+      "loss": 0.8428,
+      "step": 71889
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.842433745257908e-07,
+      "loss": 0.5112,
+      "step": 71890
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.842171198072517e-07,
+      "loss": 0.5815,
+      "step": 71891
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.841908653304933e-07,
+      "loss": 0.6211,
+      "step": 71892
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.841646110955354e-07,
+      "loss": 0.75,
+      "step": 71893
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.841383571023982e-07,
+      "loss": 0.6797,
+      "step": 71894
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.841121033511015e-07,
+      "loss": 0.6025,
+      "step": 71895
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.840858498416664e-07,
+      "loss": 0.5977,
+      "step": 71896
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.840595965741115e-07,
+      "loss": 0.4937,
+      "step": 71897
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.840333435484578e-07,
+      "loss": 0.46,
+      "step": 71898
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.840070907647253e-07,
+      "loss": 0.8057,
+      "step": 71899
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.83980838222934e-07,
+      "loss": 0.624,
+      "step": 71900
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.83954585923104e-07,
+      "loss": 0.7393,
+      "step": 71901
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.839283338652555e-07,
+      "loss": 0.7061,
+      "step": 71902
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.839020820494083e-07,
+      "loss": 0.574,
+      "step": 71903
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.838758304755831e-07,
+      "loss": 0.7075,
+      "step": 71904
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.838495791437993e-07,
+      "loss": 0.5332,
+      "step": 71905
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.838233280540778e-07,
+      "loss": 0.7002,
+      "step": 71906
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.837970772064381e-07,
+      "loss": 0.4966,
+      "step": 71907
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.837708266009002e-07,
+      "loss": 0.6743,
+      "step": 71908
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.837445762374845e-07,
+      "loss": 0.749,
+      "step": 71909
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.83718326116211e-07,
+      "loss": 0.6543,
+      "step": 71910
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.836920762371e-07,
+      "loss": 0.6865,
+      "step": 71911
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.836658266001713e-07,
+      "loss": 0.5815,
+      "step": 71912
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.836395772054452e-07,
+      "loss": 0.6675,
+      "step": 71913
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.836133280529415e-07,
+      "loss": 0.6729,
+      "step": 71914
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.835870791426809e-07,
+      "loss": 0.5566,
+      "step": 71915
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.835608304746833e-07,
+      "loss": 0.5083,
+      "step": 71916
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.835345820489684e-07,
+      "loss": 0.6865,
+      "step": 71917
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.835083338655562e-07,
+      "loss": 0.8174,
+      "step": 71918
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.834820859244675e-07,
+      "loss": 0.4978,
+      "step": 71919
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.834558382257216e-07,
+      "loss": 0.6104,
+      "step": 71920
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.834295907693392e-07,
+      "loss": 0.7432,
+      "step": 71921
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.834033435553403e-07,
+      "loss": 0.6758,
+      "step": 71922
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.833770965837448e-07,
+      "loss": 0.6675,
+      "step": 71923
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.833508498545729e-07,
+      "loss": 0.481,
+      "step": 71924
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.833246033678448e-07,
+      "loss": 0.5518,
+      "step": 71925
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.832983571235808e-07,
+      "loss": 0.7959,
+      "step": 71926
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.832721111218001e-07,
+      "loss": 0.6743,
+      "step": 71927
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.832458653625237e-07,
+      "loss": 0.6377,
+      "step": 71928
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.832196198457712e-07,
+      "loss": 0.6484,
+      "step": 71929
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.83193374571563e-07,
+      "loss": 0.6953,
+      "step": 71930
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.831671295399189e-07,
+      "loss": 0.6572,
+      "step": 71931
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.831408847508593e-07,
+      "loss": 0.6128,
+      "step": 71932
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.83114640204404e-07,
+      "loss": 0.625,
+      "step": 71933
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.830883959005736e-07,
+      "loss": 0.5781,
+      "step": 71934
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.830621518393875e-07,
+      "loss": 0.7441,
+      "step": 71935
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.830359080208666e-07,
+      "loss": 0.7012,
+      "step": 71936
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.8300966444503e-07,
+      "loss": 0.4656,
+      "step": 71937
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.829834211118986e-07,
+      "loss": 0.9277,
+      "step": 71938
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.829571780214921e-07,
+      "loss": 0.5957,
+      "step": 71939
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.829309351738308e-07,
+      "loss": 0.7393,
+      "step": 71940
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.829046925689346e-07,
+      "loss": 0.5449,
+      "step": 71941
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.828784502068238e-07,
+      "loss": 0.6914,
+      "step": 71942
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.828522080875181e-07,
+      "loss": 0.627,
+      "step": 71943
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.828259662110383e-07,
+      "loss": 0.4475,
+      "step": 71944
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.827997245774037e-07,
+      "loss": 0.6816,
+      "step": 71945
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.827734831866352e-07,
+      "loss": 0.7812,
+      "step": 71946
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.827472420387523e-07,
+      "loss": 0.6045,
+      "step": 71947
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.827210011337751e-07,
+      "loss": 0.5146,
+      "step": 71948
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.82694760471724e-07,
+      "loss": 0.6953,
+      "step": 71949
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.826685200526186e-07,
+      "loss": 0.7793,
+      "step": 71950
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.826422798764797e-07,
+      "loss": 0.6992,
+      "step": 71951
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.826160399433269e-07,
+      "loss": 0.623,
+      "step": 71952
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.825898002531803e-07,
+      "loss": 0.6719,
+      "step": 71953
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.825635608060603e-07,
+      "loss": 0.6309,
+      "step": 71954
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.825373216019865e-07,
+      "loss": 0.5884,
+      "step": 71955
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.825110826409797e-07,
+      "loss": 0.7344,
+      "step": 71956
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.824848439230592e-07,
+      "loss": 0.5674,
+      "step": 71957
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.824586054482455e-07,
+      "loss": 0.7544,
+      "step": 71958
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.824323672165588e-07,
+      "loss": 0.4609,
+      "step": 71959
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.824061292280187e-07,
+      "loss": 0.6768,
+      "step": 71960
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.823798914826461e-07,
+      "loss": 0.5635,
+      "step": 71961
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.823536539804602e-07,
+      "loss": 0.6738,
+      "step": 71962
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.823274167214817e-07,
+      "loss": 0.6548,
+      "step": 71963
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.823011797057303e-07,
+      "loss": 0.584,
+      "step": 71964
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.822749429332267e-07,
+      "loss": 0.5879,
+      "step": 71965
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.822487064039904e-07,
+      "loss": 0.6084,
+      "step": 71966
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.822224701180414e-07,
+      "loss": 0.7021,
+      "step": 71967
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.821962340754004e-07,
+      "loss": 0.584,
+      "step": 71968
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.821699982760869e-07,
+      "loss": 0.6006,
+      "step": 71969
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.821437627201212e-07,
+      "loss": 0.582,
+      "step": 71970
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.821175274075235e-07,
+      "loss": 0.5576,
+      "step": 71971
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.820912923383136e-07,
+      "loss": 0.6348,
+      "step": 71972
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.82065057512512e-07,
+      "loss": 0.7881,
+      "step": 71973
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.820388229301384e-07,
+      "loss": 0.7041,
+      "step": 71974
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.820125885912132e-07,
+      "loss": 0.666,
+      "step": 71975
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.819863544957566e-07,
+      "loss": 0.4943,
+      "step": 71976
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.81960120643788e-07,
+      "loss": 0.626,
+      "step": 71977
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.819338870353281e-07,
+      "loss": 0.5532,
+      "step": 71978
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.819076536703966e-07,
+      "loss": 0.5605,
+      "step": 71979
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.81881420549014e-07,
+      "loss": 0.6033,
+      "step": 71980
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.818551876712e-07,
+      "loss": 0.8535,
+      "step": 71981
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.818289550369749e-07,
+      "loss": 0.6143,
+      "step": 71982
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.818027226463587e-07,
+      "loss": 0.5767,
+      "step": 71983
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 6.817764904993717e-07,
+      "loss": 0.5742,
+      "step": 71984
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.817502585960335e-07,
+      "loss": 0.6631,
+      "step": 71985
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.817240269363652e-07,
+      "loss": 0.3938,
+      "step": 71986
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.816977955203855e-07,
+      "loss": 0.4531,
+      "step": 71987
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.816715643481155e-07,
+      "loss": 0.707,
+      "step": 71988
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.816453334195747e-07,
+      "loss": 0.6211,
+      "step": 71989
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.816191027347835e-07,
+      "loss": 0.5486,
+      "step": 71990
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.815928722937619e-07,
+      "loss": 0.5894,
+      "step": 71991
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.815666420965301e-07,
+      "loss": 0.666,
+      "step": 71992
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.815404121431076e-07,
+      "loss": 0.7461,
+      "step": 71993
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.815141824335155e-07,
+      "loss": 0.5216,
+      "step": 71994
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.81487952967773e-07,
+      "loss": 0.6152,
+      "step": 71995
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.814617237459011e-07,
+      "loss": 0.7944,
+      "step": 71996
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.814354947679191e-07,
+      "loss": 0.7275,
+      "step": 71997
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.814092660338471e-07,
+      "loss": 0.4966,
+      "step": 71998
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.813830375437054e-07,
+      "loss": 0.5352,
+      "step": 71999
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.813568092975139e-07,
+      "loss": 0.6914,
+      "step": 72000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.813305812952932e-07,
+      "loss": 0.5698,
+      "step": 72001
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.813043535370626e-07,
+      "loss": 0.6689,
+      "step": 72002
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.81278126022843e-07,
+      "loss": 0.6182,
+      "step": 72003
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.812518987526537e-07,
+      "loss": 0.5825,
+      "step": 72004
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.812256717265155e-07,
+      "loss": 0.7329,
+      "step": 72005
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.811994449444483e-07,
+      "loss": 0.403,
+      "step": 72006
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.811732184064718e-07,
+      "loss": 0.6594,
+      "step": 72007
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.811469921126063e-07,
+      "loss": 0.5713,
+      "step": 72008
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.811207660628718e-07,
+      "loss": 0.6953,
+      "step": 72009
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.810945402572884e-07,
+      "loss": 0.5952,
+      "step": 72010
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.810683146958764e-07,
+      "loss": 0.4596,
+      "step": 72011
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.810420893786557e-07,
+      "loss": 0.6787,
+      "step": 72012
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.810158643056464e-07,
+      "loss": 0.5884,
+      "step": 72013
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.809896394768685e-07,
+      "loss": 0.6992,
+      "step": 72014
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.809634148923423e-07,
+      "loss": 0.5386,
+      "step": 72015
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80937190552088e-07,
+      "loss": 0.7705,
+      "step": 72016
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80910966456125e-07,
+      "loss": 0.5146,
+      "step": 72017
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.808847426044739e-07,
+      "loss": 0.4995,
+      "step": 72018
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.808585189971546e-07,
+      "loss": 0.6206,
+      "step": 72019
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.808322956341874e-07,
+      "loss": 0.635,
+      "step": 72020
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.808060725155921e-07,
+      "loss": 0.718,
+      "step": 72021
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80779849641389e-07,
+      "loss": 0.5549,
+      "step": 72022
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.807536270115981e-07,
+      "loss": 0.5674,
+      "step": 72023
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.807274046262395e-07,
+      "loss": 0.3816,
+      "step": 72024
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.807011824853333e-07,
+      "loss": 0.7437,
+      "step": 72025
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.806749605888998e-07,
+      "loss": 0.4771,
+      "step": 72026
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.806487389369582e-07,
+      "loss": 0.6289,
+      "step": 72027
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.806225175295295e-07,
+      "loss": 0.9443,
+      "step": 72028
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.805962963666331e-07,
+      "loss": 0.6846,
+      "step": 72029
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.805700754482899e-07,
+      "loss": 0.5627,
+      "step": 72030
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.805438547745191e-07,
+      "loss": 0.6914,
+      "step": 72031
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.805176343453415e-07,
+      "loss": 0.6631,
+      "step": 72032
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.804914141607766e-07,
+      "loss": 0.4561,
+      "step": 72033
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.804651942208449e-07,
+      "loss": 0.7021,
+      "step": 72034
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.804389745255661e-07,
+      "loss": 0.6973,
+      "step": 72035
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80412755074961e-07,
+      "loss": 0.4999,
+      "step": 72036
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80386535869049e-07,
+      "loss": 0.6582,
+      "step": 72037
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.8036031690785e-07,
+      "loss": 0.5742,
+      "step": 72038
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.803340981913845e-07,
+      "loss": 0.7861,
+      "step": 72039
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.803078797196724e-07,
+      "loss": 0.6074,
+      "step": 72040
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.802816614927341e-07,
+      "loss": 0.6807,
+      "step": 72041
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.802554435105894e-07,
+      "loss": 0.4827,
+      "step": 72042
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.802292257732583e-07,
+      "loss": 0.5864,
+      "step": 72043
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80203008280761e-07,
+      "loss": 0.6426,
+      "step": 72044
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.801767910331176e-07,
+      "loss": 0.7129,
+      "step": 72045
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.801505740303483e-07,
+      "loss": 0.7314,
+      "step": 72046
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.801243572724728e-07,
+      "loss": 0.5811,
+      "step": 72047
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.800981407595112e-07,
+      "loss": 0.6271,
+      "step": 72048
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80071924491484e-07,
+      "loss": 0.4873,
+      "step": 72049
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.800457084684108e-07,
+      "loss": 0.6445,
+      "step": 72050
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.80019492690312e-07,
+      "loss": 0.6206,
+      "step": 72051
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.799932771572074e-07,
+      "loss": 0.7197,
+      "step": 72052
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.799670618691175e-07,
+      "loss": 0.6152,
+      "step": 72053
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.799408468260619e-07,
+      "loss": 0.6343,
+      "step": 72054
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.799146320280611e-07,
+      "loss": 0.6465,
+      "step": 72055
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.798884174751351e-07,
+      "loss": 0.6406,
+      "step": 72056
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.798622031673035e-07,
+      "loss": 0.5859,
+      "step": 72057
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.798359891045868e-07,
+      "loss": 0.9307,
+      "step": 72058
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.798097752870049e-07,
+      "loss": 0.5825,
+      "step": 72059
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.797835617145779e-07,
+      "loss": 0.1942,
+      "step": 72060
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.79757348387326e-07,
+      "loss": 0.6782,
+      "step": 72061
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.797311353052691e-07,
+      "loss": 0.5708,
+      "step": 72062
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.797049224684272e-07,
+      "loss": 0.474,
+      "step": 72063
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.796787098768206e-07,
+      "loss": 0.4849,
+      "step": 72064
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.796524975304695e-07,
+      "loss": 0.6631,
+      "step": 72065
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.796262854293939e-07,
+      "loss": 0.7188,
+      "step": 72066
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.796000735736132e-07,
+      "loss": 0.6621,
+      "step": 72067
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.795738619631483e-07,
+      "loss": 0.584,
+      "step": 72068
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.795476505980188e-07,
+      "loss": 0.7383,
+      "step": 72069
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.795214394782451e-07,
+      "loss": 0.48,
+      "step": 72070
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.794952286038469e-07,
+      "loss": 0.6978,
+      "step": 72071
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.794690179748446e-07,
+      "loss": 0.4983,
+      "step": 72072
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.79442807591258e-07,
+      "loss": 0.6396,
+      "step": 72073
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.794165974531074e-07,
+      "loss": 0.7197,
+      "step": 72074
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.793903875604127e-07,
+      "loss": 0.6221,
+      "step": 72075
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.793641779131947e-07,
+      "loss": 0.7065,
+      "step": 72076
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.793379685114719e-07,
+      "loss": 0.5474,
+      "step": 72077
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.793117593552657e-07,
+      "loss": 0.6147,
+      "step": 72078
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.792855504445955e-07,
+      "loss": 0.668,
+      "step": 72079
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.792593417794819e-07,
+      "loss": 0.7031,
+      "step": 72080
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.792331333599443e-07,
+      "loss": 0.6182,
+      "step": 72081
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.792069251860035e-07,
+      "loss": 0.6621,
+      "step": 72082
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.791807172576789e-07,
+      "loss": 0.4097,
+      "step": 72083
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.791545095749912e-07,
+      "loss": 0.5762,
+      "step": 72084
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.791283021379599e-07,
+      "loss": 0.6875,
+      "step": 72085
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.791020949466055e-07,
+      "loss": 0.4912,
+      "step": 72086
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.790758880009479e-07,
+      "loss": 0.5928,
+      "step": 72087
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.790496813010069e-07,
+      "loss": 0.5127,
+      "step": 72088
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.790234748468029e-07,
+      "loss": 0.5928,
+      "step": 72089
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.789972686383558e-07,
+      "loss": 0.5767,
+      "step": 72090
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.78971062675686e-07,
+      "loss": 0.6436,
+      "step": 72091
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.789448569588128e-07,
+      "loss": 0.6445,
+      "step": 72092
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.789186514877572e-07,
+      "loss": 0.5986,
+      "step": 72093
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.788924462625385e-07,
+      "loss": 0.5586,
+      "step": 72094
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.788662412831774e-07,
+      "loss": 0.6533,
+      "step": 72095
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.788400365496937e-07,
+      "loss": 0.6152,
+      "step": 72096
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.788138320621073e-07,
+      "loss": 0.6213,
+      "step": 72097
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.787876278204382e-07,
+      "loss": 0.5542,
+      "step": 72098
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.787614238247069e-07,
+      "loss": 0.5347,
+      "step": 72099
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.787352200749328e-07,
+      "loss": 0.48,
+      "step": 72100
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.787090165711368e-07,
+      "loss": 0.5083,
+      "step": 72101
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.786828133133382e-07,
+      "loss": 0.5618,
+      "step": 72102
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.786566103015576e-07,
+      "loss": 0.6455,
+      "step": 72103
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.786304075358146e-07,
+      "loss": 0.4907,
+      "step": 72104
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.786042050161299e-07,
+      "loss": 0.4834,
+      "step": 72105
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.785780027425233e-07,
+      "loss": 0.6133,
+      "step": 72106
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.785518007150144e-07,
+      "loss": 0.7734,
+      "step": 72107
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.785255989336236e-07,
+      "loss": 0.538,
+      "step": 72108
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.78499397398371e-07,
+      "loss": 0.5981,
+      "step": 72109
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.784731961092767e-07,
+      "loss": 0.7373,
+      "step": 72110
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.784469950663605e-07,
+      "loss": 0.7007,
+      "step": 72111
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.78420794269643e-07,
+      "loss": 0.7119,
+      "step": 72112
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.783945937191436e-07,
+      "loss": 0.6279,
+      "step": 72113
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.783683934148827e-07,
+      "loss": 0.5635,
+      "step": 72114
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.783421933568807e-07,
+      "loss": 0.4153,
+      "step": 72115
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.783159935451572e-07,
+      "loss": 0.431,
+      "step": 72116
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.782897939797319e-07,
+      "loss": 0.627,
+      "step": 72117
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.782635946606257e-07,
+      "loss": 0.6426,
+      "step": 72118
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.78237395587858e-07,
+      "loss": 0.7334,
+      "step": 72119
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.782111967614495e-07,
+      "loss": 0.6953,
+      "step": 72120
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.781849981814193e-07,
+      "loss": 0.4045,
+      "step": 72121
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.781587998477887e-07,
+      "loss": 0.5488,
+      "step": 72122
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.781326017605766e-07,
+      "loss": 0.5894,
+      "step": 72123
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.781064039198039e-07,
+      "loss": 0.6777,
+      "step": 72124
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.780802063254904e-07,
+      "loss": 0.5144,
+      "step": 72125
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.780540089776559e-07,
+      "loss": 0.5483,
+      "step": 72126
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.780278118763207e-07,
+      "loss": 0.6387,
+      "step": 72127
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.780016150215046e-07,
+      "loss": 0.5122,
+      "step": 72128
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.77975418413228e-07,
+      "loss": 0.7075,
+      "step": 72129
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.779492220515109e-07,
+      "loss": 0.6558,
+      "step": 72130
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.779230259363731e-07,
+      "loss": 0.6934,
+      "step": 72131
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.778968300678349e-07,
+      "loss": 0.8057,
+      "step": 72132
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.778706344459163e-07,
+      "loss": 0.6133,
+      "step": 72133
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.778444390706375e-07,
+      "loss": 0.71,
+      "step": 72134
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.778182439420184e-07,
+      "loss": 0.5015,
+      "step": 72135
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.777920490600788e-07,
+      "loss": 0.5278,
+      "step": 72136
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.777658544248392e-07,
+      "loss": 0.665,
+      "step": 72137
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.777396600363193e-07,
+      "loss": 0.6289,
+      "step": 72138
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.777134658945393e-07,
+      "loss": 0.7441,
+      "step": 72139
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.776872719995195e-07,
+      "loss": 0.4614,
+      "step": 72140
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.776610783512796e-07,
+      "loss": 0.4663,
+      "step": 72141
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.776348849498397e-07,
+      "loss": 0.7773,
+      "step": 72142
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.776086917952202e-07,
+      "loss": 0.5244,
+      "step": 72143
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.775824988874406e-07,
+      "loss": 0.7305,
+      "step": 72144
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.775563062265217e-07,
+      "loss": 0.6875,
+      "step": 72145
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.775301138124829e-07,
+      "loss": 0.624,
+      "step": 72146
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.775039216453442e-07,
+      "loss": 0.3892,
+      "step": 72147
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.774777297251261e-07,
+      "loss": 0.5801,
+      "step": 72148
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.774515380518486e-07,
+      "loss": 0.5425,
+      "step": 72149
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.774253466255315e-07,
+      "loss": 0.6113,
+      "step": 72150
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.773991554461951e-07,
+      "loss": 0.6602,
+      "step": 72151
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.77372964513859e-07,
+      "loss": 0.6309,
+      "step": 72152
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.773467738285439e-07,
+      "loss": 0.6758,
+      "step": 72153
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.773205833902693e-07,
+      "loss": 0.6592,
+      "step": 72154
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.772943931990559e-07,
+      "loss": 0.5018,
+      "step": 72155
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.772682032549232e-07,
+      "loss": 0.6357,
+      "step": 72156
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.772420135578911e-07,
+      "loss": 0.5083,
+      "step": 72157
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.772158241079801e-07,
+      "loss": 0.6201,
+      "step": 72158
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.771896349052102e-07,
+      "loss": 0.6738,
+      "step": 72159
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.771634459496012e-07,
+      "loss": 0.6309,
+      "step": 72160
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.771372572411734e-07,
+      "loss": 0.7139,
+      "step": 72161
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.771110687799468e-07,
+      "loss": 0.6807,
+      "step": 72162
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.770848805659413e-07,
+      "loss": 0.499,
+      "step": 72163
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.77058692599177e-07,
+      "loss": 0.7451,
+      "step": 72164
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.770325048796745e-07,
+      "loss": 0.7559,
+      "step": 72165
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.770063174074531e-07,
+      "loss": 0.6592,
+      "step": 72166
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.769801301825329e-07,
+      "loss": 0.7578,
+      "step": 72167
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.769539432049343e-07,
+      "loss": 0.5806,
+      "step": 72168
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.76927756474677e-07,
+      "loss": 0.6621,
+      "step": 72169
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.769015699917815e-07,
+      "loss": 0.5825,
+      "step": 72170
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.768753837562674e-07,
+      "loss": 0.5085,
+      "step": 72171
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.768491977681552e-07,
+      "loss": 0.5522,
+      "step": 72172
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.768230120274645e-07,
+      "loss": 0.5962,
+      "step": 72173
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.767968265342157e-07,
+      "loss": 0.8477,
+      "step": 72174
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.767706412884289e-07,
+      "loss": 0.6377,
+      "step": 72175
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.767444562901236e-07,
+      "loss": 0.4,
+      "step": 72176
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.767182715393205e-07,
+      "loss": 0.5615,
+      "step": 72177
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.76692087036039e-07,
+      "loss": 0.5874,
+      "step": 72178
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.766659027802999e-07,
+      "loss": 0.6304,
+      "step": 72179
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.766397187721224e-07,
+      "loss": 0.6343,
+      "step": 72180
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.766135350115274e-07,
+      "loss": 0.5796,
+      "step": 72181
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.765873514985343e-07,
+      "loss": 0.52,
+      "step": 72182
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.765611682331636e-07,
+      "loss": 0.5525,
+      "step": 72183
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.765349852154349e-07,
+      "loss": 0.5918,
+      "step": 72184
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.765088024453691e-07,
+      "loss": 0.6362,
+      "step": 72185
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.76482619922985e-07,
+      "loss": 0.6221,
+      "step": 72186
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.764564376483035e-07,
+      "loss": 0.6455,
+      "step": 72187
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.764302556213442e-07,
+      "loss": 0.5659,
+      "step": 72188
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.764040738421277e-07,
+      "loss": 0.7861,
+      "step": 72189
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.763778923106733e-07,
+      "loss": 0.5981,
+      "step": 72190
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.763517110270019e-07,
+      "loss": 0.4326,
+      "step": 72191
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.763255299911328e-07,
+      "loss": 0.5728,
+      "step": 72192
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.762993492030866e-07,
+      "loss": 0.4995,
+      "step": 72193
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.762731686628829e-07,
+      "loss": 0.5723,
+      "step": 72194
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.762469883705424e-07,
+      "loss": 0.6006,
+      "step": 72195
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.762208083260843e-07,
+      "loss": 0.6216,
+      "step": 72196
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.76194628529529e-07,
+      "loss": 0.7324,
+      "step": 72197
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.761684489808967e-07,
+      "loss": 0.6167,
+      "step": 72198
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.761422696802073e-07,
+      "loss": 0.6074,
+      "step": 72199
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.761160906274808e-07,
+      "loss": 0.6445,
+      "step": 72200
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.760899118227373e-07,
+      "loss": 0.582,
+      "step": 72201
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.760637332659971e-07,
+      "loss": 0.5103,
+      "step": 72202
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.760375549572796e-07,
+      "loss": 0.479,
+      "step": 72203
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.760113768966057e-07,
+      "loss": 0.6885,
+      "step": 72204
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.759851990839951e-07,
+      "loss": 0.5352,
+      "step": 72205
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.759590215194675e-07,
+      "loss": 0.6318,
+      "step": 72206
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75932844203043e-07,
+      "loss": 0.498,
+      "step": 72207
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75906667134742e-07,
+      "loss": 0.6299,
+      "step": 72208
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.758804903145841e-07,
+      "loss": 0.6436,
+      "step": 72209
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.758543137425898e-07,
+      "loss": 0.5034,
+      "step": 72210
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.758281374187789e-07,
+      "loss": 0.627,
+      "step": 72211
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.758019613431716e-07,
+      "loss": 0.5159,
+      "step": 72212
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.757757855157877e-07,
+      "loss": 0.6265,
+      "step": 72213
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.757496099366476e-07,
+      "loss": 0.667,
+      "step": 72214
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.757234346057712e-07,
+      "loss": 0.7324,
+      "step": 72215
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75697259523178e-07,
+      "loss": 0.6045,
+      "step": 72216
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.756710846888888e-07,
+      "loss": 0.6689,
+      "step": 72217
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.756449101029231e-07,
+      "loss": 0.6475,
+      "step": 72218
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.756187357653014e-07,
+      "loss": 0.8057,
+      "step": 72219
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.755925616760433e-07,
+      "loss": 0.5098,
+      "step": 72220
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.755663878351693e-07,
+      "loss": 0.6641,
+      "step": 72221
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.755402142426991e-07,
+      "loss": 0.4058,
+      "step": 72222
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.755140408986529e-07,
+      "loss": 0.6992,
+      "step": 72223
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.754878678030505e-07,
+      "loss": 0.7188,
+      "step": 72224
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.754616949559126e-07,
+      "loss": 0.4904,
+      "step": 72225
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.754355223572582e-07,
+      "loss": 0.6665,
+      "step": 72226
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.754093500071082e-07,
+      "loss": 0.7207,
+      "step": 72227
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75383177905482e-07,
+      "loss": 0.7021,
+      "step": 72228
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.753570060524003e-07,
+      "loss": 0.5737,
+      "step": 72229
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.753308344478826e-07,
+      "loss": 0.637,
+      "step": 72230
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.753046630919493e-07,
+      "loss": 0.4878,
+      "step": 72231
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.752784919846201e-07,
+      "loss": 0.5542,
+      "step": 72232
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.752523211259155e-07,
+      "loss": 0.5391,
+      "step": 72233
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.752261505158549e-07,
+      "loss": 0.5688,
+      "step": 72234
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.751999801544595e-07,
+      "loss": 0.5654,
+      "step": 72235
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.751738100417479e-07,
+      "loss": 0.5557,
+      "step": 72236
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.751476401777408e-07,
+      "loss": 0.7617,
+      "step": 72237
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.751214705624584e-07,
+      "loss": 0.7559,
+      "step": 72238
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.750953011959204e-07,
+      "loss": 0.623,
+      "step": 72239
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75069132078147e-07,
+      "loss": 0.6777,
+      "step": 72240
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.750429632091584e-07,
+      "loss": 0.6953,
+      "step": 72241
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.75016794588974e-07,
+      "loss": 0.6587,
+      "step": 72242
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.749906262176147e-07,
+      "loss": 0.7109,
+      "step": 72243
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.749644580950999e-07,
+      "loss": 0.5513,
+      "step": 72244
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.749382902214503e-07,
+      "loss": 0.632,
+      "step": 72245
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.749121225966853e-07,
+      "loss": 0.5181,
+      "step": 72246
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.74885955220825e-07,
+      "loss": 0.6934,
+      "step": 72247
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.748597880938896e-07,
+      "loss": 0.5186,
+      "step": 72248
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.748336212158989e-07,
+      "loss": 0.6279,
+      "step": 72249
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.748074545868736e-07,
+      "loss": 0.7354,
+      "step": 72250
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.747812882068329e-07,
+      "loss": 0.6377,
+      "step": 72251
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.747551220757974e-07,
+      "loss": 0.6885,
+      "step": 72252
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.747289561937868e-07,
+      "loss": 0.8076,
+      "step": 72253
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.747027905608216e-07,
+      "loss": 0.5942,
+      "step": 72254
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.746766251769216e-07,
+      "loss": 0.6641,
+      "step": 72255
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.746504600421065e-07,
+      "loss": 0.6099,
+      "step": 72256
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.746242951563963e-07,
+      "loss": 0.6641,
+      "step": 72257
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.745981305198117e-07,
+      "loss": 0.6162,
+      "step": 72258
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.745719661323721e-07,
+      "loss": 0.6213,
+      "step": 72259
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.74545801994098e-07,
+      "loss": 0.6953,
+      "step": 72260
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.745196381050089e-07,
+      "loss": 0.7695,
+      "step": 72261
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.744934744651255e-07,
+      "loss": 0.7451,
+      "step": 72262
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.744673110744672e-07,
+      "loss": 0.8271,
+      "step": 72263
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.744411479330545e-07,
+      "loss": 0.6123,
+      "step": 72264
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.744149850409076e-07,
+      "loss": 0.5061,
+      "step": 72265
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.743888223980455e-07,
+      "loss": 0.7695,
+      "step": 72266
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.743626600044893e-07,
+      "loss": 0.6079,
+      "step": 72267
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.743364978602584e-07,
+      "loss": 0.3853,
+      "step": 72268
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.743103359653732e-07,
+      "loss": 0.5908,
+      "step": 72269
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.742841743198535e-07,
+      "loss": 0.6846,
+      "step": 72270
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.742580129237196e-07,
+      "loss": 0.4497,
+      "step": 72271
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.742318517769911e-07,
+      "loss": 0.5127,
+      "step": 72272
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.742056908796884e-07,
+      "loss": 0.6992,
+      "step": 72273
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.741795302318314e-07,
+      "loss": 0.5464,
+      "step": 72274
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.741533698334408e-07,
+      "loss": 0.5869,
+      "step": 72275
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.741272096845354e-07,
+      "loss": 0.6094,
+      "step": 72276
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.741010497851358e-07,
+      "loss": 0.6562,
+      "step": 72277
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.74074890135262e-07,
+      "loss": 0.6079,
+      "step": 72278
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.740487307349343e-07,
+      "loss": 0.5122,
+      "step": 72279
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.740225715841721e-07,
+      "loss": 0.3499,
+      "step": 72280
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.739964126829963e-07,
+      "loss": 0.7139,
+      "step": 72281
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73970254031426e-07,
+      "loss": 0.5991,
+      "step": 72282
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73944095629482e-07,
+      "loss": 0.5159,
+      "step": 72283
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.739179374771837e-07,
+      "loss": 0.6914,
+      "step": 72284
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73891779574552e-07,
+      "loss": 0.749,
+      "step": 72285
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.738656219216061e-07,
+      "loss": 0.7656,
+      "step": 72286
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.738394645183662e-07,
+      "loss": 0.3787,
+      "step": 72287
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.738133073648525e-07,
+      "loss": 0.7334,
+      "step": 72288
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.737871504610848e-07,
+      "loss": 0.7256,
+      "step": 72289
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.737609938070833e-07,
+      "loss": 0.667,
+      "step": 72290
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.737348374028681e-07,
+      "loss": 0.5234,
+      "step": 72291
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.737086812484592e-07,
+      "loss": 0.5002,
+      "step": 72292
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.736825253438764e-07,
+      "loss": 0.75,
+      "step": 72293
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.7365636968914e-07,
+      "loss": 0.4915,
+      "step": 72294
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.736302142842702e-07,
+      "loss": 0.6436,
+      "step": 72295
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.736040591292865e-07,
+      "loss": 0.437,
+      "step": 72296
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.735779042242089e-07,
+      "loss": 0.6523,
+      "step": 72297
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73551749569058e-07,
+      "loss": 0.6182,
+      "step": 72298
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.735255951638532e-07,
+      "loss": 0.6284,
+      "step": 72299
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.734994410086153e-07,
+      "loss": 0.6265,
+      "step": 72300
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.734732871033634e-07,
+      "loss": 0.668,
+      "step": 72301
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.734471334481182e-07,
+      "loss": 0.5464,
+      "step": 72302
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.734209800428994e-07,
+      "loss": 0.7344,
+      "step": 72303
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.733948268877273e-07,
+      "loss": 0.7803,
+      "step": 72304
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73368673982622e-07,
+      "loss": 0.5859,
+      "step": 72305
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73342521327603e-07,
+      "loss": 0.5728,
+      "step": 72306
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.733163689226906e-07,
+      "loss": 0.6133,
+      "step": 72307
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.732902167679047e-07,
+      "loss": 0.6284,
+      "step": 72308
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.732640648632657e-07,
+      "loss": 0.6953,
+      "step": 72309
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73237913208793e-07,
+      "loss": 0.6768,
+      "step": 72310
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.732117618045074e-07,
+      "loss": 0.6226,
+      "step": 72311
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.731856106504285e-07,
+      "loss": 0.5443,
+      "step": 72312
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.731594597465762e-07,
+      "loss": 0.6289,
+      "step": 72313
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.731333090929707e-07,
+      "loss": 0.4641,
+      "step": 72314
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.731071586896324e-07,
+      "loss": 0.5935,
+      "step": 72315
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.730810085365804e-07,
+      "loss": 0.6738,
+      "step": 72316
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.730548586338355e-07,
+      "loss": 0.6377,
+      "step": 72317
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.730287089814173e-07,
+      "loss": 0.5405,
+      "step": 72318
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.73002559579346e-07,
+      "loss": 0.6577,
+      "step": 72319
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.729764104276416e-07,
+      "loss": 0.5342,
+      "step": 72320
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.729502615263242e-07,
+      "loss": 0.5596,
+      "step": 72321
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.729241128754135e-07,
+      "loss": 0.7031,
+      "step": 72322
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.728979644749302e-07,
+      "loss": 0.6182,
+      "step": 72323
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.728718163248934e-07,
+      "loss": 0.5918,
+      "step": 72324
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.728456684253241e-07,
+      "loss": 0.6201,
+      "step": 72325
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.728195207762418e-07,
+      "loss": 0.5015,
+      "step": 72326
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.727933733776663e-07,
+      "loss": 0.7217,
+      "step": 72327
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.727672262296179e-07,
+      "loss": 0.6084,
+      "step": 72328
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.727410793321166e-07,
+      "loss": 0.7148,
+      "step": 72329
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.727149326851824e-07,
+      "loss": 0.686,
+      "step": 72330
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.726887862888354e-07,
+      "loss": 0.709,
+      "step": 72331
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.726626401430953e-07,
+      "loss": 0.4263,
+      "step": 72332
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.726364942479827e-07,
+      "loss": 0.7207,
+      "step": 72333
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.726103486035171e-07,
+      "loss": 0.6934,
+      "step": 72334
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.725842032097189e-07,
+      "loss": 0.624,
+      "step": 72335
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.72558058066608e-07,
+      "loss": 0.6553,
+      "step": 72336
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.725319131742038e-07,
+      "loss": 0.7046,
+      "step": 72337
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.725057685325273e-07,
+      "loss": 0.5062,
+      "step": 72338
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.724796241415978e-07,
+      "loss": 0.6431,
+      "step": 72339
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.724534800014358e-07,
+      "loss": 0.5957,
+      "step": 72340
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.724273361120609e-07,
+      "loss": 0.5601,
+      "step": 72341
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.724011924734936e-07,
+      "loss": 0.5449,
+      "step": 72342
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.723750490857533e-07,
+      "loss": 0.5278,
+      "step": 72343
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.723489059488607e-07,
+      "loss": 0.4077,
+      "step": 72344
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.723227630628356e-07,
+      "loss": 0.3584,
+      "step": 72345
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.722966204276977e-07,
+      "loss": 0.6973,
+      "step": 72346
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.722704780434669e-07,
+      "loss": 0.5469,
+      "step": 72347
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.722443359101639e-07,
+      "loss": 0.6729,
+      "step": 72348
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.722181940278081e-07,
+      "loss": 0.7686,
+      "step": 72349
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.7219205239642e-07,
+      "loss": 0.75,
+      "step": 72350
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.721659110160191e-07,
+      "loss": 0.7803,
+      "step": 72351
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.721397698866257e-07,
+      "loss": 0.6201,
+      "step": 72352
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.721136290082599e-07,
+      "loss": 0.6152,
+      "step": 72353
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.720874883809418e-07,
+      "loss": 0.4314,
+      "step": 72354
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.720613480046914e-07,
+      "loss": 0.6147,
+      "step": 72355
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.720352078795281e-07,
+      "loss": 0.6904,
+      "step": 72356
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.720090680054725e-07,
+      "loss": 0.793,
+      "step": 72357
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.719829283825444e-07,
+      "loss": 0.7021,
+      "step": 72358
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.71956789010764e-07,
+      "loss": 0.6357,
+      "step": 72359
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.71930649890151e-07,
+      "loss": 0.4707,
+      "step": 72360
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.719045110207259e-07,
+      "loss": 0.7822,
+      "step": 72361
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.718783724025081e-07,
+      "loss": 0.5244,
+      "step": 72362
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.718522340355182e-07,
+      "loss": 0.5996,
+      "step": 72363
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.718260959197759e-07,
+      "loss": 0.71,
+      "step": 72364
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.717999580553017e-07,
+      "loss": 0.5701,
+      "step": 72365
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.717738204421146e-07,
+      "loss": 0.5991,
+      "step": 72366
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.717476830802354e-07,
+      "loss": 0.5459,
+      "step": 72367
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.717215459696838e-07,
+      "loss": 0.6543,
+      "step": 72368
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.7169540911048e-07,
+      "loss": 0.5728,
+      "step": 72369
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.716692725026438e-07,
+      "loss": 0.4938,
+      "step": 72370
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.716431361461955e-07,
+      "loss": 0.5396,
+      "step": 72371
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.716170000411548e-07,
+      "loss": 0.4902,
+      "step": 72372
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.715908641875419e-07,
+      "loss": 0.6523,
+      "step": 72373
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 6.715647285853768e-07,
+      "loss": 0.6875,
+      "step": 72374
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.715385932346798e-07,
+      "loss": 0.5908,
+      "step": 72375
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.715124581354704e-07,
+      "loss": 0.6602,
+      "step": 72376
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.714863232877685e-07,
+      "loss": 0.6196,
+      "step": 72377
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.714601886915947e-07,
+      "loss": 0.5996,
+      "step": 72378
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.714340543469685e-07,
+      "loss": 0.4058,
+      "step": 72379
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.714079202539105e-07,
+      "loss": 0.5723,
+      "step": 72380
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.713817864124401e-07,
+      "loss": 0.5405,
+      "step": 72381
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.713556528225775e-07,
+      "loss": 0.6543,
+      "step": 72382
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.713295194843428e-07,
+      "loss": 0.5952,
+      "step": 72383
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.713033863977561e-07,
+      "loss": 0.6909,
+      "step": 72384
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.712772535628375e-07,
+      "loss": 0.6787,
+      "step": 72385
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.712511209796066e-07,
+      "loss": 0.6855,
+      "step": 72386
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.712249886480834e-07,
+      "loss": 0.6436,
+      "step": 72387
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.711988565682883e-07,
+      "loss": 0.6362,
+      "step": 72388
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.711727247402409e-07,
+      "loss": 0.603,
+      "step": 72389
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.711465931639616e-07,
+      "loss": 0.6704,
+      "step": 72390
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.711204618394701e-07,
+      "loss": 0.5063,
+      "step": 72391
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.710943307667866e-07,
+      "loss": 0.6797,
+      "step": 72392
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.710681999459311e-07,
+      "loss": 0.5791,
+      "step": 72393
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.710420693769237e-07,
+      "loss": 0.6597,
+      "step": 72394
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.710159390597842e-07,
+      "loss": 0.5111,
+      "step": 72395
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.709898089945325e-07,
+      "loss": 0.6338,
+      "step": 72396
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.709636791811889e-07,
+      "loss": 0.6328,
+      "step": 72397
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.709375496197731e-07,
+      "loss": 0.6675,
+      "step": 72398
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.709114203103055e-07,
+      "loss": 0.7456,
+      "step": 72399
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.708852912528057e-07,
+      "loss": 0.6592,
+      "step": 72400
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.708591624472942e-07,
+      "loss": 0.5566,
+      "step": 72401
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.708330338937904e-07,
+      "loss": 0.7295,
+      "step": 72402
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.708069055923146e-07,
+      "loss": 0.6279,
+      "step": 72403
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.707807775428873e-07,
+      "loss": 0.6777,
+      "step": 72404
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.707546497455276e-07,
+      "loss": 0.6353,
+      "step": 72405
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.707285222002559e-07,
+      "loss": 0.564,
+      "step": 72406
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.707023949070924e-07,
+      "loss": 0.5684,
+      "step": 72407
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.706762678660567e-07,
+      "loss": 0.7627,
+      "step": 72408
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.706501410771693e-07,
+      "loss": 0.583,
+      "step": 72409
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.706240145404498e-07,
+      "loss": 0.7568,
+      "step": 72410
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.705978882559184e-07,
+      "loss": 0.5776,
+      "step": 72411
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.705717622235949e-07,
+      "loss": 0.6479,
+      "step": 72412
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.705456364434996e-07,
+      "loss": 0.5845,
+      "step": 72413
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.705195109156527e-07,
+      "loss": 0.4561,
+      "step": 72414
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.704933856400733e-07,
+      "loss": 0.5708,
+      "step": 72415
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.704672606167823e-07,
+      "loss": 0.6973,
+      "step": 72416
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.704411358457991e-07,
+      "loss": 0.4966,
+      "step": 72417
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.70415011327144e-07,
+      "loss": 0.6426,
+      "step": 72418
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.703888870608373e-07,
+      "loss": 0.6187,
+      "step": 72419
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.703627630468983e-07,
+      "loss": 0.7188,
+      "step": 72420
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.703366392853476e-07,
+      "loss": 0.5962,
+      "step": 72421
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.703105157762048e-07,
+      "loss": 0.7676,
+      "step": 72422
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.702843925194902e-07,
+      "loss": 0.7852,
+      "step": 72423
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.702582695152239e-07,
+      "loss": 0.4451,
+      "step": 72424
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.702321467634253e-07,
+      "loss": 0.7073,
+      "step": 72425
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.70206024264115e-07,
+      "loss": 0.4739,
+      "step": 72426
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.701799020173127e-07,
+      "loss": 0.6758,
+      "step": 72427
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.701537800230385e-07,
+      "loss": 0.6089,
+      "step": 72428
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.701276582813122e-07,
+      "loss": 0.7363,
+      "step": 72429
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.701015367921543e-07,
+      "loss": 0.5386,
+      "step": 72430
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.700754155555841e-07,
+      "loss": 0.6787,
+      "step": 72431
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.700492945716223e-07,
+      "loss": 0.3884,
+      "step": 72432
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.700231738402884e-07,
+      "loss": 0.6348,
+      "step": 72433
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.699970533616031e-07,
+      "loss": 0.7393,
+      "step": 72434
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.699709331355854e-07,
+      "loss": 0.71,
+      "step": 72435
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.699448131622558e-07,
+      "loss": 0.6465,
+      "step": 72436
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.699186934416343e-07,
+      "loss": 0.7158,
+      "step": 72437
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.698925739737411e-07,
+      "loss": 0.5649,
+      "step": 72438
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.698664547585956e-07,
+      "loss": 0.5659,
+      "step": 72439
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.698403357962183e-07,
+      "loss": 0.5698,
+      "step": 72440
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.698142170866291e-07,
+      "loss": 0.7227,
+      "step": 72441
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.69788098629848e-07,
+      "loss": 0.7314,
+      "step": 72442
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.697619804258949e-07,
+      "loss": 0.7188,
+      "step": 72443
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.697358624747903e-07,
+      "loss": 0.5253,
+      "step": 72444
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.697097447765534e-07,
+      "loss": 0.7075,
+      "step": 72445
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.696836273312045e-07,
+      "loss": 0.6099,
+      "step": 72446
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.696575101387638e-07,
+      "loss": 0.5579,
+      "step": 72447
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.69631393199251e-07,
+      "loss": 0.7739,
+      "step": 72448
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.696052765126864e-07,
+      "loss": 0.5381,
+      "step": 72449
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.695791600790897e-07,
+      "loss": 0.6416,
+      "step": 72450
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.695530438984811e-07,
+      "loss": 0.748,
+      "step": 72451
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.695269279708803e-07,
+      "loss": 0.6309,
+      "step": 72452
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.695008122963081e-07,
+      "loss": 0.4941,
+      "step": 72453
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.694746968747839e-07,
+      "loss": 0.6055,
+      "step": 72454
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.694485817063276e-07,
+      "loss": 0.6514,
+      "step": 72455
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.694224667909589e-07,
+      "loss": 0.7666,
+      "step": 72456
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.693963521286986e-07,
+      "loss": 0.422,
+      "step": 72457
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.693702377195661e-07,
+      "loss": 0.4949,
+      "step": 72458
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.693441235635817e-07,
+      "loss": 0.6865,
+      "step": 72459
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.693180096607653e-07,
+      "loss": 0.6396,
+      "step": 72460
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.692918960111369e-07,
+      "loss": 0.6328,
+      "step": 72461
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.692657826147164e-07,
+      "loss": 0.5317,
+      "step": 72462
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.692396694715241e-07,
+      "loss": 0.5537,
+      "step": 72463
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.692135565815799e-07,
+      "loss": 0.6787,
+      "step": 72464
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.691874439449033e-07,
+      "loss": 0.4907,
+      "step": 72465
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.691613315615149e-07,
+      "loss": 0.6299,
+      "step": 72466
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.691352194314342e-07,
+      "loss": 0.8018,
+      "step": 72467
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.691091075546817e-07,
+      "loss": 0.8047,
+      "step": 72468
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.69082995931277e-07,
+      "loss": 0.4463,
+      "step": 72469
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.690568845612403e-07,
+      "loss": 0.4878,
+      "step": 72470
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.690307734445913e-07,
+      "loss": 0.6914,
+      "step": 72471
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.690046625813506e-07,
+      "loss": 0.4956,
+      "step": 72472
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.689785519715375e-07,
+      "loss": 0.6323,
+      "step": 72473
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.689524416151729e-07,
+      "loss": 0.6562,
+      "step": 72474
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.689263315122755e-07,
+      "loss": 0.5186,
+      "step": 72475
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.689002216628663e-07,
+      "loss": 0.5969,
+      "step": 72476
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.688741120669648e-07,
+      "loss": 0.7329,
+      "step": 72477
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.688480027245913e-07,
+      "loss": 0.7031,
+      "step": 72478
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.688218936357654e-07,
+      "loss": 0.6084,
+      "step": 72479
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.687957848005077e-07,
+      "loss": 0.7109,
+      "step": 72480
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.687696762188375e-07,
+      "loss": 0.7363,
+      "step": 72481
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.687435678907754e-07,
+      "loss": 0.5862,
+      "step": 72482
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.68717459816341e-07,
+      "loss": 0.6797,
+      "step": 72483
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.686913519955547e-07,
+      "loss": 0.7061,
+      "step": 72484
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.68665244428436e-07,
+      "loss": 0.666,
+      "step": 72485
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.686391371150049e-07,
+      "loss": 0.5879,
+      "step": 72486
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.686130300552818e-07,
+      "loss": 0.5918,
+      "step": 72487
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.685869232492862e-07,
+      "loss": 0.5889,
+      "step": 72488
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.685608166970387e-07,
+      "loss": 0.6562,
+      "step": 72489
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.685347103985586e-07,
+      "loss": 0.6816,
+      "step": 72490
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.685086043538665e-07,
+      "loss": 0.6855,
+      "step": 72491
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.684824985629821e-07,
+      "loss": 0.6787,
+      "step": 72492
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.684563930259253e-07,
+      "loss": 0.5024,
+      "step": 72493
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.684302877427164e-07,
+      "loss": 0.5303,
+      "step": 72494
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.68404182713375e-07,
+      "loss": 0.5913,
+      "step": 72495
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.683780779379212e-07,
+      "loss": 0.478,
+      "step": 72496
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.683519734163751e-07,
+      "loss": 0.8115,
+      "step": 72497
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.683258691487567e-07,
+      "loss": 0.8516,
+      "step": 72498
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.68299765135086e-07,
+      "loss": 0.4941,
+      "step": 72499
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.682736613753826e-07,
+      "loss": 0.7891,
+      "step": 72500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.682475578696671e-07,
+      "loss": 0.6958,
+      "step": 72501
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.68221454617959e-07,
+      "loss": 0.5457,
+      "step": 72502
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.681953516202787e-07,
+      "loss": 0.5081,
+      "step": 72503
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.681692488766461e-07,
+      "loss": 0.6455,
+      "step": 72504
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.681431463870809e-07,
+      "loss": 0.6489,
+      "step": 72505
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.681170441516031e-07,
+      "loss": 0.5795,
+      "step": 72506
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.680909421702327e-07,
+      "loss": 0.708,
+      "step": 72507
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.6806484044299e-07,
+      "loss": 0.5077,
+      "step": 72508
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.680387389698949e-07,
+      "loss": 0.3953,
+      "step": 72509
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.680126377509669e-07,
+      "loss": 0.6665,
+      "step": 72510
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.679865367862268e-07,
+      "loss": 0.7363,
+      "step": 72511
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.679604360756939e-07,
+      "loss": 0.6523,
+      "step": 72512
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.679343356193885e-07,
+      "loss": 0.7402,
+      "step": 72513
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.679082354173307e-07,
+      "loss": 0.6006,
+      "step": 72514
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.678821354695399e-07,
+      "loss": 0.7002,
+      "step": 72515
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.678560357760367e-07,
+      "loss": 0.708,
+      "step": 72516
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.678299363368407e-07,
+      "loss": 0.7109,
+      "step": 72517
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.678038371519721e-07,
+      "loss": 0.5308,
+      "step": 72518
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.677777382214508e-07,
+      "loss": 0.6621,
+      "step": 72519
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.67751639545297e-07,
+      "loss": 0.6455,
+      "step": 72520
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.677255411235303e-07,
+      "loss": 0.6104,
+      "step": 72521
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.67699442956171e-07,
+      "loss": 0.7207,
+      "step": 72522
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.676733450432388e-07,
+      "loss": 0.5693,
+      "step": 72523
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.676472473847542e-07,
+      "loss": 0.5435,
+      "step": 72524
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.676211499807367e-07,
+      "loss": 0.627,
+      "step": 72525
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.67595052831206e-07,
+      "loss": 0.6165,
+      "step": 72526
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.675689559361829e-07,
+      "loss": 0.7026,
+      "step": 72527
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.675428592956869e-07,
+      "loss": 0.4971,
+      "step": 72528
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.675167629097378e-07,
+      "loss": 0.6174,
+      "step": 72529
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.67490666778356e-07,
+      "loss": 0.6904,
+      "step": 72530
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.674645709015613e-07,
+      "loss": 0.6914,
+      "step": 72531
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.674384752793737e-07,
+      "loss": 0.5244,
+      "step": 72532
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.674123799118129e-07,
+      "loss": 0.5688,
+      "step": 72533
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.673862847988998e-07,
+      "loss": 0.499,
+      "step": 72534
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.673601899406533e-07,
+      "loss": 0.6362,
+      "step": 72535
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.673340953370939e-07,
+      "loss": 0.5764,
+      "step": 72536
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.673080009882414e-07,
+      "loss": 0.5898,
+      "step": 72537
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.672819068941158e-07,
+      "loss": 0.7119,
+      "step": 72538
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.672558130547372e-07,
+      "loss": 0.5725,
+      "step": 72539
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.672297194701255e-07,
+      "loss": 0.5645,
+      "step": 72540
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.672036261403009e-07,
+      "loss": 0.5492,
+      "step": 72541
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.671775330652829e-07,
+      "loss": 0.5366,
+      "step": 72542
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.671514402450919e-07,
+      "loss": 0.793,
+      "step": 72543
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.67125347679748e-07,
+      "loss": 0.4509,
+      "step": 72544
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.670992553692708e-07,
+      "loss": 0.6104,
+      "step": 72545
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.670731633136801e-07,
+      "loss": 0.5771,
+      "step": 72546
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.670470715129963e-07,
+      "loss": 0.626,
+      "step": 72547
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.670209799672393e-07,
+      "loss": 0.6348,
+      "step": 72548
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.66994888676429e-07,
+      "loss": 0.6279,
+      "step": 72549
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.669687976405853e-07,
+      "loss": 0.4951,
+      "step": 72550
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.669427068597284e-07,
+      "loss": 0.8623,
+      "step": 72551
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.66916616333878e-07,
+      "loss": 0.6885,
+      "step": 72552
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.668905260630544e-07,
+      "loss": 0.5444,
+      "step": 72553
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.668644360472778e-07,
+      "loss": 0.6519,
+      "step": 72554
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.668383462865671e-07,
+      "loss": 0.6846,
+      "step": 72555
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.668122567809433e-07,
+      "loss": 0.7666,
+      "step": 72556
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.667861675304258e-07,
+      "loss": 0.7012,
+      "step": 72557
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.667600785350351e-07,
+      "loss": 0.4573,
+      "step": 72558
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.667339897947906e-07,
+      "loss": 0.5539,
+      "step": 72559
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.667079013097127e-07,
+      "loss": 0.5806,
+      "step": 72560
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.666818130798211e-07,
+      "loss": 0.5981,
+      "step": 72561
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.666557251051361e-07,
+      "loss": 0.7324,
+      "step": 72562
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.666296373856772e-07,
+      "loss": 0.4966,
+      "step": 72563
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.666035499214655e-07,
+      "loss": 0.5232,
+      "step": 72564
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.665774627125193e-07,
+      "loss": 0.6934,
+      "step": 72565
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.665513757588597e-07,
+      "loss": 0.623,
+      "step": 72566
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.665252890605062e-07,
+      "loss": 0.5616,
+      "step": 72567
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.664992026174791e-07,
+      "loss": 0.5869,
+      "step": 72568
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.664731164297981e-07,
+      "loss": 0.4741,
+      "step": 72569
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.664470304974834e-07,
+      "loss": 0.4473,
+      "step": 72570
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.664209448205546e-07,
+      "loss": 0.5197,
+      "step": 72571
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.663948593990324e-07,
+      "loss": 0.623,
+      "step": 72572
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.663687742329361e-07,
+      "loss": 0.606,
+      "step": 72573
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.663426893222861e-07,
+      "loss": 0.6084,
+      "step": 72574
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.66316604667102e-07,
+      "loss": 0.6406,
+      "step": 72575
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.662905202674037e-07,
+      "loss": 0.5752,
+      "step": 72576
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.662644361232117e-07,
+      "loss": 0.5947,
+      "step": 72577
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.662383522345454e-07,
+      "loss": 0.7363,
+      "step": 72578
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.662122686014254e-07,
+      "loss": 0.6104,
+      "step": 72579
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.66186185223871e-07,
+      "loss": 0.6323,
+      "step": 72580
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.661601021019027e-07,
+      "loss": 0.7075,
+      "step": 72581
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.661340192355403e-07,
+      "loss": 0.5674,
+      "step": 72582
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.661079366248034e-07,
+      "loss": 0.7529,
+      "step": 72583
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.660818542697129e-07,
+      "loss": 0.6377,
+      "step": 72584
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.660557721702878e-07,
+      "loss": 0.6372,
+      "step": 72585
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.660296903265483e-07,
+      "loss": 0.623,
+      "step": 72586
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.660036087385146e-07,
+      "loss": 0.7812,
+      "step": 72587
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.659775274062065e-07,
+      "loss": 0.5723,
+      "step": 72588
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.659514463296443e-07,
+      "loss": 0.5391,
+      "step": 72589
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.659253655088475e-07,
+      "loss": 0.4863,
+      "step": 72590
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.658992849438363e-07,
+      "loss": 0.668,
+      "step": 72591
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.658732046346307e-07,
+      "loss": 0.8318,
+      "step": 72592
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.658471245812507e-07,
+      "loss": 0.7617,
+      "step": 72593
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.658210447837163e-07,
+      "loss": 0.6313,
+      "step": 72594
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.657949652420472e-07,
+      "loss": 0.6611,
+      "step": 72595
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.657688859562636e-07,
+      "loss": 0.5864,
+      "step": 72596
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.657428069263851e-07,
+      "loss": 0.6074,
+      "step": 72597
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.657167281524323e-07,
+      "loss": 0.6885,
+      "step": 72598
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.656906496344246e-07,
+      "loss": 0.6772,
+      "step": 72599
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.656645713723823e-07,
+      "loss": 0.7227,
+      "step": 72600
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.656384933663255e-07,
+      "loss": 0.7222,
+      "step": 72601
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.656124156162735e-07,
+      "loss": 0.667,
+      "step": 72602
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.65586338122247e-07,
+      "loss": 0.6816,
+      "step": 72603
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.655602608842658e-07,
+      "loss": 0.7705,
+      "step": 72604
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.655341839023494e-07,
+      "loss": 0.6299,
+      "step": 72605
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.655081071765183e-07,
+      "loss": 0.667,
+      "step": 72606
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.65482030706792e-07,
+      "loss": 0.6523,
+      "step": 72607
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.654559544931909e-07,
+      "loss": 0.6479,
+      "step": 72608
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.654298785357347e-07,
+      "loss": 0.7168,
+      "step": 72609
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.654038028344435e-07,
+      "loss": 0.6367,
+      "step": 72610
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.653777273893373e-07,
+      "loss": 0.5669,
+      "step": 72611
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.653516522004359e-07,
+      "loss": 0.5569,
+      "step": 72612
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.653255772677593e-07,
+      "loss": 0.5625,
+      "step": 72613
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.65299502591328e-07,
+      "loss": 0.6973,
+      "step": 72614
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.652734281711612e-07,
+      "loss": 0.5093,
+      "step": 72615
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.652473540072787e-07,
+      "loss": 0.7051,
+      "step": 72616
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.652212800997015e-07,
+      "loss": 0.5913,
+      "step": 72617
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.651952064484488e-07,
+      "loss": 0.5488,
+      "step": 72618
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.651691330535405e-07,
+      "loss": 0.7241,
+      "step": 72619
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.65143059914997e-07,
+      "loss": 0.4707,
+      "step": 72620
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.651169870328379e-07,
+      "loss": 0.5001,
+      "step": 72621
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.650909144070836e-07,
+      "loss": 0.5977,
+      "step": 72622
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.650648420377534e-07,
+      "loss": 0.6553,
+      "step": 72623
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.650387699248681e-07,
+      "loss": 0.6367,
+      "step": 72624
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.650126980684471e-07,
+      "loss": 0.3621,
+      "step": 72625
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.649866264685102e-07,
+      "loss": 0.6436,
+      "step": 72626
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.649605551250779e-07,
+      "loss": 0.5352,
+      "step": 72627
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.649344840381697e-07,
+      "loss": 0.4587,
+      "step": 72628
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.64908413207806e-07,
+      "loss": 0.7383,
+      "step": 72629
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.648823426340063e-07,
+      "loss": 0.6553,
+      "step": 72630
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.648562723167907e-07,
+      "loss": 0.7207,
+      "step": 72631
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.648302022561793e-07,
+      "loss": 0.6816,
+      "step": 72632
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.648041324521923e-07,
+      "loss": 0.4636,
+      "step": 72633
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.647780629048494e-07,
+      "loss": 0.6035,
+      "step": 72634
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.647519936141704e-07,
+      "loss": 0.73,
+      "step": 72635
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.647259245801751e-07,
+      "loss": 0.6123,
+      "step": 72636
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.64699855802884e-07,
+      "loss": 0.7598,
+      "step": 72637
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.646737872823166e-07,
+      "loss": 0.6348,
+      "step": 72638
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.646477190184932e-07,
+      "loss": 0.5322,
+      "step": 72639
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.646216510114335e-07,
+      "loss": 0.6738,
+      "step": 72640
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.645955832611579e-07,
+      "loss": 0.5957,
+      "step": 72641
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.645695157676856e-07,
+      "loss": 0.7646,
+      "step": 72642
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.645434485310373e-07,
+      "loss": 0.6631,
+      "step": 72643
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.645173815512328e-07,
+      "loss": 0.7188,
+      "step": 72644
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.644913148282916e-07,
+      "loss": 0.5757,
+      "step": 72645
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.644652483622342e-07,
+      "loss": 0.2648,
+      "step": 72646
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.6443918215308e-07,
+      "loss": 0.49,
+      "step": 72647
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.644131162008496e-07,
+      "loss": 0.6021,
+      "step": 72648
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.643870505055624e-07,
+      "loss": 0.8311,
+      "step": 72649
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.643609850672388e-07,
+      "loss": 0.6562,
+      "step": 72650
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.643349198858983e-07,
+      "loss": 0.6479,
+      "step": 72651
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.643088549615614e-07,
+      "loss": 0.6621,
+      "step": 72652
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.642827902942476e-07,
+      "loss": 0.6587,
+      "step": 72653
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.642567258839776e-07,
+      "loss": 0.6494,
+      "step": 72654
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.642306617307702e-07,
+      "loss": 0.5764,
+      "step": 72655
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.642045978346461e-07,
+      "loss": 0.7375,
+      "step": 72656
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.64178534195625e-07,
+      "loss": 0.6953,
+      "step": 72657
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.641524708137271e-07,
+      "loss": 0.5991,
+      "step": 72658
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.641264076889718e-07,
+      "loss": 0.6328,
+      "step": 72659
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.6410034482138e-07,
+      "loss": 0.6348,
+      "step": 72660
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.640742822109707e-07,
+      "loss": 0.5576,
+      "step": 72661
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.640482198577646e-07,
+      "loss": 0.2839,
+      "step": 72662
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.64022157761781e-07,
+      "loss": 0.5571,
+      "step": 72663
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.639960959230407e-07,
+      "loss": 0.6758,
+      "step": 72664
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.639700343415629e-07,
+      "loss": 0.5918,
+      "step": 72665
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.639439730173677e-07,
+      "loss": 0.7705,
+      "step": 72666
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.639179119504752e-07,
+      "loss": 0.7188,
+      "step": 72667
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.638918511409051e-07,
+      "loss": 0.6465,
+      "step": 72668
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.638657905886778e-07,
+      "loss": 0.583,
+      "step": 72669
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.638397302938128e-07,
+      "loss": 0.7061,
+      "step": 72670
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.638136702563306e-07,
+      "loss": 0.7637,
+      "step": 72671
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.637876104762507e-07,
+      "loss": 0.5796,
+      "step": 72672
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.637615509535928e-07,
+      "loss": 0.6377,
+      "step": 72673
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.637354916883778e-07,
+      "loss": 0.583,
+      "step": 72674
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63709432680625e-07,
+      "loss": 0.3926,
+      "step": 72675
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63683373930354e-07,
+      "loss": 0.5391,
+      "step": 72676
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.636573154375854e-07,
+      "loss": 0.6299,
+      "step": 72677
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.636312572023388e-07,
+      "loss": 0.4751,
+      "step": 72678
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.636051992246345e-07,
+      "loss": 0.7344,
+      "step": 72679
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63579141504492e-07,
+      "loss": 0.5386,
+      "step": 72680
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.635530840419317e-07,
+      "loss": 0.6992,
+      "step": 72681
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63527026836973e-07,
+      "loss": 0.5151,
+      "step": 72682
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.635009698896366e-07,
+      "loss": 0.6387,
+      "step": 72683
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63474913199942e-07,
+      "loss": 0.5312,
+      "step": 72684
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.634488567679088e-07,
+      "loss": 0.6289,
+      "step": 72685
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.634228005935578e-07,
+      "loss": 0.5479,
+      "step": 72686
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.633967446769081e-07,
+      "loss": 0.7637,
+      "step": 72687
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.633706890179802e-07,
+      "loss": 0.5801,
+      "step": 72688
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.633446336167938e-07,
+      "loss": 0.7793,
+      "step": 72689
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.63318578473369e-07,
+      "loss": 0.502,
+      "step": 72690
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.632925235877258e-07,
+      "loss": 0.5989,
+      "step": 72691
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.632664689598839e-07,
+      "loss": 0.5151,
+      "step": 72692
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.632404145898634e-07,
+      "loss": 0.5405,
+      "step": 72693
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.632143604776845e-07,
+      "loss": 0.7969,
+      "step": 72694
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.631883066233666e-07,
+      "loss": 0.5181,
+      "step": 72695
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.6316225302693e-07,
+      "loss": 0.5952,
+      "step": 72696
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.631361996883944e-07,
+      "loss": 0.668,
+      "step": 72697
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.631101466077799e-07,
+      "loss": 0.5503,
+      "step": 72698
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.630840937851065e-07,
+      "loss": 0.6738,
+      "step": 72699
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.630580412203945e-07,
+      "loss": 0.6299,
+      "step": 72700
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.630319889136629e-07,
+      "loss": 0.5896,
+      "step": 72701
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.630059368649326e-07,
+      "loss": 0.6553,
+      "step": 72702
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.629798850742234e-07,
+      "loss": 0.4983,
+      "step": 72703
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.629538335415545e-07,
+      "loss": 0.6934,
+      "step": 72704
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.629277822669466e-07,
+      "loss": 0.6016,
+      "step": 72705
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.629017312504192e-07,
+      "loss": 0.6914,
+      "step": 72706
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.628756804919926e-07,
+      "loss": 0.5796,
+      "step": 72707
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.628496299916867e-07,
+      "loss": 0.6221,
+      "step": 72708
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.628235797495209e-07,
+      "loss": 0.5127,
+      "step": 72709
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62797529765516e-07,
+      "loss": 0.7009,
+      "step": 72710
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.627714800396913e-07,
+      "loss": 0.7715,
+      "step": 72711
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.627454305720672e-07,
+      "loss": 0.5884,
+      "step": 72712
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.627193813626634e-07,
+      "loss": 0.6738,
+      "step": 72713
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.626933324114998e-07,
+      "loss": 0.3793,
+      "step": 72714
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.626672837185964e-07,
+      "loss": 0.7051,
+      "step": 72715
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62641235283973e-07,
+      "loss": 0.53,
+      "step": 72716
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.626151871076498e-07,
+      "loss": 0.6099,
+      "step": 72717
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.625891391896466e-07,
+      "loss": 0.5244,
+      "step": 72718
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.625630915299834e-07,
+      "loss": 0.5703,
+      "step": 72719
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.625370441286799e-07,
+      "loss": 0.6113,
+      "step": 72720
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.625109969857567e-07,
+      "loss": 0.7314,
+      "step": 72721
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62484950101233e-07,
+      "loss": 0.6914,
+      "step": 72722
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.624589034751294e-07,
+      "loss": 0.6479,
+      "step": 72723
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.624328571074654e-07,
+      "loss": 0.5288,
+      "step": 72724
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62406810998261e-07,
+      "loss": 0.7266,
+      "step": 72725
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62380765147536e-07,
+      "loss": 0.6777,
+      "step": 72726
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.623547195553105e-07,
+      "loss": 0.7334,
+      "step": 72727
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.623286742216045e-07,
+      "loss": 0.6436,
+      "step": 72728
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.623026291464381e-07,
+      "loss": 0.6138,
+      "step": 72729
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.622765843298308e-07,
+      "loss": 0.5007,
+      "step": 72730
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62250539771803e-07,
+      "loss": 0.5212,
+      "step": 72731
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.622244954723743e-07,
+      "loss": 0.5923,
+      "step": 72732
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.621984514315651e-07,
+      "loss": 0.6094,
+      "step": 72733
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.621724076493947e-07,
+      "loss": 0.7383,
+      "step": 72734
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.621463641258833e-07,
+      "loss": 0.52,
+      "step": 72735
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62120320861051e-07,
+      "loss": 0.7446,
+      "step": 72736
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.620942778549176e-07,
+      "loss": 0.6765,
+      "step": 72737
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.62068235107503e-07,
+      "loss": 0.6147,
+      "step": 72738
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.620421926188274e-07,
+      "loss": 0.6689,
+      "step": 72739
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.620161503889105e-07,
+      "loss": 0.6562,
+      "step": 72740
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.619901084177721e-07,
+      "loss": 0.666,
+      "step": 72741
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.619640667054326e-07,
+      "loss": 0.8018,
+      "step": 72742
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.619380252519118e-07,
+      "loss": 0.5845,
+      "step": 72743
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.619119840572294e-07,
+      "loss": 0.5898,
+      "step": 72744
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.618859431214052e-07,
+      "loss": 0.4019,
+      "step": 72745
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.618599024444596e-07,
+      "loss": 0.6738,
+      "step": 72746
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.618338620264122e-07,
+      "loss": 0.5337,
+      "step": 72747
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.61807821867283e-07,
+      "loss": 0.6138,
+      "step": 72748
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.61781781967092e-07,
+      "loss": 0.6211,
+      "step": 72749
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.617557423258593e-07,
+      "loss": 0.6919,
+      "step": 72750
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.617297029436045e-07,
+      "loss": 0.6118,
+      "step": 72751
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.617036638203479e-07,
+      "loss": 0.6494,
+      "step": 72752
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.616776249561094e-07,
+      "loss": 0.6504,
+      "step": 72753
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.616515863509085e-07,
+      "loss": 0.6572,
+      "step": 72754
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.616255480047655e-07,
+      "loss": 0.7231,
+      "step": 72755
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.615995099177001e-07,
+      "loss": 0.7354,
+      "step": 72756
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.615734720897326e-07,
+      "loss": 0.5935,
+      "step": 72757
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.615474345208824e-07,
+      "loss": 0.6182,
+      "step": 72758
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.615213972111701e-07,
+      "loss": 0.6343,
+      "step": 72759
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.614953601606153e-07,
+      "loss": 0.5747,
+      "step": 72760
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.614693233692379e-07,
+      "loss": 0.6011,
+      "step": 72761
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.614432868370577e-07,
+      "loss": 0.8232,
+      "step": 72762
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.614172505640954e-07,
+      "loss": 0.5654,
+      "step": 72763
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 6.613912145503697e-07,
+      "loss": 0.4834,
+      "step": 72764
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.613651787959016e-07,
+      "loss": 0.5537,
+      "step": 72765
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.613391433007103e-07,
+      "loss": 0.5396,
+      "step": 72766
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.613131080648161e-07,
+      "loss": 0.6924,
+      "step": 72767
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.612870730882387e-07,
+      "loss": 0.6265,
+      "step": 72768
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.612610383709987e-07,
+      "loss": 0.5967,
+      "step": 72769
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.61235003913115e-07,
+      "loss": 0.5635,
+      "step": 72770
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.612089697146084e-07,
+      "loss": 0.6448,
+      "step": 72771
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.611829357754983e-07,
+      "loss": 0.8379,
+      "step": 72772
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.611569020958054e-07,
+      "loss": 0.5952,
+      "step": 72773
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.611308686755488e-07,
+      "loss": 0.7109,
+      "step": 72774
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.611048355147484e-07,
+      "loss": 0.7061,
+      "step": 72775
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.610788026134247e-07,
+      "loss": 0.5256,
+      "step": 72776
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.610527699715973e-07,
+      "loss": 0.4905,
+      "step": 72777
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.610267375892862e-07,
+      "loss": 0.6729,
+      "step": 72778
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.610007054665112e-07,
+      "loss": 0.5112,
+      "step": 72779
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.609746736032927e-07,
+      "loss": 0.5962,
+      "step": 72780
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.609486419996501e-07,
+      "loss": 0.5493,
+      "step": 72781
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.609226106556037e-07,
+      "loss": 0.6641,
+      "step": 72782
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.608965795711733e-07,
+      "loss": 0.5388,
+      "step": 72783
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.608705487463788e-07,
+      "loss": 0.8301,
+      "step": 72784
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.608445181812399e-07,
+      "loss": 0.7266,
+      "step": 72785
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.608184878757767e-07,
+      "loss": 0.5632,
+      "step": 72786
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.607924578300092e-07,
+      "loss": 0.7432,
+      "step": 72787
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.607664280439576e-07,
+      "loss": 0.6426,
+      "step": 72788
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.607403985176412e-07,
+      "loss": 0.6196,
+      "step": 72789
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.607143692510806e-07,
+      "loss": 0.5183,
+      "step": 72790
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.606883402442951e-07,
+      "loss": 0.7227,
+      "step": 72791
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.606623114973052e-07,
+      "loss": 0.5376,
+      "step": 72792
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.606362830101307e-07,
+      "loss": 0.6777,
+      "step": 72793
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.606102547827911e-07,
+      "loss": 0.6411,
+      "step": 72794
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.605842268153068e-07,
+      "loss": 0.5161,
+      "step": 72795
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.605581991076972e-07,
+      "loss": 0.6943,
+      "step": 72796
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.60532171659983e-07,
+      "loss": 0.5386,
+      "step": 72797
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.605061444721835e-07,
+      "loss": 0.3782,
+      "step": 72798
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.604801175443188e-07,
+      "loss": 0.5801,
+      "step": 72799
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.604540908764089e-07,
+      "loss": 0.7295,
+      "step": 72800
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.604280644684736e-07,
+      "loss": 0.6445,
+      "step": 72801
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.604020383205331e-07,
+      "loss": 0.7437,
+      "step": 72802
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.603760124326074e-07,
+      "loss": 0.6641,
+      "step": 72803
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.603499868047157e-07,
+      "loss": 0.71,
+      "step": 72804
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.603239614368787e-07,
+      "loss": 0.615,
+      "step": 72805
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.602979363291157e-07,
+      "loss": 0.623,
+      "step": 72806
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.602719114814472e-07,
+      "loss": 0.535,
+      "step": 72807
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.602458868938927e-07,
+      "loss": 0.7168,
+      "step": 72808
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.602198625664723e-07,
+      "loss": 0.6885,
+      "step": 72809
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.601938384992061e-07,
+      "loss": 0.6353,
+      "step": 72810
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.601678146921137e-07,
+      "loss": 0.4685,
+      "step": 72811
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.601417911452151e-07,
+      "loss": 0.5566,
+      "step": 72812
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.601157678585307e-07,
+      "loss": 0.7188,
+      "step": 72813
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.6008974483208e-07,
+      "loss": 0.6572,
+      "step": 72814
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.600637220658828e-07,
+      "loss": 0.7656,
+      "step": 72815
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.600376995599591e-07,
+      "loss": 0.5088,
+      "step": 72816
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.600116773143288e-07,
+      "loss": 0.7451,
+      "step": 72817
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.599856553290119e-07,
+      "loss": 0.6089,
+      "step": 72818
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.599596336040287e-07,
+      "loss": 0.4497,
+      "step": 72819
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.599336121393983e-07,
+      "loss": 0.479,
+      "step": 72820
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.599075909351415e-07,
+      "loss": 0.4338,
+      "step": 72821
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.598815699912776e-07,
+      "loss": 0.7764,
+      "step": 72822
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.598555493078271e-07,
+      "loss": 0.5742,
+      "step": 72823
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.598295288848092e-07,
+      "loss": 0.7334,
+      "step": 72824
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.598035087222442e-07,
+      "loss": 0.5493,
+      "step": 72825
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.597774888201521e-07,
+      "loss": 0.6787,
+      "step": 72826
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.597514691785527e-07,
+      "loss": 0.585,
+      "step": 72827
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.597254497974661e-07,
+      "loss": 0.5352,
+      "step": 72828
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.596994306769118e-07,
+      "loss": 0.6943,
+      "step": 72829
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.596734118169102e-07,
+      "loss": 0.6255,
+      "step": 72830
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.596473932174809e-07,
+      "loss": 0.6113,
+      "step": 72831
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.596213748786441e-07,
+      "loss": 0.5369,
+      "step": 72832
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.595953568004198e-07,
+      "loss": 0.6982,
+      "step": 72833
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.595693389828274e-07,
+      "loss": 0.6055,
+      "step": 72834
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.595433214258869e-07,
+      "loss": 0.5996,
+      "step": 72835
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.595173041296187e-07,
+      "loss": 0.5496,
+      "step": 72836
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.594912870940421e-07,
+      "loss": 0.623,
+      "step": 72837
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.594652703191777e-07,
+      "loss": 0.666,
+      "step": 72838
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.594392538050449e-07,
+      "loss": 0.5923,
+      "step": 72839
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.594132375516639e-07,
+      "loss": 0.749,
+      "step": 72840
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.593872215590544e-07,
+      "loss": 0.5527,
+      "step": 72841
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.593612058272366e-07,
+      "loss": 0.6196,
+      "step": 72842
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.593351903562306e-07,
+      "loss": 0.7197,
+      "step": 72843
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.593091751460553e-07,
+      "loss": 0.6938,
+      "step": 72844
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.592831601967316e-07,
+      "loss": 0.5894,
+      "step": 72845
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.592571455082791e-07,
+      "loss": 0.6616,
+      "step": 72846
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.592311310807178e-07,
+      "loss": 0.6553,
+      "step": 72847
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.592051169140673e-07,
+      "loss": 0.5273,
+      "step": 72848
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.591791030083481e-07,
+      "loss": 0.9189,
+      "step": 72849
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.591530893635795e-07,
+      "loss": 0.6792,
+      "step": 72850
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.59127075979782e-07,
+      "loss": 0.5725,
+      "step": 72851
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.591010628569749e-07,
+      "loss": 0.5498,
+      "step": 72852
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.590750499951792e-07,
+      "loss": 0.7207,
+      "step": 72853
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.590490373944133e-07,
+      "loss": 0.4761,
+      "step": 72854
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.590230250546982e-07,
+      "loss": 0.5864,
+      "step": 72855
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.589970129760531e-07,
+      "loss": 0.6797,
+      "step": 72856
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.589710011584988e-07,
+      "loss": 0.4883,
+      "step": 72857
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.589449896020543e-07,
+      "loss": 0.6943,
+      "step": 72858
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.589189783067401e-07,
+      "loss": 0.6504,
+      "step": 72859
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.588929672725759e-07,
+      "loss": 0.7764,
+      "step": 72860
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.588669564995819e-07,
+      "loss": 0.5869,
+      "step": 72861
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.588409459877776e-07,
+      "loss": 0.666,
+      "step": 72862
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.588149357371833e-07,
+      "loss": 0.5303,
+      "step": 72863
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.587889257478186e-07,
+      "loss": 0.6328,
+      "step": 72864
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.587629160197034e-07,
+      "loss": 0.6641,
+      "step": 72865
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.587369065528579e-07,
+      "loss": 0.6387,
+      "step": 72866
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.587108973473017e-07,
+      "loss": 0.5205,
+      "step": 72867
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.586848884030549e-07,
+      "loss": 0.6689,
+      "step": 72868
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.586588797201373e-07,
+      "loss": 0.7471,
+      "step": 72869
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.586328712985691e-07,
+      "loss": 0.5962,
+      "step": 72870
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.586068631383701e-07,
+      "loss": 0.4563,
+      "step": 72871
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.5858085523956e-07,
+      "loss": 0.4651,
+      "step": 72872
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.585548476021589e-07,
+      "loss": 0.5105,
+      "step": 72873
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.585288402261868e-07,
+      "loss": 0.7266,
+      "step": 72874
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.585028331116631e-07,
+      "loss": 0.6533,
+      "step": 72875
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.584768262586082e-07,
+      "loss": 0.464,
+      "step": 72876
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.584508196670419e-07,
+      "loss": 0.6216,
+      "step": 72877
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.584248133369841e-07,
+      "loss": 0.7295,
+      "step": 72878
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.583988072684546e-07,
+      "loss": 0.605,
+      "step": 72879
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.583728014614737e-07,
+      "loss": 0.7026,
+      "step": 72880
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.583467959160607e-07,
+      "loss": 0.7773,
+      "step": 72881
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.583207906322361e-07,
+      "loss": 0.6455,
+      "step": 72882
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.582947856100198e-07,
+      "loss": 0.5537,
+      "step": 72883
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.582687808494312e-07,
+      "loss": 0.665,
+      "step": 72884
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.582427763504905e-07,
+      "loss": 0.6191,
+      "step": 72885
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.582167721132174e-07,
+      "loss": 0.4097,
+      "step": 72886
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.581907681376322e-07,
+      "loss": 0.7061,
+      "step": 72887
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.581647644237547e-07,
+      "loss": 0.6113,
+      "step": 72888
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.581387609716044e-07,
+      "loss": 0.6709,
+      "step": 72889
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.581127577812018e-07,
+      "loss": 0.7578,
+      "step": 72890
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.580867548525664e-07,
+      "loss": 0.5547,
+      "step": 72891
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.580607521857183e-07,
+      "loss": 0.4016,
+      "step": 72892
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.580347497806776e-07,
+      "loss": 0.5879,
+      "step": 72893
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.580087476374636e-07,
+      "loss": 0.5601,
+      "step": 72894
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.579827457560969e-07,
+      "loss": 0.7266,
+      "step": 72895
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.579567441365968e-07,
+      "loss": 0.6934,
+      "step": 72896
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.579307427789835e-07,
+      "loss": 0.625,
+      "step": 72897
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.579047416832769e-07,
+      "loss": 0.7012,
+      "step": 72898
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.578787408494971e-07,
+      "loss": 0.4702,
+      "step": 72899
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.578527402776636e-07,
+      "loss": 0.8262,
+      "step": 72900
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.578267399677967e-07,
+      "loss": 0.7305,
+      "step": 72901
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.578007399199159e-07,
+      "loss": 0.6382,
+      "step": 72902
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.577747401340418e-07,
+      "loss": 0.7529,
+      "step": 72903
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.577487406101937e-07,
+      "loss": 0.8042,
+      "step": 72904
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.577227413483915e-07,
+      "loss": 0.4878,
+      "step": 72905
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.576967423486553e-07,
+      "loss": 0.4431,
+      "step": 72906
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.576707436110049e-07,
+      "loss": 0.7524,
+      "step": 72907
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.576447451354602e-07,
+      "loss": 0.6484,
+      "step": 72908
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.576187469220413e-07,
+      "loss": 0.4956,
+      "step": 72909
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.575927489707679e-07,
+      "loss": 0.584,
+      "step": 72910
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.575667512816601e-07,
+      "loss": 0.6094,
+      "step": 72911
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.575407538547375e-07,
+      "loss": 0.8506,
+      "step": 72912
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.575147566900207e-07,
+      "loss": 0.4673,
+      "step": 72913
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.574887597875287e-07,
+      "loss": 0.749,
+      "step": 72914
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.574627631472817e-07,
+      "loss": 0.5801,
+      "step": 72915
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.574367667693e-07,
+      "loss": 0.668,
+      "step": 72916
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.57410770653603e-07,
+      "loss": 0.6904,
+      "step": 72917
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.573847748002111e-07,
+      "loss": 0.7422,
+      "step": 72918
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.573587792091435e-07,
+      "loss": 0.6431,
+      "step": 72919
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.573327838804208e-07,
+      "loss": 0.5645,
+      "step": 72920
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.573067888140624e-07,
+      "loss": 0.7412,
+      "step": 72921
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.572807940100887e-07,
+      "loss": 0.8018,
+      "step": 72922
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.572547994685196e-07,
+      "loss": 0.6895,
+      "step": 72923
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.572288051893744e-07,
+      "loss": 0.6777,
+      "step": 72924
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.572028111726732e-07,
+      "loss": 0.7754,
+      "step": 72925
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.571768174184363e-07,
+      "loss": 0.6416,
+      "step": 72926
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.571508239266832e-07,
+      "loss": 0.7134,
+      "step": 72927
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.57124830697434e-07,
+      "loss": 0.4634,
+      "step": 72928
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.570988377307084e-07,
+      "loss": 0.7559,
+      "step": 72929
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.570728450265266e-07,
+      "loss": 0.5898,
+      "step": 72930
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.570468525849083e-07,
+      "loss": 0.6377,
+      "step": 72931
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.570208604058736e-07,
+      "loss": 0.668,
+      "step": 72932
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.569948684894425e-07,
+      "loss": 0.6865,
+      "step": 72933
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.569688768356341e-07,
+      "loss": 0.5981,
+      "step": 72934
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.56942885444469e-07,
+      "loss": 0.6304,
+      "step": 72935
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.56916894315967e-07,
+      "loss": 0.7656,
+      "step": 72936
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.56890903450148e-07,
+      "loss": 0.5898,
+      "step": 72937
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.568649128470318e-07,
+      "loss": 0.5591,
+      "step": 72938
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.568389225066384e-07,
+      "loss": 0.7515,
+      "step": 72939
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.568129324289875e-07,
+      "loss": 0.6265,
+      "step": 72940
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.567869426140995e-07,
+      "loss": 0.7388,
+      "step": 72941
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.567609530619935e-07,
+      "loss": 0.5747,
+      "step": 72942
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.567349637726906e-07,
+      "loss": 0.541,
+      "step": 72943
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.567089747462095e-07,
+      "loss": 0.6079,
+      "step": 72944
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.566829859825706e-07,
+      "loss": 0.6709,
+      "step": 72945
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.566569974817935e-07,
+      "loss": 0.7188,
+      "step": 72946
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.566310092438987e-07,
+      "loss": 0.6714,
+      "step": 72947
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.566050212689054e-07,
+      "loss": 0.6641,
+      "step": 72948
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.565790335568342e-07,
+      "loss": 0.6079,
+      "step": 72949
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.565530461077042e-07,
+      "loss": 0.7339,
+      "step": 72950
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.565270589215362e-07,
+      "loss": 0.6943,
+      "step": 72951
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.565010719983493e-07,
+      "loss": 0.5405,
+      "step": 72952
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.564750853381642e-07,
+      "loss": 0.6851,
+      "step": 72953
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.564490989410001e-07,
+      "loss": 0.5151,
+      "step": 72954
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.564231128068769e-07,
+      "loss": 0.6797,
+      "step": 72955
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.563971269358148e-07,
+      "loss": 0.7119,
+      "step": 72956
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.563711413278336e-07,
+      "loss": 0.8271,
+      "step": 72957
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.563451559829534e-07,
+      "loss": 0.4683,
+      "step": 72958
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.563191709011937e-07,
+      "loss": 0.6396,
+      "step": 72959
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.562931860825749e-07,
+      "loss": 0.6416,
+      "step": 72960
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.562672015271164e-07,
+      "loss": 0.8184,
+      "step": 72961
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.562412172348383e-07,
+      "loss": 0.731,
+      "step": 72962
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.562152332057608e-07,
+      "loss": 0.5073,
+      "step": 72963
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.561892494399033e-07,
+      "loss": 0.6816,
+      "step": 72964
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.561632659372856e-07,
+      "loss": 0.7158,
+      "step": 72965
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.561372826979281e-07,
+      "loss": 0.6519,
+      "step": 72966
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.561112997218503e-07,
+      "loss": 0.6904,
+      "step": 72967
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.560853170090725e-07,
+      "loss": 0.7197,
+      "step": 72968
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.560593345596142e-07,
+      "loss": 0.7227,
+      "step": 72969
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.560333523734956e-07,
+      "loss": 0.519,
+      "step": 72970
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.560073704507363e-07,
+      "loss": 0.4561,
+      "step": 72971
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.559813887913565e-07,
+      "loss": 0.6147,
+      "step": 72972
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.559554073953761e-07,
+      "loss": 0.5532,
+      "step": 72973
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.559294262628144e-07,
+      "loss": 0.5171,
+      "step": 72974
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.559034453936921e-07,
+      "loss": 0.3677,
+      "step": 72975
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.558774647880282e-07,
+      "loss": 0.7568,
+      "step": 72976
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.558514844458434e-07,
+      "loss": 0.6709,
+      "step": 72977
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.558255043671575e-07,
+      "loss": 0.4902,
+      "step": 72978
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.557995245519899e-07,
+      "loss": 0.6021,
+      "step": 72979
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.557735450003607e-07,
+      "loss": 0.667,
+      "step": 72980
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.5574756571229e-07,
+      "loss": 0.6284,
+      "step": 72981
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.557215866877976e-07,
+      "loss": 0.751,
+      "step": 72982
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.556956079269036e-07,
+      "loss": 0.6172,
+      "step": 72983
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.556696294296274e-07,
+      "loss": 0.7051,
+      "step": 72984
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.556436511959891e-07,
+      "loss": 0.6484,
+      "step": 72985
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.556176732260085e-07,
+      "loss": 0.5757,
+      "step": 72986
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.555916955197058e-07,
+      "loss": 0.6543,
+      "step": 72987
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.555657180771006e-07,
+      "loss": 0.4897,
+      "step": 72988
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.555397408982129e-07,
+      "loss": 0.7109,
+      "step": 72989
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.555137639830626e-07,
+      "loss": 0.7246,
+      "step": 72990
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.554877873316696e-07,
+      "loss": 0.606,
+      "step": 72991
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.554618109440536e-07,
+      "loss": 0.5371,
+      "step": 72992
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.554358348202351e-07,
+      "loss": 0.6246,
+      "step": 72993
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.554098589602333e-07,
+      "loss": 0.6797,
+      "step": 72994
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.553838833640684e-07,
+      "loss": 0.7607,
+      "step": 72995
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.553579080317598e-07,
+      "loss": 0.583,
+      "step": 72996
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.553319329633282e-07,
+      "loss": 0.7764,
+      "step": 72997
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.553059581587929e-07,
+      "loss": 0.5552,
+      "step": 72998
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.552799836181741e-07,
+      "loss": 0.6162,
+      "step": 72999
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.552540093414913e-07,
+      "loss": 0.5366,
+      "step": 73000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.552280353287649e-07,
+      "loss": 0.6738,
+      "step": 73001
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.552020615800149e-07,
+      "loss": 0.6426,
+      "step": 73002
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.551760880952603e-07,
+      "loss": 0.7192,
+      "step": 73003
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.551501148745218e-07,
+      "loss": 0.5874,
+      "step": 73004
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.551241419178186e-07,
+      "loss": 0.6826,
+      "step": 73005
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.550981692251713e-07,
+      "loss": 0.4592,
+      "step": 73006
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.550721967965993e-07,
+      "loss": 0.4948,
+      "step": 73007
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.550462246321228e-07,
+      "loss": 0.6904,
+      "step": 73008
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.550202527317615e-07,
+      "loss": 0.6895,
+      "step": 73009
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.549942810955353e-07,
+      "loss": 0.5996,
+      "step": 73010
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.549683097234641e-07,
+      "loss": 0.6294,
+      "step": 73011
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.549423386155684e-07,
+      "loss": 0.7188,
+      "step": 73012
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.549163677718668e-07,
+      "loss": 0.7217,
+      "step": 73013
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.5489039719238e-07,
+      "loss": 0.6626,
+      "step": 73014
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.548644268771276e-07,
+      "loss": 0.5952,
+      "step": 73015
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.548384568261299e-07,
+      "loss": 0.7178,
+      "step": 73016
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.548124870394062e-07,
+      "loss": 0.7344,
+      "step": 73017
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.547865175169771e-07,
+      "loss": 0.6978,
+      "step": 73018
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.547605482588618e-07,
+      "loss": 0.5879,
+      "step": 73019
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.547345792650807e-07,
+      "loss": 0.6357,
+      "step": 73020
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.547086105356532e-07,
+      "loss": 0.6182,
+      "step": 73021
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.546826420705999e-07,
+      "loss": 0.5808,
+      "step": 73022
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.546566738699399e-07,
+      "loss": 0.6343,
+      "step": 73023
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.546307059336933e-07,
+      "loss": 0.6279,
+      "step": 73024
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.546047382618803e-07,
+      "loss": 0.7939,
+      "step": 73025
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.545787708545204e-07,
+      "loss": 0.6982,
+      "step": 73026
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.545528037116338e-07,
+      "loss": 0.5859,
+      "step": 73027
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.5452683683324e-07,
+      "loss": 0.6133,
+      "step": 73028
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.545008702193594e-07,
+      "loss": 0.7031,
+      "step": 73029
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.544749038700112e-07,
+      "loss": 0.5161,
+      "step": 73030
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.544489377852162e-07,
+      "loss": 0.6147,
+      "step": 73031
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.544229719649938e-07,
+      "loss": 0.6506,
+      "step": 73032
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.543970064093636e-07,
+      "loss": 0.7227,
+      "step": 73033
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.543710411183457e-07,
+      "loss": 0.6455,
+      "step": 73034
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.543450760919599e-07,
+      "loss": 0.5193,
+      "step": 73035
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.543191113302261e-07,
+      "loss": 0.5071,
+      "step": 73036
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.542931468331645e-07,
+      "loss": 0.6646,
+      "step": 73037
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.542671826007946e-07,
+      "loss": 0.5537,
+      "step": 73038
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.542412186331365e-07,
+      "loss": 0.6699,
+      "step": 73039
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.542152549302098e-07,
+      "loss": 0.7705,
+      "step": 73040
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.541892914920348e-07,
+      "loss": 0.6011,
+      "step": 73041
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.541633283186315e-07,
+      "loss": 0.5327,
+      "step": 73042
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.541373654100189e-07,
+      "loss": 0.7324,
+      "step": 73043
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.541114027662178e-07,
+      "loss": 0.7402,
+      "step": 73044
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.540854403872472e-07,
+      "loss": 0.7305,
+      "step": 73045
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.540594782731279e-07,
+      "loss": 0.6787,
+      "step": 73046
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.54033516423879e-07,
+      "loss": 0.6987,
+      "step": 73047
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.54007554839521e-07,
+      "loss": 0.5684,
+      "step": 73048
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.539815935200733e-07,
+      "loss": 0.6572,
+      "step": 73049
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.539556324655562e-07,
+      "loss": 0.6123,
+      "step": 73050
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.539296716759894e-07,
+      "loss": 0.7256,
+      "step": 73051
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.539037111513929e-07,
+      "loss": 0.7754,
+      "step": 73052
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.53877750891786e-07,
+      "loss": 0.5615,
+      "step": 73053
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.538517908971892e-07,
+      "loss": 0.6025,
+      "step": 73054
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.538258311676219e-07,
+      "loss": 0.7383,
+      "step": 73055
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.537998717031045e-07,
+      "loss": 0.6543,
+      "step": 73056
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.537739125036564e-07,
+      "loss": 0.5923,
+      "step": 73057
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.53747953569298e-07,
+      "loss": 0.5806,
+      "step": 73058
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.537219949000486e-07,
+      "loss": 0.6777,
+      "step": 73059
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.536960364959285e-07,
+      "loss": 0.7686,
+      "step": 73060
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.536700783569572e-07,
+      "loss": 0.6387,
+      "step": 73061
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.536441204831554e-07,
+      "loss": 0.6543,
+      "step": 73062
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.536181628745418e-07,
+      "loss": 0.5049,
+      "step": 73063
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.53592205531137e-07,
+      "loss": 0.7529,
+      "step": 73064
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.535662484529608e-07,
+      "loss": 0.751,
+      "step": 73065
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.535402916400328e-07,
+      "loss": 0.4752,
+      "step": 73066
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.535143350923732e-07,
+      "loss": 0.6128,
+      "step": 73067
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.534883788100019e-07,
+      "loss": 0.7524,
+      "step": 73068
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.534624227929382e-07,
+      "loss": 0.6553,
+      "step": 73069
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.534364670412026e-07,
+      "loss": 0.6748,
+      "step": 73070
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.534105115548147e-07,
+      "loss": 0.4878,
+      "step": 73071
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.533845563337949e-07,
+      "loss": 0.6279,
+      "step": 73072
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.533586013781622e-07,
+      "loss": 0.5688,
+      "step": 73073
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.533326466879368e-07,
+      "loss": 0.6416,
+      "step": 73074
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.533066922631388e-07,
+      "loss": 0.4858,
+      "step": 73075
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.532807381037879e-07,
+      "loss": 0.8701,
+      "step": 73076
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.53254784209904e-07,
+      "loss": 0.6226,
+      "step": 73077
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.532288305815068e-07,
+      "loss": 0.6055,
+      "step": 73078
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.532028772186166e-07,
+      "loss": 0.5859,
+      "step": 73079
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.531769241212526e-07,
+      "loss": 0.5715,
+      "step": 73080
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.531509712894357e-07,
+      "loss": 0.6802,
+      "step": 73081
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.531250187231849e-07,
+      "loss": 0.3105,
+      "step": 73082
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.530990664225202e-07,
+      "loss": 0.5066,
+      "step": 73083
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.530731143874617e-07,
+      "loss": 0.5715,
+      "step": 73084
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.530471626180291e-07,
+      "loss": 0.793,
+      "step": 73085
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.530212111142423e-07,
+      "loss": 0.5571,
+      "step": 73086
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.529952598761213e-07,
+      "loss": 0.5376,
+      "step": 73087
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.529693089036858e-07,
+      "loss": 0.8291,
+      "step": 73088
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.529433581969558e-07,
+      "loss": 0.5723,
+      "step": 73089
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.529174077559509e-07,
+      "loss": 0.521,
+      "step": 73090
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.528914575806914e-07,
+      "loss": 0.5474,
+      "step": 73091
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.528655076711973e-07,
+      "loss": 0.6826,
+      "step": 73092
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.528395580274875e-07,
+      "loss": 0.7871,
+      "step": 73093
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.528136086495829e-07,
+      "loss": 0.7646,
+      "step": 73094
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.527876595375026e-07,
+      "loss": 0.5576,
+      "step": 73095
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.52761710691267e-07,
+      "loss": 0.6621,
+      "step": 73096
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.527357621108957e-07,
+      "loss": 0.8037,
+      "step": 73097
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.527098137964089e-07,
+      "loss": 0.5957,
+      "step": 73098
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.526838657478259e-07,
+      "loss": 0.5596,
+      "step": 73099
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.526579179651672e-07,
+      "loss": 0.6128,
+      "step": 73100
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.526319704484521e-07,
+      "loss": 0.5469,
+      "step": 73101
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.526060231977011e-07,
+      "loss": 0.5586,
+      "step": 73102
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.525800762129336e-07,
+      "loss": 0.5684,
+      "step": 73103
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.525541294941694e-07,
+      "loss": 0.4819,
+      "step": 73104
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.525281830414284e-07,
+      "loss": 0.4351,
+      "step": 73105
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.525022368547308e-07,
+      "loss": 0.7217,
+      "step": 73106
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.524762909340961e-07,
+      "loss": 0.6112,
+      "step": 73107
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.524503452795446e-07,
+      "loss": 0.5796,
+      "step": 73108
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.524243998910956e-07,
+      "loss": 0.6128,
+      "step": 73109
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.523984547687693e-07,
+      "loss": 0.6289,
+      "step": 73110
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.523725099125856e-07,
+      "loss": 0.7803,
+      "step": 73111
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.523465653225645e-07,
+      "loss": 0.4839,
+      "step": 73112
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.523206209987256e-07,
+      "loss": 0.6006,
+      "step": 73113
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.522946769410884e-07,
+      "loss": 0.5435,
+      "step": 73114
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.522687331496734e-07,
+      "loss": 0.5122,
+      "step": 73115
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.522427896245002e-07,
+      "loss": 0.7002,
+      "step": 73116
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.522168463655888e-07,
+      "loss": 0.7188,
+      "step": 73117
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.521909033729588e-07,
+      "loss": 0.6611,
+      "step": 73118
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.521649606466304e-07,
+      "loss": 0.7119,
+      "step": 73119
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.521390181866231e-07,
+      "loss": 0.833,
+      "step": 73120
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.521130759929572e-07,
+      "loss": 0.7559,
+      "step": 73121
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.520871340656527e-07,
+      "loss": 0.6509,
+      "step": 73122
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.520611924047286e-07,
+      "loss": 0.6484,
+      "step": 73123
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.520352510102052e-07,
+      "loss": 0.7373,
+      "step": 73124
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.520093098821025e-07,
+      "loss": 0.7256,
+      "step": 73125
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.519833690204401e-07,
+      "loss": 0.3957,
+      "step": 73126
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.519574284252383e-07,
+      "loss": 0.7061,
+      "step": 73127
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.519314880965164e-07,
+      "loss": 0.605,
+      "step": 73128
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.519055480342947e-07,
+      "loss": 0.623,
+      "step": 73129
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.518796082385929e-07,
+      "loss": 0.5977,
+      "step": 73130
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.518536687094309e-07,
+      "loss": 0.584,
+      "step": 73131
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.518277294468289e-07,
+      "loss": 0.7637,
+      "step": 73132
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.51801790450806e-07,
+      "loss": 0.5757,
+      "step": 73133
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.517758517213825e-07,
+      "loss": 0.7402,
+      "step": 73134
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.517499132585782e-07,
+      "loss": 0.7046,
+      "step": 73135
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.517239750624131e-07,
+      "loss": 0.7637,
+      "step": 73136
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.516980371329066e-07,
+      "loss": 0.7031,
+      "step": 73137
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.516720994700793e-07,
+      "loss": 0.4678,
+      "step": 73138
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.516461620739503e-07,
+      "loss": 0.562,
+      "step": 73139
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.516202249445401e-07,
+      "loss": 0.5093,
+      "step": 73140
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.515942880818683e-07,
+      "loss": 0.7251,
+      "step": 73141
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.515683514859549e-07,
+      "loss": 0.6582,
+      "step": 73142
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.515424151568192e-07,
+      "loss": 0.7144,
+      "step": 73143
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.515164790944816e-07,
+      "loss": 0.5574,
+      "step": 73144
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.514905432989617e-07,
+      "loss": 0.564,
+      "step": 73145
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.514646077702796e-07,
+      "loss": 0.7446,
+      "step": 73146
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.514386725084548e-07,
+      "loss": 0.6826,
+      "step": 73147
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.514127375135076e-07,
+      "loss": 0.6436,
+      "step": 73148
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.513868027854575e-07,
+      "loss": 0.4497,
+      "step": 73149
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.513608683243247e-07,
+      "loss": 0.6279,
+      "step": 73150
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.513349341301286e-07,
+      "loss": 0.6284,
+      "step": 73151
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.513090002028897e-07,
+      "loss": 0.7383,
+      "step": 73152
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.512830665426272e-07,
+      "loss": 0.7051,
+      "step": 73153
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 6.512571331493612e-07,
+      "loss": 0.6191,
+      "step": 73154
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.512312000231117e-07,
+      "loss": 0.751,
+      "step": 73155
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.512052671638981e-07,
+      "loss": 0.6748,
+      "step": 73156
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.511793345717409e-07,
+      "loss": 0.6436,
+      "step": 73157
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.511534022466598e-07,
+      "loss": 0.5356,
+      "step": 73158
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.511274701886741e-07,
+      "loss": 0.7119,
+      "step": 73159
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.511015383978042e-07,
+      "loss": 0.6387,
+      "step": 73160
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.510756068740697e-07,
+      "loss": 0.6465,
+      "step": 73161
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.51049675617491e-07,
+      "loss": 0.6484,
+      "step": 73162
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.510237446280873e-07,
+      "loss": 0.6328,
+      "step": 73163
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.509978139058784e-07,
+      "loss": 0.5557,
+      "step": 73164
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.509718834508848e-07,
+      "loss": 0.6973,
+      "step": 73165
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.509459532631256e-07,
+      "loss": 0.6465,
+      "step": 73166
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.509200233426214e-07,
+      "loss": 0.5635,
+      "step": 73167
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.508940936893914e-07,
+      "loss": 0.6187,
+      "step": 73168
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.508681643034559e-07,
+      "loss": 0.5918,
+      "step": 73169
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.508422351848344e-07,
+      "loss": 0.4517,
+      "step": 73170
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.508163063335472e-07,
+      "loss": 0.6914,
+      "step": 73171
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.507903777496141e-07,
+      "loss": 0.75,
+      "step": 73172
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.507644494330544e-07,
+      "loss": 0.5488,
+      "step": 73173
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.507385213838883e-07,
+      "loss": 0.7783,
+      "step": 73174
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.507125936021358e-07,
+      "loss": 0.6465,
+      "step": 73175
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.506866660878164e-07,
+      "loss": 0.6216,
+      "step": 73176
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.506607388409504e-07,
+      "loss": 0.6172,
+      "step": 73177
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.506348118615572e-07,
+      "loss": 0.5483,
+      "step": 73178
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.50608885149657e-07,
+      "loss": 0.7832,
+      "step": 73179
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.505829587052693e-07,
+      "loss": 0.5718,
+      "step": 73180
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.505570325284144e-07,
+      "loss": 0.5344,
+      "step": 73181
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.505311066191121e-07,
+      "loss": 0.48,
+      "step": 73182
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.505051809773816e-07,
+      "loss": 0.5918,
+      "step": 73183
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.504792556032435e-07,
+      "loss": 0.7178,
+      "step": 73184
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.50453330496717e-07,
+      "loss": 0.5757,
+      "step": 73185
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.504274056578228e-07,
+      "loss": 0.4709,
+      "step": 73186
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.504014810865799e-07,
+      "loss": 0.5366,
+      "step": 73187
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.503755567830087e-07,
+      "loss": 0.7412,
+      "step": 73188
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.503496327471286e-07,
+      "loss": 0.5759,
+      "step": 73189
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.503237089789601e-07,
+      "loss": 0.5486,
+      "step": 73190
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.502977854785223e-07,
+      "loss": 0.6333,
+      "step": 73191
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.502718622458359e-07,
+      "loss": 0.8115,
+      "step": 73192
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.502459392809199e-07,
+      "loss": 0.5938,
+      "step": 73193
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.502200165837947e-07,
+      "loss": 0.4142,
+      "step": 73194
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.501940941544796e-07,
+      "loss": 0.6655,
+      "step": 73195
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.50168171992995e-07,
+      "loss": 0.7041,
+      "step": 73196
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.501422500993605e-07,
+      "loss": 0.834,
+      "step": 73197
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.50116328473596e-07,
+      "loss": 0.5856,
+      "step": 73198
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.500904071157213e-07,
+      "loss": 0.5259,
+      "step": 73199
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.500644860257564e-07,
+      "loss": 0.6313,
+      "step": 73200
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.500385652037208e-07,
+      "loss": 0.5471,
+      "step": 73201
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.500126446496351e-07,
+      "loss": 0.7539,
+      "step": 73202
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.499867243635184e-07,
+      "loss": 0.7236,
+      "step": 73203
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.499608043453904e-07,
+      "loss": 0.6235,
+      "step": 73204
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.499348845952717e-07,
+      "loss": 0.8154,
+      "step": 73205
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.499089651131815e-07,
+      "loss": 0.7012,
+      "step": 73206
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.498830458991402e-07,
+      "loss": 0.4778,
+      "step": 73207
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.498571269531669e-07,
+      "loss": 0.7734,
+      "step": 73208
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.498312082752824e-07,
+      "loss": 0.6279,
+      "step": 73209
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.498052898655057e-07,
+      "loss": 0.7607,
+      "step": 73210
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.497793717238571e-07,
+      "loss": 0.533,
+      "step": 73211
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.497534538503567e-07,
+      "loss": 0.7598,
+      "step": 73212
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.497275362450237e-07,
+      "loss": 0.5806,
+      "step": 73213
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.497016189078781e-07,
+      "loss": 0.7627,
+      "step": 73214
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.496757018389399e-07,
+      "loss": 0.6387,
+      "step": 73215
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.496497850382288e-07,
+      "loss": 0.5977,
+      "step": 73216
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.496238685057649e-07,
+      "loss": 0.5737,
+      "step": 73217
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.495979522415677e-07,
+      "loss": 0.7314,
+      "step": 73218
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.495720362456576e-07,
+      "loss": 0.7734,
+      "step": 73219
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.495461205180537e-07,
+      "loss": 0.584,
+      "step": 73220
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.495202050587765e-07,
+      "loss": 0.6045,
+      "step": 73221
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.494942898678457e-07,
+      "loss": 0.6396,
+      "step": 73222
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.494683749452808e-07,
+      "loss": 0.6885,
+      "step": 73223
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.494424602911018e-07,
+      "loss": 0.6201,
+      "step": 73224
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.494165459053285e-07,
+      "loss": 0.5488,
+      "step": 73225
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.49390631787981e-07,
+      "loss": 0.6504,
+      "step": 73226
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.493647179390786e-07,
+      "loss": 0.5718,
+      "step": 73227
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.49338804358642e-07,
+      "loss": 0.6016,
+      "step": 73228
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.493128910466903e-07,
+      "loss": 0.4812,
+      "step": 73229
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.492869780032438e-07,
+      "loss": 0.5796,
+      "step": 73230
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.492610652283219e-07,
+      "loss": 0.6973,
+      "step": 73231
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.49235152721945e-07,
+      "loss": 0.5562,
+      "step": 73232
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.492092404841324e-07,
+      "loss": 0.4819,
+      "step": 73233
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.491833285149042e-07,
+      "loss": 0.5122,
+      "step": 73234
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.4915741681428e-07,
+      "loss": 0.6416,
+      "step": 73235
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.491315053822802e-07,
+      "loss": 0.5776,
+      "step": 73236
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.491055942189237e-07,
+      "loss": 0.6011,
+      "step": 73237
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.490796833242313e-07,
+      "loss": 0.5376,
+      "step": 73238
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.490537726982224e-07,
+      "loss": 0.7461,
+      "step": 73239
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.490278623409169e-07,
+      "loss": 0.793,
+      "step": 73240
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.490019522523345e-07,
+      "loss": 0.4685,
+      "step": 73241
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.489760424324957e-07,
+      "loss": 0.5508,
+      "step": 73242
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.489501328814194e-07,
+      "loss": 0.6714,
+      "step": 73243
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.489242235991257e-07,
+      "loss": 0.6973,
+      "step": 73244
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.488983145856348e-07,
+      "loss": 0.582,
+      "step": 73245
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.488724058409661e-07,
+      "loss": 0.7178,
+      "step": 73246
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.488464973651399e-07,
+      "loss": 0.5361,
+      "step": 73247
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.488205891581758e-07,
+      "loss": 0.4722,
+      "step": 73248
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.487946812200933e-07,
+      "loss": 0.5649,
+      "step": 73249
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.487687735509128e-07,
+      "loss": 0.6816,
+      "step": 73250
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.487428661506538e-07,
+      "loss": 0.6289,
+      "step": 73251
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.487169590193365e-07,
+      "loss": 0.7295,
+      "step": 73252
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.486910521569804e-07,
+      "loss": 0.6953,
+      "step": 73253
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.486651455636052e-07,
+      "loss": 0.667,
+      "step": 73254
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.486392392392311e-07,
+      "loss": 0.7129,
+      "step": 73255
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.486133331838775e-07,
+      "loss": 0.3975,
+      "step": 73256
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.485874273975649e-07,
+      "loss": 0.5171,
+      "step": 73257
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.485615218803124e-07,
+      "loss": 0.4629,
+      "step": 73258
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.485356166321404e-07,
+      "loss": 0.585,
+      "step": 73259
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.485097116530683e-07,
+      "loss": 0.7207,
+      "step": 73260
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.484838069431165e-07,
+      "loss": 0.7588,
+      "step": 73261
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.484579025023045e-07,
+      "loss": 0.6426,
+      "step": 73262
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.484319983306518e-07,
+      "loss": 0.3848,
+      "step": 73263
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.484060944281788e-07,
+      "loss": 0.6738,
+      "step": 73264
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.483801907949048e-07,
+      "loss": 0.7295,
+      "step": 73265
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.483542874308502e-07,
+      "loss": 0.668,
+      "step": 73266
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.483283843360345e-07,
+      "loss": 0.645,
+      "step": 73267
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.483024815104774e-07,
+      "loss": 0.833,
+      "step": 73268
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.482765789541991e-07,
+      "loss": 0.6973,
+      "step": 73269
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.482506766672191e-07,
+      "loss": 0.6602,
+      "step": 73270
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.482247746495575e-07,
+      "loss": 0.8447,
+      "step": 73271
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.481988729012343e-07,
+      "loss": 0.5337,
+      "step": 73272
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.481729714222686e-07,
+      "loss": 0.6816,
+      "step": 73273
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.481470702126809e-07,
+      "loss": 0.6152,
+      "step": 73274
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.481211692724906e-07,
+      "loss": 0.7656,
+      "step": 73275
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.480952686017179e-07,
+      "loss": 0.6406,
+      "step": 73276
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.480693682003824e-07,
+      "loss": 0.4062,
+      "step": 73277
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.480434680685041e-07,
+      "loss": 0.5596,
+      "step": 73278
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.480175682061026e-07,
+      "loss": 0.4443,
+      "step": 73279
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.479916686131979e-07,
+      "loss": 0.6494,
+      "step": 73280
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.479657692898097e-07,
+      "loss": 0.4412,
+      "step": 73281
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.479398702359585e-07,
+      "loss": 0.6528,
+      "step": 73282
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47913971451663e-07,
+      "loss": 0.5409,
+      "step": 73283
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.478880729369438e-07,
+      "loss": 0.6367,
+      "step": 73284
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.478621746918203e-07,
+      "loss": 0.5505,
+      "step": 73285
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.478362767163128e-07,
+      "loss": 0.6436,
+      "step": 73286
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.478103790104406e-07,
+      "loss": 0.457,
+      "step": 73287
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47784481574224e-07,
+      "loss": 0.5942,
+      "step": 73288
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.477585844076822e-07,
+      "loss": 0.5674,
+      "step": 73289
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47732687510836e-07,
+      "loss": 0.6582,
+      "step": 73290
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.477067908837047e-07,
+      "loss": 0.5112,
+      "step": 73291
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.476808945263078e-07,
+      "loss": 0.6392,
+      "step": 73292
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.476549984386656e-07,
+      "loss": 0.7021,
+      "step": 73293
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.476291026207976e-07,
+      "loss": 0.7002,
+      "step": 73294
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47603207072724e-07,
+      "loss": 0.585,
+      "step": 73295
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.475773117944641e-07,
+      "loss": 0.54,
+      "step": 73296
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.475514167860383e-07,
+      "loss": 0.5522,
+      "step": 73297
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47525522047466e-07,
+      "loss": 0.6375,
+      "step": 73298
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.474996275787674e-07,
+      "loss": 0.5422,
+      "step": 73299
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.47473733379962e-07,
+      "loss": 0.8408,
+      "step": 73300
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.474478394510703e-07,
+      "loss": 0.4768,
+      "step": 73301
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.474219457921109e-07,
+      "loss": 0.7104,
+      "step": 73302
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.473960524031046e-07,
+      "loss": 0.5327,
+      "step": 73303
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.473701592840708e-07,
+      "loss": 0.6064,
+      "step": 73304
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.473442664350294e-07,
+      "loss": 0.752,
+      "step": 73305
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.473183738560002e-07,
+      "loss": 0.502,
+      "step": 73306
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.472924815470035e-07,
+      "loss": 0.6143,
+      "step": 73307
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.472665895080583e-07,
+      "loss": 0.5874,
+      "step": 73308
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.472406977391851e-07,
+      "loss": 0.6846,
+      "step": 73309
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.472148062404032e-07,
+      "loss": 0.8076,
+      "step": 73310
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.471889150117334e-07,
+      "loss": 0.6074,
+      "step": 73311
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.471630240531943e-07,
+      "loss": 0.4858,
+      "step": 73312
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.471371333648062e-07,
+      "loss": 0.7266,
+      "step": 73313
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.471112429465892e-07,
+      "loss": 0.5276,
+      "step": 73314
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.470853527985627e-07,
+      "loss": 0.6489,
+      "step": 73315
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.470594629207468e-07,
+      "loss": 0.8096,
+      "step": 73316
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.470335733131611e-07,
+      "loss": 0.6377,
+      "step": 73317
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.470076839758257e-07,
+      "loss": 0.5698,
+      "step": 73318
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.469817949087602e-07,
+      "loss": 0.4629,
+      "step": 73319
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.469559061119847e-07,
+      "loss": 0.6899,
+      "step": 73320
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.469300175855189e-07,
+      "loss": 0.6665,
+      "step": 73321
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.469041293293825e-07,
+      "loss": 0.6978,
+      "step": 73322
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.468782413435951e-07,
+      "loss": 0.6108,
+      "step": 73323
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.46852353628177e-07,
+      "loss": 0.6963,
+      "step": 73324
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.468264661831475e-07,
+      "loss": 0.4211,
+      "step": 73325
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.468005790085272e-07,
+      "loss": 0.834,
+      "step": 73326
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.46774692104335e-07,
+      "loss": 0.4854,
+      "step": 73327
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.467488054705915e-07,
+      "loss": 0.5557,
+      "step": 73328
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.46722919107316e-07,
+      "loss": 0.5728,
+      "step": 73329
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.466970330145288e-07,
+      "loss": 0.6025,
+      "step": 73330
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.466711471922495e-07,
+      "loss": 0.6855,
+      "step": 73331
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.466452616404974e-07,
+      "loss": 0.8311,
+      "step": 73332
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.466193763592931e-07,
+      "loss": 0.7441,
+      "step": 73333
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.465934913486558e-07,
+      "loss": 0.6318,
+      "step": 73334
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.465676066086059e-07,
+      "loss": 0.364,
+      "step": 73335
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.465417221391627e-07,
+      "loss": 0.7065,
+      "step": 73336
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.465158379403465e-07,
+      "loss": 0.5386,
+      "step": 73337
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.464899540121768e-07,
+      "loss": 0.6699,
+      "step": 73338
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.464640703546733e-07,
+      "loss": 0.7754,
+      "step": 73339
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.464381869678563e-07,
+      "loss": 0.5996,
+      "step": 73340
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.464123038517454e-07,
+      "loss": 0.5503,
+      "step": 73341
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.463864210063599e-07,
+      "loss": 0.7109,
+      "step": 73342
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.463605384317204e-07,
+      "loss": 0.7764,
+      "step": 73343
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.463346561278461e-07,
+      "loss": 0.8223,
+      "step": 73344
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.463087740947573e-07,
+      "loss": 0.6826,
+      "step": 73345
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.462828923324735e-07,
+      "loss": 0.6694,
+      "step": 73346
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.462570108410145e-07,
+      "loss": 0.6396,
+      "step": 73347
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.462311296204004e-07,
+      "loss": 0.7148,
+      "step": 73348
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.46205248670651e-07,
+      "loss": 0.8818,
+      "step": 73349
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.461793679917858e-07,
+      "loss": 0.6533,
+      "step": 73350
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.461534875838251e-07,
+      "loss": 0.5173,
+      "step": 73351
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.461276074467881e-07,
+      "loss": 0.668,
+      "step": 73352
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.461017275806948e-07,
+      "loss": 0.7041,
+      "step": 73353
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.460758479855654e-07,
+      "loss": 0.5286,
+      "step": 73354
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.460499686614192e-07,
+      "loss": 0.6592,
+      "step": 73355
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.460240896082765e-07,
+      "loss": 0.6641,
+      "step": 73356
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.459982108261568e-07,
+      "loss": 0.4723,
+      "step": 73357
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.459723323150799e-07,
+      "loss": 0.6621,
+      "step": 73358
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.459464540750658e-07,
+      "loss": 0.6104,
+      "step": 73359
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45920576106134e-07,
+      "loss": 0.7998,
+      "step": 73360
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.458946984083051e-07,
+      "loss": 0.7842,
+      "step": 73361
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45868820981598e-07,
+      "loss": 0.5771,
+      "step": 73362
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.458429438260326e-07,
+      "loss": 0.5238,
+      "step": 73363
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.458170669416294e-07,
+      "loss": 0.5542,
+      "step": 73364
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.457911903284074e-07,
+      "loss": 0.6201,
+      "step": 73365
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.457653139863872e-07,
+      "loss": 0.5225,
+      "step": 73366
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.457394379155878e-07,
+      "loss": 0.6191,
+      "step": 73367
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.457135621160296e-07,
+      "loss": 0.8135,
+      "step": 73368
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.456876865877323e-07,
+      "loss": 0.5073,
+      "step": 73369
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.456618113307157e-07,
+      "loss": 0.4872,
+      "step": 73370
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.456359363449997e-07,
+      "loss": 0.6152,
+      "step": 73371
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.456100616306036e-07,
+      "loss": 0.7056,
+      "step": 73372
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.455841871875478e-07,
+      "loss": 0.7168,
+      "step": 73373
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45558313015852e-07,
+      "loss": 0.7227,
+      "step": 73374
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.455324391155354e-07,
+      "loss": 0.5786,
+      "step": 73375
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.455065654866187e-07,
+      "loss": 0.7065,
+      "step": 73376
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.454806921291211e-07,
+      "loss": 0.6377,
+      "step": 73377
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45454819043063e-07,
+      "loss": 0.709,
+      "step": 73378
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.454289462284635e-07,
+      "loss": 0.6729,
+      "step": 73379
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.454030736853429e-07,
+      "loss": 0.4507,
+      "step": 73380
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45377201413721e-07,
+      "loss": 0.6875,
+      "step": 73381
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.453513294136172e-07,
+      "loss": 0.7637,
+      "step": 73382
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.453254576850519e-07,
+      "loss": 0.5747,
+      "step": 73383
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.452995862280441e-07,
+      "loss": 0.4875,
+      "step": 73384
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.452737150426146e-07,
+      "loss": 0.6523,
+      "step": 73385
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.452478441287823e-07,
+      "loss": 0.5962,
+      "step": 73386
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.452219734865677e-07,
+      "loss": 0.5903,
+      "step": 73387
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.451961031159901e-07,
+      "loss": 0.6387,
+      "step": 73388
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.451702330170696e-07,
+      "loss": 0.5649,
+      "step": 73389
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.451443631898259e-07,
+      "loss": 0.5244,
+      "step": 73390
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.451184936342795e-07,
+      "loss": 0.5691,
+      "step": 73391
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.450926243504489e-07,
+      "loss": 0.6113,
+      "step": 73392
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.450667553383546e-07,
+      "loss": 0.7012,
+      "step": 73393
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.450408865980163e-07,
+      "loss": 0.6084,
+      "step": 73394
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.45015018129454e-07,
+      "loss": 0.4329,
+      "step": 73395
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.449891499326872e-07,
+      "loss": 0.6289,
+      "step": 73396
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.449632820077362e-07,
+      "loss": 0.5203,
+      "step": 73397
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.449374143546202e-07,
+      "loss": 0.6499,
+      "step": 73398
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.449115469733595e-07,
+      "loss": 0.5742,
+      "step": 73399
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.448856798639735e-07,
+      "loss": 0.6182,
+      "step": 73400
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.448598130264827e-07,
+      "loss": 0.5918,
+      "step": 73401
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44833946460906e-07,
+      "loss": 0.5503,
+      "step": 73402
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.448080801672635e-07,
+      "loss": 0.5938,
+      "step": 73403
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.447822141455754e-07,
+      "loss": 0.6074,
+      "step": 73404
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44756348395861e-07,
+      "loss": 0.5542,
+      "step": 73405
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.447304829181404e-07,
+      "loss": 0.4692,
+      "step": 73406
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.447046177124333e-07,
+      "loss": 0.7607,
+      "step": 73407
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.446787527787596e-07,
+      "loss": 0.3619,
+      "step": 73408
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.446528881171389e-07,
+      "loss": 0.5212,
+      "step": 73409
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.446270237275915e-07,
+      "loss": 0.6719,
+      "step": 73410
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.446011596101368e-07,
+      "loss": 0.6382,
+      "step": 73411
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.445752957647945e-07,
+      "loss": 0.5791,
+      "step": 73412
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.445494321915845e-07,
+      "loss": 0.5762,
+      "step": 73413
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.445235688905268e-07,
+      "loss": 0.7178,
+      "step": 73414
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.444977058616407e-07,
+      "loss": 0.7588,
+      "step": 73415
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.444718431049467e-07,
+      "loss": 0.6797,
+      "step": 73416
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.444459806204641e-07,
+      "loss": 0.627,
+      "step": 73417
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44420118408213e-07,
+      "loss": 0.6143,
+      "step": 73418
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44394256468213e-07,
+      "loss": 0.6221,
+      "step": 73419
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.443683948004839e-07,
+      "loss": 0.5376,
+      "step": 73420
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44342533405046e-07,
+      "loss": 0.7305,
+      "step": 73421
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.443166722819182e-07,
+      "loss": 0.4888,
+      "step": 73422
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44290811431121e-07,
+      "loss": 0.6548,
+      "step": 73423
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.442649508526736e-07,
+      "loss": 0.4561,
+      "step": 73424
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.442390905465965e-07,
+      "loss": 0.5139,
+      "step": 73425
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44213230512909e-07,
+      "loss": 0.6577,
+      "step": 73426
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.441873707516312e-07,
+      "loss": 0.668,
+      "step": 73427
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.441615112627826e-07,
+      "loss": 0.627,
+      "step": 73428
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.441356520463835e-07,
+      "loss": 0.2308,
+      "step": 73429
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.44109793102453e-07,
+      "loss": 0.708,
+      "step": 73430
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.440839344310118e-07,
+      "loss": 0.6354,
+      "step": 73431
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.440580760320786e-07,
+      "loss": 0.7871,
+      "step": 73432
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.440322179056741e-07,
+      "loss": 0.5664,
+      "step": 73433
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.440063600518175e-07,
+      "loss": 0.6357,
+      "step": 73434
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.439805024705291e-07,
+      "loss": 0.769,
+      "step": 73435
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.439546451618282e-07,
+      "loss": 0.6484,
+      "step": 73436
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.439287881257352e-07,
+      "loss": 0.6816,
+      "step": 73437
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.43902931362269e-07,
+      "loss": 0.6729,
+      "step": 73438
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.438770748714505e-07,
+      "loss": 0.6699,
+      "step": 73439
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.438512186532986e-07,
+      "loss": 0.6626,
+      "step": 73440
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.438253627078339e-07,
+      "loss": 0.5413,
+      "step": 73441
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.437995070350754e-07,
+      "loss": 0.5659,
+      "step": 73442
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.437736516350433e-07,
+      "loss": 0.7568,
+      "step": 73443
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.437477965077573e-07,
+      "loss": 0.5513,
+      "step": 73444
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.437219416532372e-07,
+      "loss": 0.4735,
+      "step": 73445
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.436960870715029e-07,
+      "loss": 0.5645,
+      "step": 73446
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.436702327625742e-07,
+      "loss": 0.6909,
+      "step": 73447
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.436443787264706e-07,
+      "loss": 0.5566,
+      "step": 73448
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.436185249632122e-07,
+      "loss": 0.6221,
+      "step": 73449
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.435926714728187e-07,
+      "loss": 0.6855,
+      "step": 73450
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.435668182553102e-07,
+      "loss": 0.7393,
+      "step": 73451
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.43540965310706e-07,
+      "loss": 0.502,
+      "step": 73452
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.435151126390259e-07,
+      "loss": 0.6826,
+      "step": 73453
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.434892602402902e-07,
+      "loss": 0.6655,
+      "step": 73454
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.43463408114518e-07,
+      "loss": 0.6284,
+      "step": 73455
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.434375562617298e-07,
+      "loss": 0.7607,
+      "step": 73456
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.434117046819447e-07,
+      "loss": 0.4937,
+      "step": 73457
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.433858533751833e-07,
+      "loss": 0.6182,
+      "step": 73458
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.433600023414645e-07,
+      "loss": 0.6748,
+      "step": 73459
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.43334151580809e-07,
+      "loss": 0.6807,
+      "step": 73460
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.433083010932362e-07,
+      "loss": 0.7354,
+      "step": 73461
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.432824508787655e-07,
+      "loss": 0.5703,
+      "step": 73462
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.432566009374171e-07,
+      "loss": 0.4487,
+      "step": 73463
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.432307512692108e-07,
+      "loss": 0.5107,
+      "step": 73464
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.432049018741661e-07,
+      "loss": 0.5125,
+      "step": 73465
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.431790527523032e-07,
+      "loss": 0.6963,
+      "step": 73466
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.431532039036414e-07,
+      "loss": 0.6177,
+      "step": 73467
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.431273553282012e-07,
+      "loss": 0.6641,
+      "step": 73468
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.431015070260017e-07,
+      "loss": 0.8887,
+      "step": 73469
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.430756589970631e-07,
+      "loss": 0.7354,
+      "step": 73470
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.430498112414052e-07,
+      "loss": 0.8076,
+      "step": 73471
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.430239637590473e-07,
+      "loss": 0.5566,
+      "step": 73472
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.429981165500098e-07,
+      "loss": 0.582,
+      "step": 73473
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.42972269614312e-07,
+      "loss": 0.6465,
+      "step": 73474
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.429464229519741e-07,
+      "loss": 0.6157,
+      "step": 73475
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.429205765630155e-07,
+      "loss": 0.7593,
+      "step": 73476
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.428947304474563e-07,
+      "loss": 0.5635,
+      "step": 73477
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.428688846053162e-07,
+      "loss": 0.624,
+      "step": 73478
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.428430390366151e-07,
+      "loss": 0.6924,
+      "step": 73479
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.428171937413725e-07,
+      "loss": 0.8018,
+      "step": 73480
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.427913487196089e-07,
+      "loss": 0.6597,
+      "step": 73481
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.427655039713428e-07,
+      "loss": 0.6436,
+      "step": 73482
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.427396594965951e-07,
+      "loss": 0.6777,
+      "step": 73483
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.427138152953848e-07,
+      "loss": 0.4271,
+      "step": 73484
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.426879713677326e-07,
+      "loss": 0.585,
+      "step": 73485
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.426621277136575e-07,
+      "loss": 0.563,
+      "step": 73486
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.426362843331796e-07,
+      "loss": 0.7617,
+      "step": 73487
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.426104412263187e-07,
+      "loss": 0.4236,
+      "step": 73488
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.425845983930947e-07,
+      "loss": 0.6875,
+      "step": 73489
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.425587558335268e-07,
+      "loss": 0.5864,
+      "step": 73490
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.425329135476359e-07,
+      "loss": 0.6514,
+      "step": 73491
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.425070715354408e-07,
+      "loss": 0.5698,
+      "step": 73492
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.424812297969615e-07,
+      "loss": 0.7559,
+      "step": 73493
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.424553883322179e-07,
+      "loss": 0.3063,
+      "step": 73494
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.424295471412297e-07,
+      "loss": 0.5562,
+      "step": 73495
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.424037062240169e-07,
+      "loss": 0.6965,
+      "step": 73496
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.423778655805989e-07,
+      "loss": 0.8184,
+      "step": 73497
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.42352025210996e-07,
+      "loss": 0.5745,
+      "step": 73498
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.423261851152275e-07,
+      "loss": 0.7383,
+      "step": 73499
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.423003452933137e-07,
+      "loss": 0.6406,
+      "step": 73500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.42274505745274e-07,
+      "loss": 0.7422,
+      "step": 73501
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.422486664711284e-07,
+      "loss": 0.4399,
+      "step": 73502
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.422228274708962e-07,
+      "loss": 0.6084,
+      "step": 73503
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.421969887445977e-07,
+      "loss": 0.668,
+      "step": 73504
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.421711502922523e-07,
+      "loss": 0.5659,
+      "step": 73505
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.421453121138802e-07,
+      "loss": 0.6807,
+      "step": 73506
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.421194742095009e-07,
+      "loss": 0.5923,
+      "step": 73507
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.420936365791343e-07,
+      "loss": 0.6846,
+      "step": 73508
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.420677992228e-07,
+      "loss": 0.6831,
+      "step": 73509
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.420419621405182e-07,
+      "loss": 0.5408,
+      "step": 73510
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.420161253323084e-07,
+      "loss": 0.5663,
+      "step": 73511
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.419902887981903e-07,
+      "loss": 0.6553,
+      "step": 73512
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.419644525381837e-07,
+      "loss": 0.71,
+      "step": 73513
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.419386165523084e-07,
+      "loss": 0.5161,
+      "step": 73514
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.419127808405844e-07,
+      "loss": 0.6621,
+      "step": 73515
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.418869454030311e-07,
+      "loss": 0.709,
+      "step": 73516
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.418611102396687e-07,
+      "loss": 0.6416,
+      "step": 73517
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.418352753505167e-07,
+      "loss": 0.7061,
+      "step": 73518
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.418094407355949e-07,
+      "loss": 0.665,
+      "step": 73519
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.417836063949235e-07,
+      "loss": 0.6445,
+      "step": 73520
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.417577723285219e-07,
+      "loss": 0.6924,
+      "step": 73521
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.417319385364095e-07,
+      "loss": 0.6455,
+      "step": 73522
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.417061050186068e-07,
+      "loss": 0.4741,
+      "step": 73523
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.416802717751328e-07,
+      "loss": 0.7803,
+      "step": 73524
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.416544388060081e-07,
+      "loss": 0.6255,
+      "step": 73525
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.41628606111252e-07,
+      "loss": 0.5625,
+      "step": 73526
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.416027736908846e-07,
+      "loss": 0.6523,
+      "step": 73527
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.415769415449251e-07,
+      "loss": 0.6938,
+      "step": 73528
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.41551109673394e-07,
+      "loss": 0.7314,
+      "step": 73529
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.415252780763105e-07,
+      "loss": 0.3878,
+      "step": 73530
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.414994467536949e-07,
+      "loss": 0.5786,
+      "step": 73531
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.414736157055667e-07,
+      "loss": 0.832,
+      "step": 73532
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.414477849319453e-07,
+      "loss": 0.5424,
+      "step": 73533
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.414219544328512e-07,
+      "loss": 0.4399,
+      "step": 73534
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.413961242083034e-07,
+      "loss": 0.5684,
+      "step": 73535
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.413702942583225e-07,
+      "loss": 0.709,
+      "step": 73536
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.413444645829278e-07,
+      "loss": 0.4976,
+      "step": 73537
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.413186351821389e-07,
+      "loss": 0.6963,
+      "step": 73538
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.412928060559761e-07,
+      "loss": 0.4893,
+      "step": 73539
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.412669772044588e-07,
+      "loss": 0.5869,
+      "step": 73540
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.41241148627607e-07,
+      "loss": 0.5237,
+      "step": 73541
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.412153203254404e-07,
+      "loss": 0.6982,
+      "step": 73542
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.411894922979785e-07,
+      "loss": 0.7114,
+      "step": 73543
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.411636645452414e-07,
+      "loss": 0.429,
+      "step": 73544
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 6.411378370672486e-07,
+      "loss": 0.7998,
+      "step": 73545
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.411120098640203e-07,
+      "loss": 0.5688,
+      "step": 73546
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.410861829355759e-07,
+      "loss": 0.438,
+      "step": 73547
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.410603562819354e-07,
+      "loss": 0.6123,
+      "step": 73548
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.410345299031184e-07,
+      "loss": 0.6782,
+      "step": 73549
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.410087037991449e-07,
+      "loss": 0.7197,
+      "step": 73550
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.409828779700346e-07,
+      "loss": 0.6934,
+      "step": 73551
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.40957052415807e-07,
+      "loss": 0.5889,
+      "step": 73552
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.409312271364822e-07,
+      "loss": 0.5107,
+      "step": 73553
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.409054021320797e-07,
+      "loss": 0.5576,
+      "step": 73554
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.408795774026194e-07,
+      "loss": 0.749,
+      "step": 73555
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.408537529481213e-07,
+      "loss": 0.6968,
+      "step": 73556
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.408279287686047e-07,
+      "loss": 0.5166,
+      "step": 73557
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.408021048640899e-07,
+      "loss": 0.5532,
+      "step": 73558
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.407762812345962e-07,
+      "loss": 0.5825,
+      "step": 73559
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.407504578801437e-07,
+      "loss": 0.6768,
+      "step": 73560
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.407246348007522e-07,
+      "loss": 0.7158,
+      "step": 73561
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.406988119964412e-07,
+      "loss": 0.4843,
+      "step": 73562
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.406729894672306e-07,
+      "loss": 0.6748,
+      "step": 73563
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.406471672131399e-07,
+      "loss": 0.6816,
+      "step": 73564
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.406213452341896e-07,
+      "loss": 0.6421,
+      "step": 73565
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.405955235303986e-07,
+      "loss": 0.6963,
+      "step": 73566
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.405697021017873e-07,
+      "loss": 0.6035,
+      "step": 73567
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.405438809483751e-07,
+      "loss": 0.7598,
+      "step": 73568
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.405180600701821e-07,
+      "loss": 0.6665,
+      "step": 73569
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.404922394672277e-07,
+      "loss": 0.6787,
+      "step": 73570
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.404664191395325e-07,
+      "loss": 0.6865,
+      "step": 73571
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.404405990871149e-07,
+      "loss": 0.4917,
+      "step": 73572
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.404147793099957e-07,
+      "loss": 0.5801,
+      "step": 73573
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.403889598081942e-07,
+      "loss": 0.418,
+      "step": 73574
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.403631405817304e-07,
+      "loss": 0.6836,
+      "step": 73575
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.403373216306241e-07,
+      "loss": 0.4673,
+      "step": 73576
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.403115029548949e-07,
+      "loss": 0.6963,
+      "step": 73577
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.402856845545625e-07,
+      "loss": 0.6611,
+      "step": 73578
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.40259866429647e-07,
+      "loss": 0.5957,
+      "step": 73579
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.402340485801678e-07,
+      "loss": 0.5151,
+      "step": 73580
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.402082310061454e-07,
+      "loss": 0.6758,
+      "step": 73581
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.401824137075986e-07,
+      "loss": 0.6929,
+      "step": 73582
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.401565966845475e-07,
+      "loss": 0.6035,
+      "step": 73583
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.401307799370122e-07,
+      "loss": 0.6016,
+      "step": 73584
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.401049634650118e-07,
+      "loss": 0.6055,
+      "step": 73585
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.400791472685668e-07,
+      "loss": 0.6387,
+      "step": 73586
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.400533313476965e-07,
+      "loss": 0.6392,
+      "step": 73587
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.400275157024209e-07,
+      "loss": 0.8164,
+      "step": 73588
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.400017003327596e-07,
+      "loss": 0.5835,
+      "step": 73589
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.399758852387329e-07,
+      "loss": 0.7314,
+      "step": 73590
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.399500704203596e-07,
+      "loss": 0.7412,
+      "step": 73591
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.399242558776603e-07,
+      "loss": 0.6475,
+      "step": 73592
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.398984416106541e-07,
+      "loss": 0.6318,
+      "step": 73593
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.398726276193613e-07,
+      "loss": 0.6387,
+      "step": 73594
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.398468139038013e-07,
+      "loss": 0.6218,
+      "step": 73595
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.398210004639941e-07,
+      "loss": 0.7305,
+      "step": 73596
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.397951872999594e-07,
+      "loss": 0.667,
+      "step": 73597
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.397693744117171e-07,
+      "loss": 0.6235,
+      "step": 73598
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.397435617992866e-07,
+      "loss": 0.7114,
+      "step": 73599
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.397177494626883e-07,
+      "loss": 0.4692,
+      "step": 73600
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.396919374019413e-07,
+      "loss": 0.5654,
+      "step": 73601
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.396661256170654e-07,
+      "loss": 0.5991,
+      "step": 73602
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39640314108081e-07,
+      "loss": 0.5547,
+      "step": 73603
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39614502875007e-07,
+      "loss": 0.5728,
+      "step": 73604
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.395886919178639e-07,
+      "loss": 0.7979,
+      "step": 73605
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39562881236671e-07,
+      "loss": 0.7959,
+      "step": 73606
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.395370708314485e-07,
+      "loss": 0.6562,
+      "step": 73607
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.395112607022155e-07,
+      "loss": 0.6562,
+      "step": 73608
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.394854508489924e-07,
+      "loss": 0.5635,
+      "step": 73609
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39459641271799e-07,
+      "loss": 0.6382,
+      "step": 73610
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.394338319706545e-07,
+      "loss": 0.5308,
+      "step": 73611
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.394080229455787e-07,
+      "loss": 0.5305,
+      "step": 73612
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.393822141965919e-07,
+      "loss": 0.7319,
+      "step": 73613
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.393564057237133e-07,
+      "loss": 0.709,
+      "step": 73614
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.393305975269631e-07,
+      "loss": 0.7148,
+      "step": 73615
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.393047896063608e-07,
+      "loss": 0.6416,
+      "step": 73616
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.392789819619262e-07,
+      "loss": 0.5806,
+      "step": 73617
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.392531745936791e-07,
+      "loss": 0.6069,
+      "step": 73618
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.392273675016394e-07,
+      "loss": 0.6143,
+      "step": 73619
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39201560685827e-07,
+      "loss": 0.4692,
+      "step": 73620
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.391757541462608e-07,
+      "loss": 0.7344,
+      "step": 73621
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.391499478829614e-07,
+      "loss": 0.6226,
+      "step": 73622
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.391241418959481e-07,
+      "loss": 0.4333,
+      "step": 73623
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.390983361852411e-07,
+      "loss": 0.5469,
+      "step": 73624
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.390725307508597e-07,
+      "loss": 0.5669,
+      "step": 73625
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.39046725592824e-07,
+      "loss": 0.6758,
+      "step": 73626
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.390209207111536e-07,
+      "loss": 0.5576,
+      "step": 73627
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.389951161058682e-07,
+      "loss": 0.7686,
+      "step": 73628
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.389693117769879e-07,
+      "loss": 0.6006,
+      "step": 73629
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.389435077245322e-07,
+      "loss": 0.5332,
+      "step": 73630
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.389177039485205e-07,
+      "loss": 0.3975,
+      "step": 73631
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.388919004489732e-07,
+      "loss": 0.6646,
+      "step": 73632
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.388660972259095e-07,
+      "loss": 0.7729,
+      "step": 73633
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.388402942793496e-07,
+      "loss": 0.6475,
+      "step": 73634
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.38814491609313e-07,
+      "loss": 0.6562,
+      "step": 73635
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.387886892158196e-07,
+      "loss": 0.5815,
+      "step": 73636
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.387628870988891e-07,
+      "loss": 0.3828,
+      "step": 73637
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.387370852585412e-07,
+      "loss": 0.6841,
+      "step": 73638
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.387112836947956e-07,
+      "loss": 0.665,
+      "step": 73639
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.386854824076725e-07,
+      "loss": 0.6338,
+      "step": 73640
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.386596813971912e-07,
+      "loss": 0.6089,
+      "step": 73641
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.386338806633713e-07,
+      "loss": 0.5962,
+      "step": 73642
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.38608080206233e-07,
+      "loss": 0.7041,
+      "step": 73643
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.385822800257961e-07,
+      "loss": 0.603,
+      "step": 73644
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.385564801220797e-07,
+      "loss": 0.5464,
+      "step": 73645
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.385306804951043e-07,
+      "loss": 0.6279,
+      "step": 73646
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.385048811448889e-07,
+      "loss": 0.5083,
+      "step": 73647
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.384790820714542e-07,
+      "loss": 0.7676,
+      "step": 73648
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.38453283274819e-07,
+      "loss": 0.6484,
+      "step": 73649
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.384274847550042e-07,
+      "loss": 0.7627,
+      "step": 73650
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.384016865120284e-07,
+      "loss": 0.6177,
+      "step": 73651
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.383758885459117e-07,
+      "loss": 0.6479,
+      "step": 73652
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.383500908566741e-07,
+      "loss": 0.793,
+      "step": 73653
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.38324293444335e-07,
+      "loss": 0.5723,
+      "step": 73654
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.382984963089147e-07,
+      "loss": 0.625,
+      "step": 73655
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.382726994504322e-07,
+      "loss": 0.6299,
+      "step": 73656
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.38246902868908e-07,
+      "loss": 0.6797,
+      "step": 73657
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.382211065643612e-07,
+      "loss": 0.6484,
+      "step": 73658
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.381953105368121e-07,
+      "loss": 0.5947,
+      "step": 73659
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.381695147862805e-07,
+      "loss": 0.6582,
+      "step": 73660
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.381437193127856e-07,
+      "loss": 0.7734,
+      "step": 73661
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.381179241163473e-07,
+      "loss": 0.5789,
+      "step": 73662
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.380921291969856e-07,
+      "loss": 0.624,
+      "step": 73663
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.380663345547199e-07,
+      "loss": 0.7578,
+      "step": 73664
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.380405401895703e-07,
+      "loss": 0.6855,
+      "step": 73665
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.380147461015564e-07,
+      "loss": 0.7061,
+      "step": 73666
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.379889522906979e-07,
+      "loss": 0.451,
+      "step": 73667
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.379631587570147e-07,
+      "loss": 0.7793,
+      "step": 73668
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.379373655005265e-07,
+      "loss": 0.7222,
+      "step": 73669
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.379115725212532e-07,
+      "loss": 0.5112,
+      "step": 73670
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.37885779819214e-07,
+      "loss": 0.6201,
+      "step": 73671
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.378599873944292e-07,
+      "loss": 0.8799,
+      "step": 73672
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.378341952469182e-07,
+      "loss": 0.749,
+      "step": 73673
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.37808403376701e-07,
+      "loss": 0.7236,
+      "step": 73674
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.377826117837973e-07,
+      "loss": 0.6538,
+      "step": 73675
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.377568204682267e-07,
+      "loss": 0.5908,
+      "step": 73676
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.377310294300089e-07,
+      "loss": 0.6914,
+      "step": 73677
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.377052386691642e-07,
+      "loss": 0.6489,
+      "step": 73678
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.376794481857115e-07,
+      "loss": 0.5718,
+      "step": 73679
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.376536579796716e-07,
+      "loss": 0.6001,
+      "step": 73680
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.376278680510632e-07,
+      "loss": 0.5786,
+      "step": 73681
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.376020783999066e-07,
+      "loss": 0.5923,
+      "step": 73682
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.375762890262213e-07,
+      "loss": 0.7354,
+      "step": 73683
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.375504999300272e-07,
+      "loss": 0.5649,
+      "step": 73684
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.37524711111344e-07,
+      "loss": 0.6343,
+      "step": 73685
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.374989225701915e-07,
+      "loss": 0.7256,
+      "step": 73686
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.374731343065893e-07,
+      "loss": 0.5898,
+      "step": 73687
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.374473463205574e-07,
+      "loss": 0.5588,
+      "step": 73688
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.374215586121152e-07,
+      "loss": 0.8115,
+      "step": 73689
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.373957711812832e-07,
+      "loss": 0.6416,
+      "step": 73690
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.373699840280802e-07,
+      "loss": 0.6992,
+      "step": 73691
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.373441971525262e-07,
+      "loss": 0.6855,
+      "step": 73692
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.373184105546412e-07,
+      "loss": 0.6284,
+      "step": 73693
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.372926242344447e-07,
+      "loss": 0.7148,
+      "step": 73694
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.372668381919567e-07,
+      "loss": 0.5645,
+      "step": 73695
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.372410524271966e-07,
+      "loss": 0.5354,
+      "step": 73696
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.372152669401847e-07,
+      "loss": 0.6685,
+      "step": 73697
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.3718948173094e-07,
+      "loss": 0.6553,
+      "step": 73698
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.371636967994829e-07,
+      "loss": 0.6611,
+      "step": 73699
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.371379121458331e-07,
+      "loss": 0.437,
+      "step": 73700
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.371121277700098e-07,
+      "loss": 0.4724,
+      "step": 73701
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.370863436720331e-07,
+      "loss": 0.4368,
+      "step": 73702
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.370605598519226e-07,
+      "loss": 0.6772,
+      "step": 73703
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.370347763096982e-07,
+      "loss": 0.6504,
+      "step": 73704
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.370089930453796e-07,
+      "loss": 0.4858,
+      "step": 73705
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.369832100589864e-07,
+      "loss": 0.666,
+      "step": 73706
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.369574273505387e-07,
+      "loss": 0.5869,
+      "step": 73707
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.369316449200557e-07,
+      "loss": 0.7129,
+      "step": 73708
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.369058627675576e-07,
+      "loss": 0.7207,
+      "step": 73709
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.368800808930644e-07,
+      "loss": 0.6011,
+      "step": 73710
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.368542992965949e-07,
+      "loss": 0.6763,
+      "step": 73711
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.368285179781696e-07,
+      "loss": 0.7129,
+      "step": 73712
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.368027369378077e-07,
+      "loss": 0.5898,
+      "step": 73713
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.367769561755296e-07,
+      "loss": 0.7764,
+      "step": 73714
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.367511756913544e-07,
+      "loss": 0.5825,
+      "step": 73715
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.367253954853023e-07,
+      "loss": 0.6387,
+      "step": 73716
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.36699615557393e-07,
+      "loss": 0.5029,
+      "step": 73717
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.366738359076458e-07,
+      "loss": 0.5437,
+      "step": 73718
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.366480565360809e-07,
+      "loss": 0.5083,
+      "step": 73719
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.366222774427182e-07,
+      "loss": 0.479,
+      "step": 73720
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.365964986275764e-07,
+      "loss": 0.6963,
+      "step": 73721
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.365707200906765e-07,
+      "loss": 0.4619,
+      "step": 73722
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.365449418320374e-07,
+      "loss": 0.7773,
+      "step": 73723
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.365191638516793e-07,
+      "loss": 0.585,
+      "step": 73724
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.364933861496216e-07,
+      "loss": 0.6475,
+      "step": 73725
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.364676087258842e-07,
+      "loss": 0.7168,
+      "step": 73726
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.364418315804869e-07,
+      "loss": 0.7222,
+      "step": 73727
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.364160547134493e-07,
+      "loss": 0.5361,
+      "step": 73728
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.363902781247913e-07,
+      "loss": 0.6572,
+      "step": 73729
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.363645018145329e-07,
+      "loss": 0.6982,
+      "step": 73730
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.363387257826931e-07,
+      "loss": 0.5811,
+      "step": 73731
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.363129500292919e-07,
+      "loss": 0.4814,
+      "step": 73732
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.362871745543494e-07,
+      "loss": 0.5576,
+      "step": 73733
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.362613993578851e-07,
+      "loss": 0.7012,
+      "step": 73734
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.362356244399184e-07,
+      "loss": 0.6475,
+      "step": 73735
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.362098498004696e-07,
+      "loss": 0.6504,
+      "step": 73736
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.36184075439558e-07,
+      "loss": 0.4807,
+      "step": 73737
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.361583013572038e-07,
+      "loss": 0.6396,
+      "step": 73738
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.361325275534262e-07,
+      "loss": 0.6011,
+      "step": 73739
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.361067540282457e-07,
+      "loss": 0.752,
+      "step": 73740
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.360809807816811e-07,
+      "loss": 0.665,
+      "step": 73741
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.360552078137525e-07,
+      "loss": 0.7324,
+      "step": 73742
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.360294351244799e-07,
+      "loss": 0.6089,
+      "step": 73743
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.360036627138825e-07,
+      "loss": 0.6196,
+      "step": 73744
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.359778905819808e-07,
+      "loss": 0.5337,
+      "step": 73745
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.359521187287937e-07,
+      "loss": 0.6885,
+      "step": 73746
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.359263471543415e-07,
+      "loss": 0.7007,
+      "step": 73747
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.359005758586436e-07,
+      "loss": 0.709,
+      "step": 73748
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.358748048417201e-07,
+      "loss": 0.6211,
+      "step": 73749
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.358490341035908e-07,
+      "loss": 0.7773,
+      "step": 73750
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.358232636442749e-07,
+      "loss": 0.5967,
+      "step": 73751
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.357974934637922e-07,
+      "loss": 0.541,
+      "step": 73752
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.357717235621627e-07,
+      "loss": 0.6172,
+      "step": 73753
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.357459539394059e-07,
+      "loss": 0.6846,
+      "step": 73754
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.35720184595542e-07,
+      "loss": 0.2909,
+      "step": 73755
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.356944155305901e-07,
+      "loss": 0.5,
+      "step": 73756
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.356686467445704e-07,
+      "loss": 0.6875,
+      "step": 73757
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.356428782375023e-07,
+      "loss": 0.6914,
+      "step": 73758
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.356171100094059e-07,
+      "loss": 0.7373,
+      "step": 73759
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.35591342060301e-07,
+      "loss": 0.3826,
+      "step": 73760
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.355655743902064e-07,
+      "loss": 0.791,
+      "step": 73761
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.355398069991429e-07,
+      "loss": 0.5856,
+      "step": 73762
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.355140398871296e-07,
+      "loss": 0.5254,
+      "step": 73763
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.354882730541866e-07,
+      "loss": 0.5813,
+      "step": 73764
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.354625065003333e-07,
+      "loss": 0.6392,
+      "step": 73765
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.354367402255898e-07,
+      "loss": 0.7705,
+      "step": 73766
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.354109742299754e-07,
+      "loss": 0.6445,
+      "step": 73767
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.353852085135101e-07,
+      "loss": 0.665,
+      "step": 73768
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.353594430762137e-07,
+      "loss": 0.6436,
+      "step": 73769
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.353336779181062e-07,
+      "loss": 0.7339,
+      "step": 73770
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.353079130392063e-07,
+      "loss": 0.7598,
+      "step": 73771
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.352821484395346e-07,
+      "loss": 0.5432,
+      "step": 73772
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.352563841191106e-07,
+      "loss": 0.666,
+      "step": 73773
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.352306200779539e-07,
+      "loss": 0.5923,
+      "step": 73774
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.352048563160843e-07,
+      "loss": 0.593,
+      "step": 73775
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.351790928335218e-07,
+      "loss": 0.4641,
+      "step": 73776
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.351533296302856e-07,
+      "loss": 0.4824,
+      "step": 73777
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.35127566706396e-07,
+      "loss": 0.5762,
+      "step": 73778
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.351018040618721e-07,
+      "loss": 0.7275,
+      "step": 73779
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.350760416967346e-07,
+      "loss": 0.6494,
+      "step": 73780
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.350502796110023e-07,
+      "loss": 0.8096,
+      "step": 73781
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.350245178046949e-07,
+      "loss": 0.7246,
+      "step": 73782
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.349987562778327e-07,
+      "loss": 0.5757,
+      "step": 73783
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.34972995030435e-07,
+      "loss": 0.5244,
+      "step": 73784
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.349472340625219e-07,
+      "loss": 0.6611,
+      "step": 73785
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.349214733741129e-07,
+      "loss": 0.5732,
+      "step": 73786
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.348957129652277e-07,
+      "loss": 0.644,
+      "step": 73787
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.348699528358859e-07,
+      "loss": 0.6353,
+      "step": 73788
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.348441929861074e-07,
+      "loss": 0.6543,
+      "step": 73789
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.348184334159124e-07,
+      "loss": 0.4268,
+      "step": 73790
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.347926741253201e-07,
+      "loss": 0.6484,
+      "step": 73791
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.347669151143497e-07,
+      "loss": 0.7002,
+      "step": 73792
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.347411563830218e-07,
+      "loss": 0.7627,
+      "step": 73793
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.347153979313555e-07,
+      "loss": 0.6025,
+      "step": 73794
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.346896397593711e-07,
+      "loss": 0.4805,
+      "step": 73795
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.346638818670879e-07,
+      "loss": 0.6992,
+      "step": 73796
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.34638124254526e-07,
+      "loss": 0.7334,
+      "step": 73797
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.346123669217045e-07,
+      "loss": 0.5938,
+      "step": 73798
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.345866098686439e-07,
+      "loss": 0.6787,
+      "step": 73799
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.345608530953636e-07,
+      "loss": 0.5994,
+      "step": 73800
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.345350966018829e-07,
+      "loss": 0.6694,
+      "step": 73801
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.345093403882221e-07,
+      "loss": 0.5767,
+      "step": 73802
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.344835844544005e-07,
+      "loss": 0.624,
+      "step": 73803
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.344578288004382e-07,
+      "loss": 0.5693,
+      "step": 73804
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.344320734263546e-07,
+      "loss": 0.6042,
+      "step": 73805
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.344063183321695e-07,
+      "loss": 0.5977,
+      "step": 73806
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.34380563517903e-07,
+      "loss": 0.4648,
+      "step": 73807
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.34354808983574e-07,
+      "loss": 0.6016,
+      "step": 73808
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.343290547292031e-07,
+      "loss": 0.5186,
+      "step": 73809
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.343033007548097e-07,
+      "loss": 0.6592,
+      "step": 73810
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.342775470604131e-07,
+      "loss": 0.7148,
+      "step": 73811
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.342517936460337e-07,
+      "loss": 0.5376,
+      "step": 73812
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.342260405116906e-07,
+      "loss": 0.6753,
+      "step": 73813
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.34200287657404e-07,
+      "loss": 0.667,
+      "step": 73814
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.341745350831931e-07,
+      "loss": 0.7607,
+      "step": 73815
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.341487827890782e-07,
+      "loss": 0.6743,
+      "step": 73816
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.341230307750787e-07,
+      "loss": 0.563,
+      "step": 73817
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.340972790412144e-07,
+      "loss": 0.6533,
+      "step": 73818
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.340715275875049e-07,
+      "loss": 0.5505,
+      "step": 73819
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.340457764139703e-07,
+      "loss": 0.5728,
+      "step": 73820
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.340200255206299e-07,
+      "loss": 0.5889,
+      "step": 73821
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.339942749075032e-07,
+      "loss": 0.6104,
+      "step": 73822
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.339685245746106e-07,
+      "loss": 0.5894,
+      "step": 73823
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.339427745219714e-07,
+      "loss": 0.5869,
+      "step": 73824
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.339170247496053e-07,
+      "loss": 0.6099,
+      "step": 73825
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.338912752575322e-07,
+      "loss": 0.6069,
+      "step": 73826
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.338655260457715e-07,
+      "loss": 0.7168,
+      "step": 73827
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.338397771143434e-07,
+      "loss": 0.6492,
+      "step": 73828
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.338140284632671e-07,
+      "loss": 0.5469,
+      "step": 73829
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.337882800925629e-07,
+      "loss": 0.7627,
+      "step": 73830
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.337625320022501e-07,
+      "loss": 0.5396,
+      "step": 73831
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.337367841923483e-07,
+      "loss": 0.7041,
+      "step": 73832
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.337110366628775e-07,
+      "loss": 0.7188,
+      "step": 73833
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.336852894138571e-07,
+      "loss": 0.5469,
+      "step": 73834
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.336595424453073e-07,
+      "loss": 0.5039,
+      "step": 73835
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.336337957572474e-07,
+      "loss": 0.6992,
+      "step": 73836
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.336080493496974e-07,
+      "loss": 0.5503,
+      "step": 73837
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.335823032226766e-07,
+      "loss": 0.5317,
+      "step": 73838
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.335565573762053e-07,
+      "loss": 0.7383,
+      "step": 73839
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.335308118103031e-07,
+      "loss": 0.6582,
+      "step": 73840
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.335050665249892e-07,
+      "loss": 0.6514,
+      "step": 73841
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.334793215202834e-07,
+      "loss": 0.6421,
+      "step": 73842
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.334535767962059e-07,
+      "loss": 0.748,
+      "step": 73843
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.33427832352776e-07,
+      "loss": 0.5938,
+      "step": 73844
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.334020881900136e-07,
+      "loss": 0.8506,
+      "step": 73845
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.333763443079383e-07,
+      "loss": 0.709,
+      "step": 73846
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.333506007065701e-07,
+      "loss": 0.666,
+      "step": 73847
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.333248573859281e-07,
+      "loss": 0.6924,
+      "step": 73848
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.332991143460327e-07,
+      "loss": 0.4695,
+      "step": 73849
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.332733715869036e-07,
+      "loss": 0.5908,
+      "step": 73850
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.332476291085597e-07,
+      "loss": 0.6514,
+      "step": 73851
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.332218869110215e-07,
+      "loss": 0.5918,
+      "step": 73852
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.331961449943082e-07,
+      "loss": 0.5569,
+      "step": 73853
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.331704033584399e-07,
+      "loss": 0.6475,
+      "step": 73854
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.331446620034359e-07,
+      "loss": 0.6851,
+      "step": 73855
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.331189209293166e-07,
+      "loss": 0.5962,
+      "step": 73856
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.330931801361008e-07,
+      "loss": 0.6699,
+      "step": 73857
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.330674396238091e-07,
+      "loss": 0.6826,
+      "step": 73858
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.330416993924605e-07,
+      "loss": 0.4995,
+      "step": 73859
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.330159594420756e-07,
+      "loss": 0.6396,
+      "step": 73860
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.329902197726729e-07,
+      "loss": 0.5591,
+      "step": 73861
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.329644803842728e-07,
+      "loss": 0.4783,
+      "step": 73862
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.329387412768949e-07,
+      "loss": 0.5806,
+      "step": 73863
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.329130024505591e-07,
+      "loss": 0.5791,
+      "step": 73864
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.328872639052846e-07,
+      "loss": 0.6543,
+      "step": 73865
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.328615256410918e-07,
+      "loss": 0.5742,
+      "step": 73866
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.32835787658e-07,
+      "loss": 0.667,
+      "step": 73867
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.328100499560289e-07,
+      "loss": 0.4858,
+      "step": 73868
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.32784312535198e-07,
+      "loss": 0.8369,
+      "step": 73869
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.327585753955278e-07,
+      "loss": 0.5728,
+      "step": 73870
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.327328385370373e-07,
+      "loss": 0.5142,
+      "step": 73871
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.327071019597461e-07,
+      "loss": 0.6689,
+      "step": 73872
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.326813656636745e-07,
+      "loss": 0.6221,
+      "step": 73873
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.326556296488414e-07,
+      "loss": 0.5281,
+      "step": 73874
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.326298939152675e-07,
+      "loss": 0.5552,
+      "step": 73875
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.326041584629716e-07,
+      "loss": 0.4291,
+      "step": 73876
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.325784232919742e-07,
+      "loss": 0.4771,
+      "step": 73877
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.325526884022943e-07,
+      "loss": 0.7119,
+      "step": 73878
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.325269537939522e-07,
+      "loss": 0.7188,
+      "step": 73879
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.325012194669673e-07,
+      "loss": 0.6382,
+      "step": 73880
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.324754854213592e-07,
+      "loss": 0.5249,
+      "step": 73881
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.324497516571477e-07,
+      "loss": 0.5852,
+      "step": 73882
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.324240181743526e-07,
+      "loss": 0.6132,
+      "step": 73883
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.323982849729932e-07,
+      "loss": 0.709,
+      "step": 73884
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.323725520530898e-07,
+      "loss": 0.6021,
+      "step": 73885
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.323468194146618e-07,
+      "loss": 0.7979,
+      "step": 73886
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.323210870577289e-07,
+      "loss": 0.4873,
+      "step": 73887
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.322953549823109e-07,
+      "loss": 0.603,
+      "step": 73888
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.322696231884277e-07,
+      "loss": 0.71,
+      "step": 73889
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.322438916760985e-07,
+      "loss": 0.4685,
+      "step": 73890
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.32218160445343e-07,
+      "loss": 0.7881,
+      "step": 73891
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.321924294961814e-07,
+      "loss": 0.4695,
+      "step": 73892
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.321666988286329e-07,
+      "loss": 0.689,
+      "step": 73893
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.321409684427178e-07,
+      "loss": 0.4839,
+      "step": 73894
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.321152383384552e-07,
+      "loss": 0.6729,
+      "step": 73895
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.320895085158652e-07,
+      "loss": 0.6938,
+      "step": 73896
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.320637789749673e-07,
+      "loss": 0.6133,
+      "step": 73897
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.320380497157811e-07,
+      "loss": 0.5525,
+      "step": 73898
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.320123207383269e-07,
+      "loss": 0.647,
+      "step": 73899
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.319865920426238e-07,
+      "loss": 0.7539,
+      "step": 73900
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.319608636286912e-07,
+      "loss": 0.5629,
+      "step": 73901
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.319351354965496e-07,
+      "loss": 0.6201,
+      "step": 73902
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.319094076462181e-07,
+      "loss": 0.564,
+      "step": 73903
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.318836800777168e-07,
+      "loss": 0.6768,
+      "step": 73904
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.318579527910651e-07,
+      "loss": 0.7549,
+      "step": 73905
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.318322257862831e-07,
+      "loss": 0.5645,
+      "step": 73906
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.3180649906339e-07,
+      "loss": 0.5251,
+      "step": 73907
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.317807726224058e-07,
+      "loss": 0.4036,
+      "step": 73908
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.317550464633505e-07,
+      "loss": 0.6396,
+      "step": 73909
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.317293205862429e-07,
+      "loss": 0.5815,
+      "step": 73910
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.317035949911035e-07,
+      "loss": 0.6953,
+      "step": 73911
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.316778696779515e-07,
+      "loss": 0.7832,
+      "step": 73912
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.31652144646807e-07,
+      "loss": 0.7197,
+      "step": 73913
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.316264198976896e-07,
+      "loss": 0.5244,
+      "step": 73914
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.316006954306186e-07,
+      "loss": 0.7764,
+      "step": 73915
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.315749712456144e-07,
+      "loss": 0.7549,
+      "step": 73916
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.31549247342696e-07,
+      "loss": 0.5256,
+      "step": 73917
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.315235237218834e-07,
+      "loss": 0.7539,
+      "step": 73918
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.314978003831968e-07,
+      "loss": 0.7441,
+      "step": 73919
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.314720773266549e-07,
+      "loss": 0.751,
+      "step": 73920
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.314463545522779e-07,
+      "loss": 0.7568,
+      "step": 73921
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.314206320600854e-07,
+      "loss": 0.6816,
+      "step": 73922
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.313949098500975e-07,
+      "loss": 0.4771,
+      "step": 73923
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.313691879223333e-07,
+      "loss": 0.8701,
+      "step": 73924
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.313434662768129e-07,
+      "loss": 0.6631,
+      "step": 73925
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.313177449135556e-07,
+      "loss": 0.5918,
+      "step": 73926
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.312920238325816e-07,
+      "loss": 0.7319,
+      "step": 73927
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.312663030339103e-07,
+      "loss": 0.5889,
+      "step": 73928
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.312405825175617e-07,
+      "loss": 0.5659,
+      "step": 73929
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.31214862283555e-07,
+      "loss": 0.6885,
+      "step": 73930
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.311891423319103e-07,
+      "loss": 0.6348,
+      "step": 73931
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.311634226626467e-07,
+      "loss": 0.5688,
+      "step": 73932
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.311377032757846e-07,
+      "loss": 0.661,
+      "step": 73933
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.311119841713432e-07,
+      "loss": 0.4653,
+      "step": 73934
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 6.310862653493425e-07,
+      "loss": 0.7236,
+      "step": 73935
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.31060546809802e-07,
+      "loss": 0.6333,
+      "step": 73936
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.310348285527416e-07,
+      "loss": 0.604,
+      "step": 73937
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.310091105781808e-07,
+      "loss": 0.6226,
+      "step": 73938
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.309833928861397e-07,
+      "loss": 0.7412,
+      "step": 73939
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.309576754766374e-07,
+      "loss": 0.667,
+      "step": 73940
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.309319583496936e-07,
+      "loss": 0.5391,
+      "step": 73941
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.309062415053285e-07,
+      "loss": 0.6621,
+      "step": 73942
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.308805249435613e-07,
+      "loss": 0.7354,
+      "step": 73943
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30854808664412e-07,
+      "loss": 0.9414,
+      "step": 73944
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.308290926679002e-07,
+      "loss": 0.5786,
+      "step": 73945
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.308033769540457e-07,
+      "loss": 0.4644,
+      "step": 73946
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30777661522868e-07,
+      "loss": 0.6475,
+      "step": 73947
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.307519463743868e-07,
+      "loss": 0.5498,
+      "step": 73948
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.307262315086221e-07,
+      "loss": 0.6567,
+      "step": 73949
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.307005169255934e-07,
+      "loss": 0.7412,
+      "step": 73950
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.306748026253199e-07,
+      "loss": 0.4592,
+      "step": 73951
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30649088607822e-07,
+      "loss": 0.6885,
+      "step": 73952
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.306233748731189e-07,
+      "loss": 0.6074,
+      "step": 73953
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.305976614212307e-07,
+      "loss": 0.5571,
+      "step": 73954
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.305719482521768e-07,
+      "loss": 0.6191,
+      "step": 73955
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.305462353659769e-07,
+      "loss": 0.5278,
+      "step": 73956
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.305205227626508e-07,
+      "loss": 0.6333,
+      "step": 73957
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.304948104422183e-07,
+      "loss": 0.793,
+      "step": 73958
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.304690984046991e-07,
+      "loss": 0.7686,
+      "step": 73959
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.304433866501123e-07,
+      "loss": 0.6934,
+      "step": 73960
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.304176751784782e-07,
+      "loss": 0.645,
+      "step": 73961
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.303919639898162e-07,
+      "loss": 0.7393,
+      "step": 73962
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.303662530841463e-07,
+      "loss": 0.533,
+      "step": 73963
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.303405424614876e-07,
+      "loss": 0.6221,
+      "step": 73964
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.303148321218603e-07,
+      "loss": 0.6553,
+      "step": 73965
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30289122065284e-07,
+      "loss": 0.6045,
+      "step": 73966
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.302634122917783e-07,
+      "loss": 0.6768,
+      "step": 73967
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.302377028013628e-07,
+      "loss": 0.7734,
+      "step": 73968
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.302119935940578e-07,
+      "loss": 0.6562,
+      "step": 73969
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30186284669882e-07,
+      "loss": 0.5918,
+      "step": 73970
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.301605760288556e-07,
+      "loss": 0.6572,
+      "step": 73971
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.301348676709981e-07,
+      "loss": 0.7227,
+      "step": 73972
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.301091595963296e-07,
+      "loss": 0.7852,
+      "step": 73973
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.300834518048692e-07,
+      "loss": 0.5635,
+      "step": 73974
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.300577442966372e-07,
+      "loss": 0.5903,
+      "step": 73975
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.300320370716528e-07,
+      "loss": 0.708,
+      "step": 73976
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.30006330129936e-07,
+      "loss": 0.7471,
+      "step": 73977
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.299806234715062e-07,
+      "loss": 0.6086,
+      "step": 73978
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.299549170963834e-07,
+      "loss": 0.5286,
+      "step": 73979
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.299292110045871e-07,
+      "loss": 0.5879,
+      "step": 73980
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.299035051961367e-07,
+      "loss": 0.625,
+      "step": 73981
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.298777996710525e-07,
+      "loss": 0.6191,
+      "step": 73982
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.298520944293534e-07,
+      "loss": 0.3274,
+      "step": 73983
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.2982638947106e-07,
+      "loss": 0.3811,
+      "step": 73984
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.298006847961912e-07,
+      "loss": 0.6133,
+      "step": 73985
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.297749804047672e-07,
+      "loss": 0.5225,
+      "step": 73986
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.297492762968075e-07,
+      "loss": 0.6143,
+      "step": 73987
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.297235724723316e-07,
+      "loss": 0.6167,
+      "step": 73988
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.296978689313597e-07,
+      "loss": 0.6562,
+      "step": 73989
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.296721656739108e-07,
+      "loss": 0.6953,
+      "step": 73990
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.296464627000048e-07,
+      "loss": 0.7412,
+      "step": 73991
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.296207600096616e-07,
+      "loss": 0.5254,
+      "step": 73992
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.295950576029005e-07,
+      "loss": 0.583,
+      "step": 73993
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.295693554797419e-07,
+      "loss": 0.6111,
+      "step": 73994
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.295436536402046e-07,
+      "loss": 0.5415,
+      "step": 73995
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.295179520843089e-07,
+      "loss": 0.6118,
+      "step": 73996
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.294922508120741e-07,
+      "loss": 0.4399,
+      "step": 73997
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.294665498235203e-07,
+      "loss": 0.6234,
+      "step": 73998
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.294408491186671e-07,
+      "loss": 0.7148,
+      "step": 73999
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.294151486975335e-07,
+      "loss": 0.7412,
+      "step": 74000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.293894485601399e-07,
+      "loss": 0.5084,
+      "step": 74001
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.293637487065058e-07,
+      "loss": 0.6875,
+      "step": 74002
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.293380491366508e-07,
+      "loss": 0.6172,
+      "step": 74003
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.293123498505946e-07,
+      "loss": 0.6089,
+      "step": 74004
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.292866508483568e-07,
+      "loss": 0.6631,
+      "step": 74005
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.292609521299573e-07,
+      "loss": 0.7158,
+      "step": 74006
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.292352536954156e-07,
+      "loss": 0.6177,
+      "step": 74007
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.292095555447514e-07,
+      "loss": 0.4089,
+      "step": 74008
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.291838576779847e-07,
+      "loss": 0.8145,
+      "step": 74009
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.291581600951343e-07,
+      "loss": 0.6069,
+      "step": 74010
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.291324627962209e-07,
+      "loss": 0.5601,
+      "step": 74011
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.291067657812634e-07,
+      "loss": 0.7305,
+      "step": 74012
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.290810690502819e-07,
+      "loss": 0.6006,
+      "step": 74013
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.290553726032959e-07,
+      "loss": 0.4431,
+      "step": 74014
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.290296764403253e-07,
+      "loss": 0.6157,
+      "step": 74015
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.290039805613895e-07,
+      "loss": 0.6602,
+      "step": 74016
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.289782849665084e-07,
+      "loss": 0.5349,
+      "step": 74017
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.289525896557014e-07,
+      "loss": 0.644,
+      "step": 74018
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.289268946289887e-07,
+      "loss": 0.481,
+      "step": 74019
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.289011998863895e-07,
+      "loss": 0.6099,
+      "step": 74020
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.288755054279234e-07,
+      "loss": 0.6113,
+      "step": 74021
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.288498112536103e-07,
+      "loss": 0.6948,
+      "step": 74022
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.2882411736347e-07,
+      "loss": 0.6318,
+      "step": 74023
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.287984237575218e-07,
+      "loss": 0.7329,
+      "step": 74024
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.287727304357857e-07,
+      "loss": 0.7109,
+      "step": 74025
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.287470373982811e-07,
+      "loss": 0.9287,
+      "step": 74026
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.287213446450279e-07,
+      "loss": 0.6689,
+      "step": 74027
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.286956521760455e-07,
+      "loss": 0.7285,
+      "step": 74028
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.286699599913544e-07,
+      "loss": 0.6982,
+      "step": 74029
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.286442680909734e-07,
+      "loss": 0.7734,
+      "step": 74030
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.28618576474922e-07,
+      "loss": 0.707,
+      "step": 74031
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.285928851432205e-07,
+      "loss": 0.6387,
+      "step": 74032
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.285671940958883e-07,
+      "loss": 0.7207,
+      "step": 74033
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.285415033329453e-07,
+      "loss": 0.6802,
+      "step": 74034
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.285158128544107e-07,
+      "loss": 0.7363,
+      "step": 74035
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.284901226603048e-07,
+      "loss": 0.5801,
+      "step": 74036
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.284644327506466e-07,
+      "loss": 0.7812,
+      "step": 74037
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.284387431254563e-07,
+      "loss": 0.6709,
+      "step": 74038
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.284130537847536e-07,
+      "loss": 0.5488,
+      "step": 74039
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.283873647285578e-07,
+      "loss": 0.6606,
+      "step": 74040
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.283616759568883e-07,
+      "loss": 0.6748,
+      "step": 74041
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.283359874697656e-07,
+      "loss": 0.6895,
+      "step": 74042
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.283102992672087e-07,
+      "loss": 0.5371,
+      "step": 74043
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.282846113492377e-07,
+      "loss": 0.6709,
+      "step": 74044
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.282589237158717e-07,
+      "loss": 0.7236,
+      "step": 74045
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.282332363671312e-07,
+      "loss": 0.6973,
+      "step": 74046
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.282075493030352e-07,
+      "loss": 0.5442,
+      "step": 74047
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.281818625236037e-07,
+      "loss": 0.5601,
+      "step": 74048
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.281561760288565e-07,
+      "loss": 0.5066,
+      "step": 74049
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.281304898188126e-07,
+      "loss": 0.7119,
+      "step": 74050
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.281048038934922e-07,
+      "loss": 0.6621,
+      "step": 74051
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.280791182529147e-07,
+      "loss": 0.5664,
+      "step": 74052
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.280534328971001e-07,
+      "loss": 0.5586,
+      "step": 74053
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.280277478260678e-07,
+      "loss": 0.6465,
+      "step": 74054
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.280020630398377e-07,
+      "loss": 0.7407,
+      "step": 74055
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.279763785384291e-07,
+      "loss": 0.5752,
+      "step": 74056
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.27950694321862e-07,
+      "loss": 0.521,
+      "step": 74057
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.279250103901559e-07,
+      "loss": 0.5503,
+      "step": 74058
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.27899326743331e-07,
+      "loss": 0.4269,
+      "step": 74059
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.278736433814058e-07,
+      "loss": 0.5073,
+      "step": 74060
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.27847960304401e-07,
+      "loss": 0.6992,
+      "step": 74061
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.278222775123357e-07,
+      "loss": 0.7188,
+      "step": 74062
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.277965950052299e-07,
+      "loss": 0.6924,
+      "step": 74063
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.277709127831029e-07,
+      "loss": 0.498,
+      "step": 74064
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.277452308459749e-07,
+      "loss": 0.6841,
+      "step": 74065
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.277195491938649e-07,
+      "loss": 0.7002,
+      "step": 74066
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.276938678267933e-07,
+      "loss": 0.646,
+      "step": 74067
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.276681867447791e-07,
+      "loss": 0.6875,
+      "step": 74068
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.276425059478427e-07,
+      "loss": 0.4581,
+      "step": 74069
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.27616825436003e-07,
+      "loss": 0.5806,
+      "step": 74070
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.275911452092799e-07,
+      "loss": 0.7383,
+      "step": 74071
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.275654652676933e-07,
+      "loss": 0.7227,
+      "step": 74072
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.275397856112625e-07,
+      "loss": 0.4426,
+      "step": 74073
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.275141062400076e-07,
+      "loss": 0.6699,
+      "step": 74074
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.274884271539477e-07,
+      "loss": 0.6855,
+      "step": 74075
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.274627483531029e-07,
+      "loss": 0.7568,
+      "step": 74076
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.274370698374932e-07,
+      "loss": 0.6943,
+      "step": 74077
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.274113916071371e-07,
+      "loss": 0.6348,
+      "step": 74078
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.273857136620556e-07,
+      "loss": 0.7012,
+      "step": 74079
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.273600360022675e-07,
+      "loss": 0.5874,
+      "step": 74080
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.273343586277925e-07,
+      "loss": 0.5723,
+      "step": 74081
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.273086815386505e-07,
+      "loss": 0.6299,
+      "step": 74082
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.272830047348612e-07,
+      "loss": 0.6343,
+      "step": 74083
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.272573282164441e-07,
+      "loss": 0.6084,
+      "step": 74084
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.272316519834188e-07,
+      "loss": 0.5039,
+      "step": 74085
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.272059760358052e-07,
+      "loss": 0.5259,
+      "step": 74086
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.271803003736229e-07,
+      "loss": 0.6357,
+      "step": 74087
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.271546249968915e-07,
+      "loss": 0.614,
+      "step": 74088
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.271289499056309e-07,
+      "loss": 0.6934,
+      "step": 74089
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.271032750998601e-07,
+      "loss": 0.7451,
+      "step": 74090
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.270776005795995e-07,
+      "loss": 0.5732,
+      "step": 74091
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.270519263448682e-07,
+      "loss": 0.6123,
+      "step": 74092
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.270262523956861e-07,
+      "loss": 0.5808,
+      "step": 74093
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.270005787320729e-07,
+      "loss": 0.6777,
+      "step": 74094
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26974905354048e-07,
+      "loss": 0.6509,
+      "step": 74095
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.269492322616317e-07,
+      "loss": 0.5889,
+      "step": 74096
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.269235594548427e-07,
+      "loss": 0.5903,
+      "step": 74097
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.268978869337016e-07,
+      "loss": 0.4268,
+      "step": 74098
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.268722146982279e-07,
+      "loss": 0.4805,
+      "step": 74099
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.268465427484404e-07,
+      "loss": 0.707,
+      "step": 74100
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.268208710843597e-07,
+      "loss": 0.5645,
+      "step": 74101
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.267951997060047e-07,
+      "loss": 0.7051,
+      "step": 74102
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.267695286133958e-07,
+      "loss": 0.5996,
+      "step": 74103
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26743857806552e-07,
+      "loss": 0.5676,
+      "step": 74104
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.267181872854936e-07,
+      "loss": 0.686,
+      "step": 74105
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.266925170502396e-07,
+      "loss": 0.5254,
+      "step": 74106
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.266668471008104e-07,
+      "loss": 0.5034,
+      "step": 74107
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.266411774372247e-07,
+      "loss": 0.7578,
+      "step": 74108
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.266155080595032e-07,
+      "loss": 0.5913,
+      "step": 74109
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.265898389676649e-07,
+      "loss": 0.6287,
+      "step": 74110
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.265641701617294e-07,
+      "loss": 0.6826,
+      "step": 74111
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.265385016417167e-07,
+      "loss": 0.6768,
+      "step": 74112
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.265128334076463e-07,
+      "loss": 0.5752,
+      "step": 74113
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.264871654595376e-07,
+      "loss": 0.5835,
+      "step": 74114
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.264614977974107e-07,
+      "loss": 0.7549,
+      "step": 74115
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26435830421285e-07,
+      "loss": 0.5127,
+      "step": 74116
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.264101633311803e-07,
+      "loss": 0.5747,
+      "step": 74117
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26384496527116e-07,
+      "loss": 0.4912,
+      "step": 74118
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.263588300091123e-07,
+      "loss": 0.5503,
+      "step": 74119
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.263331637771883e-07,
+      "loss": 0.6152,
+      "step": 74120
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.263074978313636e-07,
+      "loss": 0.5688,
+      "step": 74121
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.262818321716582e-07,
+      "loss": 0.4856,
+      "step": 74122
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.262561667980914e-07,
+      "loss": 0.6025,
+      "step": 74123
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.262305017106833e-07,
+      "loss": 0.7041,
+      "step": 74124
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.262048369094532e-07,
+      "loss": 0.3745,
+      "step": 74125
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26179172394421e-07,
+      "loss": 0.6519,
+      "step": 74126
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.26153508165606e-07,
+      "loss": 0.7549,
+      "step": 74127
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.261278442230284e-07,
+      "loss": 0.6406,
+      "step": 74128
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.261021805667075e-07,
+      "loss": 0.6436,
+      "step": 74129
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.260765171966629e-07,
+      "loss": 0.6323,
+      "step": 74130
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.260508541129141e-07,
+      "loss": 0.5737,
+      "step": 74131
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.260251913154813e-07,
+      "loss": 0.5908,
+      "step": 74132
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.259995288043834e-07,
+      "loss": 0.4248,
+      "step": 74133
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.259738665796408e-07,
+      "loss": 0.5842,
+      "step": 74134
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.259482046412725e-07,
+      "loss": 0.6709,
+      "step": 74135
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.259225429892987e-07,
+      "loss": 0.5818,
+      "step": 74136
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.258968816237387e-07,
+      "loss": 0.5054,
+      "step": 74137
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.258712205446123e-07,
+      "loss": 0.6797,
+      "step": 74138
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.258455597519394e-07,
+      "loss": 0.5325,
+      "step": 74139
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.25819899245739e-07,
+      "loss": 0.6797,
+      "step": 74140
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.257942390260312e-07,
+      "loss": 0.6338,
+      "step": 74141
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.257685790928353e-07,
+      "loss": 0.7344,
+      "step": 74142
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.257429194461717e-07,
+      "loss": 0.5757,
+      "step": 74143
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.25717260086059e-07,
+      "loss": 0.6943,
+      "step": 74144
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.256916010125177e-07,
+      "loss": 0.6572,
+      "step": 74145
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.256659422255669e-07,
+      "loss": 0.7275,
+      "step": 74146
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.256402837252268e-07,
+      "loss": 0.6128,
+      "step": 74147
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.256146255115162e-07,
+      "loss": 0.708,
+      "step": 74148
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.255889675844562e-07,
+      "loss": 0.593,
+      "step": 74149
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.255633099440648e-07,
+      "loss": 0.748,
+      "step": 74150
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.255376525903626e-07,
+      "loss": 0.6348,
+      "step": 74151
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.255119955233687e-07,
+      "loss": 0.4338,
+      "step": 74152
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.254863387431033e-07,
+      "loss": 0.521,
+      "step": 74153
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.254606822495855e-07,
+      "loss": 0.7163,
+      "step": 74154
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.254350260428356e-07,
+      "loss": 0.6514,
+      "step": 74155
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.254093701228727e-07,
+      "loss": 0.6182,
+      "step": 74156
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.253837144897167e-07,
+      "loss": 0.575,
+      "step": 74157
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.253580591433869e-07,
+      "loss": 0.6021,
+      "step": 74158
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.253324040839038e-07,
+      "loss": 0.6523,
+      "step": 74159
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.253067493112861e-07,
+      "loss": 0.4763,
+      "step": 74160
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.252810948255536e-07,
+      "loss": 0.3882,
+      "step": 74161
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.252554406267265e-07,
+      "loss": 0.8906,
+      "step": 74162
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.252297867148237e-07,
+      "loss": 0.5908,
+      "step": 74163
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.252041330898655e-07,
+      "loss": 0.6191,
+      "step": 74164
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.251784797518709e-07,
+      "loss": 0.5605,
+      "step": 74165
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.251528267008603e-07,
+      "loss": 0.6611,
+      "step": 74166
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.251271739368529e-07,
+      "loss": 0.5957,
+      "step": 74167
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.251015214598681e-07,
+      "loss": 0.7109,
+      "step": 74168
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.250758692699264e-07,
+      "loss": 0.3484,
+      "step": 74169
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.250502173670466e-07,
+      "loss": 0.6206,
+      "step": 74170
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.250245657512482e-07,
+      "loss": 0.6719,
+      "step": 74171
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.249989144225518e-07,
+      "loss": 0.5601,
+      "step": 74172
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.249732633809761e-07,
+      "loss": 0.5903,
+      "step": 74173
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.249476126265412e-07,
+      "loss": 0.4106,
+      "step": 74174
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.249219621592667e-07,
+      "loss": 0.7207,
+      "step": 74175
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.248963119791723e-07,
+      "loss": 0.5107,
+      "step": 74176
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.248706620862773e-07,
+      "loss": 0.6328,
+      "step": 74177
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.248450124806022e-07,
+      "loss": 0.6738,
+      "step": 74178
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.248193631621656e-07,
+      "loss": 0.6465,
+      "step": 74179
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.247937141309874e-07,
+      "loss": 0.6875,
+      "step": 74180
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.247680653870877e-07,
+      "loss": 0.7197,
+      "step": 74181
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.247424169304855e-07,
+      "loss": 0.5845,
+      "step": 74182
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.247167687612011e-07,
+      "loss": 0.6622,
+      "step": 74183
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.246911208792536e-07,
+      "loss": 0.5066,
+      "step": 74184
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.246654732846629e-07,
+      "loss": 0.5454,
+      "step": 74185
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.246398259774487e-07,
+      "loss": 0.5671,
+      "step": 74186
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.246141789576304e-07,
+      "loss": 0.5615,
+      "step": 74187
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.24588532225228e-07,
+      "loss": 0.6572,
+      "step": 74188
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.245628857802609e-07,
+      "loss": 0.6211,
+      "step": 74189
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.245372396227484e-07,
+      "loss": 0.4089,
+      "step": 74190
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.245115937527106e-07,
+      "loss": 0.8613,
+      "step": 74191
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.24485948170167e-07,
+      "loss": 0.5811,
+      "step": 74192
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.244603028751373e-07,
+      "loss": 0.5542,
+      "step": 74193
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.24434657867641e-07,
+      "loss": 0.6348,
+      "step": 74194
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.244090131476978e-07,
+      "loss": 0.5801,
+      "step": 74195
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.243833687153273e-07,
+      "loss": 0.7021,
+      "step": 74196
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.243577245705494e-07,
+      "loss": 0.4941,
+      "step": 74197
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.243320807133836e-07,
+      "loss": 0.6597,
+      "step": 74198
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.243064371438492e-07,
+      "loss": 0.7432,
+      "step": 74199
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.242807938619663e-07,
+      "loss": 0.6206,
+      "step": 74200
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.242551508677541e-07,
+      "loss": 0.6768,
+      "step": 74201
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.242295081612325e-07,
+      "loss": 0.5874,
+      "step": 74202
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.242038657424211e-07,
+      "loss": 0.5493,
+      "step": 74203
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.241782236113396e-07,
+      "loss": 0.3569,
+      "step": 74204
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.241525817680073e-07,
+      "loss": 0.6304,
+      "step": 74205
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.241269402124443e-07,
+      "loss": 0.6396,
+      "step": 74206
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.2410129894467e-07,
+      "loss": 0.6963,
+      "step": 74207
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.240756579647044e-07,
+      "loss": 0.6338,
+      "step": 74208
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.240500172725662e-07,
+      "loss": 0.6494,
+      "step": 74209
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.24024376868276e-07,
+      "loss": 0.6924,
+      "step": 74210
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.239987367518526e-07,
+      "loss": 0.6338,
+      "step": 74211
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.239730969233165e-07,
+      "loss": 0.7358,
+      "step": 74212
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.239474573826866e-07,
+      "loss": 0.5327,
+      "step": 74213
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.239218181299833e-07,
+      "loss": 0.6699,
+      "step": 74214
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.238961791652253e-07,
+      "loss": 0.6836,
+      "step": 74215
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.238705404884329e-07,
+      "loss": 0.7568,
+      "step": 74216
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.238449020996254e-07,
+      "loss": 0.5845,
+      "step": 74217
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.238192639988229e-07,
+      "loss": 0.425,
+      "step": 74218
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.237936261860446e-07,
+      "loss": 0.5696,
+      "step": 74219
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.237679886613102e-07,
+      "loss": 0.5928,
+      "step": 74220
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.237423514246392e-07,
+      "loss": 0.7803,
+      "step": 74221
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.237167144760514e-07,
+      "loss": 0.6328,
+      "step": 74222
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.236910778155663e-07,
+      "loss": 0.8232,
+      "step": 74223
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.236654414432039e-07,
+      "loss": 0.6387,
+      "step": 74224
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.236398053589833e-07,
+      "loss": 0.6318,
+      "step": 74225
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.236141695629246e-07,
+      "loss": 0.7441,
+      "step": 74226
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.23588534055047e-07,
+      "loss": 0.6514,
+      "step": 74227
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.235628988353708e-07,
+      "loss": 0.6055,
+      "step": 74228
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.235372639039149e-07,
+      "loss": 0.5664,
+      "step": 74229
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.235116292606993e-07,
+      "loss": 0.6099,
+      "step": 74230
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.234859949057434e-07,
+      "loss": 0.6052,
+      "step": 74231
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.234603608390668e-07,
+      "loss": 0.5991,
+      "step": 74232
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.234347270606897e-07,
+      "loss": 0.6099,
+      "step": 74233
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.234090935706308e-07,
+      "loss": 0.6387,
+      "step": 74234
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.233834603689108e-07,
+      "loss": 0.553,
+      "step": 74235
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.233578274555482e-07,
+      "loss": 0.6758,
+      "step": 74236
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.233321948305637e-07,
+      "loss": 0.6641,
+      "step": 74237
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.233065624939764e-07,
+      "loss": 0.6343,
+      "step": 74238
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.232809304458059e-07,
+      "loss": 0.6367,
+      "step": 74239
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.232552986860716e-07,
+      "loss": 0.4219,
+      "step": 74240
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.232296672147936e-07,
+      "loss": 0.7012,
+      "step": 74241
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.232040360319911e-07,
+      "loss": 0.5051,
+      "step": 74242
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.231784051376842e-07,
+      "loss": 0.6538,
+      "step": 74243
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.231527745318919e-07,
+      "loss": 0.564,
+      "step": 74244
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.231271442146347e-07,
+      "loss": 0.668,
+      "step": 74245
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.231015141859312e-07,
+      "loss": 0.667,
+      "step": 74246
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.230758844458019e-07,
+      "loss": 0.5747,
+      "step": 74247
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.230502549942662e-07,
+      "loss": 0.4698,
+      "step": 74248
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.230246258313434e-07,
+      "loss": 0.7676,
+      "step": 74249
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.229989969570533e-07,
+      "loss": 0.5454,
+      "step": 74250
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.229733683714153e-07,
+      "loss": 0.6553,
+      "step": 74251
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.229477400744496e-07,
+      "loss": 0.7046,
+      "step": 74252
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.229221120661751e-07,
+      "loss": 0.6802,
+      "step": 74253
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.228964843466123e-07,
+      "loss": 0.54,
+      "step": 74254
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.228708569157799e-07,
+      "loss": 0.8535,
+      "step": 74255
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.228452297736982e-07,
+      "loss": 0.7031,
+      "step": 74256
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.228196029203865e-07,
+      "loss": 0.6777,
+      "step": 74257
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.227939763558647e-07,
+      "loss": 0.6421,
+      "step": 74258
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.227683500801518e-07,
+      "loss": 0.5811,
+      "step": 74259
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.227427240932681e-07,
+      "loss": 0.6143,
+      "step": 74260
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.227170983952327e-07,
+      "loss": 0.6553,
+      "step": 74261
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.226914729860657e-07,
+      "loss": 0.5781,
+      "step": 74262
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.226658478657862e-07,
+      "loss": 0.6042,
+      "step": 74263
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.226402230344144e-07,
+      "loss": 0.6763,
+      "step": 74264
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.226145984919695e-07,
+      "loss": 0.6445,
+      "step": 74265
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.225889742384713e-07,
+      "loss": 0.3785,
+      "step": 74266
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.225633502739393e-07,
+      "loss": 0.7451,
+      "step": 74267
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.225377265983935e-07,
+      "loss": 0.5654,
+      "step": 74268
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.22512103211853e-07,
+      "loss": 0.6592,
+      "step": 74269
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.224864801143374e-07,
+      "loss": 0.5684,
+      "step": 74270
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.224608573058668e-07,
+      "loss": 0.6797,
+      "step": 74271
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.224352347864604e-07,
+      "loss": 0.6777,
+      "step": 74272
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.224096125561382e-07,
+      "loss": 0.5317,
+      "step": 74273
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.223839906149194e-07,
+      "loss": 0.6611,
+      "step": 74274
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.223583689628238e-07,
+      "loss": 0.6904,
+      "step": 74275
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.223327475998712e-07,
+      "loss": 0.6611,
+      "step": 74276
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.223071265260809e-07,
+      "loss": 0.7314,
+      "step": 74277
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.222815057414729e-07,
+      "loss": 0.6392,
+      "step": 74278
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.222558852460666e-07,
+      "loss": 0.5962,
+      "step": 74279
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.222302650398812e-07,
+      "loss": 0.7422,
+      "step": 74280
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.222046451229371e-07,
+      "loss": 0.6694,
+      "step": 74281
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.221790254952532e-07,
+      "loss": 0.6431,
+      "step": 74282
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.221534061568497e-07,
+      "loss": 0.7188,
+      "step": 74283
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.221277871077457e-07,
+      "loss": 0.6807,
+      "step": 74284
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.221021683479612e-07,
+      "loss": 0.6514,
+      "step": 74285
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.220765498775156e-07,
+      "loss": 0.6602,
+      "step": 74286
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.220509316964288e-07,
+      "loss": 0.5645,
+      "step": 74287
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.220253138047205e-07,
+      "loss": 0.6914,
+      "step": 74288
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.219996962024096e-07,
+      "loss": 0.6577,
+      "step": 74289
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.219740788895163e-07,
+      "loss": 0.6895,
+      "step": 74290
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.2194846186606e-07,
+      "loss": 0.6304,
+      "step": 74291
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.219228451320604e-07,
+      "loss": 0.6494,
+      "step": 74292
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.218972286875372e-07,
+      "loss": 0.5435,
+      "step": 74293
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.218716125325097e-07,
+      "loss": 0.5566,
+      "step": 74294
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.218459966669979e-07,
+      "loss": 0.4165,
+      "step": 74295
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.218203810910209e-07,
+      "loss": 0.6646,
+      "step": 74296
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.217947658045991e-07,
+      "loss": 0.6816,
+      "step": 74297
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.217691508077517e-07,
+      "loss": 0.7007,
+      "step": 74298
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.217435361004979e-07,
+      "loss": 0.5537,
+      "step": 74299
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.217179216828579e-07,
+      "loss": 0.6953,
+      "step": 74300
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.216923075548509e-07,
+      "loss": 0.3933,
+      "step": 74301
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.216666937164969e-07,
+      "loss": 0.7686,
+      "step": 74302
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.216410801678152e-07,
+      "loss": 0.7988,
+      "step": 74303
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.216154669088255e-07,
+      "loss": 0.6484,
+      "step": 74304
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.215898539395474e-07,
+      "loss": 0.543,
+      "step": 74305
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.215642412600006e-07,
+      "loss": 0.4124,
+      "step": 74306
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.215386288702047e-07,
+      "loss": 0.6016,
+      "step": 74307
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.215130167701794e-07,
+      "loss": 0.6318,
+      "step": 74308
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.214874049599442e-07,
+      "loss": 0.6855,
+      "step": 74309
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.214617934395186e-07,
+      "loss": 0.6523,
+      "step": 74310
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.214361822089221e-07,
+      "loss": 0.5923,
+      "step": 74311
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.214105712681748e-07,
+      "loss": 0.498,
+      "step": 74312
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.213849606172957e-07,
+      "loss": 0.71,
+      "step": 74313
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.21359350256305e-07,
+      "loss": 0.6602,
+      "step": 74314
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.213337401852217e-07,
+      "loss": 0.6367,
+      "step": 74315
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.213081304040662e-07,
+      "loss": 0.7158,
+      "step": 74316
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.212825209128571e-07,
+      "loss": 0.6062,
+      "step": 74317
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.212569117116153e-07,
+      "loss": 0.604,
+      "step": 74318
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.212313028003593e-07,
+      "loss": 0.6396,
+      "step": 74319
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.212056941791088e-07,
+      "loss": 0.666,
+      "step": 74320
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.211800858478841e-07,
+      "loss": 0.6377,
+      "step": 74321
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.211544778067039e-07,
+      "loss": 0.4697,
+      "step": 74322
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.211288700555887e-07,
+      "loss": 0.6719,
+      "step": 74323
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.211032625945576e-07,
+      "loss": 0.6528,
+      "step": 74324
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 6.210776554236303e-07,
+      "loss": 0.5574,
+      "step": 74325
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.210520485428263e-07,
+      "loss": 0.4536,
+      "step": 74326
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.210264419521656e-07,
+      "loss": 0.7236,
+      "step": 74327
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.210008356516676e-07,
+      "loss": 0.6084,
+      "step": 74328
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.209752296413517e-07,
+      "loss": 0.7861,
+      "step": 74329
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.209496239212372e-07,
+      "loss": 0.6689,
+      "step": 74330
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.209240184913447e-07,
+      "loss": 0.6377,
+      "step": 74331
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.208984133516929e-07,
+      "loss": 0.4048,
+      "step": 74332
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.208728085023021e-07,
+      "loss": 0.5945,
+      "step": 74333
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.208472039431912e-07,
+      "loss": 0.6582,
+      "step": 74334
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.208215996743803e-07,
+      "loss": 0.7754,
+      "step": 74335
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.207959956958888e-07,
+      "loss": 0.6333,
+      "step": 74336
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.207703920077365e-07,
+      "loss": 0.6758,
+      "step": 74337
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20744788609943e-07,
+      "loss": 0.7051,
+      "step": 74338
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.207191855025276e-07,
+      "loss": 0.6021,
+      "step": 74339
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.206935826855101e-07,
+      "loss": 0.8379,
+      "step": 74340
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.206679801589099e-07,
+      "loss": 0.4919,
+      "step": 74341
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20642377922747e-07,
+      "loss": 0.3822,
+      "step": 74342
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.206167759770407e-07,
+      "loss": 0.6084,
+      "step": 74343
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.205911743218106e-07,
+      "loss": 0.6274,
+      "step": 74344
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.205655729570765e-07,
+      "loss": 0.6523,
+      "step": 74345
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.205399718828579e-07,
+      "loss": 0.647,
+      "step": 74346
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.205143710991743e-07,
+      "loss": 0.6509,
+      "step": 74347
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20488770606046e-07,
+      "loss": 0.4719,
+      "step": 74348
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.204631704034914e-07,
+      "loss": 0.6475,
+      "step": 74349
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.204375704915308e-07,
+      "loss": 0.5366,
+      "step": 74350
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.204119708701836e-07,
+      "loss": 0.4897,
+      "step": 74351
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.203863715394697e-07,
+      "loss": 0.6729,
+      "step": 74352
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.203607724994082e-07,
+      "loss": 0.5967,
+      "step": 74353
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.203351737500193e-07,
+      "loss": 0.635,
+      "step": 74354
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.203095752913221e-07,
+      "loss": 0.5386,
+      "step": 74355
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.202839771233366e-07,
+      "loss": 0.6211,
+      "step": 74356
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.202583792460819e-07,
+      "loss": 0.8125,
+      "step": 74357
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.202327816595784e-07,
+      "loss": 0.5254,
+      "step": 74358
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20207184363845e-07,
+      "loss": 0.6494,
+      "step": 74359
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.201815873589012e-07,
+      "loss": 0.6646,
+      "step": 74360
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.201559906447672e-07,
+      "loss": 0.5796,
+      "step": 74361
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20130394221462e-07,
+      "loss": 0.6714,
+      "step": 74362
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.201047980890057e-07,
+      "loss": 0.5796,
+      "step": 74363
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.200792022474175e-07,
+      "loss": 0.6836,
+      "step": 74364
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.200536066967175e-07,
+      "loss": 0.5085,
+      "step": 74365
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.200280114369249e-07,
+      "loss": 0.6943,
+      "step": 74366
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.20002416468059e-07,
+      "loss": 0.6128,
+      "step": 74367
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.199768217901405e-07,
+      "loss": 0.5532,
+      "step": 74368
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.19951227403188e-07,
+      "loss": 0.6436,
+      "step": 74369
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.199256333072211e-07,
+      "loss": 0.5715,
+      "step": 74370
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.199000395022597e-07,
+      "loss": 0.5654,
+      "step": 74371
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.198744459883234e-07,
+      "loss": 0.5979,
+      "step": 74372
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.198488527654319e-07,
+      "loss": 0.7109,
+      "step": 74373
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.198232598336044e-07,
+      "loss": 0.5547,
+      "step": 74374
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.19797667192861e-07,
+      "loss": 0.6987,
+      "step": 74375
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.197720748432208e-07,
+      "loss": 0.7305,
+      "step": 74376
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.197464827847038e-07,
+      "loss": 0.5723,
+      "step": 74377
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.197208910173297e-07,
+      "loss": 0.5903,
+      "step": 74378
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.196952995411175e-07,
+      "loss": 0.668,
+      "step": 74379
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.196697083560872e-07,
+      "loss": 0.5571,
+      "step": 74380
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.196441174622582e-07,
+      "loss": 0.6426,
+      "step": 74381
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.196185268596503e-07,
+      "loss": 0.8018,
+      "step": 74382
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.195929365482831e-07,
+      "loss": 0.6484,
+      "step": 74383
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.195673465281759e-07,
+      "loss": 0.4929,
+      "step": 74384
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.195417567993486e-07,
+      "loss": 0.5996,
+      "step": 74385
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.195161673618207e-07,
+      "loss": 0.7705,
+      "step": 74386
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.194905782156117e-07,
+      "loss": 0.4285,
+      "step": 74387
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.194649893607416e-07,
+      "loss": 0.5938,
+      "step": 74388
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.194394007972293e-07,
+      "loss": 0.6875,
+      "step": 74389
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.19413812525095e-07,
+      "loss": 0.7725,
+      "step": 74390
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.193882245443577e-07,
+      "loss": 0.4868,
+      "step": 74391
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.193626368550376e-07,
+      "loss": 0.5806,
+      "step": 74392
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.193370494571537e-07,
+      "loss": 0.4178,
+      "step": 74393
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.193114623507263e-07,
+      "loss": 0.71,
+      "step": 74394
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.192858755357743e-07,
+      "loss": 0.6548,
+      "step": 74395
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.192602890123179e-07,
+      "loss": 0.7031,
+      "step": 74396
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.192347027803761e-07,
+      "loss": 0.5068,
+      "step": 74397
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.192091168399691e-07,
+      "loss": 0.3925,
+      "step": 74398
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.191835311911162e-07,
+      "loss": 0.6973,
+      "step": 74399
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.191579458338368e-07,
+      "loss": 0.6562,
+      "step": 74400
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.191323607681505e-07,
+      "loss": 0.7354,
+      "step": 74401
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.191067759940772e-07,
+      "loss": 0.6387,
+      "step": 74402
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.19081191511636e-07,
+      "loss": 0.4043,
+      "step": 74403
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.190556073208472e-07,
+      "loss": 0.7236,
+      "step": 74404
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.190300234217299e-07,
+      "loss": 0.5225,
+      "step": 74405
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.190044398143038e-07,
+      "loss": 0.5444,
+      "step": 74406
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.189788564985883e-07,
+      "loss": 0.7773,
+      "step": 74407
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.189532734746037e-07,
+      "loss": 0.6338,
+      "step": 74408
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.189276907423689e-07,
+      "loss": 0.6289,
+      "step": 74409
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.189021083019032e-07,
+      "loss": 0.7744,
+      "step": 74410
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.18876526153227e-07,
+      "loss": 0.5669,
+      "step": 74411
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.188509442963592e-07,
+      "loss": 0.6836,
+      "step": 74412
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.1882536273132e-07,
+      "loss": 0.5347,
+      "step": 74413
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.187997814581285e-07,
+      "loss": 0.7373,
+      "step": 74414
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.187742004768046e-07,
+      "loss": 0.6191,
+      "step": 74415
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.187486197873677e-07,
+      "loss": 0.5537,
+      "step": 74416
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.187230393898375e-07,
+      "loss": 0.5791,
+      "step": 74417
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.186974592842338e-07,
+      "loss": 0.6602,
+      "step": 74418
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.186718794705758e-07,
+      "loss": 0.6182,
+      "step": 74419
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.186462999488829e-07,
+      "loss": 0.6387,
+      "step": 74420
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.186207207191751e-07,
+      "loss": 0.7378,
+      "step": 74421
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.185951417814719e-07,
+      "loss": 0.5996,
+      "step": 74422
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.185695631357931e-07,
+      "loss": 0.668,
+      "step": 74423
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.185439847821576e-07,
+      "loss": 0.6196,
+      "step": 74424
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.185184067205858e-07,
+      "loss": 0.6685,
+      "step": 74425
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.184928289510967e-07,
+      "loss": 0.6084,
+      "step": 74426
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.184672514737101e-07,
+      "loss": 0.5522,
+      "step": 74427
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.184416742884459e-07,
+      "loss": 0.6577,
+      "step": 74428
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.184160973953232e-07,
+      "loss": 0.7158,
+      "step": 74429
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.183905207943617e-07,
+      "loss": 0.8486,
+      "step": 74430
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.183649444855809e-07,
+      "loss": 0.6914,
+      "step": 74431
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.183393684690007e-07,
+      "loss": 0.6353,
+      "step": 74432
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.183137927446402e-07,
+      "loss": 0.5532,
+      "step": 74433
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.182882173125197e-07,
+      "loss": 0.5273,
+      "step": 74434
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.18262642172658e-07,
+      "loss": 0.5459,
+      "step": 74435
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.182370673250752e-07,
+      "loss": 0.7002,
+      "step": 74436
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.182114927697907e-07,
+      "loss": 0.5439,
+      "step": 74437
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.181859185068245e-07,
+      "loss": 0.5566,
+      "step": 74438
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.181603445361954e-07,
+      "loss": 0.6753,
+      "step": 74439
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.181347708579234e-07,
+      "loss": 0.6685,
+      "step": 74440
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.181091974720279e-07,
+      "loss": 0.5903,
+      "step": 74441
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.180836243785289e-07,
+      "loss": 0.8291,
+      "step": 74442
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.180580515774454e-07,
+      "loss": 0.5889,
+      "step": 74443
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.180324790687976e-07,
+      "loss": 0.4854,
+      "step": 74444
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.180069068526043e-07,
+      "loss": 0.5571,
+      "step": 74445
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.179813349288861e-07,
+      "loss": 0.8027,
+      "step": 74446
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.179557632976618e-07,
+      "loss": 0.6729,
+      "step": 74447
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.179301919589514e-07,
+      "loss": 0.6416,
+      "step": 74448
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.179046209127741e-07,
+      "loss": 0.6016,
+      "step": 74449
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.178790501591495e-07,
+      "loss": 0.8164,
+      "step": 74450
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.178534796980976e-07,
+      "loss": 0.7305,
+      "step": 74451
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.178279095296375e-07,
+      "loss": 0.5913,
+      "step": 74452
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.178023396537891e-07,
+      "loss": 0.5342,
+      "step": 74453
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.177767700705719e-07,
+      "loss": 0.6089,
+      "step": 74454
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.177512007800054e-07,
+      "loss": 0.6421,
+      "step": 74455
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.177256317821093e-07,
+      "loss": 0.4812,
+      "step": 74456
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.177000630769029e-07,
+      "loss": 0.5354,
+      "step": 74457
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.176744946644066e-07,
+      "loss": 0.5541,
+      "step": 74458
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.17648926544639e-07,
+      "loss": 0.6396,
+      "step": 74459
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.176233587176197e-07,
+      "loss": 0.6729,
+      "step": 74460
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.175977911833689e-07,
+      "loss": 0.7036,
+      "step": 74461
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.175722239419058e-07,
+      "loss": 0.7041,
+      "step": 74462
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.175466569932502e-07,
+      "loss": 0.5552,
+      "step": 74463
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.175210903374213e-07,
+      "loss": 0.6167,
+      "step": 74464
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.174955239744392e-07,
+      "loss": 0.7705,
+      "step": 74465
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.174699579043229e-07,
+      "loss": 0.5063,
+      "step": 74466
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.174443921270924e-07,
+      "loss": 0.7842,
+      "step": 74467
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.174188266427676e-07,
+      "loss": 0.7412,
+      "step": 74468
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.173932614513669e-07,
+      "loss": 0.5776,
+      "step": 74469
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.173676965529109e-07,
+      "loss": 0.6494,
+      "step": 74470
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.173421319474188e-07,
+      "loss": 0.7236,
+      "step": 74471
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.173165676349102e-07,
+      "loss": 0.5864,
+      "step": 74472
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.172910036154049e-07,
+      "loss": 0.7305,
+      "step": 74473
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.17265439888922e-07,
+      "loss": 0.7617,
+      "step": 74474
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.172398764554816e-07,
+      "loss": 0.5854,
+      "step": 74475
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.172143133151027e-07,
+      "loss": 0.7285,
+      "step": 74476
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.171887504678058e-07,
+      "loss": 0.498,
+      "step": 74477
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.171631879136096e-07,
+      "loss": 0.543,
+      "step": 74478
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.171376256525335e-07,
+      "loss": 0.689,
+      "step": 74479
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.171120636845981e-07,
+      "loss": 0.7139,
+      "step": 74480
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.170865020098219e-07,
+      "loss": 0.6187,
+      "step": 74481
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.170609406282255e-07,
+      "loss": 0.6172,
+      "step": 74482
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.170353795398273e-07,
+      "loss": 0.4465,
+      "step": 74483
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.17009818744648e-07,
+      "loss": 0.7178,
+      "step": 74484
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.169842582427063e-07,
+      "loss": 0.5269,
+      "step": 74485
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.169586980340224e-07,
+      "loss": 0.7256,
+      "step": 74486
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.169331381186159e-07,
+      "loss": 0.5493,
+      "step": 74487
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.169075784965057e-07,
+      "loss": 0.5801,
+      "step": 74488
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.168820191677117e-07,
+      "loss": 0.6899,
+      "step": 74489
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.168564601322537e-07,
+      "loss": 0.6484,
+      "step": 74490
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.168309013901509e-07,
+      "loss": 0.6172,
+      "step": 74491
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.168053429414233e-07,
+      "loss": 0.6963,
+      "step": 74492
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.167797847860898e-07,
+      "loss": 0.6035,
+      "step": 74493
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.167542269241709e-07,
+      "loss": 0.7109,
+      "step": 74494
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.167286693556852e-07,
+      "loss": 0.6577,
+      "step": 74495
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.167031120806531e-07,
+      "loss": 0.479,
+      "step": 74496
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.16677555099094e-07,
+      "loss": 0.6733,
+      "step": 74497
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.166519984110269e-07,
+      "loss": 0.709,
+      "step": 74498
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.166264420164719e-07,
+      "loss": 0.5669,
+      "step": 74499
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.166008859154481e-07,
+      "loss": 0.5723,
+      "step": 74500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.165753301079757e-07,
+      "loss": 0.5898,
+      "step": 74501
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.165497745940736e-07,
+      "loss": 0.6504,
+      "step": 74502
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.165242193737621e-07,
+      "loss": 0.5854,
+      "step": 74503
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.1649866444706e-07,
+      "loss": 0.6553,
+      "step": 74504
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.164731098139875e-07,
+      "loss": 0.7402,
+      "step": 74505
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.164475554745637e-07,
+      "loss": 0.6077,
+      "step": 74506
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.164220014288091e-07,
+      "loss": 0.6436,
+      "step": 74507
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.163964476767418e-07,
+      "loss": 0.6914,
+      "step": 74508
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.163708942183823e-07,
+      "loss": 0.6421,
+      "step": 74509
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.163453410537498e-07,
+      "loss": 0.6973,
+      "step": 74510
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.163197881828642e-07,
+      "loss": 0.519,
+      "step": 74511
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.162942356057447e-07,
+      "loss": 0.4932,
+      "step": 74512
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.162686833224114e-07,
+      "loss": 0.7236,
+      "step": 74513
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.162431313328831e-07,
+      "loss": 0.8545,
+      "step": 74514
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.162175796371801e-07,
+      "loss": 0.6641,
+      "step": 74515
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.161920282353214e-07,
+      "loss": 0.5244,
+      "step": 74516
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.161664771273272e-07,
+      "loss": 0.8438,
+      "step": 74517
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.161409263132167e-07,
+      "loss": 0.5356,
+      "step": 74518
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.16115375793009e-07,
+      "loss": 0.7188,
+      "step": 74519
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.160898255667244e-07,
+      "loss": 0.5474,
+      "step": 74520
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.160642756343819e-07,
+      "loss": 0.7139,
+      "step": 74521
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.160387259960016e-07,
+      "loss": 0.6416,
+      "step": 74522
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.160131766516026e-07,
+      "loss": 0.5002,
+      "step": 74523
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.159876276012048e-07,
+      "loss": 0.3545,
+      "step": 74524
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.159620788448275e-07,
+      "loss": 0.5947,
+      "step": 74525
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.159365303824905e-07,
+      "loss": 0.5635,
+      "step": 74526
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.159109822142135e-07,
+      "loss": 0.6934,
+      "step": 74527
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.158854343400155e-07,
+      "loss": 0.6235,
+      "step": 74528
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.158598867599163e-07,
+      "loss": 0.583,
+      "step": 74529
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.158343394739356e-07,
+      "loss": 0.5117,
+      "step": 74530
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.158087924820927e-07,
+      "loss": 0.6562,
+      "step": 74531
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.157832457844076e-07,
+      "loss": 0.4722,
+      "step": 74532
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.157576993808992e-07,
+      "loss": 0.7168,
+      "step": 74533
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.15732153271588e-07,
+      "loss": 0.6357,
+      "step": 74534
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.157066074564927e-07,
+      "loss": 0.6602,
+      "step": 74535
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.156810619356332e-07,
+      "loss": 0.4004,
+      "step": 74536
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.156555167090293e-07,
+      "loss": 0.6108,
+      "step": 74537
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.156299717767002e-07,
+      "loss": 0.5522,
+      "step": 74538
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.156044271386655e-07,
+      "loss": 0.6855,
+      "step": 74539
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.155788827949446e-07,
+      "loss": 0.5327,
+      "step": 74540
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.155533387455575e-07,
+      "loss": 0.749,
+      "step": 74541
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.155277949905233e-07,
+      "loss": 0.6299,
+      "step": 74542
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.15502251529862e-07,
+      "loss": 0.6528,
+      "step": 74543
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.154767083635929e-07,
+      "loss": 0.4307,
+      "step": 74544
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.154511654917356e-07,
+      "loss": 0.4058,
+      "step": 74545
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.154256229143098e-07,
+      "loss": 0.5269,
+      "step": 74546
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.154000806313351e-07,
+      "loss": 0.7178,
+      "step": 74547
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.153745386428304e-07,
+      "loss": 0.4888,
+      "step": 74548
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.153489969488159e-07,
+      "loss": 0.5731,
+      "step": 74549
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.153234555493109e-07,
+      "loss": 0.8184,
+      "step": 74550
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.152979144443352e-07,
+      "loss": 0.707,
+      "step": 74551
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.15272373633908e-07,
+      "loss": 0.709,
+      "step": 74552
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.152468331180493e-07,
+      "loss": 0.5698,
+      "step": 74553
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.152212928967781e-07,
+      "loss": 0.6719,
+      "step": 74554
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.151957529701146e-07,
+      "loss": 0.6201,
+      "step": 74555
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.151702133380776e-07,
+      "loss": 0.7236,
+      "step": 74556
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.151446740006877e-07,
+      "loss": 0.6116,
+      "step": 74557
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.151191349579636e-07,
+      "loss": 0.623,
+      "step": 74558
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.150935962099247e-07,
+      "loss": 0.7568,
+      "step": 74559
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.150680577565913e-07,
+      "loss": 0.6826,
+      "step": 74560
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.150425195979824e-07,
+      "loss": 0.6016,
+      "step": 74561
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.150169817341179e-07,
+      "loss": 0.6719,
+      "step": 74562
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.149914441650172e-07,
+      "loss": 0.6079,
+      "step": 74563
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.149659068906996e-07,
+      "loss": 0.4456,
+      "step": 74564
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.14940369911185e-07,
+      "loss": 0.8311,
+      "step": 74565
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.149148332264929e-07,
+      "loss": 0.7607,
+      "step": 74566
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.14889296836643e-07,
+      "loss": 0.5437,
+      "step": 74567
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.148637607416546e-07,
+      "loss": 0.6111,
+      "step": 74568
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.148382249415471e-07,
+      "loss": 0.5928,
+      "step": 74569
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.148126894363403e-07,
+      "loss": 0.6182,
+      "step": 74570
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.147871542260536e-07,
+      "loss": 0.5952,
+      "step": 74571
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.147616193107071e-07,
+      "loss": 0.6543,
+      "step": 74572
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.147360846903193e-07,
+      "loss": 0.5562,
+      "step": 74573
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.147105503649109e-07,
+      "loss": 0.5129,
+      "step": 74574
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.146850163345005e-07,
+      "loss": 0.6699,
+      "step": 74575
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.146594825991084e-07,
+      "loss": 0.6045,
+      "step": 74576
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.14633949158754e-07,
+      "loss": 0.6338,
+      "step": 74577
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.146084160134561e-07,
+      "loss": 0.7754,
+      "step": 74578
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.145828831632353e-07,
+      "loss": 0.6855,
+      "step": 74579
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.145573506081106e-07,
+      "loss": 0.6143,
+      "step": 74580
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.145318183481012e-07,
+      "loss": 0.75,
+      "step": 74581
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.145062863832273e-07,
+      "loss": 0.4277,
+      "step": 74582
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.144807547135082e-07,
+      "loss": 0.4888,
+      "step": 74583
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.144552233389635e-07,
+      "loss": 0.6411,
+      "step": 74584
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.144296922596127e-07,
+      "loss": 0.6216,
+      "step": 74585
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.144041614754753e-07,
+      "loss": 0.7012,
+      "step": 74586
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.143786309865712e-07,
+      "loss": 0.4536,
+      "step": 74587
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.143531007929193e-07,
+      "loss": 0.6313,
+      "step": 74588
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.143275708945397e-07,
+      "loss": 0.6538,
+      "step": 74589
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.143020412914514e-07,
+      "loss": 0.6367,
+      "step": 74590
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.142765119836748e-07,
+      "loss": 0.5945,
+      "step": 74591
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.142509829712286e-07,
+      "loss": 0.6621,
+      "step": 74592
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.142254542541327e-07,
+      "loss": 0.7207,
+      "step": 74593
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.141999258324066e-07,
+      "loss": 0.5818,
+      "step": 74594
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.1417439770607e-07,
+      "loss": 0.6499,
+      "step": 74595
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.141488698751423e-07,
+      "loss": 0.6284,
+      "step": 74596
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.141233423396434e-07,
+      "loss": 0.666,
+      "step": 74597
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.140978150995921e-07,
+      "loss": 0.7207,
+      "step": 74598
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.140722881550084e-07,
+      "loss": 0.6499,
+      "step": 74599
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.140467615059118e-07,
+      "loss": 0.2649,
+      "step": 74600
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.14021235152322e-07,
+      "loss": 0.6055,
+      "step": 74601
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.139957090942581e-07,
+      "loss": 0.4177,
+      "step": 74602
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.139701833317403e-07,
+      "loss": 0.6289,
+      "step": 74603
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.139446578647873e-07,
+      "loss": 0.4768,
+      "step": 74604
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.139191326934196e-07,
+      "loss": 0.5198,
+      "step": 74605
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.13893607817656e-07,
+      "loss": 0.5645,
+      "step": 74606
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.138680832375167e-07,
+      "loss": 0.7021,
+      "step": 74607
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.138425589530205e-07,
+      "loss": 0.3096,
+      "step": 74608
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.138170349641874e-07,
+      "loss": 0.8047,
+      "step": 74609
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.137915112710367e-07,
+      "loss": 0.6899,
+      "step": 74610
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.137659878735881e-07,
+      "loss": 0.5107,
+      "step": 74611
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.137404647718613e-07,
+      "loss": 0.6782,
+      "step": 74612
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.137149419658754e-07,
+      "loss": 0.6537,
+      "step": 74613
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.136894194556505e-07,
+      "loss": 0.6895,
+      "step": 74614
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.136638972412055e-07,
+      "loss": 0.8135,
+      "step": 74615
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.136383753225605e-07,
+      "loss": 0.6689,
+      "step": 74616
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.136128536997352e-07,
+      "loss": 0.5679,
+      "step": 74617
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.135873323727485e-07,
+      "loss": 0.5752,
+      "step": 74618
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.1356181134162e-07,
+      "loss": 0.6133,
+      "step": 74619
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.135362906063697e-07,
+      "loss": 0.7686,
+      "step": 74620
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.135107701670166e-07,
+      "loss": 0.6121,
+      "step": 74621
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.134852500235808e-07,
+      "loss": 0.6855,
+      "step": 74622
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.134597301760813e-07,
+      "loss": 0.7646,
+      "step": 74623
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.134342106245381e-07,
+      "loss": 0.4302,
+      "step": 74624
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.134086913689704e-07,
+      "loss": 0.6631,
+      "step": 74625
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.133831724093981e-07,
+      "loss": 0.6382,
+      "step": 74626
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.133576537458407e-07,
+      "loss": 0.6128,
+      "step": 74627
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.133321353783172e-07,
+      "loss": 0.6069,
+      "step": 74628
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.133066173068477e-07,
+      "loss": 0.6768,
+      "step": 74629
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.132810995314513e-07,
+      "loss": 0.79,
+      "step": 74630
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.132555820521482e-07,
+      "loss": 0.6904,
+      "step": 74631
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.13230064868957e-07,
+      "loss": 0.7275,
+      "step": 74632
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.132045479818982e-07,
+      "loss": 0.5776,
+      "step": 74633
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.131790313909904e-07,
+      "loss": 0.5234,
+      "step": 74634
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.13153515096254e-07,
+      "loss": 0.6108,
+      "step": 74635
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.131279990977081e-07,
+      "loss": 0.501,
+      "step": 74636
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.131024833953727e-07,
+      "loss": 0.6274,
+      "step": 74637
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.130769679892665e-07,
+      "loss": 0.5684,
+      "step": 74638
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.130514528794095e-07,
+      "loss": 0.6016,
+      "step": 74639
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.130259380658212e-07,
+      "loss": 0.6221,
+      "step": 74640
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.130004235485211e-07,
+      "loss": 0.5249,
+      "step": 74641
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.129749093275287e-07,
+      "loss": 0.5747,
+      "step": 74642
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.129493954028639e-07,
+      "loss": 0.5337,
+      "step": 74643
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.129238817745457e-07,
+      "loss": 0.6367,
+      "step": 74644
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.128983684425941e-07,
+      "loss": 0.6533,
+      "step": 74645
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.128728554070281e-07,
+      "loss": 0.6069,
+      "step": 74646
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.128473426678681e-07,
+      "loss": 0.5811,
+      "step": 74647
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.128218302251329e-07,
+      "loss": 0.4968,
+      "step": 74648
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.127963180788419e-07,
+      "loss": 0.7324,
+      "step": 74649
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.127708062290152e-07,
+      "loss": 0.6758,
+      "step": 74650
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.127452946756719e-07,
+      "loss": 0.7373,
+      "step": 74651
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.12719783418832e-07,
+      "loss": 0.7666,
+      "step": 74652
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.126942724585147e-07,
+      "loss": 0.5933,
+      "step": 74653
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.126687617947393e-07,
+      "loss": 0.6484,
+      "step": 74654
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.12643251427526e-07,
+      "loss": 0.5874,
+      "step": 74655
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.126177413568935e-07,
+      "loss": 0.7764,
+      "step": 74656
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.125922315828624e-07,
+      "loss": 0.5664,
+      "step": 74657
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.125667221054513e-07,
+      "loss": 0.6245,
+      "step": 74658
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.125412129246799e-07,
+      "loss": 0.6455,
+      "step": 74659
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.12515704040568e-07,
+      "loss": 0.4595,
+      "step": 74660
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.124901954531349e-07,
+      "loss": 0.6265,
+      "step": 74661
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.124646871624004e-07,
+      "loss": 0.5776,
+      "step": 74662
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.124391791683836e-07,
+      "loss": 0.5884,
+      "step": 74663
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.124136714711046e-07,
+      "loss": 0.5215,
+      "step": 74664
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.123881640705823e-07,
+      "loss": 0.6055,
+      "step": 74665
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.12362656966837e-07,
+      "loss": 0.6523,
+      "step": 74666
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.123371501598879e-07,
+      "loss": 0.7173,
+      "step": 74667
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.123116436497539e-07,
+      "loss": 0.4802,
+      "step": 74668
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.122861374364554e-07,
+      "loss": 0.7822,
+      "step": 74669
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.122606315200114e-07,
+      "loss": 0.6826,
+      "step": 74670
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.122351259004415e-07,
+      "loss": 0.6279,
+      "step": 74671
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.122096205777656e-07,
+      "loss": 0.5889,
+      "step": 74672
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.121841155520027e-07,
+      "loss": 0.5635,
+      "step": 74673
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.121586108231728e-07,
+      "loss": 0.7285,
+      "step": 74674
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.12133106391295e-07,
+      "loss": 0.6914,
+      "step": 74675
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.121076022563892e-07,
+      "loss": 0.7104,
+      "step": 74676
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.120820984184751e-07,
+      "loss": 0.5122,
+      "step": 74677
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.120565948775714e-07,
+      "loss": 0.5684,
+      "step": 74678
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.120310916336983e-07,
+      "loss": 0.7051,
+      "step": 74679
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.120055886868751e-07,
+      "loss": 0.6777,
+      "step": 74680
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.119800860371214e-07,
+      "loss": 0.6372,
+      "step": 74681
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.119545836844566e-07,
+      "loss": 0.5859,
+      "step": 74682
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.119290816289006e-07,
+      "loss": 0.5942,
+      "step": 74683
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.119035798704723e-07,
+      "loss": 0.6182,
+      "step": 74684
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.118780784091919e-07,
+      "loss": 0.6304,
+      "step": 74685
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.118525772450784e-07,
+      "loss": 0.5405,
+      "step": 74686
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.118270763781519e-07,
+      "loss": 0.542,
+      "step": 74687
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.118015758084314e-07,
+      "loss": 0.7812,
+      "step": 74688
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.117760755359365e-07,
+      "loss": 0.6504,
+      "step": 74689
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.117505755606867e-07,
+      "loss": 0.5674,
+      "step": 74690
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.117250758827018e-07,
+      "loss": 0.6523,
+      "step": 74691
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.11699576502001e-07,
+      "loss": 0.6289,
+      "step": 74692
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.116740774186041e-07,
+      "loss": 0.6709,
+      "step": 74693
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.116485786325302e-07,
+      "loss": 0.7002,
+      "step": 74694
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.116230801437996e-07,
+      "loss": 0.6631,
+      "step": 74695
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.115975819524311e-07,
+      "loss": 0.6621,
+      "step": 74696
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.115720840584447e-07,
+      "loss": 0.7197,
+      "step": 74697
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.115465864618596e-07,
+      "loss": 0.5146,
+      "step": 74698
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.115210891626952e-07,
+      "loss": 0.4893,
+      "step": 74699
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.114955921609715e-07,
+      "loss": 0.749,
+      "step": 74700
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.114700954567074e-07,
+      "loss": 0.6108,
+      "step": 74701
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.11444599049923e-07,
+      "loss": 0.6074,
+      "step": 74702
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.114191029406376e-07,
+      "loss": 0.5183,
+      "step": 74703
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.113936071288707e-07,
+      "loss": 0.5884,
+      "step": 74704
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.113681116146419e-07,
+      "loss": 0.6689,
+      "step": 74705
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.113426163979707e-07,
+      "loss": 0.7432,
+      "step": 74706
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.113171214788768e-07,
+      "loss": 0.6934,
+      "step": 74707
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.112916268573793e-07,
+      "loss": 0.5444,
+      "step": 74708
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.112661325334978e-07,
+      "loss": 0.5913,
+      "step": 74709
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.11240638507252e-07,
+      "loss": 0.7285,
+      "step": 74710
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.112151447786614e-07,
+      "loss": 0.688,
+      "step": 74711
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.111896513477455e-07,
+      "loss": 0.6709,
+      "step": 74712
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.111641582145236e-07,
+      "loss": 0.7441,
+      "step": 74713
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.111386653790157e-07,
+      "loss": 0.5571,
+      "step": 74714
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 6.111131728412408e-07,
+      "loss": 0.7227,
+      "step": 74715
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.110876806012189e-07,
+      "loss": 0.6592,
+      "step": 74716
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.110621886589695e-07,
+      "loss": 0.6343,
+      "step": 74717
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.110366970145115e-07,
+      "loss": 0.5179,
+      "step": 74718
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.110112056678649e-07,
+      "loss": 0.5132,
+      "step": 74719
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.109857146190491e-07,
+      "loss": 0.5815,
+      "step": 74720
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.109602238680838e-07,
+      "loss": 0.6196,
+      "step": 74721
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.109347334149881e-07,
+      "loss": 0.4355,
+      "step": 74722
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10909243259782e-07,
+      "loss": 0.6089,
+      "step": 74723
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.108837534024845e-07,
+      "loss": 0.6904,
+      "step": 74724
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.108582638431158e-07,
+      "loss": 0.8193,
+      "step": 74725
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10832774581695e-07,
+      "loss": 0.6748,
+      "step": 74726
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.108072856182418e-07,
+      "loss": 0.689,
+      "step": 74727
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.107817969527752e-07,
+      "loss": 0.5322,
+      "step": 74728
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.107563085853153e-07,
+      "loss": 0.6455,
+      "step": 74729
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.107308205158811e-07,
+      "loss": 0.6826,
+      "step": 74730
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.107053327444926e-07,
+      "loss": 0.6104,
+      "step": 74731
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.106798452711691e-07,
+      "loss": 0.7192,
+      "step": 74732
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.106543580959303e-07,
+      "loss": 0.5354,
+      "step": 74733
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.106288712187952e-07,
+      "loss": 0.6592,
+      "step": 74734
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10603384639784e-07,
+      "loss": 0.7383,
+      "step": 74735
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.105778983589156e-07,
+      "loss": 0.6768,
+      "step": 74736
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.105524123762102e-07,
+      "loss": 0.7744,
+      "step": 74737
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.105269266916869e-07,
+      "loss": 0.5527,
+      "step": 74738
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.105014413053649e-07,
+      "loss": 0.6226,
+      "step": 74739
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.104759562172641e-07,
+      "loss": 0.6846,
+      "step": 74740
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10450471427404e-07,
+      "loss": 0.6479,
+      "step": 74741
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10424986935804e-07,
+      "loss": 0.467,
+      "step": 74742
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10399502742484e-07,
+      "loss": 0.5811,
+      "step": 74743
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.103740188474626e-07,
+      "loss": 0.9297,
+      "step": 74744
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.103485352507604e-07,
+      "loss": 0.6401,
+      "step": 74745
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.103230519523961e-07,
+      "loss": 0.5288,
+      "step": 74746
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.1029756895239e-07,
+      "loss": 0.5596,
+      "step": 74747
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.102720862507608e-07,
+      "loss": 0.8369,
+      "step": 74748
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.102466038475282e-07,
+      "loss": 0.7158,
+      "step": 74749
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.10221121742712e-07,
+      "loss": 0.6309,
+      "step": 74750
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.101956399363315e-07,
+      "loss": 0.7676,
+      "step": 74751
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.101701584284062e-07,
+      "loss": 0.6992,
+      "step": 74752
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.101446772189557e-07,
+      "loss": 0.749,
+      "step": 74753
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.101191963079998e-07,
+      "loss": 0.6973,
+      "step": 74754
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.100937156955574e-07,
+      "loss": 0.5386,
+      "step": 74755
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.100682353816484e-07,
+      "loss": 0.7705,
+      "step": 74756
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.100427553662926e-07,
+      "loss": 0.4871,
+      "step": 74757
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.100172756495085e-07,
+      "loss": 0.7148,
+      "step": 74758
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.099917962313167e-07,
+      "loss": 0.7725,
+      "step": 74759
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.099663171117362e-07,
+      "loss": 0.5889,
+      "step": 74760
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.099408382907862e-07,
+      "loss": 0.5815,
+      "step": 74761
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.099153597684868e-07,
+      "loss": 0.6504,
+      "step": 74762
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.098898815448572e-07,
+      "loss": 0.5337,
+      "step": 74763
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.098644036199171e-07,
+      "loss": 0.6743,
+      "step": 74764
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.098389259936857e-07,
+      "loss": 0.7256,
+      "step": 74765
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.098134486661829e-07,
+      "loss": 0.7725,
+      "step": 74766
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.097879716374281e-07,
+      "loss": 0.7188,
+      "step": 74767
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.097624949074404e-07,
+      "loss": 0.4939,
+      "step": 74768
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.097370184762398e-07,
+      "loss": 0.4236,
+      "step": 74769
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.097115423438455e-07,
+      "loss": 0.5908,
+      "step": 74770
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.096860665102772e-07,
+      "loss": 0.603,
+      "step": 74771
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.096605909755543e-07,
+      "loss": 0.8271,
+      "step": 74772
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.096351157396964e-07,
+      "loss": 0.6318,
+      "step": 74773
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.096096408027227e-07,
+      "loss": 0.7021,
+      "step": 74774
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.095841661646533e-07,
+      "loss": 0.667,
+      "step": 74775
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.095586918255074e-07,
+      "loss": 0.4966,
+      "step": 74776
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.095332177853041e-07,
+      "loss": 0.6504,
+      "step": 74777
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.095077440440635e-07,
+      "loss": 0.5674,
+      "step": 74778
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.09482270601805e-07,
+      "loss": 0.47,
+      "step": 74779
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.094567974585475e-07,
+      "loss": 0.6553,
+      "step": 74780
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.094313246143112e-07,
+      "loss": 0.6885,
+      "step": 74781
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.094058520691154e-07,
+      "loss": 0.7305,
+      "step": 74782
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.093803798229797e-07,
+      "loss": 0.6343,
+      "step": 74783
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.093549078759232e-07,
+      "loss": 0.5203,
+      "step": 74784
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.093294362279659e-07,
+      "loss": 0.4421,
+      "step": 74785
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.093039648791273e-07,
+      "loss": 0.6562,
+      "step": 74786
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.092784938294261e-07,
+      "loss": 0.5613,
+      "step": 74787
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.092530230788829e-07,
+      "loss": 0.5244,
+      "step": 74788
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.092275526275163e-07,
+      "loss": 0.5005,
+      "step": 74789
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.092020824753466e-07,
+      "loss": 0.5327,
+      "step": 74790
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.091766126223925e-07,
+      "loss": 0.4995,
+      "step": 74791
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.091511430686742e-07,
+      "loss": 0.7637,
+      "step": 74792
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.091256738142106e-07,
+      "loss": 0.7295,
+      "step": 74793
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.091002048590219e-07,
+      "loss": 0.7051,
+      "step": 74794
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.090747362031267e-07,
+      "loss": 0.5503,
+      "step": 74795
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.090492678465459e-07,
+      "loss": 0.7666,
+      "step": 74796
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.090237997892973e-07,
+      "loss": 0.7302,
+      "step": 74797
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.089983320314015e-07,
+      "loss": 0.8047,
+      "step": 74798
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.089728645728774e-07,
+      "loss": 0.6143,
+      "step": 74799
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.089473974137451e-07,
+      "loss": 0.5439,
+      "step": 74800
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.089219305540235e-07,
+      "loss": 0.7725,
+      "step": 74801
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.088964639937326e-07,
+      "loss": 0.6865,
+      "step": 74802
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.088709977328916e-07,
+      "loss": 0.6494,
+      "step": 74803
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.088455317715202e-07,
+      "loss": 0.5928,
+      "step": 74804
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.088200661096377e-07,
+      "loss": 0.4731,
+      "step": 74805
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.08794600747264e-07,
+      "loss": 0.6538,
+      "step": 74806
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.08769135684418e-07,
+      "loss": 0.6094,
+      "step": 74807
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.087436709211194e-07,
+      "loss": 0.8145,
+      "step": 74808
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.08718206457388e-07,
+      "loss": 0.668,
+      "step": 74809
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.086927422932428e-07,
+      "loss": 0.6313,
+      "step": 74810
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.086672784287039e-07,
+      "loss": 0.6504,
+      "step": 74811
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.086418148637902e-07,
+      "loss": 0.6904,
+      "step": 74812
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.086163515985216e-07,
+      "loss": 0.6323,
+      "step": 74813
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.085908886329173e-07,
+      "loss": 0.5898,
+      "step": 74814
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.085654259669972e-07,
+      "loss": 0.4922,
+      "step": 74815
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.085399636007808e-07,
+      "loss": 0.5244,
+      "step": 74816
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.08514501534287e-07,
+      "loss": 0.6914,
+      "step": 74817
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.084890397675355e-07,
+      "loss": 0.4858,
+      "step": 74818
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.084635783005461e-07,
+      "loss": 0.6372,
+      "step": 74819
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.084381171333382e-07,
+      "loss": 0.498,
+      "step": 74820
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.084126562659311e-07,
+      "loss": 0.749,
+      "step": 74821
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.083871956983444e-07,
+      "loss": 0.6196,
+      "step": 74822
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.083617354305976e-07,
+      "loss": 0.6382,
+      "step": 74823
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.083362754627104e-07,
+      "loss": 0.6543,
+      "step": 74824
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.08310815794702e-07,
+      "loss": 0.489,
+      "step": 74825
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.082853564265923e-07,
+      "loss": 0.5244,
+      "step": 74826
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.082598973584001e-07,
+      "loss": 0.7119,
+      "step": 74827
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.082344385901455e-07,
+      "loss": 0.5591,
+      "step": 74828
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.082089801218475e-07,
+      "loss": 0.5029,
+      "step": 74829
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.081835219535261e-07,
+      "loss": 0.542,
+      "step": 74830
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.081580640852003e-07,
+      "loss": 0.6099,
+      "step": 74831
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.081326065168902e-07,
+      "loss": 0.4808,
+      "step": 74832
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.081071492486147e-07,
+      "loss": 0.7852,
+      "step": 74833
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.080816922803936e-07,
+      "loss": 0.7891,
+      "step": 74834
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.080562356122465e-07,
+      "loss": 0.5098,
+      "step": 74835
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.080307792441928e-07,
+      "loss": 0.5112,
+      "step": 74836
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.080053231762515e-07,
+      "loss": 0.5918,
+      "step": 74837
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.079798674084427e-07,
+      "loss": 0.3398,
+      "step": 74838
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.079544119407856e-07,
+      "loss": 0.6626,
+      "step": 74839
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.079289567732999e-07,
+      "loss": 0.4956,
+      "step": 74840
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.079035019060048e-07,
+      "loss": 0.6445,
+      "step": 74841
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.078780473389202e-07,
+      "loss": 0.6738,
+      "step": 74842
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.07852593072065e-07,
+      "loss": 0.6924,
+      "step": 74843
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.078271391054594e-07,
+      "loss": 0.7725,
+      "step": 74844
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.078016854391223e-07,
+      "loss": 0.5972,
+      "step": 74845
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.077762320730737e-07,
+      "loss": 0.5024,
+      "step": 74846
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.077507790073328e-07,
+      "loss": 0.6113,
+      "step": 74847
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.077253262419187e-07,
+      "loss": 0.562,
+      "step": 74848
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.076998737768516e-07,
+      "loss": 0.5908,
+      "step": 74849
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.076744216121505e-07,
+      "loss": 0.5923,
+      "step": 74850
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.076489697478351e-07,
+      "loss": 0.6289,
+      "step": 74851
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.076235181839248e-07,
+      "loss": 0.6777,
+      "step": 74852
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.075980669204392e-07,
+      "loss": 0.6235,
+      "step": 74853
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.075726159573977e-07,
+      "loss": 0.7109,
+      "step": 74854
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.075471652948198e-07,
+      "loss": 0.6484,
+      "step": 74855
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.075217149327252e-07,
+      "loss": 0.6104,
+      "step": 74856
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.074962648711331e-07,
+      "loss": 0.6611,
+      "step": 74857
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.074708151100627e-07,
+      "loss": 0.4219,
+      "step": 74858
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.074453656495342e-07,
+      "loss": 0.4839,
+      "step": 74859
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.074199164895664e-07,
+      "loss": 0.3911,
+      "step": 74860
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.073944676301794e-07,
+      "loss": 0.6934,
+      "step": 74861
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.073690190713922e-07,
+      "loss": 0.6963,
+      "step": 74862
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.073435708132249e-07,
+      "loss": 0.5229,
+      "step": 74863
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.073181228556961e-07,
+      "loss": 0.6494,
+      "step": 74864
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.072926751988259e-07,
+      "loss": 0.5771,
+      "step": 74865
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.072672278426341e-07,
+      "loss": 0.833,
+      "step": 74866
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.072417807871392e-07,
+      "loss": 0.749,
+      "step": 74867
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.072163340323613e-07,
+      "loss": 0.7236,
+      "step": 74868
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.0719088757832e-07,
+      "loss": 0.7002,
+      "step": 74869
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.071654414250342e-07,
+      "loss": 0.4536,
+      "step": 74870
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.071399955725241e-07,
+      "loss": 0.7178,
+      "step": 74871
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.071145500208086e-07,
+      "loss": 0.79,
+      "step": 74872
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.070891047699075e-07,
+      "loss": 0.6416,
+      "step": 74873
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.070636598198401e-07,
+      "loss": 0.6147,
+      "step": 74874
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.070382151706261e-07,
+      "loss": 0.6406,
+      "step": 74875
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.070127708222852e-07,
+      "loss": 0.5264,
+      "step": 74876
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.069873267748363e-07,
+      "loss": 0.6465,
+      "step": 74877
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.06961883028299e-07,
+      "loss": 0.4355,
+      "step": 74878
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.069364395826928e-07,
+      "loss": 0.5044,
+      "step": 74879
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.069109964380376e-07,
+      "loss": 0.4756,
+      "step": 74880
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.068855535943523e-07,
+      "loss": 0.3749,
+      "step": 74881
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.068601110516569e-07,
+      "loss": 0.7651,
+      "step": 74882
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.068346688099703e-07,
+      "loss": 0.6748,
+      "step": 74883
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.068092268693128e-07,
+      "loss": 0.5942,
+      "step": 74884
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.067837852297031e-07,
+      "loss": 0.6543,
+      "step": 74885
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.067583438911614e-07,
+      "loss": 0.5308,
+      "step": 74886
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.067329028537063e-07,
+      "loss": 0.5869,
+      "step": 74887
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.067074621173578e-07,
+      "loss": 0.5957,
+      "step": 74888
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.066820216821354e-07,
+      "loss": 0.6973,
+      "step": 74889
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.066565815480585e-07,
+      "loss": 0.6768,
+      "step": 74890
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.066311417151461e-07,
+      "loss": 0.5103,
+      "step": 74891
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.066057021834188e-07,
+      "loss": 0.3967,
+      "step": 74892
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.06580262952895e-07,
+      "loss": 0.6475,
+      "step": 74893
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.065548240235948e-07,
+      "loss": 0.6504,
+      "step": 74894
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.065293853955373e-07,
+      "loss": 0.6416,
+      "step": 74895
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.065039470687426e-07,
+      "loss": 0.5503,
+      "step": 74896
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.064785090432295e-07,
+      "loss": 0.6523,
+      "step": 74897
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.064530713190174e-07,
+      "loss": 0.6899,
+      "step": 74898
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.064276338961264e-07,
+      "loss": 0.479,
+      "step": 74899
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.064021967745756e-07,
+      "loss": 0.7646,
+      "step": 74900
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.063767599543846e-07,
+      "loss": 0.7188,
+      "step": 74901
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.063513234355726e-07,
+      "loss": 0.5977,
+      "step": 74902
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.063258872181596e-07,
+      "loss": 0.6396,
+      "step": 74903
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.063004513021645e-07,
+      "loss": 0.667,
+      "step": 74904
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.062750156876072e-07,
+      "loss": 0.6553,
+      "step": 74905
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.062495803745072e-07,
+      "loss": 0.6646,
+      "step": 74906
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.062241453628837e-07,
+      "loss": 0.4686,
+      "step": 74907
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.06198710652756e-07,
+      "loss": 0.4355,
+      "step": 74908
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.061732762441441e-07,
+      "loss": 0.707,
+      "step": 74909
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.061478421370669e-07,
+      "loss": 0.4397,
+      "step": 74910
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.061224083315443e-07,
+      "loss": 0.8457,
+      "step": 74911
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.060969748275957e-07,
+      "loss": 0.6455,
+      "step": 74912
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.060715416252407e-07,
+      "loss": 0.7529,
+      "step": 74913
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.060461087244984e-07,
+      "loss": 0.6895,
+      "step": 74914
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.060206761253886e-07,
+      "loss": 0.5718,
+      "step": 74915
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.059952438279308e-07,
+      "loss": 0.7051,
+      "step": 74916
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05969811832144e-07,
+      "loss": 0.7302,
+      "step": 74917
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.059443801380482e-07,
+      "loss": 0.6851,
+      "step": 74918
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.059189487456624e-07,
+      "loss": 0.5479,
+      "step": 74919
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.058935176550066e-07,
+      "loss": 0.6147,
+      "step": 74920
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.058680868660999e-07,
+      "loss": 0.5042,
+      "step": 74921
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05842656378962e-07,
+      "loss": 0.6616,
+      "step": 74922
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.058172261936124e-07,
+      "loss": 0.5093,
+      "step": 74923
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.057917963100699e-07,
+      "loss": 0.7314,
+      "step": 74924
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05766366728355e-07,
+      "loss": 0.8672,
+      "step": 74925
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.057409374484867e-07,
+      "loss": 0.6865,
+      "step": 74926
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.057155084704841e-07,
+      "loss": 0.4279,
+      "step": 74927
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.056900797943672e-07,
+      "loss": 0.6226,
+      "step": 74928
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05664651420155e-07,
+      "loss": 0.5361,
+      "step": 74929
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.056392233478676e-07,
+      "loss": 0.6846,
+      "step": 74930
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.056137955775238e-07,
+      "loss": 0.7803,
+      "step": 74931
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.055883681091436e-07,
+      "loss": 0.6406,
+      "step": 74932
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05562940942746e-07,
+      "loss": 0.6758,
+      "step": 74933
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.05537514078351e-07,
+      "loss": 0.6465,
+      "step": 74934
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.055120875159775e-07,
+      "loss": 0.553,
+      "step": 74935
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.054866612556457e-07,
+      "loss": 0.457,
+      "step": 74936
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.054612352973745e-07,
+      "loss": 0.4512,
+      "step": 74937
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.054358096411831e-07,
+      "loss": 0.6172,
+      "step": 74938
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.054103842870918e-07,
+      "loss": 0.5543,
+      "step": 74939
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.053849592351192e-07,
+      "loss": 0.4702,
+      "step": 74940
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.053595344852855e-07,
+      "loss": 0.6865,
+      "step": 74941
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.0533411003761e-07,
+      "loss": 0.6846,
+      "step": 74942
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.053086858921117e-07,
+      "loss": 0.709,
+      "step": 74943
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.052832620488105e-07,
+      "loss": 0.6582,
+      "step": 74944
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.052578385077256e-07,
+      "loss": 0.5762,
+      "step": 74945
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.052324152688772e-07,
+      "loss": 0.5239,
+      "step": 74946
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.052069923322838e-07,
+      "loss": 0.791,
+      "step": 74947
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.051815696979651e-07,
+      "loss": 0.4484,
+      "step": 74948
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.051561473659409e-07,
+      "loss": 0.6899,
+      "step": 74949
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.051307253362304e-07,
+      "loss": 0.5254,
+      "step": 74950
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.051053036088531e-07,
+      "loss": 0.6147,
+      "step": 74951
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.050798821838285e-07,
+      "loss": 0.6633,
+      "step": 74952
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.050544610611762e-07,
+      "loss": 0.5864,
+      "step": 74953
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.050290402409154e-07,
+      "loss": 0.6348,
+      "step": 74954
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.050036197230659e-07,
+      "loss": 0.6533,
+      "step": 74955
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.049781995076471e-07,
+      "loss": 0.7471,
+      "step": 74956
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.04952779594678e-07,
+      "loss": 0.7158,
+      "step": 74957
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.049273599841785e-07,
+      "loss": 0.6616,
+      "step": 74958
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.049019406761679e-07,
+      "loss": 0.6826,
+      "step": 74959
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.048765216706657e-07,
+      "loss": 0.6187,
+      "step": 74960
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.048511029676916e-07,
+      "loss": 0.5806,
+      "step": 74961
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.048256845672645e-07,
+      "loss": 0.6567,
+      "step": 74962
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.048002664694045e-07,
+      "loss": 0.6377,
+      "step": 74963
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.047748486741304e-07,
+      "loss": 0.5706,
+      "step": 74964
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.047494311814625e-07,
+      "loss": 0.6338,
+      "step": 74965
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.047240139914199e-07,
+      "loss": 0.4956,
+      "step": 74966
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.046985971040213e-07,
+      "loss": 0.5913,
+      "step": 74967
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.046731805192873e-07,
+      "loss": 0.6475,
+      "step": 74968
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.046477642372365e-07,
+      "loss": 0.6274,
+      "step": 74969
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.046223482578889e-07,
+      "loss": 0.5254,
+      "step": 74970
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.045969325812637e-07,
+      "loss": 0.4977,
+      "step": 74971
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.045715172073807e-07,
+      "loss": 0.7461,
+      "step": 74972
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.045461021362587e-07,
+      "loss": 0.5718,
+      "step": 74973
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.045206873679179e-07,
+      "loss": 0.6367,
+      "step": 74974
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.044952729023773e-07,
+      "loss": 0.667,
+      "step": 74975
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.04469858739657e-07,
+      "loss": 0.6562,
+      "step": 74976
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.044444448797753e-07,
+      "loss": 0.543,
+      "step": 74977
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.044190313227526e-07,
+      "loss": 0.6221,
+      "step": 74978
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.043936180686078e-07,
+      "loss": 0.6245,
+      "step": 74979
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.043682051173608e-07,
+      "loss": 0.5005,
+      "step": 74980
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.043427924690307e-07,
+      "loss": 0.5708,
+      "step": 74981
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.043173801236373e-07,
+      "loss": 0.5669,
+      "step": 74982
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.042919680811997e-07,
+      "loss": 0.4675,
+      "step": 74983
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.042665563417378e-07,
+      "loss": 0.5815,
+      "step": 74984
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.042411449052704e-07,
+      "loss": 0.6455,
+      "step": 74985
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.04215733771818e-07,
+      "loss": 0.5149,
+      "step": 74986
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.041903229413991e-07,
+      "loss": 0.5654,
+      "step": 74987
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.041649124140334e-07,
+      "loss": 0.6128,
+      "step": 74988
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.041395021897405e-07,
+      "loss": 0.5552,
+      "step": 74989
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.041140922685396e-07,
+      "loss": 0.4468,
+      "step": 74990
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.040886826504506e-07,
+      "loss": 0.6377,
+      "step": 74991
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.040632733354923e-07,
+      "loss": 0.5591,
+      "step": 74992
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.040378643236849e-07,
+      "loss": 0.6953,
+      "step": 74993
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.040124556150473e-07,
+      "loss": 0.582,
+      "step": 74994
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.039870472095993e-07,
+      "loss": 0.6987,
+      "step": 74995
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.039616391073605e-07,
+      "loss": 0.667,
+      "step": 74996
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.039362313083498e-07,
+      "loss": 0.7256,
+      "step": 74997
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.039108238125867e-07,
+      "loss": 0.494,
+      "step": 74998
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.038854166200911e-07,
+      "loss": 0.7432,
+      "step": 74999
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.03860009730882e-07,
+      "loss": 0.7031,
+      "step": 75000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.038346031449792e-07,
+      "loss": 0.6494,
+      "step": 75001
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.03809196862402e-07,
+      "loss": 0.6001,
+      "step": 75002
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.0378379088317e-07,
+      "loss": 0.7686,
+      "step": 75003
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.037583852073023e-07,
+      "loss": 0.6719,
+      "step": 75004
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.037329798348189e-07,
+      "loss": 0.5889,
+      "step": 75005
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.037075747657389e-07,
+      "loss": 0.6562,
+      "step": 75006
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.036821700000816e-07,
+      "loss": 0.6436,
+      "step": 75007
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.036567655378668e-07,
+      "loss": 0.7549,
+      "step": 75008
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.036313613791135e-07,
+      "loss": 0.6313,
+      "step": 75009
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.036059575238418e-07,
+      "loss": 0.6841,
+      "step": 75010
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.035805539720704e-07,
+      "loss": 0.48,
+      "step": 75011
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.035551507238195e-07,
+      "loss": 0.5811,
+      "step": 75012
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.035297477791079e-07,
+      "loss": 0.6523,
+      "step": 75013
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.035043451379556e-07,
+      "loss": 0.6074,
+      "step": 75014
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.034789428003818e-07,
+      "loss": 0.7329,
+      "step": 75015
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.03453540766406e-07,
+      "loss": 0.7471,
+      "step": 75016
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.034281390360474e-07,
+      "loss": 0.4795,
+      "step": 75017
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.034027376093258e-07,
+      "loss": 0.6523,
+      "step": 75018
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.033773364862602e-07,
+      "loss": 0.6748,
+      "step": 75019
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.033519356668706e-07,
+      "loss": 0.6968,
+      "step": 75020
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.033265351511759e-07,
+      "loss": 0.5688,
+      "step": 75021
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.033011349391961e-07,
+      "loss": 0.4169,
+      "step": 75022
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.032757350309501e-07,
+      "loss": 0.4905,
+      "step": 75023
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.03250335426458e-07,
+      "loss": 0.5542,
+      "step": 75024
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.032249361257385e-07,
+      "loss": 0.458,
+      "step": 75025
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.031995371288119e-07,
+      "loss": 0.5439,
+      "step": 75026
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.03174138435697e-07,
+      "loss": 0.6406,
+      "step": 75027
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.031487400464132e-07,
+      "loss": 0.5522,
+      "step": 75028
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.031233419609802e-07,
+      "loss": 0.5781,
+      "step": 75029
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.030979441794174e-07,
+      "loss": 0.6323,
+      "step": 75030
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.030725467017444e-07,
+      "loss": 0.5591,
+      "step": 75031
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.030471495279805e-07,
+      "loss": 0.5576,
+      "step": 75032
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.030217526581449e-07,
+      "loss": 0.749,
+      "step": 75033
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.029963560922576e-07,
+      "loss": 0.3771,
+      "step": 75034
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.029709598303374e-07,
+      "loss": 0.6709,
+      "step": 75035
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.029455638724046e-07,
+      "loss": 0.7144,
+      "step": 75036
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.029201682184779e-07,
+      "loss": 0.8125,
+      "step": 75037
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.028947728685767e-07,
+      "loss": 0.3254,
+      "step": 75038
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.028693778227209e-07,
+      "loss": 0.6133,
+      "step": 75039
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.028439830809296e-07,
+      "loss": 0.5962,
+      "step": 75040
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.028185886432227e-07,
+      "loss": 0.6831,
+      "step": 75041
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.027931945096189e-07,
+      "loss": 0.6885,
+      "step": 75042
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.027678006801385e-07,
+      "loss": 0.6826,
+      "step": 75043
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.027424071548003e-07,
+      "loss": 0.8037,
+      "step": 75044
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.027170139336241e-07,
+      "loss": 0.6523,
+      "step": 75045
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.026916210166295e-07,
+      "loss": 0.7598,
+      "step": 75046
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.026662284038353e-07,
+      "loss": 0.667,
+      "step": 75047
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.026408360952613e-07,
+      "loss": 0.7275,
+      "step": 75048
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.02615444090927e-07,
+      "loss": 0.6235,
+      "step": 75049
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.025900523908517e-07,
+      "loss": 0.6274,
+      "step": 75050
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.025646609950551e-07,
+      "loss": 0.6416,
+      "step": 75051
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.025392699035562e-07,
+      "loss": 0.3041,
+      "step": 75052
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.025138791163748e-07,
+      "loss": 0.5859,
+      "step": 75053
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.024884886335303e-07,
+      "loss": 0.5693,
+      "step": 75054
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.024630984550422e-07,
+      "loss": 0.6377,
+      "step": 75055
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.024377085809301e-07,
+      "loss": 0.6338,
+      "step": 75056
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.024123190112124e-07,
+      "loss": 0.7168,
+      "step": 75057
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.023869297459099e-07,
+      "loss": 0.634,
+      "step": 75058
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.023615407850411e-07,
+      "loss": 0.6128,
+      "step": 75059
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.02336152128626e-07,
+      "loss": 0.6523,
+      "step": 75060
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.023107637766836e-07,
+      "loss": 0.5693,
+      "step": 75061
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.022853757292338e-07,
+      "loss": 0.7061,
+      "step": 75062
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.022599879862955e-07,
+      "loss": 0.4993,
+      "step": 75063
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.022346005478887e-07,
+      "loss": 0.7373,
+      "step": 75064
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.022092134140329e-07,
+      "loss": 0.7012,
+      "step": 75065
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.021838265847469e-07,
+      "loss": 0.3555,
+      "step": 75066
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.021584400600501e-07,
+      "loss": 0.6592,
+      "step": 75067
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.021330538399627e-07,
+      "loss": 0.5811,
+      "step": 75068
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.021076679245035e-07,
+      "loss": 0.6416,
+      "step": 75069
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.020822823136924e-07,
+      "loss": 0.6138,
+      "step": 75070
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.020568970075484e-07,
+      "loss": 0.6147,
+      "step": 75071
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.020315120060911e-07,
+      "loss": 0.7686,
+      "step": 75072
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.020061273093399e-07,
+      "loss": 0.6245,
+      "step": 75073
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.019807429173147e-07,
+      "loss": 0.5947,
+      "step": 75074
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.019553588300346e-07,
+      "loss": 0.5327,
+      "step": 75075
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.019299750475186e-07,
+      "loss": 0.5817,
+      "step": 75076
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.019045915697867e-07,
+      "loss": 0.7061,
+      "step": 75077
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.018792083968578e-07,
+      "loss": 0.623,
+      "step": 75078
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.018538255287521e-07,
+      "loss": 0.4399,
+      "step": 75079
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.018284429654883e-07,
+      "loss": 0.6606,
+      "step": 75080
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.018030607070865e-07,
+      "loss": 0.6465,
+      "step": 75081
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.017776787535653e-07,
+      "loss": 0.5815,
+      "step": 75082
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.017522971049451e-07,
+      "loss": 0.4644,
+      "step": 75083
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.017269157612446e-07,
+      "loss": 0.4678,
+      "step": 75084
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.01701534722484e-07,
+      "loss": 0.7158,
+      "step": 75085
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.016761539886816e-07,
+      "loss": 0.6367,
+      "step": 75086
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.016507735598577e-07,
+      "loss": 0.542,
+      "step": 75087
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.016253934360314e-07,
+      "loss": 0.6631,
+      "step": 75088
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.016000136172224e-07,
+      "loss": 0.6912,
+      "step": 75089
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.015746341034496e-07,
+      "loss": 0.6528,
+      "step": 75090
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.01549254894733e-07,
+      "loss": 0.5054,
+      "step": 75091
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.015238759910916e-07,
+      "loss": 0.5266,
+      "step": 75092
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.014984973925454e-07,
+      "loss": 0.5928,
+      "step": 75093
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.014731190991131e-07,
+      "loss": 0.6191,
+      "step": 75094
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.014477411108151e-07,
+      "loss": 0.5447,
+      "step": 75095
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.0142236342767e-07,
+      "loss": 0.5544,
+      "step": 75096
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.013969860496972e-07,
+      "loss": 0.2887,
+      "step": 75097
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.013716089769166e-07,
+      "loss": 0.6631,
+      "step": 75098
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.013462322093472e-07,
+      "loss": 0.7002,
+      "step": 75099
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.013208557470089e-07,
+      "loss": 0.5479,
+      "step": 75100
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.012954795899206e-07,
+      "loss": 0.7334,
+      "step": 75101
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.012701037381024e-07,
+      "loss": 0.6982,
+      "step": 75102
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.01244728191573e-07,
+      "loss": 0.6885,
+      "step": 75103
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.012193529503524e-07,
+      "loss": 0.6611,
+      "step": 75104
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 6.011939780144602e-07,
+      "loss": 0.7378,
+      "step": 75105
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.011686033839149e-07,
+      "loss": 0.5933,
+      "step": 75106
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.011432290587365e-07,
+      "loss": 0.7217,
+      "step": 75107
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.011178550389444e-07,
+      "loss": 0.5208,
+      "step": 75108
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.01092481324558e-07,
+      "loss": 0.4487,
+      "step": 75109
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.010671079155969e-07,
+      "loss": 0.6211,
+      "step": 75110
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.010417348120802e-07,
+      "loss": 0.6348,
+      "step": 75111
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.010163620140276e-07,
+      "loss": 0.6543,
+      "step": 75112
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.009909895214582e-07,
+      "loss": 0.5645,
+      "step": 75113
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.00965617334392e-07,
+      "loss": 0.7139,
+      "step": 75114
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.009402454528483e-07,
+      "loss": 0.6074,
+      "step": 75115
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.009148738768457e-07,
+      "loss": 0.5991,
+      "step": 75116
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.008895026064046e-07,
+      "loss": 0.666,
+      "step": 75117
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.008641316415437e-07,
+      "loss": 0.4978,
+      "step": 75118
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.008387609822832e-07,
+      "loss": 0.7107,
+      "step": 75119
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.008133906286417e-07,
+      "loss": 0.5796,
+      "step": 75120
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.007880205806394e-07,
+      "loss": 0.5942,
+      "step": 75121
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.007626508382952e-07,
+      "loss": 0.6113,
+      "step": 75122
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.007372814016286e-07,
+      "loss": 0.5347,
+      "step": 75123
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.007119122706593e-07,
+      "loss": 0.6855,
+      "step": 75124
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.006865434454067e-07,
+      "loss": 0.6616,
+      "step": 75125
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.006611749258896e-07,
+      "loss": 0.6475,
+      "step": 75126
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.006358067121281e-07,
+      "loss": 0.9131,
+      "step": 75127
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.006104388041412e-07,
+      "loss": 0.6504,
+      "step": 75128
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.005850712019487e-07,
+      "loss": 0.5981,
+      "step": 75129
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.005597039055698e-07,
+      "loss": 0.6338,
+      "step": 75130
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.00534336915024e-07,
+      "loss": 0.5298,
+      "step": 75131
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.005089702303305e-07,
+      "loss": 0.6992,
+      "step": 75132
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.004836038515092e-07,
+      "loss": 0.4644,
+      "step": 75133
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.00458237778579e-07,
+      "loss": 0.4264,
+      "step": 75134
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.004328720115599e-07,
+      "loss": 0.748,
+      "step": 75135
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.004075065504707e-07,
+      "loss": 0.4724,
+      "step": 75136
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.003821413953312e-07,
+      "loss": 0.6943,
+      "step": 75137
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.003567765461608e-07,
+      "loss": 0.7266,
+      "step": 75138
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.003314120029787e-07,
+      "loss": 0.6592,
+      "step": 75139
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.003060477658044e-07,
+      "loss": 0.7471,
+      "step": 75140
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.002806838346575e-07,
+      "loss": 0.6152,
+      "step": 75141
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.00255320209557e-07,
+      "loss": 0.6182,
+      "step": 75142
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.002299568905231e-07,
+      "loss": 0.4609,
+      "step": 75143
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.002045938775743e-07,
+      "loss": 0.5679,
+      "step": 75144
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.00179231170731e-07,
+      "loss": 0.7773,
+      "step": 75145
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.001538687700119e-07,
+      "loss": 0.5854,
+      "step": 75146
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.001285066754362e-07,
+      "loss": 0.6028,
+      "step": 75147
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.001031448870241e-07,
+      "loss": 0.6201,
+      "step": 75148
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.000777834047944e-07,
+      "loss": 0.493,
+      "step": 75149
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.000524222287668e-07,
+      "loss": 0.6968,
+      "step": 75150
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.000270613589607e-07,
+      "loss": 0.6943,
+      "step": 75151
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.000017007953955e-07,
+      "loss": 0.6016,
+      "step": 75152
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.999763405380905e-07,
+      "loss": 0.7656,
+      "step": 75153
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.999509805870655e-07,
+      "loss": 0.5542,
+      "step": 75154
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.999256209423397e-07,
+      "loss": 0.585,
+      "step": 75155
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.999002616039323e-07,
+      "loss": 0.6069,
+      "step": 75156
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.998749025718627e-07,
+      "loss": 0.6953,
+      "step": 75157
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.998495438461507e-07,
+      "loss": 0.5493,
+      "step": 75158
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.998241854268153e-07,
+      "loss": 0.665,
+      "step": 75159
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.997988273138763e-07,
+      "loss": 0.6777,
+      "step": 75160
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.997734695073527e-07,
+      "loss": 0.6448,
+      "step": 75161
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.997481120072642e-07,
+      "loss": 0.6875,
+      "step": 75162
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.997227548136303e-07,
+      "loss": 0.709,
+      "step": 75163
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.996973979264704e-07,
+      "loss": 0.5142,
+      "step": 75164
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.996720413458038e-07,
+      "loss": 0.5239,
+      "step": 75165
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.996466850716497e-07,
+      "loss": 0.7959,
+      "step": 75166
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.996213291040277e-07,
+      "loss": 0.6357,
+      "step": 75167
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.995959734429573e-07,
+      "loss": 0.6055,
+      "step": 75168
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.995706180884579e-07,
+      "loss": 0.6162,
+      "step": 75169
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.995452630405488e-07,
+      "loss": 0.6113,
+      "step": 75170
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.995199082992495e-07,
+      "loss": 0.6934,
+      "step": 75171
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.994945538645792e-07,
+      "loss": 0.5815,
+      "step": 75172
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.99469199736558e-07,
+      "loss": 0.7129,
+      "step": 75173
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.994438459152042e-07,
+      "loss": 0.4111,
+      "step": 75174
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.994184924005388e-07,
+      "loss": 0.6562,
+      "step": 75175
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.993931391925795e-07,
+      "loss": 0.3896,
+      "step": 75176
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.993677862913466e-07,
+      "loss": 0.5938,
+      "step": 75177
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.993424336968591e-07,
+      "loss": 0.6719,
+      "step": 75178
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.993170814091371e-07,
+      "loss": 0.7373,
+      "step": 75179
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.992917294281992e-07,
+      "loss": 0.6377,
+      "step": 75180
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.992663777540654e-07,
+      "loss": 0.5986,
+      "step": 75181
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.992410263867548e-07,
+      "loss": 0.6709,
+      "step": 75182
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.992156753262872e-07,
+      "loss": 0.5977,
+      "step": 75183
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.991903245726813e-07,
+      "loss": 0.6343,
+      "step": 75184
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.991649741259574e-07,
+      "loss": 0.5405,
+      "step": 75185
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.991396239861344e-07,
+      "loss": 0.6162,
+      "step": 75186
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.991142741532313e-07,
+      "loss": 0.6221,
+      "step": 75187
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.990889246272683e-07,
+      "loss": 0.6372,
+      "step": 75188
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.990635754082645e-07,
+      "loss": 0.668,
+      "step": 75189
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.990382264962393e-07,
+      "loss": 0.5007,
+      "step": 75190
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.990128778912118e-07,
+      "loss": 0.7031,
+      "step": 75191
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.98987529593202e-07,
+      "loss": 0.6689,
+      "step": 75192
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.989621816022288e-07,
+      "loss": 0.7007,
+      "step": 75193
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.989368339183119e-07,
+      "loss": 0.7324,
+      "step": 75194
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.98911486541471e-07,
+      "loss": 0.5513,
+      "step": 75195
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.988861394717248e-07,
+      "loss": 0.4648,
+      "step": 75196
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.98860792709093e-07,
+      "loss": 0.5737,
+      "step": 75197
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.988354462535951e-07,
+      "loss": 0.7031,
+      "step": 75198
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.988101001052504e-07,
+      "loss": 0.7334,
+      "step": 75199
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.987847542640784e-07,
+      "loss": 0.6738,
+      "step": 75200
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.987594087300983e-07,
+      "loss": 0.4111,
+      "step": 75201
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.987340635033299e-07,
+      "loss": 0.8379,
+      "step": 75202
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.987087185837921e-07,
+      "loss": 0.6152,
+      "step": 75203
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.986833739715049e-07,
+      "loss": 0.6523,
+      "step": 75204
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.986580296664876e-07,
+      "loss": 0.5566,
+      "step": 75205
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.986326856687589e-07,
+      "loss": 0.6768,
+      "step": 75206
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.986073419783389e-07,
+      "loss": 0.623,
+      "step": 75207
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.985819985952467e-07,
+      "loss": 0.7109,
+      "step": 75208
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.985566555195019e-07,
+      "loss": 0.5991,
+      "step": 75209
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.985313127511236e-07,
+      "loss": 0.5192,
+      "step": 75210
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.985059702901316e-07,
+      "loss": 0.563,
+      "step": 75211
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.984806281365453e-07,
+      "loss": 0.6772,
+      "step": 75212
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.984552862903835e-07,
+      "loss": 0.7441,
+      "step": 75213
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.984299447516664e-07,
+      "loss": 0.5781,
+      "step": 75214
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.984046035204132e-07,
+      "loss": 0.6514,
+      "step": 75215
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.983792625966426e-07,
+      "loss": 0.7676,
+      "step": 75216
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.983539219803748e-07,
+      "loss": 0.6553,
+      "step": 75217
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.983285816716288e-07,
+      "loss": 0.4656,
+      "step": 75218
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.983032416704243e-07,
+      "loss": 0.5752,
+      "step": 75219
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.982779019767803e-07,
+      "loss": 0.7422,
+      "step": 75220
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.982525625907167e-07,
+      "loss": 0.426,
+      "step": 75221
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.982272235122523e-07,
+      "loss": 0.562,
+      "step": 75222
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.982018847414072e-07,
+      "loss": 0.75,
+      "step": 75223
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.981765462782002e-07,
+      "loss": 0.5684,
+      "step": 75224
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.981512081226513e-07,
+      "loss": 0.6041,
+      "step": 75225
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.981258702747794e-07,
+      "loss": 0.6875,
+      "step": 75226
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.981005327346038e-07,
+      "loss": 0.5034,
+      "step": 75227
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.980751955021444e-07,
+      "loss": 0.623,
+      "step": 75228
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.980498585774203e-07,
+      "loss": 0.646,
+      "step": 75229
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.980245219604507e-07,
+      "loss": 0.5801,
+      "step": 75230
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.979991856512554e-07,
+      "loss": 0.6909,
+      "step": 75231
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.979738496498536e-07,
+      "loss": 0.7334,
+      "step": 75232
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.979485139562648e-07,
+      "loss": 0.6973,
+      "step": 75233
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.979231785705081e-07,
+      "loss": 0.5957,
+      "step": 75234
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.978978434926036e-07,
+      "loss": 0.7637,
+      "step": 75235
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.978725087225702e-07,
+      "loss": 0.4849,
+      "step": 75236
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.97847174260427e-07,
+      "loss": 0.5693,
+      "step": 75237
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.978218401061938e-07,
+      "loss": 0.5482,
+      "step": 75238
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.977965062598899e-07,
+      "loss": 0.7344,
+      "step": 75239
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.977711727215349e-07,
+      "loss": 0.5107,
+      "step": 75240
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.977458394911477e-07,
+      "loss": 0.3691,
+      "step": 75241
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.977205065687483e-07,
+      "loss": 0.5996,
+      "step": 75242
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.976951739543557e-07,
+      "loss": 0.6035,
+      "step": 75243
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.976698416479894e-07,
+      "loss": 0.7476,
+      "step": 75244
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.976445096496692e-07,
+      "loss": 0.7192,
+      "step": 75245
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.976191779594138e-07,
+      "loss": 0.5801,
+      "step": 75246
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.975938465772425e-07,
+      "loss": 0.6375,
+      "step": 75247
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.975685155031755e-07,
+      "loss": 0.5776,
+      "step": 75248
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.975431847372317e-07,
+      "loss": 0.4504,
+      "step": 75249
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.975178542794306e-07,
+      "loss": 0.6489,
+      "step": 75250
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.974925241297915e-07,
+      "loss": 0.6787,
+      "step": 75251
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.974671942883339e-07,
+      "loss": 0.6177,
+      "step": 75252
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.974418647550771e-07,
+      "loss": 0.5317,
+      "step": 75253
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.974165355300406e-07,
+      "loss": 0.7568,
+      "step": 75254
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.973912066132442e-07,
+      "loss": 0.6797,
+      "step": 75255
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.973658780047062e-07,
+      "loss": 0.408,
+      "step": 75256
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.97340549704447e-07,
+      "loss": 0.5439,
+      "step": 75257
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.973152217124853e-07,
+      "loss": 0.5801,
+      "step": 75258
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.972898940288411e-07,
+      "loss": 0.5684,
+      "step": 75259
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.972645666535332e-07,
+      "loss": 0.6826,
+      "step": 75260
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.972392395865817e-07,
+      "loss": 0.5674,
+      "step": 75261
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.972139128280051e-07,
+      "loss": 0.5767,
+      "step": 75262
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.971885863778238e-07,
+      "loss": 0.6494,
+      "step": 75263
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.971632602360563e-07,
+      "loss": 0.5752,
+      "step": 75264
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.97137934402723e-07,
+      "loss": 0.6533,
+      "step": 75265
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.971126088778421e-07,
+      "loss": 0.7812,
+      "step": 75266
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.970872836614338e-07,
+      "loss": 0.4956,
+      "step": 75267
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.970619587535168e-07,
+      "loss": 0.5654,
+      "step": 75268
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.970366341541114e-07,
+      "loss": 0.6143,
+      "step": 75269
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.970113098632363e-07,
+      "loss": 0.7896,
+      "step": 75270
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.969859858809112e-07,
+      "loss": 0.6528,
+      "step": 75271
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.969606622071551e-07,
+      "loss": 0.6284,
+      "step": 75272
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.969353388419879e-07,
+      "loss": 0.5626,
+      "step": 75273
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.969100157854287e-07,
+      "loss": 0.687,
+      "step": 75274
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.968846930374975e-07,
+      "loss": 0.6523,
+      "step": 75275
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.968593705982127e-07,
+      "loss": 0.5796,
+      "step": 75276
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.968340484675943e-07,
+      "loss": 0.6079,
+      "step": 75277
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.968087266456614e-07,
+      "loss": 0.5793,
+      "step": 75278
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.967834051324334e-07,
+      "loss": 0.5107,
+      "step": 75279
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.9675808392793e-07,
+      "loss": 0.5295,
+      "step": 75280
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.967327630321701e-07,
+      "loss": 0.5933,
+      "step": 75281
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.967074424451737e-07,
+      "loss": 0.7617,
+      "step": 75282
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.966821221669597e-07,
+      "loss": 0.7617,
+      "step": 75283
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.966568021975478e-07,
+      "loss": 0.5581,
+      "step": 75284
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.966314825369574e-07,
+      "loss": 0.5601,
+      "step": 75285
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.966061631852076e-07,
+      "loss": 0.4663,
+      "step": 75286
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.965808441423177e-07,
+      "loss": 0.7695,
+      "step": 75287
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.965555254083074e-07,
+      "loss": 0.6646,
+      "step": 75288
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.965302069831959e-07,
+      "loss": 0.6392,
+      "step": 75289
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.965048888670029e-07,
+      "loss": 0.5054,
+      "step": 75290
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.964795710597472e-07,
+      "loss": 0.4314,
+      "step": 75291
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.96454253561449e-07,
+      "loss": 0.5581,
+      "step": 75292
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.964289363721267e-07,
+      "loss": 0.5674,
+      "step": 75293
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.964036194918006e-07,
+      "loss": 0.4548,
+      "step": 75294
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.963783029204898e-07,
+      "loss": 0.7207,
+      "step": 75295
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.963529866582134e-07,
+      "loss": 0.6943,
+      "step": 75296
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.963276707049909e-07,
+      "loss": 0.5239,
+      "step": 75297
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.963023550608416e-07,
+      "loss": 0.6045,
+      "step": 75298
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.962770397257852e-07,
+      "loss": 0.6113,
+      "step": 75299
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.962517246998408e-07,
+      "loss": 0.6299,
+      "step": 75300
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.96226409983028e-07,
+      "loss": 0.624,
+      "step": 75301
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.962010955753661e-07,
+      "loss": 0.7178,
+      "step": 75302
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.961757814768742e-07,
+      "loss": 0.5933,
+      "step": 75303
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.961504676875722e-07,
+      "loss": 0.7061,
+      "step": 75304
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.961251542074794e-07,
+      "loss": 0.5078,
+      "step": 75305
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.960998410366147e-07,
+      "loss": 0.5062,
+      "step": 75306
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.960745281749979e-07,
+      "loss": 0.5933,
+      "step": 75307
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.960492156226481e-07,
+      "loss": 0.5771,
+      "step": 75308
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.96023903379585e-07,
+      "loss": 0.5308,
+      "step": 75309
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.959985914458276e-07,
+      "loss": 0.6689,
+      "step": 75310
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.959732798213956e-07,
+      "loss": 0.5825,
+      "step": 75311
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.959479685063083e-07,
+      "loss": 0.6348,
+      "step": 75312
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.959226575005852e-07,
+      "loss": 0.6201,
+      "step": 75313
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.958973468042453e-07,
+      "loss": 0.7197,
+      "step": 75314
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.958720364173087e-07,
+      "loss": 0.6865,
+      "step": 75315
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.958467263397939e-07,
+      "loss": 0.5652,
+      "step": 75316
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.958214165717206e-07,
+      "loss": 0.8242,
+      "step": 75317
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.957961071131086e-07,
+      "loss": 0.6304,
+      "step": 75318
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.957707979639768e-07,
+      "loss": 0.4639,
+      "step": 75319
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.957454891243445e-07,
+      "loss": 0.5972,
+      "step": 75320
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.957201805942314e-07,
+      "loss": 0.6479,
+      "step": 75321
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.956948723736567e-07,
+      "loss": 0.6777,
+      "step": 75322
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.9566956446264e-07,
+      "loss": 0.5662,
+      "step": 75323
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.956442568612003e-07,
+      "loss": 0.6494,
+      "step": 75324
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.956189495693577e-07,
+      "loss": 0.6676,
+      "step": 75325
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.955936425871308e-07,
+      "loss": 0.6919,
+      "step": 75326
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.955683359145391e-07,
+      "loss": 0.5415,
+      "step": 75327
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.955430295516023e-07,
+      "loss": 0.4222,
+      "step": 75328
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.955177234983394e-07,
+      "loss": 0.574,
+      "step": 75329
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.954924177547702e-07,
+      "loss": 0.7266,
+      "step": 75330
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.954671123209137e-07,
+      "loss": 0.5674,
+      "step": 75331
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.954418071967895e-07,
+      "loss": 0.6086,
+      "step": 75332
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.954165023824167e-07,
+      "loss": 0.6821,
+      "step": 75333
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.953911978778152e-07,
+      "loss": 0.6387,
+      "step": 75334
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.953658936830042e-07,
+      "loss": 0.6401,
+      "step": 75335
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.953405897980029e-07,
+      "loss": 0.7129,
+      "step": 75336
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.953152862228302e-07,
+      "loss": 0.6211,
+      "step": 75337
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.952899829575063e-07,
+      "loss": 0.4875,
+      "step": 75338
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.952646800020501e-07,
+      "loss": 0.6313,
+      "step": 75339
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.952393773564813e-07,
+      "loss": 0.6387,
+      "step": 75340
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.952140750208189e-07,
+      "loss": 0.6423,
+      "step": 75341
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.951887729950826e-07,
+      "loss": 0.6816,
+      "step": 75342
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.951634712792915e-07,
+      "loss": 0.7422,
+      "step": 75343
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.951381698734655e-07,
+      "loss": 0.5142,
+      "step": 75344
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.951128687776235e-07,
+      "loss": 0.4788,
+      "step": 75345
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.950875679917847e-07,
+      "loss": 0.7549,
+      "step": 75346
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.950622675159689e-07,
+      "loss": 0.6367,
+      "step": 75347
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.950369673501952e-07,
+      "loss": 0.7021,
+      "step": 75348
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.950116674944831e-07,
+      "loss": 0.6528,
+      "step": 75349
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.949863679488518e-07,
+      "loss": 0.7188,
+      "step": 75350
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.949610687133211e-07,
+      "loss": 0.5046,
+      "step": 75351
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.949357697879097e-07,
+      "loss": 0.519,
+      "step": 75352
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.949104711726377e-07,
+      "loss": 0.7197,
+      "step": 75353
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.94885172867524e-07,
+      "loss": 0.6006,
+      "step": 75354
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.948598748725886e-07,
+      "loss": 0.4858,
+      "step": 75355
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.948345771878496e-07,
+      "loss": 0.7812,
+      "step": 75356
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.948092798133275e-07,
+      "loss": 0.7454,
+      "step": 75357
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.947839827490412e-07,
+      "loss": 0.6089,
+      "step": 75358
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.947586859950102e-07,
+      "loss": 0.3855,
+      "step": 75359
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.947333895512537e-07,
+      "loss": 0.6523,
+      "step": 75360
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.947080934177915e-07,
+      "loss": 0.4526,
+      "step": 75361
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.946827975946425e-07,
+      "loss": 0.4893,
+      "step": 75362
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.946575020818263e-07,
+      "loss": 0.6333,
+      "step": 75363
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.946322068793623e-07,
+      "loss": 0.4192,
+      "step": 75364
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.946069119872696e-07,
+      "loss": 0.6611,
+      "step": 75365
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.945816174055679e-07,
+      "loss": 0.6577,
+      "step": 75366
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.945563231342763e-07,
+      "loss": 0.6714,
+      "step": 75367
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.945310291734142e-07,
+      "loss": 0.6553,
+      "step": 75368
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.945057355230011e-07,
+      "loss": 0.5605,
+      "step": 75369
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.944804421830565e-07,
+      "loss": 0.7246,
+      "step": 75370
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.944551491535993e-07,
+      "loss": 0.6509,
+      "step": 75371
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.944298564346493e-07,
+      "loss": 0.54,
+      "step": 75372
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.944045640262256e-07,
+      "loss": 0.4839,
+      "step": 75373
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.943792719283483e-07,
+      "loss": 0.5264,
+      "step": 75374
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.943539801410354e-07,
+      "loss": 0.791,
+      "step": 75375
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.943286886643072e-07,
+      "loss": 0.6343,
+      "step": 75376
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.94303397498183e-07,
+      "loss": 0.4971,
+      "step": 75377
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.942781066426819e-07,
+      "loss": 0.6309,
+      "step": 75378
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.942528160978233e-07,
+      "loss": 0.7192,
+      "step": 75379
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.942275258636269e-07,
+      "loss": 0.5439,
+      "step": 75380
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.942022359401115e-07,
+      "loss": 0.5352,
+      "step": 75381
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.941769463272971e-07,
+      "loss": 0.6445,
+      "step": 75382
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.941516570252026e-07,
+      "loss": 0.5332,
+      "step": 75383
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.941263680338477e-07,
+      "loss": 0.5938,
+      "step": 75384
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.941010793532517e-07,
+      "loss": 0.4448,
+      "step": 75385
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.940757909834333e-07,
+      "loss": 0.6201,
+      "step": 75386
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.940505029244129e-07,
+      "loss": 0.6582,
+      "step": 75387
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.940252151762088e-07,
+      "loss": 0.6377,
+      "step": 75388
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.939999277388414e-07,
+      "loss": 0.6531,
+      "step": 75389
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.939746406123294e-07,
+      "loss": 0.7334,
+      "step": 75390
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.939493537966924e-07,
+      "loss": 0.4561,
+      "step": 75391
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.939240672919499e-07,
+      "loss": 0.6855,
+      "step": 75392
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.938987810981208e-07,
+      "loss": 0.6836,
+      "step": 75393
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.938734952152251e-07,
+      "loss": 0.6616,
+      "step": 75394
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.938482096432815e-07,
+      "loss": 0.6719,
+      "step": 75395
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.938229243823097e-07,
+      "loss": 0.8438,
+      "step": 75396
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.937976394323289e-07,
+      "loss": 0.709,
+      "step": 75397
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.937723547933585e-07,
+      "loss": 0.7695,
+      "step": 75398
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.937470704654181e-07,
+      "loss": 0.6152,
+      "step": 75399
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.937217864485268e-07,
+      "loss": 0.6177,
+      "step": 75400
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.936965027427042e-07,
+      "loss": 0.71,
+      "step": 75401
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.936712193479693e-07,
+      "loss": 0.6699,
+      "step": 75402
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.936459362643418e-07,
+      "loss": 0.585,
+      "step": 75403
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.936206534918411e-07,
+      "loss": 0.8301,
+      "step": 75404
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.935953710304862e-07,
+      "loss": 0.6719,
+      "step": 75405
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.935700888802966e-07,
+      "loss": 0.6753,
+      "step": 75406
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.935448070412916e-07,
+      "loss": 0.6099,
+      "step": 75407
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.935195255134909e-07,
+      "loss": 0.7256,
+      "step": 75408
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.934942442969135e-07,
+      "loss": 0.7119,
+      "step": 75409
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.934689633915785e-07,
+      "loss": 0.4937,
+      "step": 75410
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.934436827975063e-07,
+      "loss": 0.5542,
+      "step": 75411
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.934184025147151e-07,
+      "loss": 0.6313,
+      "step": 75412
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.933931225432248e-07,
+      "loss": 0.751,
+      "step": 75413
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.933678428830551e-07,
+      "loss": 0.5435,
+      "step": 75414
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.933425635342244e-07,
+      "loss": 0.5469,
+      "step": 75415
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.933172844967529e-07,
+      "loss": 0.7607,
+      "step": 75416
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.932920057706594e-07,
+      "loss": 0.5547,
+      "step": 75417
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.932667273559638e-07,
+      "loss": 0.6562,
+      "step": 75418
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.93241449252685e-07,
+      "loss": 0.7129,
+      "step": 75419
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.932161714608426e-07,
+      "loss": 0.5537,
+      "step": 75420
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.931908939804557e-07,
+      "loss": 0.6484,
+      "step": 75421
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.93165616811544e-07,
+      "loss": 0.6182,
+      "step": 75422
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.931403399541265e-07,
+      "loss": 0.7803,
+      "step": 75423
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.931150634082232e-07,
+      "loss": 0.623,
+      "step": 75424
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.930897871738528e-07,
+      "loss": 0.7676,
+      "step": 75425
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.930645112510346e-07,
+      "loss": 0.432,
+      "step": 75426
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.930392356397882e-07,
+      "loss": 0.5073,
+      "step": 75427
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.930139603401331e-07,
+      "loss": 0.6304,
+      "step": 75428
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.929886853520882e-07,
+      "loss": 0.6072,
+      "step": 75429
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.929634106756735e-07,
+      "loss": 0.6274,
+      "step": 75430
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.929381363109078e-07,
+      "loss": 0.8145,
+      "step": 75431
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.929128622578108e-07,
+      "loss": 0.7041,
+      "step": 75432
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.928875885164015e-07,
+      "loss": 0.5996,
+      "step": 75433
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.928623150866998e-07,
+      "loss": 0.6348,
+      "step": 75434
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.928370419687247e-07,
+      "loss": 0.6211,
+      "step": 75435
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.928117691624952e-07,
+      "loss": 0.7129,
+      "step": 75436
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.927864966680312e-07,
+      "loss": 0.6895,
+      "step": 75437
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.927612244853517e-07,
+      "loss": 0.522,
+      "step": 75438
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.927359526144764e-07,
+      "loss": 0.5024,
+      "step": 75439
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.927106810554241e-07,
+      "loss": 0.7041,
+      "step": 75440
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.926854098082149e-07,
+      "loss": 0.6187,
+      "step": 75441
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.926601388728675e-07,
+      "loss": 0.6719,
+      "step": 75442
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.926348682494018e-07,
+      "loss": 0.6436,
+      "step": 75443
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.926095979378369e-07,
+      "loss": 0.6074,
+      "step": 75444
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.92584327938192e-07,
+      "loss": 0.5903,
+      "step": 75445
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.925590582504863e-07,
+      "loss": 0.5894,
+      "step": 75446
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.925337888747395e-07,
+      "loss": 0.6343,
+      "step": 75447
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.925085198109708e-07,
+      "loss": 0.5386,
+      "step": 75448
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.924832510591996e-07,
+      "loss": 0.6875,
+      "step": 75449
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.924579826194451e-07,
+      "loss": 0.7295,
+      "step": 75450
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.924327144917272e-07,
+      "loss": 0.5205,
+      "step": 75451
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.924074466760645e-07,
+      "loss": 0.6812,
+      "step": 75452
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.923821791724767e-07,
+      "loss": 0.626,
+      "step": 75453
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.923569119809835e-07,
+      "loss": 0.668,
+      "step": 75454
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.923316451016033e-07,
+      "loss": 0.537,
+      "step": 75455
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.923063785343565e-07,
+      "loss": 0.592,
+      "step": 75456
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.922811122792617e-07,
+      "loss": 0.6973,
+      "step": 75457
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.922558463363385e-07,
+      "loss": 0.498,
+      "step": 75458
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.922305807056063e-07,
+      "loss": 0.7021,
+      "step": 75459
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.922053153870842e-07,
+      "loss": 0.4595,
+      "step": 75460
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.921800503807919e-07,
+      "loss": 0.6152,
+      "step": 75461
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.921547856867488e-07,
+      "loss": 0.6025,
+      "step": 75462
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.921295213049736e-07,
+      "loss": 0.6279,
+      "step": 75463
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.921042572354868e-07,
+      "loss": 0.4834,
+      "step": 75464
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.920789934783063e-07,
+      "loss": 0.6338,
+      "step": 75465
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.920537300334526e-07,
+      "loss": 0.5503,
+      "step": 75466
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.920284669009442e-07,
+      "loss": 0.6436,
+      "step": 75467
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.920032040808012e-07,
+      "loss": 0.6201,
+      "step": 75468
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.919779415730422e-07,
+      "loss": 0.666,
+      "step": 75469
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.919526793776872e-07,
+      "loss": 0.7881,
+      "step": 75470
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91927417494755e-07,
+      "loss": 0.6562,
+      "step": 75471
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.919021559242656e-07,
+      "loss": 0.6904,
+      "step": 75472
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.918768946662377e-07,
+      "loss": 0.585,
+      "step": 75473
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.918516337206912e-07,
+      "loss": 0.6187,
+      "step": 75474
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91826373087645e-07,
+      "loss": 0.5488,
+      "step": 75475
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.918011127671184e-07,
+      "loss": 0.5615,
+      "step": 75476
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91775852759131e-07,
+      "loss": 0.5269,
+      "step": 75477
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91750593063702e-07,
+      "loss": 0.6807,
+      "step": 75478
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91725333680851e-07,
+      "loss": 0.585,
+      "step": 75479
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.917000746105968e-07,
+      "loss": 0.7412,
+      "step": 75480
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.916748158529594e-07,
+      "loss": 0.5046,
+      "step": 75481
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.916495574079577e-07,
+      "loss": 0.6528,
+      "step": 75482
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.916242992756111e-07,
+      "loss": 0.6934,
+      "step": 75483
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.915990414559393e-07,
+      "loss": 0.457,
+      "step": 75484
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.915737839489611e-07,
+      "loss": 0.7168,
+      "step": 75485
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.91548526754696e-07,
+      "loss": 0.6631,
+      "step": 75486
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.915232698731636e-07,
+      "loss": 0.5635,
+      "step": 75487
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.914980133043827e-07,
+      "loss": 0.8467,
+      "step": 75488
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.914727570483732e-07,
+      "loss": 0.5143,
+      "step": 75489
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.914475011051543e-07,
+      "loss": 0.6628,
+      "step": 75490
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.914222454747453e-07,
+      "loss": 0.5273,
+      "step": 75491
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.913969901571651e-07,
+      "loss": 0.6484,
+      "step": 75492
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.913717351524339e-07,
+      "loss": 0.5977,
+      "step": 75493
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.913464804605708e-07,
+      "loss": 0.6045,
+      "step": 75494
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.913212260815943e-07,
+      "loss": 0.4956,
+      "step": 75495
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.912959720155246e-07,
+      "loss": 0.624,
+      "step": 75496
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.912707182623807e-07,
+      "loss": 0.7451,
+      "step": 75497
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.912454648221821e-07,
+      "loss": 0.603,
+      "step": 75498
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.912202116949481e-07,
+      "loss": 0.7007,
+      "step": 75499
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.911949588806977e-07,
+      "loss": 0.5571,
+      "step": 75500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.911697063794509e-07,
+      "loss": 0.6597,
+      "step": 75501
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.911444541912263e-07,
+      "loss": 0.5649,
+      "step": 75502
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.911192023160438e-07,
+      "loss": 0.6353,
+      "step": 75503
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.910939507539229e-07,
+      "loss": 0.5498,
+      "step": 75504
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.91068699504882e-07,
+      "loss": 0.6504,
+      "step": 75505
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.910434485689413e-07,
+      "loss": 0.6533,
+      "step": 75506
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.910181979461195e-07,
+      "loss": 0.6982,
+      "step": 75507
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.909929476364366e-07,
+      "loss": 0.7148,
+      "step": 75508
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.909676976399111e-07,
+      "loss": 0.6094,
+      "step": 75509
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.909424479565635e-07,
+      "loss": 0.7197,
+      "step": 75510
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.909171985864119e-07,
+      "loss": 0.5942,
+      "step": 75511
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.908919495294764e-07,
+      "loss": 0.5708,
+      "step": 75512
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.908667007857761e-07,
+      "loss": 0.6123,
+      "step": 75513
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.908414523553307e-07,
+      "loss": 0.6201,
+      "step": 75514
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.908162042381588e-07,
+      "loss": 0.6597,
+      "step": 75515
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.907909564342803e-07,
+      "loss": 0.5527,
+      "step": 75516
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.907657089437142e-07,
+      "loss": 0.7188,
+      "step": 75517
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.9074046176648e-07,
+      "loss": 0.6191,
+      "step": 75518
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.907152149025969e-07,
+      "loss": 0.7627,
+      "step": 75519
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.906899683520846e-07,
+      "loss": 0.4459,
+      "step": 75520
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.906647221149619e-07,
+      "loss": 0.5859,
+      "step": 75521
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.906394761912487e-07,
+      "loss": 0.6416,
+      "step": 75522
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.906142305809638e-07,
+      "loss": 0.5591,
+      "step": 75523
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.90588985284127e-07,
+      "loss": 0.5087,
+      "step": 75524
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.905637403007572e-07,
+      "loss": 0.5693,
+      "step": 75525
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.905384956308738e-07,
+      "loss": 0.6543,
+      "step": 75526
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.905132512744966e-07,
+      "loss": 0.5,
+      "step": 75527
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.904880072316443e-07,
+      "loss": 0.5913,
+      "step": 75528
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.904627635023366e-07,
+      "loss": 0.543,
+      "step": 75529
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.904375200865925e-07,
+      "loss": 0.5874,
+      "step": 75530
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.904122769844318e-07,
+      "loss": 0.6514,
+      "step": 75531
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.903870341958735e-07,
+      "loss": 0.708,
+      "step": 75532
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.903617917209371e-07,
+      "loss": 0.5967,
+      "step": 75533
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.903365495596421e-07,
+      "loss": 0.3521,
+      "step": 75534
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.903113077120073e-07,
+      "loss": 0.5946,
+      "step": 75535
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.902860661780522e-07,
+      "loss": 0.6641,
+      "step": 75536
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.902608249577962e-07,
+      "loss": 0.7148,
+      "step": 75537
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.902355840512587e-07,
+      "loss": 0.6973,
+      "step": 75538
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.902103434584591e-07,
+      "loss": 0.7881,
+      "step": 75539
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.901851031794162e-07,
+      "loss": 0.9854,
+      "step": 75540
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.901598632141502e-07,
+      "loss": 0.481,
+      "step": 75541
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.901346235626796e-07,
+      "loss": 0.4386,
+      "step": 75542
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.901093842250244e-07,
+      "loss": 0.8652,
+      "step": 75543
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.900841452012037e-07,
+      "loss": 0.6348,
+      "step": 75544
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.900589064912362e-07,
+      "loss": 0.6099,
+      "step": 75545
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.900336680951421e-07,
+      "loss": 0.6265,
+      "step": 75546
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.900084300129402e-07,
+      "loss": 0.7178,
+      "step": 75547
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.899831922446501e-07,
+      "loss": 0.7217,
+      "step": 75548
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.899579547902909e-07,
+      "loss": 0.585,
+      "step": 75549
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.899327176498822e-07,
+      "loss": 0.6309,
+      "step": 75550
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.89907480823443e-07,
+      "loss": 0.5347,
+      "step": 75551
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.898822443109928e-07,
+      "loss": 0.4399,
+      "step": 75552
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.89857008112551e-07,
+      "loss": 0.5317,
+      "step": 75553
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.898317722281372e-07,
+      "loss": 0.6201,
+      "step": 75554
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.898065366577698e-07,
+      "loss": 0.6797,
+      "step": 75555
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.897813014014689e-07,
+      "loss": 0.7339,
+      "step": 75556
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.897560664592535e-07,
+      "loss": 0.6602,
+      "step": 75557
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.897308318311431e-07,
+      "loss": 0.5291,
+      "step": 75558
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.897055975171567e-07,
+      "loss": 0.4624,
+      "step": 75559
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.896803635173142e-07,
+      "loss": 0.6367,
+      "step": 75560
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.896551298316342e-07,
+      "loss": 0.4941,
+      "step": 75561
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.896298964601367e-07,
+      "loss": 0.6074,
+      "step": 75562
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.896046634028404e-07,
+      "loss": 0.6328,
+      "step": 75563
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.895794306597656e-07,
+      "loss": 0.605,
+      "step": 75564
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.895541982309305e-07,
+      "loss": 0.6157,
+      "step": 75565
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.895289661163549e-07,
+      "loss": 0.751,
+      "step": 75566
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.895037343160582e-07,
+      "loss": 0.5903,
+      "step": 75567
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.894785028300593e-07,
+      "loss": 0.326,
+      "step": 75568
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.894532716583782e-07,
+      "loss": 0.7617,
+      "step": 75569
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.894280408010334e-07,
+      "loss": 0.6313,
+      "step": 75570
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.894028102580451e-07,
+      "loss": 0.666,
+      "step": 75571
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.893775800294323e-07,
+      "loss": 0.5742,
+      "step": 75572
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.893523501152138e-07,
+      "loss": 0.4209,
+      "step": 75573
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.893271205154097e-07,
+      "loss": 0.4971,
+      "step": 75574
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.893018912300389e-07,
+      "loss": 0.5464,
+      "step": 75575
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.892766622591205e-07,
+      "loss": 0.626,
+      "step": 75576
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.892514336026741e-07,
+      "loss": 0.6094,
+      "step": 75577
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.89226205260719e-07,
+      "loss": 0.6895,
+      "step": 75578
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.892009772332746e-07,
+      "loss": 0.6377,
+      "step": 75579
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.8917574952036e-07,
+      "loss": 0.5315,
+      "step": 75580
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.891505221219948e-07,
+      "loss": 0.6182,
+      "step": 75581
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.89125295038198e-07,
+      "loss": 0.5601,
+      "step": 75582
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.891000682689892e-07,
+      "loss": 0.606,
+      "step": 75583
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.890748418143878e-07,
+      "loss": 0.6733,
+      "step": 75584
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.890496156744125e-07,
+      "loss": 0.6377,
+      "step": 75585
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.890243898490832e-07,
+      "loss": 0.5393,
+      "step": 75586
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.889991643384189e-07,
+      "loss": 0.6924,
+      "step": 75587
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.889739391424391e-07,
+      "loss": 0.7476,
+      "step": 75588
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.889487142611631e-07,
+      "loss": 0.5605,
+      "step": 75589
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.889234896946101e-07,
+      "loss": 0.5081,
+      "step": 75590
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.888982654427996e-07,
+      "loss": 0.667,
+      "step": 75591
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.888730415057506e-07,
+      "loss": 0.6191,
+      "step": 75592
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.888478178834828e-07,
+      "loss": 0.7207,
+      "step": 75593
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.888225945760154e-07,
+      "loss": 0.4307,
+      "step": 75594
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.887973715833674e-07,
+      "loss": 0.4771,
+      "step": 75595
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.887721489055586e-07,
+      "loss": 0.666,
+      "step": 75596
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.887469265426078e-07,
+      "loss": 0.5132,
+      "step": 75597
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.887217044945346e-07,
+      "loss": 0.6064,
+      "step": 75598
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.886964827613582e-07,
+      "loss": 0.6221,
+      "step": 75599
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.886712613430982e-07,
+      "loss": 0.832,
+      "step": 75600
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.886460402397735e-07,
+      "loss": 0.6021,
+      "step": 75601
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.886208194514038e-07,
+      "loss": 0.647,
+      "step": 75602
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.88595598978008e-07,
+      "loss": 0.592,
+      "step": 75603
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.885703788196061e-07,
+      "loss": 0.6802,
+      "step": 75604
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.885451589762167e-07,
+      "loss": 0.6216,
+      "step": 75605
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.885199394478591e-07,
+      "loss": 0.6528,
+      "step": 75606
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.884947202345531e-07,
+      "loss": 0.7393,
+      "step": 75607
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.884695013363179e-07,
+      "loss": 0.7471,
+      "step": 75608
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.884442827531723e-07,
+      "loss": 0.5532,
+      "step": 75609
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.884190644851364e-07,
+      "loss": 0.6636,
+      "step": 75610
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.883938465322287e-07,
+      "loss": 0.5857,
+      "step": 75611
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.883686288944692e-07,
+      "loss": 0.6045,
+      "step": 75612
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.883434115718766e-07,
+      "loss": 0.5723,
+      "step": 75613
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.883181945644711e-07,
+      "loss": 0.6313,
+      "step": 75614
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.88292977872271e-07,
+      "loss": 0.7305,
+      "step": 75615
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.88267761495296e-07,
+      "loss": 0.7715,
+      "step": 75616
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.882425454335656e-07,
+      "loss": 0.5786,
+      "step": 75617
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.882173296870987e-07,
+      "loss": 0.6338,
+      "step": 75618
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.881921142559151e-07,
+      "loss": 0.5806,
+      "step": 75619
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.881668991400337e-07,
+      "loss": 0.6201,
+      "step": 75620
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.881416843394742e-07,
+      "loss": 0.6992,
+      "step": 75621
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.881164698542553e-07,
+      "loss": 0.7959,
+      "step": 75622
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.880912556843971e-07,
+      "loss": 0.5188,
+      "step": 75623
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.880660418299186e-07,
+      "loss": 0.54,
+      "step": 75624
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.880408282908388e-07,
+      "loss": 0.4346,
+      "step": 75625
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.880156150671769e-07,
+      "loss": 0.6416,
+      "step": 75626
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.879904021589528e-07,
+      "loss": 0.5562,
+      "step": 75627
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.879651895661853e-07,
+      "loss": 0.7139,
+      "step": 75628
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.87939977288894e-07,
+      "loss": 0.707,
+      "step": 75629
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.879147653270981e-07,
+      "loss": 0.5605,
+      "step": 75630
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.878895536808169e-07,
+      "loss": 0.7021,
+      "step": 75631
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.878643423500698e-07,
+      "loss": 0.6943,
+      "step": 75632
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.878391313348759e-07,
+      "loss": 0.7656,
+      "step": 75633
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.87813920635255e-07,
+      "loss": 0.6367,
+      "step": 75634
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.877887102512256e-07,
+      "loss": 0.5537,
+      "step": 75635
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.877635001828077e-07,
+      "loss": 0.6504,
+      "step": 75636
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.877382904300201e-07,
+      "loss": 0.5293,
+      "step": 75637
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.877130809928824e-07,
+      "loss": 0.6362,
+      "step": 75638
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.876878718714138e-07,
+      "loss": 0.6606,
+      "step": 75639
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.876626630656338e-07,
+      "loss": 0.5034,
+      "step": 75640
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.876374545755613e-07,
+      "loss": 0.5845,
+      "step": 75641
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.876122464012161e-07,
+      "loss": 0.7178,
+      "step": 75642
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.875870385426169e-07,
+      "loss": 0.582,
+      "step": 75643
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.87561830999784e-07,
+      "loss": 0.7007,
+      "step": 75644
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.875366237727355e-07,
+      "loss": 0.6543,
+      "step": 75645
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.875114168614914e-07,
+      "loss": 0.5098,
+      "step": 75646
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.874862102660707e-07,
+      "loss": 0.5845,
+      "step": 75647
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.874610039864931e-07,
+      "loss": 0.6787,
+      "step": 75648
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.874357980227773e-07,
+      "loss": 0.6631,
+      "step": 75649
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.874105923749431e-07,
+      "loss": 0.5332,
+      "step": 75650
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.873853870430097e-07,
+      "loss": 0.7334,
+      "step": 75651
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.873601820269962e-07,
+      "loss": 0.5723,
+      "step": 75652
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.87334977326922e-07,
+      "loss": 0.7153,
+      "step": 75653
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.873097729428069e-07,
+      "loss": 0.5767,
+      "step": 75654
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.872845688746695e-07,
+      "loss": 0.5884,
+      "step": 75655
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.872593651225291e-07,
+      "loss": 0.5542,
+      "step": 75656
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.872341616864054e-07,
+      "loss": 0.6138,
+      "step": 75657
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.872089585663174e-07,
+      "loss": 0.5942,
+      "step": 75658
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.871837557622846e-07,
+      "loss": 0.6279,
+      "step": 75659
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.87158553274326e-07,
+      "loss": 0.7607,
+      "step": 75660
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.871333511024615e-07,
+      "loss": 0.3912,
+      "step": 75661
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.871081492467099e-07,
+      "loss": 0.4185,
+      "step": 75662
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.870829477070906e-07,
+      "loss": 0.5806,
+      "step": 75663
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.870577464836229e-07,
+      "loss": 0.5659,
+      "step": 75664
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.870325455763259e-07,
+      "loss": 0.563,
+      "step": 75665
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.870073449852191e-07,
+      "loss": 0.6382,
+      "step": 75666
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.86982144710322e-07,
+      "loss": 0.4446,
+      "step": 75667
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.869569447516535e-07,
+      "loss": 0.5708,
+      "step": 75668
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.869317451092333e-07,
+      "loss": 0.5391,
+      "step": 75669
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.869065457830802e-07,
+      "loss": 0.7993,
+      "step": 75670
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.868813467732139e-07,
+      "loss": 0.6484,
+      "step": 75671
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.868561480796534e-07,
+      "loss": 0.6406,
+      "step": 75672
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.868309497024187e-07,
+      "loss": 0.6172,
+      "step": 75673
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.868057516415281e-07,
+      "loss": 0.7686,
+      "step": 75674
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.867805538970012e-07,
+      "loss": 0.6924,
+      "step": 75675
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.867553564688576e-07,
+      "loss": 0.627,
+      "step": 75676
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.867301593571164e-07,
+      "loss": 0.666,
+      "step": 75677
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.867049625617971e-07,
+      "loss": 0.6016,
+      "step": 75678
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.866797660829186e-07,
+      "loss": 0.5679,
+      "step": 75679
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.866545699205003e-07,
+      "loss": 0.3542,
+      "step": 75680
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.866293740745618e-07,
+      "loss": 0.563,
+      "step": 75681
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.866041785451221e-07,
+      "loss": 0.6895,
+      "step": 75682
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.865789833322008e-07,
+      "loss": 0.708,
+      "step": 75683
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.86553788435817e-07,
+      "loss": 0.6382,
+      "step": 75684
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.865285938559895e-07,
+      "loss": 0.4395,
+      "step": 75685
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.865033995927382e-07,
+      "loss": 0.7373,
+      "step": 75686
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.864782056460822e-07,
+      "loss": 0.6689,
+      "step": 75687
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.864530120160409e-07,
+      "loss": 0.7725,
+      "step": 75688
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.864278187026334e-07,
+      "loss": 0.7002,
+      "step": 75689
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.864026257058794e-07,
+      "loss": 0.645,
+      "step": 75690
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.863774330257975e-07,
+      "loss": 0.6025,
+      "step": 75691
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.863522406624077e-07,
+      "loss": 0.4135,
+      "step": 75692
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.863270486157293e-07,
+      "loss": 0.6118,
+      "step": 75693
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.863018568857805e-07,
+      "loss": 0.6382,
+      "step": 75694
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.862766654725818e-07,
+      "loss": 0.4839,
+      "step": 75695
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.862514743761517e-07,
+      "loss": 0.6729,
+      "step": 75696
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.862262835965102e-07,
+      "loss": 0.7148,
+      "step": 75697
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.86201093133676e-07,
+      "loss": 0.6523,
+      "step": 75698
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.861759029876685e-07,
+      "loss": 0.5464,
+      "step": 75699
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.861507131585074e-07,
+      "loss": 0.8037,
+      "step": 75700
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.861255236462112e-07,
+      "loss": 0.5498,
+      "step": 75701
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.861003344508002e-07,
+      "loss": 0.6689,
+      "step": 75702
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.860751455722931e-07,
+      "loss": 0.6992,
+      "step": 75703
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.860499570107089e-07,
+      "loss": 0.6992,
+      "step": 75704
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.860247687660673e-07,
+      "loss": 0.7627,
+      "step": 75705
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.859995808383875e-07,
+      "loss": 0.6689,
+      "step": 75706
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.859743932276889e-07,
+      "loss": 0.5613,
+      "step": 75707
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.859492059339904e-07,
+      "loss": 0.6357,
+      "step": 75708
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.859240189573118e-07,
+      "loss": 0.646,
+      "step": 75709
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.858988322976719e-07,
+      "loss": 0.4102,
+      "step": 75710
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.858736459550905e-07,
+      "loss": 0.6729,
+      "step": 75711
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.858484599295864e-07,
+      "loss": 0.6543,
+      "step": 75712
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.858232742211795e-07,
+      "loss": 0.4817,
+      "step": 75713
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.857980888298884e-07,
+      "loss": 0.7676,
+      "step": 75714
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.857729037557327e-07,
+      "loss": 0.7344,
+      "step": 75715
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.857477189987314e-07,
+      "loss": 0.7666,
+      "step": 75716
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.857225345589043e-07,
+      "loss": 0.7793,
+      "step": 75717
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.856973504362701e-07,
+      "loss": 0.4387,
+      "step": 75718
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.856721666308486e-07,
+      "loss": 0.6836,
+      "step": 75719
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.856469831426588e-07,
+      "loss": 0.4746,
+      "step": 75720
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.856217999717202e-07,
+      "loss": 0.6313,
+      "step": 75721
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.855966171180518e-07,
+      "loss": 0.5518,
+      "step": 75722
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.855714345816733e-07,
+      "loss": 0.6284,
+      "step": 75723
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.855462523626035e-07,
+      "loss": 0.3071,
+      "step": 75724
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.855210704608617e-07,
+      "loss": 0.5205,
+      "step": 75725
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.854958888764675e-07,
+      "loss": 0.7852,
+      "step": 75726
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.854707076094399e-07,
+      "loss": 0.7861,
+      "step": 75727
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.854455266597986e-07,
+      "loss": 0.6553,
+      "step": 75728
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.854203460275622e-07,
+      "loss": 0.7451,
+      "step": 75729
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.853951657127507e-07,
+      "loss": 0.627,
+      "step": 75730
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.853699857153829e-07,
+      "loss": 0.5942,
+      "step": 75731
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.853448060354784e-07,
+      "loss": 0.3667,
+      "step": 75732
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.853196266730565e-07,
+      "loss": 0.4402,
+      "step": 75733
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.85294447628136e-07,
+      "loss": 0.7031,
+      "step": 75734
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.852692689007365e-07,
+      "loss": 0.8857,
+      "step": 75735
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.852440904908773e-07,
+      "loss": 0.626,
+      "step": 75736
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.852189123985773e-07,
+      "loss": 0.7432,
+      "step": 75737
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.851937346238566e-07,
+      "loss": 0.6821,
+      "step": 75738
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.851685571667337e-07,
+      "loss": 0.7021,
+      "step": 75739
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.851433800272283e-07,
+      "loss": 0.6562,
+      "step": 75740
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.851182032053592e-07,
+      "loss": 0.6582,
+      "step": 75741
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.850930267011465e-07,
+      "loss": 0.4833,
+      "step": 75742
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.850678505146089e-07,
+      "loss": 0.5981,
+      "step": 75743
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.850426746457657e-07,
+      "loss": 0.6504,
+      "step": 75744
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.850174990946362e-07,
+      "loss": 0.5425,
+      "step": 75745
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.849923238612395e-07,
+      "loss": 0.6948,
+      "step": 75746
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.849671489455953e-07,
+      "loss": 0.8047,
+      "step": 75747
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.849419743477226e-07,
+      "loss": 0.519,
+      "step": 75748
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.849168000676409e-07,
+      "loss": 0.6162,
+      "step": 75749
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.848916261053691e-07,
+      "loss": 0.6025,
+      "step": 75750
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.84866452460927e-07,
+      "loss": 0.623,
+      "step": 75751
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.848412791343334e-07,
+      "loss": 0.5443,
+      "step": 75752
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.848161061256081e-07,
+      "loss": 0.7197,
+      "step": 75753
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.847909334347694e-07,
+      "loss": 0.6396,
+      "step": 75754
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.847657610618376e-07,
+      "loss": 0.5159,
+      "step": 75755
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.847405890068313e-07,
+      "loss": 0.4595,
+      "step": 75756
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.847154172697702e-07,
+      "loss": 0.6101,
+      "step": 75757
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.846902458506732e-07,
+      "loss": 0.7188,
+      "step": 75758
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.8466507474956e-07,
+      "loss": 0.3779,
+      "step": 75759
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.846399039664495e-07,
+      "loss": 0.5786,
+      "step": 75760
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.846147335013614e-07,
+      "loss": 0.6572,
+      "step": 75761
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.845895633543143e-07,
+      "loss": 0.7119,
+      "step": 75762
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.845643935253283e-07,
+      "loss": 0.6738,
+      "step": 75763
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.845392240144223e-07,
+      "loss": 0.5415,
+      "step": 75764
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.84514054821615e-07,
+      "loss": 0.5227,
+      "step": 75765
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.844888859469266e-07,
+      "loss": 0.6553,
+      "step": 75766
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.844637173903756e-07,
+      "loss": 0.6191,
+      "step": 75767
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.844385491519819e-07,
+      "loss": 0.6543,
+      "step": 75768
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.844133812317642e-07,
+      "loss": 0.4963,
+      "step": 75769
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.843882136297424e-07,
+      "loss": 0.8564,
+      "step": 75770
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.843630463459355e-07,
+      "loss": 0.6709,
+      "step": 75771
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.843378793803623e-07,
+      "loss": 0.7461,
+      "step": 75772
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.84312712733043e-07,
+      "loss": 0.6118,
+      "step": 75773
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.842875464039961e-07,
+      "loss": 0.7041,
+      "step": 75774
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.842623803932408e-07,
+      "loss": 0.5239,
+      "step": 75775
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.84237214700797e-07,
+      "loss": 0.4886,
+      "step": 75776
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.842120493266835e-07,
+      "loss": 0.6396,
+      "step": 75777
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.841868842709199e-07,
+      "loss": 0.7666,
+      "step": 75778
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.841617195335249e-07,
+      "loss": 0.7559,
+      "step": 75779
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.841365551145185e-07,
+      "loss": 0.4878,
+      "step": 75780
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.841113910139195e-07,
+      "loss": 0.6157,
+      "step": 75781
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.840862272317473e-07,
+      "loss": 0.6958,
+      "step": 75782
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.840610637680214e-07,
+      "loss": 0.6611,
+      "step": 75783
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.840359006227606e-07,
+      "loss": 0.5142,
+      "step": 75784
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.840107377959844e-07,
+      "loss": 0.7373,
+      "step": 75785
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.839855752877119e-07,
+      "loss": 0.6836,
+      "step": 75786
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.839604130979628e-07,
+      "loss": 0.6509,
+      "step": 75787
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.839352512267559e-07,
+      "loss": 0.5396,
+      "step": 75788
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.839100896741105e-07,
+      "loss": 0.6548,
+      "step": 75789
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.838849284400462e-07,
+      "loss": 0.6709,
+      "step": 75790
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.83859767524582e-07,
+      "loss": 0.5801,
+      "step": 75791
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.838346069277371e-07,
+      "loss": 0.5728,
+      "step": 75792
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.838094466495313e-07,
+      "loss": 0.6729,
+      "step": 75793
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.837842866899836e-07,
+      "loss": 0.4056,
+      "step": 75794
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.837591270491127e-07,
+      "loss": 0.6572,
+      "step": 75795
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.837339677269383e-07,
+      "loss": 0.709,
+      "step": 75796
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.837088087234796e-07,
+      "loss": 0.5156,
+      "step": 75797
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.836836500387563e-07,
+      "loss": 0.6191,
+      "step": 75798
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.83658491672787e-07,
+      "loss": 0.4673,
+      "step": 75799
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.836333336255912e-07,
+      "loss": 0.5693,
+      "step": 75800
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.836081758971882e-07,
+      "loss": 0.5142,
+      "step": 75801
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.835830184875976e-07,
+      "loss": 0.562,
+      "step": 75802
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.835578613968385e-07,
+      "loss": 0.575,
+      "step": 75803
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.835327046249295e-07,
+      "loss": 0.7236,
+      "step": 75804
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.835075481718903e-07,
+      "loss": 0.8018,
+      "step": 75805
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.834823920377407e-07,
+      "loss": 0.6802,
+      "step": 75806
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.834572362224992e-07,
+      "loss": 0.4214,
+      "step": 75807
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.834320807261852e-07,
+      "loss": 0.71,
+      "step": 75808
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.834069255488181e-07,
+      "loss": 0.4207,
+      "step": 75809
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.833817706904177e-07,
+      "loss": 0.6616,
+      "step": 75810
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.83356616151002e-07,
+      "loss": 0.5337,
+      "step": 75811
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.833314619305911e-07,
+      "loss": 0.624,
+      "step": 75812
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.833063080292046e-07,
+      "loss": 0.5747,
+      "step": 75813
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.83281154446861e-07,
+      "loss": 0.6172,
+      "step": 75814
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.832560011835803e-07,
+      "loss": 0.5051,
+      "step": 75815
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.832308482393806e-07,
+      "loss": 0.522,
+      "step": 75816
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.832056956142822e-07,
+      "loss": 0.5664,
+      "step": 75817
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.831805433083039e-07,
+      "loss": 0.5695,
+      "step": 75818
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.831553913214655e-07,
+      "loss": 0.8105,
+      "step": 75819
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.831302396537854e-07,
+      "loss": 0.5596,
+      "step": 75820
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.831050883052834e-07,
+      "loss": 0.707,
+      "step": 75821
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.830799372759785e-07,
+      "loss": 0.6221,
+      "step": 75822
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.830547865658911e-07,
+      "loss": 0.4919,
+      "step": 75823
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.830296361750386e-07,
+      "loss": 0.509,
+      "step": 75824
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.830044861034411e-07,
+      "loss": 0.4536,
+      "step": 75825
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.82979336351118e-07,
+      "loss": 0.791,
+      "step": 75826
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.829541869180887e-07,
+      "loss": 0.5352,
+      "step": 75827
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.82929037804372e-07,
+      "loss": 0.6406,
+      "step": 75828
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.829038890099872e-07,
+      "loss": 0.6162,
+      "step": 75829
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.828787405349537e-07,
+      "loss": 0.666,
+      "step": 75830
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.828535923792914e-07,
+      "loss": 0.7559,
+      "step": 75831
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.828284445430183e-07,
+      "loss": 0.6274,
+      "step": 75832
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.828032970261547e-07,
+      "loss": 0.9492,
+      "step": 75833
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.827781498287191e-07,
+      "loss": 0.5581,
+      "step": 75834
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.82753002950731e-07,
+      "loss": 0.5591,
+      "step": 75835
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.827278563922102e-07,
+      "loss": 0.6304,
+      "step": 75836
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.827027101531751e-07,
+      "loss": 0.6523,
+      "step": 75837
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.826775642336453e-07,
+      "loss": 0.564,
+      "step": 75838
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.826524186336401e-07,
+      "loss": 0.6865,
+      "step": 75839
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.826272733531793e-07,
+      "loss": 0.4863,
+      "step": 75840
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.826021283922811e-07,
+      "loss": 0.6453,
+      "step": 75841
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.825769837509652e-07,
+      "loss": 0.7354,
+      "step": 75842
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.825518394292513e-07,
+      "loss": 0.4583,
+      "step": 75843
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.825266954271584e-07,
+      "loss": 0.5693,
+      "step": 75844
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.825015517447052e-07,
+      "loss": 0.5054,
+      "step": 75845
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.824764083819112e-07,
+      "loss": 0.6885,
+      "step": 75846
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.82451265338796e-07,
+      "loss": 0.8838,
+      "step": 75847
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.824261226153789e-07,
+      "loss": 0.5469,
+      "step": 75848
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.824009802116786e-07,
+      "loss": 0.8281,
+      "step": 75849
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.823758381277146e-07,
+      "loss": 0.5581,
+      "step": 75850
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.823506963635063e-07,
+      "loss": 0.4688,
+      "step": 75851
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.823255549190733e-07,
+      "loss": 0.6147,
+      "step": 75852
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.823004137944344e-07,
+      "loss": 0.7129,
+      "step": 75853
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.822752729896084e-07,
+      "loss": 0.4068,
+      "step": 75854
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.822501325046151e-07,
+      "loss": 0.4951,
+      "step": 75855
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.822249923394735e-07,
+      "loss": 0.5449,
+      "step": 75856
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.821998524942036e-07,
+      "loss": 0.7197,
+      "step": 75857
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.821747129688237e-07,
+      "loss": 0.7422,
+      "step": 75858
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.821495737633532e-07,
+      "loss": 0.3264,
+      "step": 75859
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.82124434877812e-07,
+      "loss": 0.3811,
+      "step": 75860
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.820992963122187e-07,
+      "loss": 0.6392,
+      "step": 75861
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.820741580665925e-07,
+      "loss": 0.5271,
+      "step": 75862
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.820490201409534e-07,
+      "loss": 0.5732,
+      "step": 75863
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.820238825353199e-07,
+      "loss": 0.502,
+      "step": 75864
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.819987452497117e-07,
+      "loss": 0.5811,
+      "step": 75865
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.819736082841476e-07,
+      "loss": 0.5991,
+      "step": 75866
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.819484716386469e-07,
+      "loss": 0.5308,
+      "step": 75867
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.819233353132291e-07,
+      "loss": 0.5,
+      "step": 75868
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.818981993079139e-07,
+      "loss": 0.5591,
+      "step": 75869
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.818730636227195e-07,
+      "loss": 0.4521,
+      "step": 75870
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.818479282576656e-07,
+      "loss": 0.5942,
+      "step": 75871
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.818227932127717e-07,
+      "loss": 0.5747,
+      "step": 75872
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.817976584880571e-07,
+      "loss": 0.8018,
+      "step": 75873
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.817725240835409e-07,
+      "loss": 0.6104,
+      "step": 75874
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.817473899992417e-07,
+      "loss": 0.5762,
+      "step": 75875
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.817222562351796e-07,
+      "loss": 0.5913,
+      "step": 75876
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.816971227913736e-07,
+      "loss": 0.5308,
+      "step": 75877
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.816719896678427e-07,
+      "loss": 0.5464,
+      "step": 75878
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.816468568646062e-07,
+      "loss": 0.6539,
+      "step": 75879
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.816217243816835e-07,
+      "loss": 0.5557,
+      "step": 75880
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.815965922190943e-07,
+      "loss": 0.6367,
+      "step": 75881
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.815714603768569e-07,
+      "loss": 0.7451,
+      "step": 75882
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.815463288549913e-07,
+      "loss": 0.5367,
+      "step": 75883
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.815211976535162e-07,
+      "loss": 0.582,
+      "step": 75884
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.814960667724509e-07,
+      "loss": 0.6138,
+      "step": 75885
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.814709362118153e-07,
+      "loss": 0.6055,
+      "step": 75886
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.814458059716278e-07,
+      "loss": 0.5305,
+      "step": 75887
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.814206760519078e-07,
+      "loss": 0.4893,
+      "step": 75888
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.813955464526749e-07,
+      "loss": 0.7039,
+      "step": 75889
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.813704171739486e-07,
+      "loss": 0.6211,
+      "step": 75890
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.813452882157473e-07,
+      "loss": 0.6523,
+      "step": 75891
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.813201595780908e-07,
+      "loss": 0.4622,
+      "step": 75892
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.812950312609983e-07,
+      "loss": 0.6797,
+      "step": 75893
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.812699032644892e-07,
+      "loss": 0.6367,
+      "step": 75894
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.81244775588582e-07,
+      "loss": 0.501,
+      "step": 75895
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.812196482332964e-07,
+      "loss": 0.6221,
+      "step": 75896
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.811945211986516e-07,
+      "loss": 0.6514,
+      "step": 75897
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.811693944846674e-07,
+      "loss": 0.4612,
+      "step": 75898
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.81144268091362e-07,
+      "loss": 0.7793,
+      "step": 75899
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.811191420187553e-07,
+      "loss": 0.4712,
+      "step": 75900
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.810940162668664e-07,
+      "loss": 0.791,
+      "step": 75901
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.81068890835715e-07,
+      "loss": 0.6943,
+      "step": 75902
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.810437657253199e-07,
+      "loss": 0.5625,
+      "step": 75903
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.810186409356998e-07,
+      "loss": 0.6777,
+      "step": 75904
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.809935164668744e-07,
+      "loss": 0.6748,
+      "step": 75905
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.809683923188633e-07,
+      "loss": 0.6162,
+      "step": 75906
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.809432684916855e-07,
+      "loss": 0.4971,
+      "step": 75907
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.8091814498536e-07,
+      "loss": 0.667,
+      "step": 75908
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.808930217999062e-07,
+      "loss": 0.5571,
+      "step": 75909
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.808678989353432e-07,
+      "loss": 0.8145,
+      "step": 75910
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.808427763916908e-07,
+      "loss": 0.4998,
+      "step": 75911
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.808176541689675e-07,
+      "loss": 0.6182,
+      "step": 75912
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.807925322671932e-07,
+      "loss": 0.707,
+      "step": 75913
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.807674106863863e-07,
+      "loss": 0.833,
+      "step": 75914
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80742289426567e-07,
+      "loss": 0.6699,
+      "step": 75915
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.807171684877537e-07,
+      "loss": 0.3777,
+      "step": 75916
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80692047869966e-07,
+      "loss": 0.4114,
+      "step": 75917
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80666927573223e-07,
+      "loss": 0.7588,
+      "step": 75918
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.806418075975445e-07,
+      "loss": 0.6255,
+      "step": 75919
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.806166879429488e-07,
+      "loss": 0.6655,
+      "step": 75920
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.805915686094558e-07,
+      "loss": 0.5771,
+      "step": 75921
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.805664495970845e-07,
+      "loss": 0.606,
+      "step": 75922
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.805413309058544e-07,
+      "loss": 0.6387,
+      "step": 75923
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.805162125357848e-07,
+      "loss": 0.6807,
+      "step": 75924
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.804910944868941e-07,
+      "loss": 0.793,
+      "step": 75925
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80465976759202e-07,
+      "loss": 0.7451,
+      "step": 75926
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.804408593527279e-07,
+      "loss": 0.5879,
+      "step": 75927
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.804157422674913e-07,
+      "loss": 0.626,
+      "step": 75928
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80390625503511e-07,
+      "loss": 0.667,
+      "step": 75929
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.803655090608059e-07,
+      "loss": 0.5713,
+      "step": 75930
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.803403929393957e-07,
+      "loss": 0.7119,
+      "step": 75931
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.803152771393001e-07,
+      "loss": 0.6328,
+      "step": 75932
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.802901616605375e-07,
+      "loss": 0.6572,
+      "step": 75933
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.802650465031273e-07,
+      "loss": 0.4995,
+      "step": 75934
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.802399316670889e-07,
+      "loss": 0.4304,
+      "step": 75935
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.802148171524417e-07,
+      "loss": 0.6201,
+      "step": 75936
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.801897029592044e-07,
+      "loss": 0.5262,
+      "step": 75937
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.801645890873965e-07,
+      "loss": 0.7842,
+      "step": 75938
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.801394755370373e-07,
+      "loss": 0.5615,
+      "step": 75939
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.801143623081465e-07,
+      "loss": 0.7305,
+      "step": 75940
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.800892494007423e-07,
+      "loss": 0.7339,
+      "step": 75941
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.800641368148444e-07,
+      "loss": 0.6729,
+      "step": 75942
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.800390245504726e-07,
+      "loss": 0.5503,
+      "step": 75943
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.80013912607645e-07,
+      "loss": 0.7568,
+      "step": 75944
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.799888009863822e-07,
+      "loss": 0.6953,
+      "step": 75945
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79963689686702e-07,
+      "loss": 0.4664,
+      "step": 75946
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.799385787086244e-07,
+      "loss": 0.7607,
+      "step": 75947
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.799134680521684e-07,
+      "loss": 0.563,
+      "step": 75948
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.798883577173539e-07,
+      "loss": 0.833,
+      "step": 75949
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.798632477041991e-07,
+      "loss": 0.6641,
+      "step": 75950
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.798381380127236e-07,
+      "loss": 0.7031,
+      "step": 75951
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.798130286429472e-07,
+      "loss": 0.6118,
+      "step": 75952
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.797879195948887e-07,
+      "loss": 0.751,
+      "step": 75953
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.797628108685667e-07,
+      "loss": 0.6228,
+      "step": 75954
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.797377024640012e-07,
+      "loss": 0.7969,
+      "step": 75955
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79712594381211e-07,
+      "loss": 0.6636,
+      "step": 75956
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79687486620216e-07,
+      "loss": 0.6689,
+      "step": 75957
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.796623791810347e-07,
+      "loss": 0.6631,
+      "step": 75958
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.796372720636865e-07,
+      "loss": 0.707,
+      "step": 75959
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.796121652681906e-07,
+      "loss": 0.708,
+      "step": 75960
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79587058794567e-07,
+      "loss": 0.4731,
+      "step": 75961
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.795619526428341e-07,
+      "loss": 0.6143,
+      "step": 75962
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.795368468130109e-07,
+      "loss": 0.5244,
+      "step": 75963
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.795117413051169e-07,
+      "loss": 0.5212,
+      "step": 75964
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.794866361191714e-07,
+      "loss": 0.5513,
+      "step": 75965
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.794615312551942e-07,
+      "loss": 0.71,
+      "step": 75966
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.794364267132037e-07,
+      "loss": 0.6104,
+      "step": 75967
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.794113224932191e-07,
+      "loss": 0.5854,
+      "step": 75968
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.793862185952604e-07,
+      "loss": 0.6387,
+      "step": 75969
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79361115019346e-07,
+      "loss": 0.6235,
+      "step": 75970
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.793360117654955e-07,
+      "loss": 0.5283,
+      "step": 75971
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.793109088337285e-07,
+      "loss": 0.791,
+      "step": 75972
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.792858062240631e-07,
+      "loss": 0.7412,
+      "step": 75973
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.792607039365198e-07,
+      "loss": 0.7002,
+      "step": 75974
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.792356019711169e-07,
+      "loss": 0.5908,
+      "step": 75975
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.792105003278739e-07,
+      "loss": 0.5647,
+      "step": 75976
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.791853990068101e-07,
+      "loss": 0.6191,
+      "step": 75977
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.791602980079449e-07,
+      "loss": 0.7305,
+      "step": 75978
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.791351973312971e-07,
+      "loss": 0.6621,
+      "step": 75979
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.79110096976886e-07,
+      "loss": 0.4565,
+      "step": 75980
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.790849969447311e-07,
+      "loss": 0.583,
+      "step": 75981
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.790598972348518e-07,
+      "loss": 0.6809,
+      "step": 75982
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.790347978472669e-07,
+      "loss": 0.6455,
+      "step": 75983
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.790096987819954e-07,
+      "loss": 0.5952,
+      "step": 75984
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.789846000390568e-07,
+      "loss": 0.6387,
+      "step": 75985
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.789595016184708e-07,
+      "loss": 0.8271,
+      "step": 75986
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.789344035202557e-07,
+      "loss": 0.5781,
+      "step": 75987
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.789093057444311e-07,
+      "loss": 0.7031,
+      "step": 75988
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.788842082910164e-07,
+      "loss": 0.5017,
+      "step": 75989
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78859111160031e-07,
+      "loss": 0.6836,
+      "step": 75990
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.788340143514934e-07,
+      "loss": 0.709,
+      "step": 75991
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.788089178654236e-07,
+      "loss": 0.6465,
+      "step": 75992
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.787838217018401e-07,
+      "loss": 0.6753,
+      "step": 75993
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.787587258607625e-07,
+      "loss": 0.6162,
+      "step": 75994
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.787336303422103e-07,
+      "loss": 0.71,
+      "step": 75995
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78708535146202e-07,
+      "loss": 0.666,
+      "step": 75996
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.786834402727572e-07,
+      "loss": 0.6282,
+      "step": 75997
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.786583457218952e-07,
+      "loss": 0.6074,
+      "step": 75998
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.786332514936355e-07,
+      "loss": 0.5522,
+      "step": 75999
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.786081575879965e-07,
+      "loss": 0.6432,
+      "step": 76000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78583064004998e-07,
+      "loss": 0.5576,
+      "step": 76001
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.785579707446593e-07,
+      "loss": 0.7744,
+      "step": 76002
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.785328778069993e-07,
+      "loss": 0.3103,
+      "step": 76003
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.785077851920371e-07,
+      "loss": 0.6162,
+      "step": 76004
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.784826928997921e-07,
+      "loss": 0.6401,
+      "step": 76005
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.784576009302835e-07,
+      "loss": 0.6416,
+      "step": 76006
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.784325092835309e-07,
+      "loss": 0.5818,
+      "step": 76007
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.784074179595526e-07,
+      "loss": 0.7422,
+      "step": 76008
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.783823269583685e-07,
+      "loss": 0.8398,
+      "step": 76009
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.783572362799978e-07,
+      "loss": 0.7012,
+      "step": 76010
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.783321459244597e-07,
+      "loss": 0.833,
+      "step": 76011
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.783070558917736e-07,
+      "loss": 0.5537,
+      "step": 76012
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.782819661819577e-07,
+      "loss": 0.5835,
+      "step": 76013
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78256876795032e-07,
+      "loss": 0.7754,
+      "step": 76014
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.782317877310156e-07,
+      "loss": 0.6382,
+      "step": 76015
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.782066989899281e-07,
+      "loss": 0.7256,
+      "step": 76016
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78181610571788e-07,
+      "loss": 0.4604,
+      "step": 76017
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.781565224766149e-07,
+      "loss": 0.5903,
+      "step": 76018
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.78131434704428e-07,
+      "loss": 0.562,
+      "step": 76019
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.781063472552468e-07,
+      "loss": 0.4475,
+      "step": 76020
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.780812601290896e-07,
+      "loss": 0.5044,
+      "step": 76021
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.780561733259768e-07,
+      "loss": 0.6292,
+      "step": 76022
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.780310868459265e-07,
+      "loss": 0.6616,
+      "step": 76023
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.780060006889588e-07,
+      "loss": 0.5771,
+      "step": 76024
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.779809148550921e-07,
+      "loss": 0.4622,
+      "step": 76025
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.77955829344346e-07,
+      "loss": 0.54,
+      "step": 76026
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.779307441567397e-07,
+      "loss": 0.4917,
+      "step": 76027
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.779056592922929e-07,
+      "loss": 0.5771,
+      "step": 76028
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.778805747510238e-07,
+      "loss": 0.5732,
+      "step": 76029
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.778554905329522e-07,
+      "loss": 0.4766,
+      "step": 76030
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.778304066380972e-07,
+      "loss": 0.708,
+      "step": 76031
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.778053230664786e-07,
+      "loss": 0.5493,
+      "step": 76032
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.777802398181148e-07,
+      "loss": 0.7202,
+      "step": 76033
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.77755156893025e-07,
+      "loss": 0.4607,
+      "step": 76034
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.777300742912288e-07,
+      "loss": 0.5284,
+      "step": 76035
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.77704992012745e-07,
+      "loss": 0.5508,
+      "step": 76036
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.776799100575936e-07,
+      "loss": 0.7363,
+      "step": 76037
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.776548284257929e-07,
+      "loss": 0.5913,
+      "step": 76038
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.776297471173626e-07,
+      "loss": 0.6455,
+      "step": 76039
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.77604666132322e-07,
+      "loss": 0.5142,
+      "step": 76040
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.775795854706896e-07,
+      "loss": 0.7041,
+      "step": 76041
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.775545051324856e-07,
+      "loss": 0.875,
+      "step": 76042
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.775294251177283e-07,
+      "loss": 0.6465,
+      "step": 76043
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.775043454264373e-07,
+      "loss": 0.603,
+      "step": 76044
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.774792660586322e-07,
+      "loss": 0.5283,
+      "step": 76045
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.774541870143314e-07,
+      "loss": 0.6392,
+      "step": 76046
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.774291082935545e-07,
+      "loss": 0.5762,
+      "step": 76047
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.774040298963206e-07,
+      "loss": 0.687,
+      "step": 76048
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.773789518226495e-07,
+      "loss": 0.5469,
+      "step": 76049
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.773538740725594e-07,
+      "loss": 0.6611,
+      "step": 76050
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.773287966460703e-07,
+      "loss": 0.8037,
+      "step": 76051
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.773037195432011e-07,
+      "loss": 0.606,
+      "step": 76052
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.772786427639709e-07,
+      "loss": 0.4766,
+      "step": 76053
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.772535663083993e-07,
+      "loss": 0.7549,
+      "step": 76054
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.772284901765048e-07,
+      "loss": 0.5293,
+      "step": 76055
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.772034143683071e-07,
+      "loss": 0.5493,
+      "step": 76056
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.771783388838256e-07,
+      "loss": 0.5288,
+      "step": 76057
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.771532637230788e-07,
+      "loss": 0.6143,
+      "step": 76058
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.771281888860864e-07,
+      "loss": 0.4795,
+      "step": 76059
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.771031143728673e-07,
+      "loss": 0.5146,
+      "step": 76060
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.770780401834414e-07,
+      "loss": 0.6221,
+      "step": 76061
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.770529663178275e-07,
+      "loss": 0.6641,
+      "step": 76062
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.770278927760443e-07,
+      "loss": 0.7578,
+      "step": 76063
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.770028195581114e-07,
+      "loss": 0.5522,
+      "step": 76064
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.769777466640478e-07,
+      "loss": 0.5527,
+      "step": 76065
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.769526740938734e-07,
+      "loss": 0.6655,
+      "step": 76066
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.769276018476065e-07,
+      "loss": 0.6279,
+      "step": 76067
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.769025299252667e-07,
+      "loss": 0.4587,
+      "step": 76068
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.768774583268732e-07,
+      "loss": 0.708,
+      "step": 76069
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.768523870524454e-07,
+      "loss": 0.6289,
+      "step": 76070
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.76827316102002e-07,
+      "loss": 0.4746,
+      "step": 76071
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.768022454755629e-07,
+      "loss": 0.5112,
+      "step": 76072
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.767771751731464e-07,
+      "loss": 0.5986,
+      "step": 76073
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.767521051947721e-07,
+      "loss": 0.7373,
+      "step": 76074
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.767270355404599e-07,
+      "loss": 0.5391,
+      "step": 76075
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.767019662102277e-07,
+      "loss": 0.5317,
+      "step": 76076
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.766768972040954e-07,
+      "loss": 0.6067,
+      "step": 76077
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.766518285220825e-07,
+      "loss": 0.6807,
+      "step": 76078
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.766267601642075e-07,
+      "loss": 0.7178,
+      "step": 76079
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.766016921304899e-07,
+      "loss": 0.4617,
+      "step": 76080
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.765766244209489e-07,
+      "loss": 0.5818,
+      "step": 76081
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.765515570356041e-07,
+      "loss": 0.5383,
+      "step": 76082
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.765264899744744e-07,
+      "loss": 0.7178,
+      "step": 76083
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.765014232375784e-07,
+      "loss": 0.6587,
+      "step": 76084
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.764763568249358e-07,
+      "loss": 0.6743,
+      "step": 76085
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.764512907365658e-07,
+      "loss": 0.7744,
+      "step": 76086
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.764262249724881e-07,
+      "loss": 0.5315,
+      "step": 76087
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.764011595327207e-07,
+      "loss": 0.5547,
+      "step": 76088
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.763760944172835e-07,
+      "loss": 0.873,
+      "step": 76089
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.763510296261957e-07,
+      "loss": 0.5581,
+      "step": 76090
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.76325965159477e-07,
+      "loss": 0.5757,
+      "step": 76091
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.763009010171458e-07,
+      "loss": 0.7861,
+      "step": 76092
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.762758371992213e-07,
+      "loss": 0.6978,
+      "step": 76093
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.762507737057227e-07,
+      "loss": 0.52,
+      "step": 76094
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.7622571053667e-07,
+      "loss": 0.5903,
+      "step": 76095
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.762006476920813e-07,
+      "loss": 0.5857,
+      "step": 76096
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.761755851719763e-07,
+      "loss": 0.4719,
+      "step": 76097
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.761505229763741e-07,
+      "loss": 0.7715,
+      "step": 76098
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.761254611052944e-07,
+      "loss": 0.6475,
+      "step": 76099
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.761003995587556e-07,
+      "loss": 0.5776,
+      "step": 76100
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.760753383367771e-07,
+      "loss": 0.6997,
+      "step": 76101
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.760502774393788e-07,
+      "loss": 0.6641,
+      "step": 76102
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.760252168665787e-07,
+      "loss": 0.5449,
+      "step": 76103
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.760001566183972e-07,
+      "loss": 0.5278,
+      "step": 76104
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.759750966948523e-07,
+      "loss": 0.5149,
+      "step": 76105
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.759500370959639e-07,
+      "loss": 0.6753,
+      "step": 76106
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.75924977821751e-07,
+      "loss": 0.5439,
+      "step": 76107
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.758999188722333e-07,
+      "loss": 0.6709,
+      "step": 76108
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.75874860247429e-07,
+      "loss": 0.6543,
+      "step": 76109
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.758498019473579e-07,
+      "loss": 0.5518,
+      "step": 76110
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.758247439720392e-07,
+      "loss": 0.6919,
+      "step": 76111
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.757996863214927e-07,
+      "loss": 0.7402,
+      "step": 76112
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.757746289957362e-07,
+      "loss": 0.583,
+      "step": 76113
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.757495719947894e-07,
+      "loss": 0.6953,
+      "step": 76114
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.757245153186716e-07,
+      "loss": 0.5818,
+      "step": 76115
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.756994589674025e-07,
+      "loss": 0.6973,
+      "step": 76116
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.756744029410004e-07,
+      "loss": 0.6514,
+      "step": 76117
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.75649347239485e-07,
+      "loss": 0.5527,
+      "step": 76118
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.756242918628753e-07,
+      "loss": 0.6611,
+      "step": 76119
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.755992368111911e-07,
+      "loss": 0.5676,
+      "step": 76120
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.755741820844504e-07,
+      "loss": 0.6475,
+      "step": 76121
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.755491276826736e-07,
+      "loss": 0.6355,
+      "step": 76122
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.755240736058789e-07,
+      "loss": 0.7959,
+      "step": 76123
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.754990198540857e-07,
+      "loss": 0.5996,
+      "step": 76124
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.754739664273141e-07,
+      "loss": 0.5728,
+      "step": 76125
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.75448913325582e-07,
+      "loss": 0.6309,
+      "step": 76126
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.754238605489091e-07,
+      "loss": 0.6226,
+      "step": 76127
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.753988080973147e-07,
+      "loss": 0.6792,
+      "step": 76128
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.753737559708182e-07,
+      "loss": 0.5693,
+      "step": 76129
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.753487041694382e-07,
+      "loss": 0.7744,
+      "step": 76130
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.753236526931942e-07,
+      "loss": 0.4854,
+      "step": 76131
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.752986015421058e-07,
+      "loss": 0.6787,
+      "step": 76132
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.752735507161917e-07,
+      "loss": 0.7236,
+      "step": 76133
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.752485002154705e-07,
+      "loss": 0.4492,
+      "step": 76134
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.752234500399622e-07,
+      "loss": 0.6025,
+      "step": 76135
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.751984001896858e-07,
+      "loss": 0.7598,
+      "step": 76136
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.751733506646607e-07,
+      "loss": 0.562,
+      "step": 76137
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.751483014649056e-07,
+      "loss": 0.6963,
+      "step": 76138
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.751232525904398e-07,
+      "loss": 0.6719,
+      "step": 76139
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.750982040412826e-07,
+      "loss": 0.5342,
+      "step": 76140
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.750731558174535e-07,
+      "loss": 0.6758,
+      "step": 76141
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.750481079189715e-07,
+      "loss": 0.6455,
+      "step": 76142
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.750230603458552e-07,
+      "loss": 0.4878,
+      "step": 76143
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.749980130981241e-07,
+      "loss": 0.6313,
+      "step": 76144
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.749729661757976e-07,
+      "loss": 0.5903,
+      "step": 76145
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.749479195788951e-07,
+      "loss": 0.584,
+      "step": 76146
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.749228733074349e-07,
+      "loss": 0.6641,
+      "step": 76147
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.748978273614368e-07,
+      "loss": 0.6484,
+      "step": 76148
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.748727817409205e-07,
+      "loss": 0.7598,
+      "step": 76149
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.748477364459039e-07,
+      "loss": 0.4956,
+      "step": 76150
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.748226914764069e-07,
+      "loss": 0.7065,
+      "step": 76151
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.747976468324491e-07,
+      "loss": 0.6758,
+      "step": 76152
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.747726025140486e-07,
+      "loss": 0.6758,
+      "step": 76153
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.747475585212258e-07,
+      "loss": 0.6416,
+      "step": 76154
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.747225148539987e-07,
+      "loss": 0.7793,
+      "step": 76155
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.74697471512387e-07,
+      "loss": 0.5332,
+      "step": 76156
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.746724284964099e-07,
+      "loss": 0.5193,
+      "step": 76157
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.74647385806087e-07,
+      "loss": 0.6753,
+      "step": 76158
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.746223434414366e-07,
+      "loss": 0.4722,
+      "step": 76159
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.745973014024784e-07,
+      "loss": 0.7178,
+      "step": 76160
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.745722596892315e-07,
+      "loss": 0.6372,
+      "step": 76161
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.745472183017154e-07,
+      "loss": 0.7192,
+      "step": 76162
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.745221772399489e-07,
+      "loss": 0.6602,
+      "step": 76163
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.744971365039508e-07,
+      "loss": 0.7637,
+      "step": 76164
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.744720960937407e-07,
+      "loss": 0.2976,
+      "step": 76165
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.744470560093382e-07,
+      "loss": 0.5718,
+      "step": 76166
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.744220162507615e-07,
+      "loss": 0.7695,
+      "step": 76167
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.743969768180304e-07,
+      "loss": 0.5923,
+      "step": 76168
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.74371937711164e-07,
+      "loss": 0.623,
+      "step": 76169
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.743468989301818e-07,
+      "loss": 0.6133,
+      "step": 76170
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.743218604751021e-07,
+      "loss": 0.5854,
+      "step": 76171
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.742968223459452e-07,
+      "loss": 0.5396,
+      "step": 76172
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.742717845427289e-07,
+      "loss": 0.6348,
+      "step": 76173
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.742467470654735e-07,
+      "loss": 0.7871,
+      "step": 76174
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.742217099141979e-07,
+      "loss": 0.6943,
+      "step": 76175
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.741966730889209e-07,
+      "loss": 0.5601,
+      "step": 76176
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.74171636589662e-07,
+      "loss": 0.6426,
+      "step": 76177
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.741466004164401e-07,
+      "loss": 0.7188,
+      "step": 76178
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.741215645692752e-07,
+      "loss": 0.771,
+      "step": 76179
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.740965290481852e-07,
+      "loss": 0.6396,
+      "step": 76180
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.740714938531901e-07,
+      "loss": 0.564,
+      "step": 76181
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.740464589843091e-07,
+      "loss": 0.6851,
+      "step": 76182
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.740214244415613e-07,
+      "loss": 0.7461,
+      "step": 76183
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.739963902249651e-07,
+      "loss": 0.7002,
+      "step": 76184
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.739713563345405e-07,
+      "loss": 0.6328,
+      "step": 76185
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.739463227703063e-07,
+      "loss": 0.6841,
+      "step": 76186
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.739212895322822e-07,
+      "loss": 0.8096,
+      "step": 76187
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.738962566204868e-07,
+      "loss": 0.7271,
+      "step": 76188
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.738712240349393e-07,
+      "loss": 0.7236,
+      "step": 76189
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.73846191775659e-07,
+      "loss": 0.5615,
+      "step": 76190
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.738211598426654e-07,
+      "loss": 0.6885,
+      "step": 76191
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.737961282359773e-07,
+      "loss": 0.5752,
+      "step": 76192
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.737710969556138e-07,
+      "loss": 0.5923,
+      "step": 76193
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.737460660015938e-07,
+      "loss": 0.6621,
+      "step": 76194
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.737210353739372e-07,
+      "loss": 0.6602,
+      "step": 76195
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.73696005072663e-07,
+      "loss": 0.6836,
+      "step": 76196
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.736709750977897e-07,
+      "loss": 0.7295,
+      "step": 76197
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.736459454493371e-07,
+      "loss": 0.6128,
+      "step": 76198
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.736209161273242e-07,
+      "loss": 0.7056,
+      "step": 76199
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.735958871317703e-07,
+      "loss": 0.6729,
+      "step": 76200
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.735708584626942e-07,
+      "loss": 0.5762,
+      "step": 76201
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.735458301201157e-07,
+      "loss": 0.6216,
+      "step": 76202
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.735208021040532e-07,
+      "loss": 0.593,
+      "step": 76203
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.734957744145266e-07,
+      "loss": 0.5786,
+      "step": 76204
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.734707470515542e-07,
+      "loss": 0.6484,
+      "step": 76205
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.734457200151557e-07,
+      "loss": 0.6963,
+      "step": 76206
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.734206933053501e-07,
+      "loss": 0.7461,
+      "step": 76207
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.733956669221571e-07,
+      "loss": 0.6299,
+      "step": 76208
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.73370640865595e-07,
+      "loss": 0.5834,
+      "step": 76209
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.733456151356835e-07,
+      "loss": 0.6162,
+      "step": 76210
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.733205897324416e-07,
+      "loss": 0.6436,
+      "step": 76211
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.732955646558888e-07,
+      "loss": 0.709,
+      "step": 76212
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.73270539906044e-07,
+      "loss": 0.5908,
+      "step": 76213
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.732455154829261e-07,
+      "loss": 0.6548,
+      "step": 76214
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.732204913865544e-07,
+      "loss": 0.7578,
+      "step": 76215
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.731954676169479e-07,
+      "loss": 0.6338,
+      "step": 76216
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.731704441741267e-07,
+      "loss": 0.6572,
+      "step": 76217
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.731454210581088e-07,
+      "loss": 0.585,
+      "step": 76218
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.731203982689137e-07,
+      "loss": 0.5369,
+      "step": 76219
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.730953758065613e-07,
+      "loss": 0.5271,
+      "step": 76220
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.730703536710695e-07,
+      "loss": 0.7178,
+      "step": 76221
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.730453318624587e-07,
+      "loss": 0.6836,
+      "step": 76222
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.73020310380747e-07,
+      "loss": 0.624,
+      "step": 76223
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.72995289225954e-07,
+      "loss": 0.6267,
+      "step": 76224
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.729702683980992e-07,
+      "loss": 0.4958,
+      "step": 76225
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.729452478972011e-07,
+      "loss": 0.5366,
+      "step": 76226
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.729202277232792e-07,
+      "loss": 0.6787,
+      "step": 76227
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.728952078763526e-07,
+      "loss": 0.6895,
+      "step": 76228
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.728701883564409e-07,
+      "loss": 0.46,
+      "step": 76229
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.728451691635625e-07,
+      "loss": 0.791,
+      "step": 76230
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.728201502977369e-07,
+      "loss": 0.6538,
+      "step": 76231
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.727951317589836e-07,
+      "loss": 0.5542,
+      "step": 76232
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.727701135473211e-07,
+      "loss": 0.6592,
+      "step": 76233
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.727450956627693e-07,
+      "loss": 0.5845,
+      "step": 76234
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.727200781053465e-07,
+      "loss": 0.6851,
+      "step": 76235
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.726950608750723e-07,
+      "loss": 0.6353,
+      "step": 76236
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.726700439719659e-07,
+      "loss": 0.4795,
+      "step": 76237
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.726450273960467e-07,
+      "loss": 0.6553,
+      "step": 76238
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.726200111473331e-07,
+      "loss": 0.7129,
+      "step": 76239
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.725949952258449e-07,
+      "loss": 0.9033,
+      "step": 76240
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.725699796316012e-07,
+      "loss": 0.5867,
+      "step": 76241
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.725449643646212e-07,
+      "loss": 0.4976,
+      "step": 76242
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.725199494249236e-07,
+      "loss": 0.5566,
+      "step": 76243
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.724949348125275e-07,
+      "loss": 0.5703,
+      "step": 76244
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.724699205274526e-07,
+      "loss": 0.7119,
+      "step": 76245
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.724449065697181e-07,
+      "loss": 0.4985,
+      "step": 76246
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.724198929393425e-07,
+      "loss": 0.4556,
+      "step": 76247
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.723948796363453e-07,
+      "loss": 0.7021,
+      "step": 76248
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.723698666607457e-07,
+      "loss": 0.6787,
+      "step": 76249
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.723448540125633e-07,
+      "loss": 0.7012,
+      "step": 76250
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.723198416918168e-07,
+      "loss": 0.4834,
+      "step": 76251
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.722948296985249e-07,
+      "loss": 0.7607,
+      "step": 76252
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.72269818032707e-07,
+      "loss": 0.7354,
+      "step": 76253
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.722448066943827e-07,
+      "loss": 0.707,
+      "step": 76254
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.722197956835712e-07,
+      "loss": 0.624,
+      "step": 76255
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.721947850002909e-07,
+      "loss": 0.3051,
+      "step": 76256
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.721697746445613e-07,
+      "loss": 0.5493,
+      "step": 76257
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.721447646164021e-07,
+      "loss": 0.4824,
+      "step": 76258
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.721197549158317e-07,
+      "loss": 0.708,
+      "step": 76259
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.720947455428693e-07,
+      "loss": 0.7627,
+      "step": 76260
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.720697364975349e-07,
+      "loss": 0.6396,
+      "step": 76261
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.720447277798464e-07,
+      "loss": 0.6328,
+      "step": 76262
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.720197193898242e-07,
+      "loss": 0.4656,
+      "step": 76263
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.719947113274863e-07,
+      "loss": 0.6162,
+      "step": 76264
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.719697035928523e-07,
+      "loss": 0.5586,
+      "step": 76265
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.719446961859415e-07,
+      "loss": 0.541,
+      "step": 76266
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.719196891067734e-07,
+      "loss": 0.514,
+      "step": 76267
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.718946823553663e-07,
+      "loss": 0.6982,
+      "step": 76268
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.718696759317398e-07,
+      "loss": 0.4802,
+      "step": 76269
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.718446698359129e-07,
+      "loss": 0.6543,
+      "step": 76270
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.718196640679054e-07,
+      "loss": 0.5066,
+      "step": 76271
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.717946586277357e-07,
+      "loss": 0.6855,
+      "step": 76272
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.717696535154228e-07,
+      "loss": 0.519,
+      "step": 76273
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.717446487309863e-07,
+      "loss": 0.5811,
+      "step": 76274
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.717196442744454e-07,
+      "loss": 0.5303,
+      "step": 76275
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 5.71694640145819e-07,
+      "loss": 0.8125,
+      "step": 76276
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.716696363451262e-07,
+      "loss": 0.4849,
+      "step": 76277
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.716446328723861e-07,
+      "loss": 0.4553,
+      "step": 76278
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.716196297276185e-07,
+      "loss": 0.6943,
+      "step": 76279
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.715946269108416e-07,
+      "loss": 0.6885,
+      "step": 76280
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.715696244220754e-07,
+      "loss": 0.5811,
+      "step": 76281
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.715446222613384e-07,
+      "loss": 0.4639,
+      "step": 76282
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.715196204286499e-07,
+      "loss": 0.627,
+      "step": 76283
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.714946189240295e-07,
+      "loss": 0.6504,
+      "step": 76284
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.714696177474956e-07,
+      "loss": 0.5247,
+      "step": 76285
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.714446168990677e-07,
+      "loss": 0.665,
+      "step": 76286
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.714196163787648e-07,
+      "loss": 0.5455,
+      "step": 76287
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.713946161866068e-07,
+      "loss": 0.7598,
+      "step": 76288
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.713696163226117e-07,
+      "loss": 0.45,
+      "step": 76289
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.713446167867993e-07,
+      "loss": 0.5854,
+      "step": 76290
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.713196175791889e-07,
+      "loss": 0.7061,
+      "step": 76291
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.712946186997993e-07,
+      "loss": 0.5183,
+      "step": 76292
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.712696201486496e-07,
+      "loss": 0.6543,
+      "step": 76293
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.712446219257587e-07,
+      "loss": 0.5669,
+      "step": 76294
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.712196240311462e-07,
+      "loss": 0.75,
+      "step": 76295
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.711946264648316e-07,
+      "loss": 0.6064,
+      "step": 76296
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.711696292268331e-07,
+      "loss": 0.6953,
+      "step": 76297
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.711446323171703e-07,
+      "loss": 0.6816,
+      "step": 76298
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.711196357358623e-07,
+      "loss": 0.575,
+      "step": 76299
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.710946394829288e-07,
+      "loss": 0.6519,
+      "step": 76300
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.710696435583883e-07,
+      "loss": 0.7383,
+      "step": 76301
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.710446479622597e-07,
+      "loss": 0.6345,
+      "step": 76302
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.710196526945623e-07,
+      "loss": 0.3068,
+      "step": 76303
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.709946577553157e-07,
+      "loss": 0.536,
+      "step": 76304
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.709696631445391e-07,
+      "loss": 0.6631,
+      "step": 76305
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.709446688622508e-07,
+      "loss": 0.6782,
+      "step": 76306
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.709196749084705e-07,
+      "loss": 0.5273,
+      "step": 76307
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.708946812832172e-07,
+      "loss": 0.6582,
+      "step": 76308
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.708696879865107e-07,
+      "loss": 0.7144,
+      "step": 76309
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.70844695018369e-07,
+      "loss": 0.5322,
+      "step": 76310
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.708197023788122e-07,
+      "loss": 0.7061,
+      "step": 76311
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.707947100678585e-07,
+      "loss": 0.6738,
+      "step": 76312
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.707697180855283e-07,
+      "loss": 0.6582,
+      "step": 76313
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.707447264318392e-07,
+      "loss": 0.5322,
+      "step": 76314
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.707197351068114e-07,
+      "loss": 0.4272,
+      "step": 76315
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.706947441104636e-07,
+      "loss": 0.6274,
+      "step": 76316
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.706697534428156e-07,
+      "loss": 0.6123,
+      "step": 76317
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.706447631038856e-07,
+      "loss": 0.6289,
+      "step": 76318
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.70619773093693e-07,
+      "loss": 0.6743,
+      "step": 76319
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.705947834122573e-07,
+      "loss": 0.707,
+      "step": 76320
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.705697940595978e-07,
+      "loss": 0.6172,
+      "step": 76321
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.705448050357333e-07,
+      "loss": 0.4956,
+      "step": 76322
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.705198163406823e-07,
+      "loss": 0.5723,
+      "step": 76323
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.704948279744648e-07,
+      "loss": 0.5889,
+      "step": 76324
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.704698399370995e-07,
+      "loss": 0.7798,
+      "step": 76325
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.704448522286061e-07,
+      "loss": 0.4175,
+      "step": 76326
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.704198648490029e-07,
+      "loss": 0.7588,
+      "step": 76327
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.703948777983095e-07,
+      "loss": 0.6611,
+      "step": 76328
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.703698910765453e-07,
+      "loss": 0.4756,
+      "step": 76329
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.703449046837288e-07,
+      "loss": 0.7891,
+      "step": 76330
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.703199186198799e-07,
+      "loss": 0.6123,
+      "step": 76331
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.702949328850169e-07,
+      "loss": 0.5437,
+      "step": 76332
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.702699474791592e-07,
+      "loss": 0.5449,
+      "step": 76333
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.702449624023265e-07,
+      "loss": 0.7393,
+      "step": 76334
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.702199776545371e-07,
+      "loss": 0.5288,
+      "step": 76335
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.701949932358105e-07,
+      "loss": 0.4708,
+      "step": 76336
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.701700091461657e-07,
+      "loss": 0.6641,
+      "step": 76337
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.701450253856225e-07,
+      "loss": 0.6562,
+      "step": 76338
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.701200419541989e-07,
+      "loss": 0.5479,
+      "step": 76339
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.700950588519148e-07,
+      "loss": 0.7354,
+      "step": 76340
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.700700760787896e-07,
+      "loss": 0.6621,
+      "step": 76341
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.700450936348414e-07,
+      "loss": 0.5952,
+      "step": 76342
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.700201115200905e-07,
+      "loss": 0.4946,
+      "step": 76343
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.699951297345549e-07,
+      "loss": 0.7422,
+      "step": 76344
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.699701482782542e-07,
+      "loss": 0.665,
+      "step": 76345
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.699451671512081e-07,
+      "loss": 0.7842,
+      "step": 76346
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.699201863534348e-07,
+      "loss": 0.6045,
+      "step": 76347
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.698952058849538e-07,
+      "loss": 0.6475,
+      "step": 76348
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.698702257457842e-07,
+      "loss": 0.6714,
+      "step": 76349
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.698452459359456e-07,
+      "loss": 0.5635,
+      "step": 76350
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.698202664554568e-07,
+      "loss": 0.751,
+      "step": 76351
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.697952873043365e-07,
+      "loss": 0.605,
+      "step": 76352
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69770308482604e-07,
+      "loss": 0.3298,
+      "step": 76353
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.697453299902787e-07,
+      "loss": 0.6543,
+      "step": 76354
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.6972035182738e-07,
+      "loss": 0.8066,
+      "step": 76355
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.696953739939262e-07,
+      "loss": 0.7539,
+      "step": 76356
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69670396489937e-07,
+      "loss": 0.606,
+      "step": 76357
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.696454193154312e-07,
+      "loss": 0.6724,
+      "step": 76358
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.696204424704285e-07,
+      "loss": 0.5645,
+      "step": 76359
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.695954659549475e-07,
+      "loss": 0.6235,
+      "step": 76360
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.695704897690077e-07,
+      "loss": 0.7031,
+      "step": 76361
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.695455139126275e-07,
+      "loss": 0.6289,
+      "step": 76362
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.695205383858271e-07,
+      "loss": 0.7539,
+      "step": 76363
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.694955631886245e-07,
+      "loss": 0.6777,
+      "step": 76364
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.694705883210393e-07,
+      "loss": 0.7803,
+      "step": 76365
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.694456137830907e-07,
+      "loss": 0.4316,
+      "step": 76366
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.694206395747983e-07,
+      "loss": 0.7349,
+      "step": 76367
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.693956656961802e-07,
+      "loss": 0.7939,
+      "step": 76368
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69370692147256e-07,
+      "loss": 0.6855,
+      "step": 76369
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69345718928045e-07,
+      "loss": 0.6499,
+      "step": 76370
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.693207460385668e-07,
+      "loss": 0.6904,
+      "step": 76371
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.692957734788395e-07,
+      "loss": 0.7461,
+      "step": 76372
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.692708012488824e-07,
+      "loss": 0.4565,
+      "step": 76373
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.692458293487148e-07,
+      "loss": 0.5996,
+      "step": 76374
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.692208577783558e-07,
+      "loss": 0.5718,
+      "step": 76375
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69195886537825e-07,
+      "loss": 0.6035,
+      "step": 76376
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.691709156271409e-07,
+      "loss": 0.6167,
+      "step": 76377
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.691459450463226e-07,
+      "loss": 0.6143,
+      "step": 76378
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.691209747953896e-07,
+      "loss": 0.6313,
+      "step": 76379
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.690960048743612e-07,
+      "loss": 0.55,
+      "step": 76380
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.69071035283256e-07,
+      "loss": 0.6758,
+      "step": 76381
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.690460660220931e-07,
+      "loss": 0.5581,
+      "step": 76382
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.690210970908918e-07,
+      "loss": 0.4736,
+      "step": 76383
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.689961284896715e-07,
+      "loss": 0.7383,
+      "step": 76384
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.689711602184507e-07,
+      "loss": 0.666,
+      "step": 76385
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.689461922772489e-07,
+      "loss": 0.4299,
+      "step": 76386
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.689212246660852e-07,
+      "loss": 0.6533,
+      "step": 76387
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.688962573849791e-07,
+      "loss": 0.6572,
+      "step": 76388
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.688712904339487e-07,
+      "loss": 0.5532,
+      "step": 76389
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.688463238130141e-07,
+      "loss": 0.6167,
+      "step": 76390
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.688213575221942e-07,
+      "loss": 0.5942,
+      "step": 76391
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.687963915615076e-07,
+      "loss": 0.5898,
+      "step": 76392
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.687714259309742e-07,
+      "loss": 0.5879,
+      "step": 76393
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.687464606306122e-07,
+      "loss": 0.5557,
+      "step": 76394
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.687214956604413e-07,
+      "loss": 0.6787,
+      "step": 76395
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.686965310204806e-07,
+      "loss": 0.6855,
+      "step": 76396
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.686715667107493e-07,
+      "loss": 0.5105,
+      "step": 76397
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.686466027312662e-07,
+      "loss": 0.6812,
+      "step": 76398
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.686216390820504e-07,
+      "loss": 0.5654,
+      "step": 76399
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.685966757631213e-07,
+      "loss": 0.6138,
+      "step": 76400
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.685717127744986e-07,
+      "loss": 0.5986,
+      "step": 76401
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.685467501161998e-07,
+      "loss": 0.5947,
+      "step": 76402
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.685217877882451e-07,
+      "loss": 0.6592,
+      "step": 76403
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.684968257906532e-07,
+      "loss": 0.6992,
+      "step": 76404
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.684718641234442e-07,
+      "loss": 0.8682,
+      "step": 76405
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.684469027866359e-07,
+      "loss": 0.5171,
+      "step": 76406
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.684219417802478e-07,
+      "loss": 0.7822,
+      "step": 76407
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.683969811042993e-07,
+      "loss": 0.5869,
+      "step": 76408
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.683720207588098e-07,
+      "loss": 0.4561,
+      "step": 76409
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.683470607437975e-07,
+      "loss": 0.5913,
+      "step": 76410
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.683221010592825e-07,
+      "loss": 0.5288,
+      "step": 76411
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.68297141705283e-07,
+      "loss": 0.5007,
+      "step": 76412
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.682721826818186e-07,
+      "loss": 0.6074,
+      "step": 76413
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.682472239889085e-07,
+      "loss": 0.6748,
+      "step": 76414
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.682222656265714e-07,
+      "loss": 0.5879,
+      "step": 76415
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.681973075948268e-07,
+      "loss": 0.5435,
+      "step": 76416
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.681723498936934e-07,
+      "loss": 0.7432,
+      "step": 76417
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.681473925231913e-07,
+      "loss": 0.5645,
+      "step": 76418
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.681224354833383e-07,
+      "loss": 0.6392,
+      "step": 76419
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.680974787741539e-07,
+      "loss": 0.6611,
+      "step": 76420
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.680725223956581e-07,
+      "loss": 0.6963,
+      "step": 76421
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.680475663478692e-07,
+      "loss": 0.645,
+      "step": 76422
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.680226106308059e-07,
+      "loss": 0.7812,
+      "step": 76423
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.679976552444881e-07,
+      "loss": 0.6348,
+      "step": 76424
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.679727001889344e-07,
+      "loss": 0.4902,
+      "step": 76425
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.679477454641646e-07,
+      "loss": 0.4763,
+      "step": 76426
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.679227910701969e-07,
+      "loss": 0.7529,
+      "step": 76427
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67897837007051e-07,
+      "loss": 0.5468,
+      "step": 76428
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.678728832747457e-07,
+      "loss": 0.7949,
+      "step": 76429
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.678479298733008e-07,
+      "loss": 0.7231,
+      "step": 76430
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.678229768027347e-07,
+      "loss": 0.7256,
+      "step": 76431
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.677980240630664e-07,
+      "loss": 0.502,
+      "step": 76432
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.677730716543152e-07,
+      "loss": 0.501,
+      "step": 76433
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.677481195765004e-07,
+      "loss": 0.7207,
+      "step": 76434
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.677231678296412e-07,
+      "loss": 0.5815,
+      "step": 76435
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.676982164137563e-07,
+      "loss": 0.5083,
+      "step": 76436
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67673265328865e-07,
+      "loss": 0.6133,
+      "step": 76437
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.676483145749866e-07,
+      "loss": 0.5962,
+      "step": 76438
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.676233641521397e-07,
+      "loss": 0.2526,
+      "step": 76439
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.675984140603438e-07,
+      "loss": 0.6582,
+      "step": 76440
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.675734642996183e-07,
+      "loss": 0.5905,
+      "step": 76441
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.675485148699815e-07,
+      "loss": 0.7065,
+      "step": 76442
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.675235657714533e-07,
+      "loss": 0.666,
+      "step": 76443
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.674986170040519e-07,
+      "loss": 0.6875,
+      "step": 76444
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.674736685677969e-07,
+      "loss": 0.5938,
+      "step": 76445
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.674487204627076e-07,
+      "loss": 0.4949,
+      "step": 76446
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.674237726888034e-07,
+      "loss": 0.5356,
+      "step": 76447
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.673988252461023e-07,
+      "loss": 0.5752,
+      "step": 76448
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.673738781346241e-07,
+      "loss": 0.7578,
+      "step": 76449
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.673489313543879e-07,
+      "loss": 0.6738,
+      "step": 76450
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67323984905413e-07,
+      "loss": 0.4731,
+      "step": 76451
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.672990387877184e-07,
+      "loss": 0.5957,
+      "step": 76452
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.672740930013224e-07,
+      "loss": 0.8057,
+      "step": 76453
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67249147546245e-07,
+      "loss": 0.5112,
+      "step": 76454
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.672242024225053e-07,
+      "loss": 0.6577,
+      "step": 76455
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.671992576301217e-07,
+      "loss": 0.7451,
+      "step": 76456
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.671743131691138e-07,
+      "loss": 0.7373,
+      "step": 76457
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.671493690395005e-07,
+      "loss": 0.5938,
+      "step": 76458
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.671244252413016e-07,
+      "loss": 0.8135,
+      "step": 76459
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67099481774535e-07,
+      "loss": 0.4883,
+      "step": 76460
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.67074538639221e-07,
+      "loss": 0.7031,
+      "step": 76461
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.670495958353777e-07,
+      "loss": 0.625,
+      "step": 76462
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.670246533630246e-07,
+      "loss": 0.5034,
+      "step": 76463
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.669997112221811e-07,
+      "loss": 0.6104,
+      "step": 76464
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.669747694128657e-07,
+      "loss": 0.6392,
+      "step": 76465
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.669498279350979e-07,
+      "loss": 0.5122,
+      "step": 76466
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.669248867888966e-07,
+      "loss": 0.7529,
+      "step": 76467
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.668999459742815e-07,
+      "loss": 0.6592,
+      "step": 76468
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.668750054912706e-07,
+      "loss": 0.4417,
+      "step": 76469
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.668500653398838e-07,
+      "loss": 0.6294,
+      "step": 76470
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.668251255201403e-07,
+      "loss": 0.5957,
+      "step": 76471
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.668001860320588e-07,
+      "loss": 0.7012,
+      "step": 76472
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.667752468756582e-07,
+      "loss": 0.7676,
+      "step": 76473
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.667503080509579e-07,
+      "loss": 0.4302,
+      "step": 76474
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.66725369557977e-07,
+      "loss": 0.6079,
+      "step": 76475
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.667004313967348e-07,
+      "loss": 0.6416,
+      "step": 76476
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.666754935672498e-07,
+      "loss": 0.7295,
+      "step": 76477
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.666505560695415e-07,
+      "loss": 0.5419,
+      "step": 76478
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.666256189036289e-07,
+      "loss": 0.6572,
+      "step": 76479
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.666006820695315e-07,
+      "loss": 0.5571,
+      "step": 76480
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.665757455672681e-07,
+      "loss": 0.6787,
+      "step": 76481
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.665508093968574e-07,
+      "loss": 0.5806,
+      "step": 76482
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.665258735583187e-07,
+      "loss": 0.6743,
+      "step": 76483
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.665009380516712e-07,
+      "loss": 0.7588,
+      "step": 76484
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.664760028769345e-07,
+      "loss": 0.6094,
+      "step": 76485
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.664510680341266e-07,
+      "loss": 0.563,
+      "step": 76486
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.664261335232674e-07,
+      "loss": 0.6143,
+      "step": 76487
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.664011993443755e-07,
+      "loss": 0.5957,
+      "step": 76488
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.663762654974709e-07,
+      "loss": 0.7031,
+      "step": 76489
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.663513319825717e-07,
+      "loss": 0.6787,
+      "step": 76490
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.663263987996976e-07,
+      "loss": 0.5237,
+      "step": 76491
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.66301465948867e-07,
+      "loss": 0.6162,
+      "step": 76492
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.662765334300999e-07,
+      "loss": 0.5503,
+      "step": 76493
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.662516012434146e-07,
+      "loss": 0.4863,
+      "step": 76494
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.662266693888303e-07,
+      "loss": 0.6958,
+      "step": 76495
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.662017378663665e-07,
+      "loss": 0.5942,
+      "step": 76496
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.661768066760424e-07,
+      "loss": 0.5059,
+      "step": 76497
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.661518758178763e-07,
+      "loss": 0.5947,
+      "step": 76498
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.661269452918879e-07,
+      "loss": 0.79,
+      "step": 76499
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.661020150980961e-07,
+      "loss": 0.6445,
+      "step": 76500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.660770852365205e-07,
+      "loss": 0.5664,
+      "step": 76501
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.660521557071796e-07,
+      "loss": 0.4646,
+      "step": 76502
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.660272265100922e-07,
+      "loss": 0.5557,
+      "step": 76503
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.660022976452779e-07,
+      "loss": 0.3943,
+      "step": 76504
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.659773691127555e-07,
+      "loss": 0.707,
+      "step": 76505
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.659524409125448e-07,
+      "loss": 0.5864,
+      "step": 76506
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65927513044664e-07,
+      "loss": 0.5276,
+      "step": 76507
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.659025855091324e-07,
+      "loss": 0.6211,
+      "step": 76508
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.658776583059698e-07,
+      "loss": 0.7666,
+      "step": 76509
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.658527314351942e-07,
+      "loss": 0.5854,
+      "step": 76510
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.658278048968257e-07,
+      "loss": 0.5098,
+      "step": 76511
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.658028786908824e-07,
+      "loss": 0.6499,
+      "step": 76512
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.657779528173839e-07,
+      "loss": 0.6636,
+      "step": 76513
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.657530272763498e-07,
+      "loss": 0.6113,
+      "step": 76514
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65728102067798e-07,
+      "loss": 0.5908,
+      "step": 76515
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.657031771917483e-07,
+      "loss": 0.5328,
+      "step": 76516
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.656782526482197e-07,
+      "loss": 0.5796,
+      "step": 76517
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.656533284372318e-07,
+      "loss": 0.7617,
+      "step": 76518
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.656284045588027e-07,
+      "loss": 0.5605,
+      "step": 76519
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.656034810129519e-07,
+      "loss": 0.5586,
+      "step": 76520
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65578557799699e-07,
+      "loss": 0.7354,
+      "step": 76521
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.655536349190622e-07,
+      "loss": 0.5752,
+      "step": 76522
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.655287123710615e-07,
+      "loss": 0.8037,
+      "step": 76523
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65503790155715e-07,
+      "loss": 0.6885,
+      "step": 76524
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.654788682730423e-07,
+      "loss": 0.6875,
+      "step": 76525
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.654539467230628e-07,
+      "loss": 0.5112,
+      "step": 76526
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.654290255057949e-07,
+      "loss": 0.5498,
+      "step": 76527
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.654041046212579e-07,
+      "loss": 0.6299,
+      "step": 76528
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65379184069471e-07,
+      "loss": 0.709,
+      "step": 76529
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.653542638504538e-07,
+      "loss": 0.5508,
+      "step": 76530
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.653293439642249e-07,
+      "loss": 0.6221,
+      "step": 76531
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.653044244108027e-07,
+      "loss": 0.5698,
+      "step": 76532
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.652795051902071e-07,
+      "loss": 0.6621,
+      "step": 76533
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.65254586302457e-07,
+      "loss": 0.708,
+      "step": 76534
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.652296677475718e-07,
+      "loss": 0.5479,
+      "step": 76535
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.652047495255698e-07,
+      "loss": 0.6948,
+      "step": 76536
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.651798316364706e-07,
+      "loss": 0.6885,
+      "step": 76537
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.651549140802933e-07,
+      "loss": 0.7969,
+      "step": 76538
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.651299968570571e-07,
+      "loss": 0.8223,
+      "step": 76539
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.651050799667807e-07,
+      "loss": 0.4761,
+      "step": 76540
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.650801634094835e-07,
+      "loss": 0.6484,
+      "step": 76541
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.650552471851841e-07,
+      "loss": 0.5181,
+      "step": 76542
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.650303312939024e-07,
+      "loss": 0.4188,
+      "step": 76543
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.650054157356565e-07,
+      "loss": 0.6353,
+      "step": 76544
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64980500510466e-07,
+      "loss": 0.5269,
+      "step": 76545
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.649555856183499e-07,
+      "loss": 0.6221,
+      "step": 76546
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.649306710593275e-07,
+      "loss": 0.7119,
+      "step": 76547
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.649057568334175e-07,
+      "loss": 0.624,
+      "step": 76548
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.648808429406391e-07,
+      "loss": 0.5654,
+      "step": 76549
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.648559293810119e-07,
+      "loss": 0.4025,
+      "step": 76550
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.648310161545541e-07,
+      "loss": 0.7305,
+      "step": 76551
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.648061032612855e-07,
+      "loss": 0.67,
+      "step": 76552
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.647811907012245e-07,
+      "loss": 0.5811,
+      "step": 76553
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.647562784743904e-07,
+      "loss": 0.7998,
+      "step": 76554
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.647313665808027e-07,
+      "loss": 0.6841,
+      "step": 76555
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.647064550204803e-07,
+      "loss": 0.5918,
+      "step": 76556
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64681543793442e-07,
+      "loss": 0.6709,
+      "step": 76557
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.646566328997067e-07,
+      "loss": 0.8232,
+      "step": 76558
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64631722339294e-07,
+      "loss": 0.6533,
+      "step": 76559
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.646068121122236e-07,
+      "loss": 0.813,
+      "step": 76560
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64581902218513e-07,
+      "loss": 0.5967,
+      "step": 76561
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.645569926581819e-07,
+      "loss": 0.6929,
+      "step": 76562
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.645320834312494e-07,
+      "loss": 0.6631,
+      "step": 76563
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.645071745377352e-07,
+      "loss": 0.6025,
+      "step": 76564
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.644822659776574e-07,
+      "loss": 0.5103,
+      "step": 76565
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.644573577510355e-07,
+      "loss": 0.4298,
+      "step": 76566
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.644324498578885e-07,
+      "loss": 0.5413,
+      "step": 76567
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.644075422982362e-07,
+      "loss": 0.4941,
+      "step": 76568
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.643826350720964e-07,
+      "loss": 0.2236,
+      "step": 76569
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.643577281794892e-07,
+      "loss": 0.6299,
+      "step": 76570
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.643328216204329e-07,
+      "loss": 0.6099,
+      "step": 76571
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.643079153949469e-07,
+      "loss": 0.6074,
+      "step": 76572
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.642830095030507e-07,
+      "loss": 0.6479,
+      "step": 76573
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.642581039447624e-07,
+      "loss": 0.5516,
+      "step": 76574
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.642331987201019e-07,
+      "loss": 0.5977,
+      "step": 76575
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64208293829088e-07,
+      "loss": 0.5703,
+      "step": 76576
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.641833892717401e-07,
+      "loss": 0.5103,
+      "step": 76577
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.641584850480764e-07,
+      "loss": 0.688,
+      "step": 76578
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.641335811581166e-07,
+      "loss": 0.6426,
+      "step": 76579
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.641086776018801e-07,
+      "loss": 0.5234,
+      "step": 76580
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.640837743793856e-07,
+      "loss": 0.7178,
+      "step": 76581
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.640588714906516e-07,
+      "loss": 0.7363,
+      "step": 76582
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.640339689356977e-07,
+      "loss": 0.5449,
+      "step": 76583
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.64009066714543e-07,
+      "loss": 0.5854,
+      "step": 76584
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.639841648272068e-07,
+      "loss": 0.4961,
+      "step": 76585
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.639592632737075e-07,
+      "loss": 0.7061,
+      "step": 76586
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.639343620540647e-07,
+      "loss": 0.5356,
+      "step": 76587
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.639094611682972e-07,
+      "loss": 0.5967,
+      "step": 76588
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.638845606164245e-07,
+      "loss": 0.5293,
+      "step": 76589
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.638596603984656e-07,
+      "loss": 0.5781,
+      "step": 76590
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.638347605144386e-07,
+      "loss": 0.5591,
+      "step": 76591
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.638098609643633e-07,
+      "loss": 0.6406,
+      "step": 76592
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.637849617482588e-07,
+      "loss": 0.4321,
+      "step": 76593
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.637600628661446e-07,
+      "loss": 0.5894,
+      "step": 76594
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.637351643180388e-07,
+      "loss": 0.4649,
+      "step": 76595
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.637102661039609e-07,
+      "loss": 0.4824,
+      "step": 76596
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.6368536822393e-07,
+      "loss": 0.584,
+      "step": 76597
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.636604706779656e-07,
+      "loss": 0.605,
+      "step": 76598
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.636355734660858e-07,
+      "loss": 0.6377,
+      "step": 76599
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.636106765883106e-07,
+      "loss": 0.4424,
+      "step": 76600
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.635857800446583e-07,
+      "loss": 0.7197,
+      "step": 76601
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.635608838351486e-07,
+      "loss": 0.75,
+      "step": 76602
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.635359879598001e-07,
+      "loss": 0.4219,
+      "step": 76603
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.635110924186317e-07,
+      "loss": 0.7891,
+      "step": 76604
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.63486197211663e-07,
+      "loss": 0.7788,
+      "step": 76605
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.634613023389132e-07,
+      "loss": 0.5713,
+      "step": 76606
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.634364078004006e-07,
+      "loss": 0.6897,
+      "step": 76607
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.634115135961447e-07,
+      "loss": 0.4546,
+      "step": 76608
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.633866197261646e-07,
+      "loss": 0.5361,
+      "step": 76609
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.633617261904795e-07,
+      "loss": 0.6631,
+      "step": 76610
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.633368329891084e-07,
+      "loss": 0.605,
+      "step": 76611
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.633119401220696e-07,
+      "loss": 0.6982,
+      "step": 76612
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.63287047589383e-07,
+      "loss": 0.6958,
+      "step": 76613
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.632621553910674e-07,
+      "loss": 0.6035,
+      "step": 76614
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.632372635271422e-07,
+      "loss": 0.6738,
+      "step": 76615
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.632123719976258e-07,
+      "loss": 0.521,
+      "step": 76616
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.631874808025375e-07,
+      "loss": 0.7583,
+      "step": 76617
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.63162589941897e-07,
+      "loss": 0.6729,
+      "step": 76618
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.631376994157223e-07,
+      "loss": 0.7402,
+      "step": 76619
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.631128092240335e-07,
+      "loss": 0.509,
+      "step": 76620
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.630879193668487e-07,
+      "loss": 0.7329,
+      "step": 76621
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.630630298441873e-07,
+      "loss": 0.5579,
+      "step": 76622
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.63038140656069e-07,
+      "loss": 0.5864,
+      "step": 76623
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.630132518025117e-07,
+      "loss": 0.5452,
+      "step": 76624
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.629883632835352e-07,
+      "loss": 0.5737,
+      "step": 76625
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.629634750991583e-07,
+      "loss": 0.6685,
+      "step": 76626
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.629385872494006e-07,
+      "loss": 0.5991,
+      "step": 76627
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.629136997342804e-07,
+      "loss": 0.5991,
+      "step": 76628
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.628888125538172e-07,
+      "loss": 0.7627,
+      "step": 76629
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.628639257080301e-07,
+      "loss": 0.6768,
+      "step": 76630
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.628390391969378e-07,
+      "loss": 0.5522,
+      "step": 76631
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.628141530205597e-07,
+      "loss": 0.875,
+      "step": 76632
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.627892671789145e-07,
+      "loss": 0.4773,
+      "step": 76633
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.627643816720214e-07,
+      "loss": 0.6504,
+      "step": 76634
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.627394964998999e-07,
+      "loss": 0.7256,
+      "step": 76635
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.627146116625683e-07,
+      "loss": 0.7041,
+      "step": 76636
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.62689727160046e-07,
+      "loss": 0.6089,
+      "step": 76637
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.626648429923519e-07,
+      "loss": 0.75,
+      "step": 76638
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.626399591595059e-07,
+      "loss": 0.6621,
+      "step": 76639
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.626150756615262e-07,
+      "loss": 0.5215,
+      "step": 76640
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.625901924984318e-07,
+      "loss": 0.5061,
+      "step": 76641
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.625653096702418e-07,
+      "loss": 0.5487,
+      "step": 76642
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.625404271769754e-07,
+      "loss": 0.625,
+      "step": 76643
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.625155450186522e-07,
+      "loss": 0.501,
+      "step": 76644
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.624906631952902e-07,
+      "loss": 0.5605,
+      "step": 76645
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.624657817069092e-07,
+      "loss": 0.6611,
+      "step": 76646
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.624409005535278e-07,
+      "loss": 0.6475,
+      "step": 76647
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.624160197351657e-07,
+      "loss": 0.5947,
+      "step": 76648
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.623911392518414e-07,
+      "loss": 0.7588,
+      "step": 76649
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.623662591035742e-07,
+      "loss": 0.6357,
+      "step": 76650
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.623413792903828e-07,
+      "loss": 0.7339,
+      "step": 76651
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.623164998122867e-07,
+      "loss": 0.7046,
+      "step": 76652
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.622916206693045e-07,
+      "loss": 0.5225,
+      "step": 76653
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.622667418614554e-07,
+      "loss": 0.7402,
+      "step": 76654
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.622418633887586e-07,
+      "loss": 0.5791,
+      "step": 76655
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.622169852512334e-07,
+      "loss": 0.7783,
+      "step": 76656
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.621921074488982e-07,
+      "loss": 0.7183,
+      "step": 76657
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.621672299817724e-07,
+      "loss": 0.7188,
+      "step": 76658
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.621423528498747e-07,
+      "loss": 0.5728,
+      "step": 76659
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.621174760532253e-07,
+      "loss": 0.7227,
+      "step": 76660
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.620925995918422e-07,
+      "loss": 0.8545,
+      "step": 76661
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.620677234657442e-07,
+      "loss": 0.5615,
+      "step": 76662
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.620428476749509e-07,
+      "loss": 0.665,
+      "step": 76663
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.620179722194813e-07,
+      "loss": 0.7441,
+      "step": 76664
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.619930970993548e-07,
+      "loss": 0.5923,
+      "step": 76665
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 5.619682223145895e-07,
+      "loss": 0.7979,
+      "step": 76666
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.619433478652052e-07,
+      "loss": 0.4624,
+      "step": 76667
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.619184737512206e-07,
+      "loss": 0.4714,
+      "step": 76668
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.618935999726553e-07,
+      "loss": 0.3795,
+      "step": 76669
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.618687265295279e-07,
+      "loss": 0.7041,
+      "step": 76670
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.61843853421857e-07,
+      "loss": 0.6152,
+      "step": 76671
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.618189806496623e-07,
+      "loss": 0.5244,
+      "step": 76672
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.617941082129629e-07,
+      "loss": 0.543,
+      "step": 76673
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.617692361117773e-07,
+      "loss": 0.6973,
+      "step": 76674
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.617443643461248e-07,
+      "loss": 0.4561,
+      "step": 76675
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.617194929160245e-07,
+      "loss": 0.563,
+      "step": 76676
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.616946218214959e-07,
+      "loss": 0.686,
+      "step": 76677
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.616697510625571e-07,
+      "loss": 0.6768,
+      "step": 76678
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.616448806392275e-07,
+      "loss": 0.5337,
+      "step": 76679
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.61620010551527e-07,
+      "loss": 0.344,
+      "step": 76680
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.615951407994732e-07,
+      "loss": 0.6826,
+      "step": 76681
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.615702713830863e-07,
+      "loss": 0.5981,
+      "step": 76682
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.615454023023844e-07,
+      "loss": 0.8037,
+      "step": 76683
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.615205335573873e-07,
+      "loss": 0.71,
+      "step": 76684
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.614956651481135e-07,
+      "loss": 0.5957,
+      "step": 76685
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.614707970745829e-07,
+      "loss": 0.5659,
+      "step": 76686
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.614459293368134e-07,
+      "loss": 0.7852,
+      "step": 76687
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.614210619348247e-07,
+      "loss": 0.6641,
+      "step": 76688
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.613961948686361e-07,
+      "loss": 0.709,
+      "step": 76689
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.613713281382662e-07,
+      "loss": 0.5874,
+      "step": 76690
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.613464617437338e-07,
+      "loss": 0.7002,
+      "step": 76691
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.613215956850581e-07,
+      "loss": 0.4722,
+      "step": 76692
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.612967299622584e-07,
+      "loss": 0.5635,
+      "step": 76693
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.61271864575354e-07,
+      "loss": 0.6499,
+      "step": 76694
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.612469995243632e-07,
+      "loss": 0.5859,
+      "step": 76695
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.612221348093053e-07,
+      "loss": 0.647,
+      "step": 76696
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.611972704301995e-07,
+      "loss": 0.707,
+      "step": 76697
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.611724063870651e-07,
+      "loss": 0.7197,
+      "step": 76698
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.611475426799204e-07,
+      "loss": 0.3895,
+      "step": 76699
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.611226793087855e-07,
+      "loss": 0.5698,
+      "step": 76700
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.610978162736782e-07,
+      "loss": 0.6816,
+      "step": 76701
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.61072953574618e-07,
+      "loss": 0.5552,
+      "step": 76702
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.610480912116247e-07,
+      "loss": 0.477,
+      "step": 76703
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.610232291847161e-07,
+      "loss": 0.8164,
+      "step": 76704
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.609983674939118e-07,
+      "loss": 0.5835,
+      "step": 76705
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.609735061392314e-07,
+      "loss": 0.5503,
+      "step": 76706
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60948645120693e-07,
+      "loss": 0.4946,
+      "step": 76707
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60923784438316e-07,
+      "loss": 0.8047,
+      "step": 76708
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.608989240921195e-07,
+      "loss": 0.521,
+      "step": 76709
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.608740640821227e-07,
+      "loss": 0.4736,
+      "step": 76710
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.608492044083448e-07,
+      "loss": 0.7822,
+      "step": 76711
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.608243450708039e-07,
+      "loss": 0.6592,
+      "step": 76712
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.607994860695195e-07,
+      "loss": 0.5359,
+      "step": 76713
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.607746274045109e-07,
+      "loss": 0.6865,
+      "step": 76714
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.607497690757972e-07,
+      "loss": 0.7305,
+      "step": 76715
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.607249110833969e-07,
+      "loss": 0.6162,
+      "step": 76716
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.607000534273294e-07,
+      "loss": 0.5898,
+      "step": 76717
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.606751961076136e-07,
+      "loss": 0.5806,
+      "step": 76718
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60650339124269e-07,
+      "loss": 0.6179,
+      "step": 76719
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.606254824773143e-07,
+      "loss": 0.6719,
+      "step": 76720
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60600626166768e-07,
+      "loss": 0.4124,
+      "step": 76721
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.605757701926496e-07,
+      "loss": 0.8906,
+      "step": 76722
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.605509145549785e-07,
+      "loss": 0.7207,
+      "step": 76723
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60526059253773e-07,
+      "loss": 0.751,
+      "step": 76724
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.605012042890524e-07,
+      "loss": 0.3896,
+      "step": 76725
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60476349660836e-07,
+      "loss": 0.6376,
+      "step": 76726
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60451495369143e-07,
+      "loss": 0.5952,
+      "step": 76727
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.604266414139915e-07,
+      "loss": 0.6865,
+      "step": 76728
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.604017877954013e-07,
+      "loss": 0.6733,
+      "step": 76729
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.603769345133915e-07,
+      "loss": 0.6528,
+      "step": 76730
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.603520815679806e-07,
+      "loss": 0.5574,
+      "step": 76731
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.603272289591883e-07,
+      "loss": 0.7988,
+      "step": 76732
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.603023766870328e-07,
+      "loss": 0.5762,
+      "step": 76733
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.602775247515334e-07,
+      "loss": 0.3634,
+      "step": 76734
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.602526731527094e-07,
+      "loss": 0.6553,
+      "step": 76735
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.6022782189058e-07,
+      "loss": 0.6582,
+      "step": 76736
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.602029709651637e-07,
+      "loss": 0.6816,
+      "step": 76737
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.601781203764797e-07,
+      "loss": 0.603,
+      "step": 76738
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60153270124547e-07,
+      "loss": 0.6328,
+      "step": 76739
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.601284202093853e-07,
+      "loss": 0.6524,
+      "step": 76740
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.601035706310128e-07,
+      "loss": 0.667,
+      "step": 76741
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.600787213894486e-07,
+      "loss": 0.7998,
+      "step": 76742
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60053872484712e-07,
+      "loss": 0.7075,
+      "step": 76743
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.60029023916822e-07,
+      "loss": 0.6377,
+      "step": 76744
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.600041756857973e-07,
+      "loss": 0.4976,
+      "step": 76745
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.599793277916571e-07,
+      "loss": 0.5234,
+      "step": 76746
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.599544802344206e-07,
+      "loss": 0.6309,
+      "step": 76747
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.59929633014107e-07,
+      "loss": 0.7002,
+      "step": 76748
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.599047861307348e-07,
+      "loss": 0.5815,
+      "step": 76749
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.598799395843236e-07,
+      "loss": 0.5269,
+      "step": 76750
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.598550933748916e-07,
+      "loss": 0.5752,
+      "step": 76751
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.598302475024583e-07,
+      "loss": 0.5115,
+      "step": 76752
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.598054019670432e-07,
+      "loss": 0.5417,
+      "step": 76753
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.597805567686646e-07,
+      "loss": 0.4683,
+      "step": 76754
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.597557119073417e-07,
+      "loss": 0.8467,
+      "step": 76755
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.597308673830936e-07,
+      "loss": 0.625,
+      "step": 76756
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.597060231959396e-07,
+      "loss": 0.5374,
+      "step": 76757
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.596811793458982e-07,
+      "loss": 0.6685,
+      "step": 76758
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.596563358329886e-07,
+      "loss": 0.6016,
+      "step": 76759
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.596314926572305e-07,
+      "loss": 0.5479,
+      "step": 76760
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.596066498186422e-07,
+      "loss": 0.5674,
+      "step": 76761
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.595818073172423e-07,
+      "loss": 0.748,
+      "step": 76762
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.595569651530506e-07,
+      "loss": 0.6011,
+      "step": 76763
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.595321233260857e-07,
+      "loss": 0.6104,
+      "step": 76764
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.59507281836367e-07,
+      "loss": 0.4882,
+      "step": 76765
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.594824406839134e-07,
+      "loss": 0.3604,
+      "step": 76766
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.594575998687434e-07,
+      "loss": 0.5764,
+      "step": 76767
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.594327593908767e-07,
+      "loss": 0.5132,
+      "step": 76768
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.594079192503325e-07,
+      "loss": 0.5718,
+      "step": 76769
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.593830794471292e-07,
+      "loss": 0.6318,
+      "step": 76770
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.593582399812859e-07,
+      "loss": 0.5781,
+      "step": 76771
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.593334008528216e-07,
+      "loss": 0.6748,
+      "step": 76772
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.593085620617554e-07,
+      "loss": 0.7197,
+      "step": 76773
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.592837236081067e-07,
+      "loss": 0.5208,
+      "step": 76774
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.59258885491894e-07,
+      "loss": 0.6963,
+      "step": 76775
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.592340477131363e-07,
+      "loss": 0.646,
+      "step": 76776
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.592092102718529e-07,
+      "loss": 0.3879,
+      "step": 76777
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.591843731680632e-07,
+      "loss": 0.6787,
+      "step": 76778
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.591595364017854e-07,
+      "loss": 0.5625,
+      "step": 76779
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.591346999730391e-07,
+      "loss": 0.5215,
+      "step": 76780
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.591098638818428e-07,
+      "loss": 0.7754,
+      "step": 76781
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.590850281282162e-07,
+      "loss": 0.5737,
+      "step": 76782
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.590601927121775e-07,
+      "loss": 0.6777,
+      "step": 76783
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.590353576337461e-07,
+      "loss": 0.5144,
+      "step": 76784
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.590105228929412e-07,
+      "loss": 0.4541,
+      "step": 76785
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.589856884897819e-07,
+      "loss": 0.7812,
+      "step": 76786
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.589608544242866e-07,
+      "loss": 0.6387,
+      "step": 76787
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.589360206964747e-07,
+      "loss": 0.5979,
+      "step": 76788
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.589111873063652e-07,
+      "loss": 0.6914,
+      "step": 76789
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.588863542539776e-07,
+      "loss": 0.4382,
+      "step": 76790
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.588615215393304e-07,
+      "loss": 0.6338,
+      "step": 76791
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.588366891624423e-07,
+      "loss": 0.5732,
+      "step": 76792
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.588118571233325e-07,
+      "loss": 0.584,
+      "step": 76793
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.587870254220204e-07,
+      "loss": 0.6553,
+      "step": 76794
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.58762194058525e-07,
+      "loss": 0.5918,
+      "step": 76795
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.587373630328648e-07,
+      "loss": 0.6211,
+      "step": 76796
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.58712532345059e-07,
+      "loss": 0.7188,
+      "step": 76797
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.586877019951272e-07,
+      "loss": 0.3022,
+      "step": 76798
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.586628719830875e-07,
+      "loss": 0.5942,
+      "step": 76799
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.586380423089598e-07,
+      "loss": 0.5674,
+      "step": 76800
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.586132129727623e-07,
+      "loss": 0.7549,
+      "step": 76801
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.585883839745142e-07,
+      "loss": 0.6162,
+      "step": 76802
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.585635553142352e-07,
+      "loss": 0.6421,
+      "step": 76803
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.585387269919435e-07,
+      "loss": 0.5317,
+      "step": 76804
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.585138990076582e-07,
+      "loss": 0.6748,
+      "step": 76805
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.584890713613986e-07,
+      "loss": 0.3796,
+      "step": 76806
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.58464244053184e-07,
+      "loss": 0.7471,
+      "step": 76807
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.584394170830327e-07,
+      "loss": 0.6777,
+      "step": 76808
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.584145904509638e-07,
+      "loss": 0.71,
+      "step": 76809
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.583897641569971e-07,
+      "loss": 0.4844,
+      "step": 76810
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.583649382011506e-07,
+      "loss": 0.6594,
+      "step": 76811
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.583401125834443e-07,
+      "loss": 0.4932,
+      "step": 76812
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.583152873038961e-07,
+      "loss": 0.7104,
+      "step": 76813
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.582904623625258e-07,
+      "loss": 0.6846,
+      "step": 76814
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.582656377593523e-07,
+      "loss": 0.6968,
+      "step": 76815
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.582408134943943e-07,
+      "loss": 0.6328,
+      "step": 76816
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.582159895676708e-07,
+      "loss": 0.6406,
+      "step": 76817
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.581911659792012e-07,
+      "loss": 0.541,
+      "step": 76818
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.581663427290046e-07,
+      "loss": 0.5962,
+      "step": 76819
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.581415198170998e-07,
+      "loss": 0.5996,
+      "step": 76820
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.581166972435051e-07,
+      "loss": 0.7344,
+      "step": 76821
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.580918750082404e-07,
+      "loss": 0.562,
+      "step": 76822
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.580670531113243e-07,
+      "loss": 0.5996,
+      "step": 76823
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.580422315527764e-07,
+      "loss": 0.6685,
+      "step": 76824
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.580174103326148e-07,
+      "loss": 0.6318,
+      "step": 76825
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.57992589450859e-07,
+      "loss": 0.7832,
+      "step": 76826
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.579677689075279e-07,
+      "loss": 0.709,
+      "step": 76827
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.57942948702641e-07,
+      "loss": 0.5801,
+      "step": 76828
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.579181288362163e-07,
+      "loss": 0.6807,
+      "step": 76829
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.578933093082739e-07,
+      "loss": 0.5198,
+      "step": 76830
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.578684901188319e-07,
+      "loss": 0.665,
+      "step": 76831
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.5784367126791e-07,
+      "loss": 0.5557,
+      "step": 76832
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.578188527555265e-07,
+      "loss": 0.8174,
+      "step": 76833
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.577940345817007e-07,
+      "loss": 0.4609,
+      "step": 76834
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.577692167464517e-07,
+      "loss": 0.5557,
+      "step": 76835
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.57744399249799e-07,
+      "loss": 0.752,
+      "step": 76836
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.577195820917605e-07,
+      "loss": 0.6479,
+      "step": 76837
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.576947652723559e-07,
+      "loss": 0.8135,
+      "step": 76838
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.576699487916043e-07,
+      "loss": 0.6953,
+      "step": 76839
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.576451326495243e-07,
+      "loss": 0.6729,
+      "step": 76840
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.576203168461354e-07,
+      "loss": 0.6929,
+      "step": 76841
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.575955013814559e-07,
+      "loss": 0.7471,
+      "step": 76842
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.575706862555052e-07,
+      "loss": 0.7129,
+      "step": 76843
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.575458714683021e-07,
+      "loss": 0.7314,
+      "step": 76844
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.575210570198664e-07,
+      "loss": 0.6104,
+      "step": 76845
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.57496242910216e-07,
+      "loss": 0.5918,
+      "step": 76846
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.574714291393706e-07,
+      "loss": 0.748,
+      "step": 76847
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.574466157073485e-07,
+      "loss": 0.458,
+      "step": 76848
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.574218026141704e-07,
+      "loss": 0.5032,
+      "step": 76849
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.57396989859853e-07,
+      "loss": 0.5989,
+      "step": 76850
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.573721774444165e-07,
+      "loss": 0.7583,
+      "step": 76851
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.573473653678798e-07,
+      "loss": 0.6641,
+      "step": 76852
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.573225536302623e-07,
+      "loss": 0.7002,
+      "step": 76853
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.572977422315821e-07,
+      "loss": 0.6533,
+      "step": 76854
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.572729311718586e-07,
+      "loss": 0.5601,
+      "step": 76855
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.572481204511109e-07,
+      "loss": 0.6562,
+      "step": 76856
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.572233100693584e-07,
+      "loss": 0.6387,
+      "step": 76857
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.571985000266193e-07,
+      "loss": 0.5981,
+      "step": 76858
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.571736903229132e-07,
+      "loss": 0.7109,
+      "step": 76859
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.571488809582585e-07,
+      "loss": 0.5977,
+      "step": 76860
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.571240719326746e-07,
+      "loss": 0.5898,
+      "step": 76861
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.570992632461808e-07,
+      "loss": 0.6597,
+      "step": 76862
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.570744548987954e-07,
+      "loss": 0.5042,
+      "step": 76863
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.570496468905375e-07,
+      "loss": 0.7598,
+      "step": 76864
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.570248392214265e-07,
+      "loss": 0.7085,
+      "step": 76865
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.570000318914816e-07,
+      "loss": 0.5933,
+      "step": 76866
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56975224900721e-07,
+      "loss": 0.4553,
+      "step": 76867
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.569504182491641e-07,
+      "loss": 0.45,
+      "step": 76868
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.569256119368302e-07,
+      "loss": 0.7461,
+      "step": 76869
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.569008059637381e-07,
+      "loss": 0.583,
+      "step": 76870
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.568760003299062e-07,
+      "loss": 0.7104,
+      "step": 76871
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56851195035354e-07,
+      "loss": 0.7607,
+      "step": 76872
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.568263900801005e-07,
+      "loss": 0.5176,
+      "step": 76873
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56801585464165e-07,
+      "loss": 0.6875,
+      "step": 76874
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.567767811875659e-07,
+      "loss": 0.5811,
+      "step": 76875
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.567519772503221e-07,
+      "loss": 0.5874,
+      "step": 76876
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.567271736524533e-07,
+      "loss": 0.5436,
+      "step": 76877
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.567023703939783e-07,
+      "loss": 0.5388,
+      "step": 76878
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56677567474916e-07,
+      "loss": 0.5718,
+      "step": 76879
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56652764895285e-07,
+      "loss": 0.4988,
+      "step": 76880
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.566279626551045e-07,
+      "loss": 0.5779,
+      "step": 76881
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.566031607543935e-07,
+      "loss": 0.5996,
+      "step": 76882
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.565783591931714e-07,
+      "loss": 0.7939,
+      "step": 76883
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.565535579714567e-07,
+      "loss": 0.5198,
+      "step": 76884
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.565287570892685e-07,
+      "loss": 0.7812,
+      "step": 76885
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.565039565466261e-07,
+      "loss": 0.6963,
+      "step": 76886
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56479156343548e-07,
+      "loss": 0.3719,
+      "step": 76887
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.564543564800533e-07,
+      "loss": 0.6489,
+      "step": 76888
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.564295569561614e-07,
+      "loss": 0.793,
+      "step": 76889
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.564047577718907e-07,
+      "loss": 0.6084,
+      "step": 76890
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.56379958927261e-07,
+      "loss": 0.6494,
+      "step": 76891
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.563551604222902e-07,
+      "loss": 0.5435,
+      "step": 76892
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.563303622569979e-07,
+      "loss": 0.6479,
+      "step": 76893
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.563055644314032e-07,
+      "loss": 0.6465,
+      "step": 76894
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.562807669455252e-07,
+      "loss": 0.6084,
+      "step": 76895
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.562559697993821e-07,
+      "loss": 0.6111,
+      "step": 76896
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.562311729929936e-07,
+      "loss": 0.5789,
+      "step": 76897
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.562063765263783e-07,
+      "loss": 0.6484,
+      "step": 76898
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.561815803995559e-07,
+      "loss": 0.7725,
+      "step": 76899
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.561567846125448e-07,
+      "loss": 0.5596,
+      "step": 76900
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.561319891653637e-07,
+      "loss": 0.6284,
+      "step": 76901
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.561071940580318e-07,
+      "loss": 0.5615,
+      "step": 76902
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.560823992905684e-07,
+      "loss": 0.5085,
+      "step": 76903
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.560576048629925e-07,
+      "loss": 0.5342,
+      "step": 76904
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.560328107753226e-07,
+      "loss": 0.6514,
+      "step": 76905
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.560080170275779e-07,
+      "loss": 0.6904,
+      "step": 76906
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.559832236197779e-07,
+      "loss": 0.6313,
+      "step": 76907
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.559584305519407e-07,
+      "loss": 0.5269,
+      "step": 76908
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.559336378240862e-07,
+      "loss": 0.5242,
+      "step": 76909
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.559088454362324e-07,
+      "loss": 0.54,
+      "step": 76910
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.558840533883987e-07,
+      "loss": 0.6299,
+      "step": 76911
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.558592616806047e-07,
+      "loss": 0.5718,
+      "step": 76912
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.558344703128684e-07,
+      "loss": 0.6665,
+      "step": 76913
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.558096792852091e-07,
+      "loss": 0.7461,
+      "step": 76914
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.557848885976462e-07,
+      "loss": 0.5552,
+      "step": 76915
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.557600982501985e-07,
+      "loss": 0.6387,
+      "step": 76916
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.557353082428846e-07,
+      "loss": 0.5342,
+      "step": 76917
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.557105185757237e-07,
+      "loss": 0.6426,
+      "step": 76918
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.556857292487353e-07,
+      "loss": 0.5,
+      "step": 76919
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.556609402619374e-07,
+      "loss": 0.6453,
+      "step": 76920
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.556361516153499e-07,
+      "loss": 0.4128,
+      "step": 76921
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.55611363308991e-07,
+      "loss": 0.5352,
+      "step": 76922
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.555865753428801e-07,
+      "loss": 0.7764,
+      "step": 76923
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.555617877170365e-07,
+      "loss": 0.4731,
+      "step": 76924
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.555370004314784e-07,
+      "loss": 0.6191,
+      "step": 76925
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.555122134862251e-07,
+      "loss": 0.71,
+      "step": 76926
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.554874268812957e-07,
+      "loss": 0.6392,
+      "step": 76927
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.554626406167096e-07,
+      "loss": 0.668,
+      "step": 76928
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.554378546924853e-07,
+      "loss": 0.7988,
+      "step": 76929
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.554130691086414e-07,
+      "loss": 0.6465,
+      "step": 76930
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.553882838651972e-07,
+      "loss": 0.3818,
+      "step": 76931
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.553634989621718e-07,
+      "loss": 0.6958,
+      "step": 76932
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.553387143995844e-07,
+      "loss": 0.6421,
+      "step": 76933
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.553139301774534e-07,
+      "loss": 0.6943,
+      "step": 76934
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.552891462957981e-07,
+      "loss": 0.481,
+      "step": 76935
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.552643627546375e-07,
+      "loss": 0.7432,
+      "step": 76936
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.552395795539908e-07,
+      "loss": 0.4428,
+      "step": 76937
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.552147966938765e-07,
+      "loss": 0.5747,
+      "step": 76938
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.55190014174314e-07,
+      "loss": 0.6572,
+      "step": 76939
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.551652319953217e-07,
+      "loss": 0.7666,
+      "step": 76940
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.551404501569193e-07,
+      "loss": 0.6147,
+      "step": 76941
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.551156686591251e-07,
+      "loss": 0.6094,
+      "step": 76942
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.550908875019583e-07,
+      "loss": 0.6436,
+      "step": 76943
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.550661066854379e-07,
+      "loss": 0.5479,
+      "step": 76944
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.550413262095834e-07,
+      "loss": 0.6426,
+      "step": 76945
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.55016546074413e-07,
+      "loss": 0.665,
+      "step": 76946
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.549917662799457e-07,
+      "loss": 0.7109,
+      "step": 76947
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.549669868262009e-07,
+      "loss": 0.4946,
+      "step": 76948
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.549422077131978e-07,
+      "loss": 0.5474,
+      "step": 76949
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.54917428940955e-07,
+      "loss": 0.5811,
+      "step": 76950
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.54892650509491e-07,
+      "loss": 0.6504,
+      "step": 76951
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.548678724188254e-07,
+      "loss": 0.6709,
+      "step": 76952
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.548430946689768e-07,
+      "loss": 0.3877,
+      "step": 76953
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.548183172599648e-07,
+      "loss": 0.6953,
+      "step": 76954
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.547935401918077e-07,
+      "loss": 0.4966,
+      "step": 76955
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.547687634645246e-07,
+      "loss": 0.6504,
+      "step": 76956
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.547439870781345e-07,
+      "loss": 0.6982,
+      "step": 76957
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.54719211032657e-07,
+      "loss": 0.5251,
+      "step": 76958
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.546944353281106e-07,
+      "loss": 0.6113,
+      "step": 76959
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.546696599645137e-07,
+      "loss": 0.6357,
+      "step": 76960
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.546448849418856e-07,
+      "loss": 0.7627,
+      "step": 76961
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.546201102602461e-07,
+      "loss": 0.6572,
+      "step": 76962
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.545953359196129e-07,
+      "loss": 0.7422,
+      "step": 76963
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.545705619200056e-07,
+      "loss": 0.4785,
+      "step": 76964
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.545457882614433e-07,
+      "loss": 0.521,
+      "step": 76965
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.54521014943945e-07,
+      "loss": 0.7373,
+      "step": 76966
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.544962419675292e-07,
+      "loss": 0.5654,
+      "step": 76967
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.544714693322152e-07,
+      "loss": 0.5615,
+      "step": 76968
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.544466970380222e-07,
+      "loss": 0.6035,
+      "step": 76969
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.544219250849685e-07,
+      "loss": 0.6191,
+      "step": 76970
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.543971534730738e-07,
+      "loss": 0.7051,
+      "step": 76971
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.543723822023564e-07,
+      "loss": 0.7295,
+      "step": 76972
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.543476112728356e-07,
+      "loss": 0.8623,
+      "step": 76973
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.543228406845302e-07,
+      "loss": 0.6689,
+      "step": 76974
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.542980704374598e-07,
+      "loss": 0.623,
+      "step": 76975
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.542733005316424e-07,
+      "loss": 0.6758,
+      "step": 76976
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.542485309670974e-07,
+      "loss": 0.7666,
+      "step": 76977
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.542237617438445e-07,
+      "loss": 0.5957,
+      "step": 76978
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.541989928619017e-07,
+      "loss": 0.5566,
+      "step": 76979
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.541742243212879e-07,
+      "loss": 0.6924,
+      "step": 76980
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.541494561220224e-07,
+      "loss": 0.5718,
+      "step": 76981
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.541246882641241e-07,
+      "loss": 0.5312,
+      "step": 76982
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.540999207476125e-07,
+      "loss": 0.3888,
+      "step": 76983
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.540751535725056e-07,
+      "loss": 0.4756,
+      "step": 76984
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.54050386738823e-07,
+      "loss": 0.6978,
+      "step": 76985
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.540256202465833e-07,
+      "loss": 0.2551,
+      "step": 76986
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.540008540958063e-07,
+      "loss": 0.6392,
+      "step": 76987
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.539760882865098e-07,
+      "loss": 0.6753,
+      "step": 76988
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.539513228187138e-07,
+      "loss": 0.5372,
+      "step": 76989
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.539265576924363e-07,
+      "loss": 0.6387,
+      "step": 76990
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.539017929076968e-07,
+      "loss": 0.525,
+      "step": 76991
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.538770284645145e-07,
+      "loss": 0.667,
+      "step": 76992
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.538522643629077e-07,
+      "loss": 0.6787,
+      "step": 76993
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.538275006028956e-07,
+      "loss": 0.5654,
+      "step": 76994
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.538027371844977e-07,
+      "loss": 0.6377,
+      "step": 76995
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.537779741077324e-07,
+      "loss": 0.5723,
+      "step": 76996
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.537532113726185e-07,
+      "loss": 0.5981,
+      "step": 76997
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.537284489791752e-07,
+      "loss": 0.6074,
+      "step": 76998
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.537036869274221e-07,
+      "loss": 0.6521,
+      "step": 76999
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.536789252173774e-07,
+      "loss": 0.4858,
+      "step": 77000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.5365416384906e-07,
+      "loss": 0.4443,
+      "step": 77001
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.536294028224889e-07,
+      "loss": 0.6035,
+      "step": 77002
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.536046421376833e-07,
+      "loss": 0.7705,
+      "step": 77003
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.535798817946625e-07,
+      "loss": 0.6133,
+      "step": 77004
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.535551217934447e-07,
+      "loss": 0.7012,
+      "step": 77005
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.535303621340492e-07,
+      "loss": 0.6309,
+      "step": 77006
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.53505602816495e-07,
+      "loss": 0.7944,
+      "step": 77007
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.534808438408014e-07,
+      "loss": 0.7002,
+      "step": 77008
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.534560852069869e-07,
+      "loss": 0.6943,
+      "step": 77009
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.534313269150703e-07,
+      "loss": 0.6509,
+      "step": 77010
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.534065689650708e-07,
+      "loss": 0.7949,
+      "step": 77011
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.533818113570077e-07,
+      "loss": 0.5728,
+      "step": 77012
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.533570540908992e-07,
+      "loss": 0.7207,
+      "step": 77013
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.533322971667646e-07,
+      "loss": 0.7764,
+      "step": 77014
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.533075405846231e-07,
+      "loss": 0.6572,
+      "step": 77015
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.532827843444936e-07,
+      "loss": 0.6846,
+      "step": 77016
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.532580284463947e-07,
+      "loss": 0.6182,
+      "step": 77017
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.532332728903456e-07,
+      "loss": 0.6431,
+      "step": 77018
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.532085176763657e-07,
+      "loss": 0.5869,
+      "step": 77019
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.531837628044728e-07,
+      "loss": 0.4917,
+      "step": 77020
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.531590082746873e-07,
+      "loss": 0.5366,
+      "step": 77021
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.531342540870267e-07,
+      "loss": 0.5227,
+      "step": 77022
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.531095002415109e-07,
+      "loss": 0.6436,
+      "step": 77023
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.530847467381585e-07,
+      "loss": 0.5994,
+      "step": 77024
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.530599935769891e-07,
+      "loss": 0.4883,
+      "step": 77025
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.530352407580206e-07,
+      "loss": 0.6738,
+      "step": 77026
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.530104882812724e-07,
+      "loss": 0.5015,
+      "step": 77027
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.529857361467636e-07,
+      "loss": 0.6228,
+      "step": 77028
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.529609843545137e-07,
+      "loss": 0.6943,
+      "step": 77029
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.529362329045404e-07,
+      "loss": 0.6206,
+      "step": 77030
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.529114817968632e-07,
+      "loss": 0.7109,
+      "step": 77031
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.528867310315013e-07,
+      "loss": 0.6035,
+      "step": 77032
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.528619806084737e-07,
+      "loss": 0.6953,
+      "step": 77033
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.528372305277986e-07,
+      "loss": 0.7007,
+      "step": 77034
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.528124807894957e-07,
+      "loss": 0.6621,
+      "step": 77035
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.527877313935836e-07,
+      "loss": 0.3552,
+      "step": 77036
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.527629823400819e-07,
+      "loss": 0.6396,
+      "step": 77037
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.527382336290085e-07,
+      "loss": 0.5215,
+      "step": 77038
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.527134852603832e-07,
+      "loss": 0.5703,
+      "step": 77039
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.526887372342243e-07,
+      "loss": 0.5386,
+      "step": 77040
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.526639895505511e-07,
+      "loss": 0.5911,
+      "step": 77041
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.52639242209383e-07,
+      "loss": 0.7334,
+      "step": 77042
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.526144952107378e-07,
+      "loss": 0.519,
+      "step": 77043
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.525897485546354e-07,
+      "loss": 0.5273,
+      "step": 77044
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.525650022410943e-07,
+      "loss": 0.9131,
+      "step": 77045
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.52540256270134e-07,
+      "loss": 0.6465,
+      "step": 77046
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.525155106417727e-07,
+      "loss": 0.4917,
+      "step": 77047
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.524907653560297e-07,
+      "loss": 0.6777,
+      "step": 77048
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.524660204129243e-07,
+      "loss": 0.5942,
+      "step": 77049
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.52441275812475e-07,
+      "loss": 0.7041,
+      "step": 77050
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.524165315547007e-07,
+      "loss": 0.6748,
+      "step": 77051
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.523917876396203e-07,
+      "loss": 0.6562,
+      "step": 77052
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.523670440672529e-07,
+      "loss": 0.8428,
+      "step": 77053
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.523423008376179e-07,
+      "loss": 0.7158,
+      "step": 77054
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.523175579507334e-07,
+      "loss": 0.6919,
+      "step": 77055
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 5.522928154066189e-07,
+      "loss": 0.5325,
+      "step": 77056
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.522680732052932e-07,
+      "loss": 0.6387,
+      "step": 77057
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.522433313467753e-07,
+      "loss": 0.5669,
+      "step": 77058
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.522185898310844e-07,
+      "loss": 0.668,
+      "step": 77059
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.521938486582387e-07,
+      "loss": 0.6387,
+      "step": 77060
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.521691078282576e-07,
+      "loss": 0.5737,
+      "step": 77061
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.5214436734116e-07,
+      "loss": 0.5684,
+      "step": 77062
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.521196271969653e-07,
+      "loss": 0.834,
+      "step": 77063
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.520948873956914e-07,
+      "loss": 0.5713,
+      "step": 77064
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.520701479373582e-07,
+      "loss": 0.6938,
+      "step": 77065
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.52045408821984e-07,
+      "loss": 0.5295,
+      "step": 77066
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.520206700495885e-07,
+      "loss": 0.5789,
+      "step": 77067
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.519959316201898e-07,
+      "loss": 0.6035,
+      "step": 77068
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.519711935338077e-07,
+      "loss": 0.625,
+      "step": 77069
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.519464557904601e-07,
+      "loss": 0.615,
+      "step": 77070
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.51921718390167e-07,
+      "loss": 0.5249,
+      "step": 77071
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.518969813329463e-07,
+      "loss": 0.6865,
+      "step": 77072
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.518722446188178e-07,
+      "loss": 0.7578,
+      "step": 77073
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.518475082477998e-07,
+      "loss": 0.3537,
+      "step": 77074
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.518227722199121e-07,
+      "loss": 0.5996,
+      "step": 77075
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.517980365351727e-07,
+      "loss": 0.6533,
+      "step": 77076
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.517733011936008e-07,
+      "loss": 0.5376,
+      "step": 77077
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.517485661952156e-07,
+      "loss": 0.5142,
+      "step": 77078
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.517238315400363e-07,
+      "loss": 0.5737,
+      "step": 77079
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.516990972280814e-07,
+      "loss": 0.2742,
+      "step": 77080
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.516743632593694e-07,
+      "loss": 0.6924,
+      "step": 77081
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.5164962963392e-07,
+      "loss": 0.6641,
+      "step": 77082
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.516248963517515e-07,
+      "loss": 0.7725,
+      "step": 77083
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.51600163412884e-07,
+      "loss": 0.6636,
+      "step": 77084
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.515754308173351e-07,
+      "loss": 0.6338,
+      "step": 77085
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.515506985651241e-07,
+      "loss": 0.6235,
+      "step": 77086
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.515259666562706e-07,
+      "loss": 0.5854,
+      "step": 77087
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.515012350907926e-07,
+      "loss": 0.5981,
+      "step": 77088
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.5147650386871e-07,
+      "loss": 0.6416,
+      "step": 77089
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.514517729900408e-07,
+      "loss": 0.5267,
+      "step": 77090
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.514270424548044e-07,
+      "loss": 0.6255,
+      "step": 77091
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.514023122630199e-07,
+      "loss": 0.6206,
+      "step": 77092
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.513775824147058e-07,
+      "loss": 0.6245,
+      "step": 77093
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.513528529098812e-07,
+      "loss": 0.6465,
+      "step": 77094
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.51328123748565e-07,
+      "loss": 0.6904,
+      "step": 77095
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.513033949307767e-07,
+      "loss": 0.6357,
+      "step": 77096
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.512786664565343e-07,
+      "loss": 0.665,
+      "step": 77097
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.512539383258574e-07,
+      "loss": 0.5625,
+      "step": 77098
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.512292105387649e-07,
+      "loss": 0.5713,
+      "step": 77099
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.512044830952753e-07,
+      "loss": 0.6641,
+      "step": 77100
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.511797559954081e-07,
+      "loss": 0.4601,
+      "step": 77101
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.511550292391816e-07,
+      "loss": 0.6196,
+      "step": 77102
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.51130302826615e-07,
+      "loss": 0.5566,
+      "step": 77103
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.511055767577277e-07,
+      "loss": 0.6934,
+      "step": 77104
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.510808510325378e-07,
+      "loss": 0.6201,
+      "step": 77105
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.510561256510647e-07,
+      "loss": 0.7305,
+      "step": 77106
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.510314006133272e-07,
+      "loss": 0.5735,
+      "step": 77107
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.510066759193447e-07,
+      "loss": 0.4673,
+      "step": 77108
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.509819515691359e-07,
+      "loss": 0.6543,
+      "step": 77109
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50957227562719e-07,
+      "loss": 0.6289,
+      "step": 77110
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.509325039001136e-07,
+      "loss": 0.5212,
+      "step": 77111
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.509077805813385e-07,
+      "loss": 0.5913,
+      "step": 77112
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.508830576064131e-07,
+      "loss": 0.6709,
+      "step": 77113
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.508583349753555e-07,
+      "loss": 0.7549,
+      "step": 77114
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50833612688185e-07,
+      "loss": 0.582,
+      "step": 77115
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.508088907449204e-07,
+      "loss": 0.6758,
+      "step": 77116
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.507841691455815e-07,
+      "loss": 0.7539,
+      "step": 77117
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.507594478901858e-07,
+      "loss": 0.3563,
+      "step": 77118
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.507347269787534e-07,
+      "loss": 0.7012,
+      "step": 77119
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.507100064113025e-07,
+      "loss": 0.4678,
+      "step": 77120
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.506852861878526e-07,
+      "loss": 0.397,
+      "step": 77121
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.506605663084219e-07,
+      "loss": 0.7407,
+      "step": 77122
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.506358467730299e-07,
+      "loss": 0.6953,
+      "step": 77123
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.506111275816951e-07,
+      "loss": 0.7568,
+      "step": 77124
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.505864087344373e-07,
+      "loss": 0.5464,
+      "step": 77125
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.505616902312743e-07,
+      "loss": 0.6758,
+      "step": 77126
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.505369720722257e-07,
+      "loss": 0.6108,
+      "step": 77127
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.505122542573102e-07,
+      "loss": 0.7305,
+      "step": 77128
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.504875367865471e-07,
+      "loss": 0.7432,
+      "step": 77129
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50462819659955e-07,
+      "loss": 0.6895,
+      "step": 77130
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.504381028775527e-07,
+      "loss": 0.6514,
+      "step": 77131
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50413386439359e-07,
+      "loss": 0.6279,
+      "step": 77132
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.503886703453933e-07,
+      "loss": 0.5259,
+      "step": 77133
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.503639545956746e-07,
+      "loss": 0.4927,
+      "step": 77134
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.503392391902211e-07,
+      "loss": 0.6104,
+      "step": 77135
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.503145241290524e-07,
+      "loss": 0.7383,
+      "step": 77136
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50289809412187e-07,
+      "loss": 0.6978,
+      "step": 77137
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.502650950396448e-07,
+      "loss": 0.6514,
+      "step": 77138
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.502403810114433e-07,
+      "loss": 0.7422,
+      "step": 77139
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.50215667327602e-07,
+      "loss": 0.6777,
+      "step": 77140
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.501909539881398e-07,
+      "loss": 0.6333,
+      "step": 77141
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.501662409930761e-07,
+      "loss": 0.584,
+      "step": 77142
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.501415283424291e-07,
+      "loss": 0.5605,
+      "step": 77143
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.501168160362179e-07,
+      "loss": 0.5718,
+      "step": 77144
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.500921040744617e-07,
+      "loss": 0.8623,
+      "step": 77145
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.500673924571797e-07,
+      "loss": 0.7217,
+      "step": 77146
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.5004268118439e-07,
+      "loss": 0.5684,
+      "step": 77147
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.500179702561123e-07,
+      "loss": 0.709,
+      "step": 77148
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.499932596723647e-07,
+      "loss": 0.7695,
+      "step": 77149
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.499685494331667e-07,
+      "loss": 0.7148,
+      "step": 77150
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.499438395385374e-07,
+      "loss": 0.7236,
+      "step": 77151
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.499191299884951e-07,
+      "loss": 0.7656,
+      "step": 77152
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.498944207830589e-07,
+      "loss": 0.7217,
+      "step": 77153
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.498697119222479e-07,
+      "loss": 0.6143,
+      "step": 77154
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.498450034060814e-07,
+      "loss": 0.5962,
+      "step": 77155
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.498202952345774e-07,
+      "loss": 0.7393,
+      "step": 77156
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.497955874077553e-07,
+      "loss": 0.6816,
+      "step": 77157
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.497708799256346e-07,
+      "loss": 0.5503,
+      "step": 77158
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.497461727882334e-07,
+      "loss": 0.6887,
+      "step": 77159
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.497214659955706e-07,
+      "loss": 0.6182,
+      "step": 77160
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.496967595476654e-07,
+      "loss": 0.5029,
+      "step": 77161
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.496720534445365e-07,
+      "loss": 0.603,
+      "step": 77162
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.496473476862035e-07,
+      "loss": 0.6279,
+      "step": 77163
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.496226422726844e-07,
+      "loss": 0.7422,
+      "step": 77164
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.495979372039986e-07,
+      "loss": 0.6611,
+      "step": 77165
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.49573232480165e-07,
+      "loss": 0.4907,
+      "step": 77166
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.495485281012027e-07,
+      "loss": 0.6152,
+      "step": 77167
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.495238240671304e-07,
+      "loss": 0.5903,
+      "step": 77168
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.494991203779666e-07,
+      "loss": 0.5129,
+      "step": 77169
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.494744170337307e-07,
+      "loss": 0.4463,
+      "step": 77170
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.494497140344415e-07,
+      "loss": 0.3987,
+      "step": 77171
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.494250113801183e-07,
+      "loss": 0.6484,
+      "step": 77172
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.494003090707793e-07,
+      "loss": 0.7236,
+      "step": 77173
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.493756071064437e-07,
+      "loss": 0.5713,
+      "step": 77174
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.49350905487131e-07,
+      "loss": 0.5098,
+      "step": 77175
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.49326204212859e-07,
+      "loss": 0.5698,
+      "step": 77176
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.493015032836473e-07,
+      "loss": 0.6064,
+      "step": 77177
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.492768026995152e-07,
+      "loss": 0.7168,
+      "step": 77178
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.492521024604805e-07,
+      "loss": 0.6475,
+      "step": 77179
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.492274025665633e-07,
+      "loss": 0.6196,
+      "step": 77180
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.492027030177816e-07,
+      "loss": 0.6973,
+      "step": 77181
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.491780038141547e-07,
+      "loss": 0.5511,
+      "step": 77182
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.491533049557013e-07,
+      "loss": 0.6699,
+      "step": 77183
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.49128606442441e-07,
+      "loss": 0.6045,
+      "step": 77184
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.491039082743917e-07,
+      "loss": 0.5703,
+      "step": 77185
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.490792104515729e-07,
+      "loss": 0.5801,
+      "step": 77186
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.490545129740033e-07,
+      "loss": 0.5298,
+      "step": 77187
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.490298158417024e-07,
+      "loss": 0.5967,
+      "step": 77188
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.490051190546886e-07,
+      "loss": 0.5859,
+      "step": 77189
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.489804226129805e-07,
+      "loss": 0.6299,
+      "step": 77190
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.489557265165974e-07,
+      "loss": 0.7109,
+      "step": 77191
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.489310307655584e-07,
+      "loss": 0.4836,
+      "step": 77192
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.489063353598818e-07,
+      "loss": 0.6836,
+      "step": 77193
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.48881640299587e-07,
+      "loss": 0.6348,
+      "step": 77194
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.488569455846928e-07,
+      "loss": 0.7905,
+      "step": 77195
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.488322512152184e-07,
+      "loss": 0.437,
+      "step": 77196
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.488075571911818e-07,
+      "loss": 0.5488,
+      "step": 77197
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.487828635126033e-07,
+      "loss": 0.626,
+      "step": 77198
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.487581701795003e-07,
+      "loss": 0.575,
+      "step": 77199
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.487334771918926e-07,
+      "loss": 0.6533,
+      "step": 77200
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.487087845497993e-07,
+      "loss": 0.4927,
+      "step": 77201
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.486840922532385e-07,
+      "loss": 0.6058,
+      "step": 77202
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.486594003022295e-07,
+      "loss": 0.667,
+      "step": 77203
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.486347086967913e-07,
+      "loss": 0.5737,
+      "step": 77204
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.486100174369432e-07,
+      "loss": 0.573,
+      "step": 77205
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.485853265227031e-07,
+      "loss": 0.7114,
+      "step": 77206
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.485606359540907e-07,
+      "loss": 0.6396,
+      "step": 77207
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.485359457311247e-07,
+      "loss": 0.6475,
+      "step": 77208
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.485112558538243e-07,
+      "loss": 0.6958,
+      "step": 77209
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.484865663222076e-07,
+      "loss": 0.5474,
+      "step": 77210
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.484618771362939e-07,
+      "loss": 0.873,
+      "step": 77211
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.484371882961022e-07,
+      "loss": 0.4346,
+      "step": 77212
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.484124998016517e-07,
+      "loss": 0.7227,
+      "step": 77213
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.483878116529608e-07,
+      "loss": 0.5977,
+      "step": 77214
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.483631238500484e-07,
+      "loss": 0.5269,
+      "step": 77215
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.483384363929336e-07,
+      "loss": 0.5432,
+      "step": 77216
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.483137492816357e-07,
+      "loss": 0.5942,
+      "step": 77217
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.482890625161733e-07,
+      "loss": 0.7549,
+      "step": 77218
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.482643760965648e-07,
+      "loss": 0.4741,
+      "step": 77219
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.482396900228294e-07,
+      "loss": 0.6309,
+      "step": 77220
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.48215004294986e-07,
+      "loss": 0.6533,
+      "step": 77221
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.481903189130542e-07,
+      "loss": 0.7275,
+      "step": 77222
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.481656338770519e-07,
+      "loss": 0.5219,
+      "step": 77223
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.481409491869982e-07,
+      "loss": 0.667,
+      "step": 77224
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.481162648429123e-07,
+      "loss": 0.5431,
+      "step": 77225
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.480915808448135e-07,
+      "loss": 0.5464,
+      "step": 77226
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.480668971927196e-07,
+      "loss": 0.519,
+      "step": 77227
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.480422138866507e-07,
+      "loss": 0.7197,
+      "step": 77228
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.480175309266247e-07,
+      "loss": 0.4556,
+      "step": 77229
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.47992848312661e-07,
+      "loss": 0.6416,
+      "step": 77230
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.479681660447782e-07,
+      "loss": 0.7002,
+      "step": 77231
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.479434841229954e-07,
+      "loss": 0.5225,
+      "step": 77232
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.479188025473315e-07,
+      "loss": 0.7949,
+      "step": 77233
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.478941213178057e-07,
+      "loss": 0.7051,
+      "step": 77234
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.478694404344363e-07,
+      "loss": 0.5889,
+      "step": 77235
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.478447598972424e-07,
+      "loss": 0.5239,
+      "step": 77236
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.47820079706243e-07,
+      "loss": 0.4941,
+      "step": 77237
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.477953998614573e-07,
+      "loss": 0.4749,
+      "step": 77238
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.477707203629039e-07,
+      "loss": 0.6484,
+      "step": 77239
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.477460412106014e-07,
+      "loss": 0.615,
+      "step": 77240
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.477213624045689e-07,
+      "loss": 0.6523,
+      "step": 77241
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.476966839448252e-07,
+      "loss": 0.6406,
+      "step": 77242
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.4767200583139e-07,
+      "loss": 0.666,
+      "step": 77243
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.476473280642809e-07,
+      "loss": 0.3945,
+      "step": 77244
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.476226506435176e-07,
+      "loss": 0.522,
+      "step": 77245
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.475979735691187e-07,
+      "loss": 0.9072,
+      "step": 77246
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.475732968411038e-07,
+      "loss": 0.6826,
+      "step": 77247
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.475486204594912e-07,
+      "loss": 0.6602,
+      "step": 77248
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.475239444242993e-07,
+      "loss": 0.4819,
+      "step": 77249
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.474992687355476e-07,
+      "loss": 0.7246,
+      "step": 77250
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.474745933932552e-07,
+      "loss": 0.511,
+      "step": 77251
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.474499183974404e-07,
+      "loss": 0.751,
+      "step": 77252
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.474252437481224e-07,
+      "loss": 0.6587,
+      "step": 77253
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.474005694453199e-07,
+      "loss": 0.7104,
+      "step": 77254
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.473758954890526e-07,
+      "loss": 0.6289,
+      "step": 77255
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.473512218793384e-07,
+      "loss": 0.5615,
+      "step": 77256
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.473265486161963e-07,
+      "loss": 0.6611,
+      "step": 77257
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.473018756996459e-07,
+      "loss": 0.8057,
+      "step": 77258
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.472772031297054e-07,
+      "loss": 0.6147,
+      "step": 77259
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.472525309063942e-07,
+      "loss": 0.7168,
+      "step": 77260
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.472278590297306e-07,
+      "loss": 0.3347,
+      "step": 77261
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.472031874997338e-07,
+      "loss": 0.6719,
+      "step": 77262
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.471785163164226e-07,
+      "loss": 0.4277,
+      "step": 77263
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.471538454798165e-07,
+      "loss": 0.5652,
+      "step": 77264
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.471291749899335e-07,
+      "loss": 0.5962,
+      "step": 77265
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.471045048467929e-07,
+      "loss": 0.6362,
+      "step": 77266
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.470798350504137e-07,
+      "loss": 0.4558,
+      "step": 77267
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.470551656008149e-07,
+      "loss": 0.5801,
+      "step": 77268
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.470304964980146e-07,
+      "loss": 0.5273,
+      "step": 77269
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.470058277420323e-07,
+      "loss": 0.5122,
+      "step": 77270
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.469811593328868e-07,
+      "loss": 0.6997,
+      "step": 77271
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.469564912705973e-07,
+      "loss": 0.6836,
+      "step": 77272
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.469318235551821e-07,
+      "loss": 0.6272,
+      "step": 77273
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.469071561866603e-07,
+      "loss": 0.6699,
+      "step": 77274
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.468824891650508e-07,
+      "loss": 0.6797,
+      "step": 77275
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.46857822490373e-07,
+      "loss": 0.7012,
+      "step": 77276
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.46833156162645e-07,
+      "loss": 0.7646,
+      "step": 77277
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.468084901818863e-07,
+      "loss": 0.834,
+      "step": 77278
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.467838245481153e-07,
+      "loss": 0.7871,
+      "step": 77279
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.467591592613509e-07,
+      "loss": 0.5361,
+      "step": 77280
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.467344943216127e-07,
+      "loss": 0.8242,
+      "step": 77281
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.467098297289185e-07,
+      "loss": 0.6738,
+      "step": 77282
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.466851654832878e-07,
+      "loss": 0.6277,
+      "step": 77283
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.466605015847398e-07,
+      "loss": 0.5669,
+      "step": 77284
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.466358380332927e-07,
+      "loss": 0.6284,
+      "step": 77285
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.466111748289657e-07,
+      "loss": 0.5552,
+      "step": 77286
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.465865119717777e-07,
+      "loss": 0.6328,
+      "step": 77287
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.465618494617478e-07,
+      "loss": 0.5496,
+      "step": 77288
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.465371872988946e-07,
+      "loss": 0.7148,
+      "step": 77289
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.465125254832368e-07,
+      "loss": 0.6289,
+      "step": 77290
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.464878640147935e-07,
+      "loss": 0.5825,
+      "step": 77291
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.464632028935837e-07,
+      "loss": 0.5264,
+      "step": 77292
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.464385421196263e-07,
+      "loss": 0.5437,
+      "step": 77293
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.464138816929398e-07,
+      "loss": 0.5386,
+      "step": 77294
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.463892216135433e-07,
+      "loss": 0.5276,
+      "step": 77295
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.463645618814557e-07,
+      "loss": 0.4785,
+      "step": 77296
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.463399024966964e-07,
+      "loss": 0.7207,
+      "step": 77297
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.463152434592836e-07,
+      "loss": 0.4358,
+      "step": 77298
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.46290584769236e-07,
+      "loss": 0.6938,
+      "step": 77299
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.46265926426573e-07,
+      "loss": 0.4707,
+      "step": 77300
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.462412684313136e-07,
+      "loss": 0.5117,
+      "step": 77301
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.462166107834759e-07,
+      "loss": 0.6953,
+      "step": 77302
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.461919534830795e-07,
+      "loss": 0.6235,
+      "step": 77303
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.461672965301428e-07,
+      "loss": 0.7969,
+      "step": 77304
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.461426399246856e-07,
+      "loss": 0.4142,
+      "step": 77305
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.461179836667254e-07,
+      "loss": 0.7871,
+      "step": 77306
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.46093327756282e-07,
+      "loss": 0.6235,
+      "step": 77307
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.460686721933745e-07,
+      "loss": 0.6377,
+      "step": 77308
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.460440169780208e-07,
+      "loss": 0.5347,
+      "step": 77309
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.460193621102409e-07,
+      "loss": 0.5886,
+      "step": 77310
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.459947075900526e-07,
+      "loss": 0.5557,
+      "step": 77311
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.459700534174752e-07,
+      "loss": 0.562,
+      "step": 77312
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.459453995925278e-07,
+      "loss": 0.6133,
+      "step": 77313
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.459207461152294e-07,
+      "loss": 0.6499,
+      "step": 77314
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.458960929855983e-07,
+      "loss": 0.7529,
+      "step": 77315
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.458714402036536e-07,
+      "loss": 0.4699,
+      "step": 77316
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.458467877694144e-07,
+      "loss": 0.6885,
+      "step": 77317
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.458221356829001e-07,
+      "loss": 0.5562,
+      "step": 77318
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.457974839441281e-07,
+      "loss": 0.52,
+      "step": 77319
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.457728325531181e-07,
+      "loss": 0.353,
+      "step": 77320
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.457481815098889e-07,
+      "loss": 0.7314,
+      "step": 77321
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.457235308144601e-07,
+      "loss": 0.5615,
+      "step": 77322
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.456988804668492e-07,
+      "loss": 0.7217,
+      "step": 77323
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.456742304670759e-07,
+      "loss": 0.6582,
+      "step": 77324
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.45649580815159e-07,
+      "loss": 0.5679,
+      "step": 77325
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.456249315111177e-07,
+      "loss": 0.8516,
+      "step": 77326
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.4560028255497e-07,
+      "loss": 0.6992,
+      "step": 77327
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.455756339467356e-07,
+      "loss": 0.48,
+      "step": 77328
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.455509856864328e-07,
+      "loss": 0.7217,
+      "step": 77329
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.455263377740809e-07,
+      "loss": 0.707,
+      "step": 77330
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.455016902096987e-07,
+      "loss": 0.4633,
+      "step": 77331
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.454770429933047e-07,
+      "loss": 0.519,
+      "step": 77332
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.454523961249179e-07,
+      "loss": 0.5396,
+      "step": 77333
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.454277496045573e-07,
+      "loss": 0.5693,
+      "step": 77334
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.454031034322424e-07,
+      "loss": 0.5283,
+      "step": 77335
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.453784576079907e-07,
+      "loss": 0.3632,
+      "step": 77336
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.453538121318221e-07,
+      "loss": 0.7744,
+      "step": 77337
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.453291670037554e-07,
+      "loss": 0.6377,
+      "step": 77338
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.453045222238093e-07,
+      "loss": 0.5151,
+      "step": 77339
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.452798777920022e-07,
+      "loss": 0.7715,
+      "step": 77340
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.452552337083536e-07,
+      "loss": 0.563,
+      "step": 77341
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.452305899728819e-07,
+      "loss": 0.4612,
+      "step": 77342
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.452059465856067e-07,
+      "loss": 0.7334,
+      "step": 77343
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.45181303546546e-07,
+      "loss": 0.6616,
+      "step": 77344
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.451566608557191e-07,
+      "loss": 0.6699,
+      "step": 77345
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.451320185131447e-07,
+      "loss": 0.5947,
+      "step": 77346
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.451073765188424e-07,
+      "loss": 0.5708,
+      "step": 77347
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.450827348728304e-07,
+      "loss": 0.4886,
+      "step": 77348
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.450580935751271e-07,
+      "loss": 0.7783,
+      "step": 77349
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.45033452625752e-07,
+      "loss": 0.6143,
+      "step": 77350
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.450088120247238e-07,
+      "loss": 0.5083,
+      "step": 77351
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.44984171772062e-07,
+      "loss": 0.8047,
+      "step": 77352
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.449595318677844e-07,
+      "loss": 0.627,
+      "step": 77353
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.449348923119104e-07,
+      "loss": 0.7021,
+      "step": 77354
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.449102531044592e-07,
+      "loss": 0.5977,
+      "step": 77355
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.448856142454487e-07,
+      "loss": 0.6084,
+      "step": 77356
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.448609757348986e-07,
+      "loss": 0.6699,
+      "step": 77357
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.448363375728279e-07,
+      "loss": 0.6675,
+      "step": 77358
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.448116997592546e-07,
+      "loss": 0.6724,
+      "step": 77359
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.447870622941985e-07,
+      "loss": 0.4236,
+      "step": 77360
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.447624251776777e-07,
+      "loss": 0.8408,
+      "step": 77361
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.447377884097112e-07,
+      "loss": 0.5161,
+      "step": 77362
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.447131519903181e-07,
+      "loss": 0.543,
+      "step": 77363
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.446885159195176e-07,
+      "loss": 0.7373,
+      "step": 77364
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.446638801973278e-07,
+      "loss": 0.6094,
+      "step": 77365
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.44639244823768e-07,
+      "loss": 0.6306,
+      "step": 77366
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.446146097988569e-07,
+      "loss": 0.4473,
+      "step": 77367
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.445899751226139e-07,
+      "loss": 0.6313,
+      "step": 77368
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.445653407950574e-07,
+      "loss": 0.585,
+      "step": 77369
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.445407068162058e-07,
+      "loss": 0.4501,
+      "step": 77370
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.445160731860785e-07,
+      "loss": 0.6494,
+      "step": 77371
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.444914399046947e-07,
+      "loss": 0.7041,
+      "step": 77372
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.444668069720726e-07,
+      "loss": 0.5259,
+      "step": 77373
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.444421743882311e-07,
+      "loss": 0.6577,
+      "step": 77374
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.444175421531893e-07,
+      "loss": 0.6504,
+      "step": 77375
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.443929102669666e-07,
+      "loss": 0.5508,
+      "step": 77376
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.443682787295808e-07,
+      "loss": 0.8145,
+      "step": 77377
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.443436475410517e-07,
+      "loss": 0.707,
+      "step": 77378
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.443190167013972e-07,
+      "loss": 0.6587,
+      "step": 77379
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.442943862106369e-07,
+      "loss": 0.5723,
+      "step": 77380
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.442697560687897e-07,
+      "loss": 0.4524,
+      "step": 77381
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.442451262758737e-07,
+      "loss": 0.542,
+      "step": 77382
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.442204968319084e-07,
+      "loss": 0.6111,
+      "step": 77383
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.441958677369124e-07,
+      "loss": 0.6377,
+      "step": 77384
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.44171238990905e-07,
+      "loss": 0.6572,
+      "step": 77385
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.441466105939046e-07,
+      "loss": 0.7783,
+      "step": 77386
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.4412198254593e-07,
+      "loss": 0.6172,
+      "step": 77387
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.440973548470006e-07,
+      "loss": 0.5981,
+      "step": 77388
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.440727274971348e-07,
+      "loss": 0.5089,
+      "step": 77389
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.440481004963515e-07,
+      "loss": 0.4863,
+      "step": 77390
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.440234738446693e-07,
+      "loss": 0.5737,
+      "step": 77391
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.439988475421075e-07,
+      "loss": 0.7559,
+      "step": 77392
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.439742215886851e-07,
+      "loss": 0.6089,
+      "step": 77393
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.439495959844203e-07,
+      "loss": 0.5903,
+      "step": 77394
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.439249707293324e-07,
+      "loss": 0.5654,
+      "step": 77395
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.439003458234401e-07,
+      "loss": 0.5693,
+      "step": 77396
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.438757212667629e-07,
+      "loss": 0.6123,
+      "step": 77397
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.438510970593188e-07,
+      "loss": 0.5398,
+      "step": 77398
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.438264732011268e-07,
+      "loss": 0.6758,
+      "step": 77399
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.438018496922059e-07,
+      "loss": 0.6611,
+      "step": 77400
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.437772265325748e-07,
+      "loss": 0.6641,
+      "step": 77401
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.437526037222528e-07,
+      "loss": 0.5757,
+      "step": 77402
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.437279812612583e-07,
+      "loss": 0.7539,
+      "step": 77403
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.437033591496101e-07,
+      "loss": 0.667,
+      "step": 77404
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.436787373873276e-07,
+      "loss": 0.6953,
+      "step": 77405
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.436541159744293e-07,
+      "loss": 0.4512,
+      "step": 77406
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.436294949109339e-07,
+      "loss": 0.6777,
+      "step": 77407
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.436048741968606e-07,
+      "loss": 0.5339,
+      "step": 77408
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.435802538322278e-07,
+      "loss": 0.5557,
+      "step": 77409
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.43555633817055e-07,
+      "loss": 0.8867,
+      "step": 77410
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.435310141513604e-07,
+      "loss": 0.6113,
+      "step": 77411
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.43506394835163e-07,
+      "loss": 0.8477,
+      "step": 77412
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.434817758684818e-07,
+      "loss": 0.6431,
+      "step": 77413
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.434571572513359e-07,
+      "loss": 0.8354,
+      "step": 77414
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.434325389837436e-07,
+      "loss": 0.6123,
+      "step": 77415
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.434079210657238e-07,
+      "loss": 0.7852,
+      "step": 77416
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.433833034972957e-07,
+      "loss": 0.6328,
+      "step": 77417
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.433586862784785e-07,
+      "loss": 0.6323,
+      "step": 77418
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.433340694092905e-07,
+      "loss": 0.5576,
+      "step": 77419
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.433094528897501e-07,
+      "loss": 0.363,
+      "step": 77420
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.432848367198768e-07,
+      "loss": 0.3209,
+      "step": 77421
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.432602208996891e-07,
+      "loss": 0.7588,
+      "step": 77422
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.432356054292067e-07,
+      "loss": 0.6719,
+      "step": 77423
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.432109903084471e-07,
+      "loss": 0.7026,
+      "step": 77424
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.431863755374302e-07,
+      "loss": 0.6621,
+      "step": 77425
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.431617611161742e-07,
+      "loss": 0.5767,
+      "step": 77426
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.431371470446987e-07,
+      "loss": 0.6724,
+      "step": 77427
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.431125333230222e-07,
+      "loss": 0.4878,
+      "step": 77428
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.430879199511629e-07,
+      "loss": 0.4692,
+      "step": 77429
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.430633069291401e-07,
+      "loss": 0.582,
+      "step": 77430
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.430386942569734e-07,
+      "loss": 0.6279,
+      "step": 77431
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.430140819346803e-07,
+      "loss": 0.6597,
+      "step": 77432
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.429894699622803e-07,
+      "loss": 0.5659,
+      "step": 77433
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.429648583397924e-07,
+      "loss": 0.5488,
+      "step": 77434
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.429402470672356e-07,
+      "loss": 0.6748,
+      "step": 77435
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.429156361446281e-07,
+      "loss": 0.5913,
+      "step": 77436
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.428910255719894e-07,
+      "loss": 0.7842,
+      "step": 77437
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.428664153493376e-07,
+      "loss": 0.6484,
+      "step": 77438
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.428418054766921e-07,
+      "loss": 0.573,
+      "step": 77439
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.428171959540719e-07,
+      "loss": 0.5605,
+      "step": 77440
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.427925867814953e-07,
+      "loss": 0.5916,
+      "step": 77441
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.427679779589812e-07,
+      "loss": 0.6133,
+      "step": 77442
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.427433694865487e-07,
+      "loss": 0.6982,
+      "step": 77443
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.427187613642171e-07,
+      "loss": 0.7188,
+      "step": 77444
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.426941535920042e-07,
+      "loss": 0.71,
+      "step": 77445
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.426695461699293e-07,
+      "loss": 0.3694,
+      "step": 77446
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.426449390980118e-07,
+      "loss": 0.6313,
+      "step": 77447
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.4262033237627e-07,
+      "loss": 0.4385,
+      "step": 77448
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.425957260047225e-07,
+      "loss": 0.6245,
+      "step": 77449
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.425711199833883e-07,
+      "loss": 0.7266,
+      "step": 77450
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.425465143122863e-07,
+      "loss": 0.4424,
+      "step": 77451
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.425219089914359e-07,
+      "loss": 0.6465,
+      "step": 77452
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.42497304020855e-07,
+      "loss": 0.5132,
+      "step": 77453
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.424726994005628e-07,
+      "loss": 0.646,
+      "step": 77454
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.424480951305783e-07,
+      "loss": 0.6699,
+      "step": 77455
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.424234912109207e-07,
+      "loss": 0.6196,
+      "step": 77456
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.423988876416084e-07,
+      "loss": 0.6768,
+      "step": 77457
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.423742844226598e-07,
+      "loss": 0.2644,
+      "step": 77458
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.423496815540942e-07,
+      "loss": 0.6118,
+      "step": 77459
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.423250790359304e-07,
+      "loss": 0.406,
+      "step": 77460
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.423004768681876e-07,
+      "loss": 0.6328,
+      "step": 77461
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.422758750508838e-07,
+      "loss": 0.7812,
+      "step": 77462
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.422512735840384e-07,
+      "loss": 0.5962,
+      "step": 77463
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.422266724676706e-07,
+      "loss": 0.3366,
+      "step": 77464
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.422020717017983e-07,
+      "loss": 0.7158,
+      "step": 77465
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.421774712864409e-07,
+      "loss": 0.5972,
+      "step": 77466
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.421528712216175e-07,
+      "loss": 0.5586,
+      "step": 77467
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.421282715073462e-07,
+      "loss": 0.6738,
+      "step": 77468
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.421036721436466e-07,
+      "loss": 0.6587,
+      "step": 77469
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.420790731305368e-07,
+      "loss": 0.7188,
+      "step": 77470
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.42054474468036e-07,
+      "loss": 0.5344,
+      "step": 77471
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.42029876156163e-07,
+      "loss": 0.5662,
+      "step": 77472
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.420052781949372e-07,
+      "loss": 0.5908,
+      "step": 77473
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.419806805843763e-07,
+      "loss": 0.4761,
+      "step": 77474
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.419560833244999e-07,
+      "loss": 0.6782,
+      "step": 77475
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.419314864153266e-07,
+      "loss": 0.708,
+      "step": 77476
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.419068898568757e-07,
+      "loss": 0.5996,
+      "step": 77477
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.418822936491656e-07,
+      "loss": 0.5894,
+      "step": 77478
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.418576977922149e-07,
+      "loss": 0.6924,
+      "step": 77479
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.418331022860426e-07,
+      "loss": 0.7471,
+      "step": 77480
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.41808507130668e-07,
+      "loss": 0.5391,
+      "step": 77481
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.417839123261091e-07,
+      "loss": 0.7324,
+      "step": 77482
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.417593178723852e-07,
+      "loss": 0.5864,
+      "step": 77483
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.417347237695152e-07,
+      "loss": 0.4648,
+      "step": 77484
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.417101300175182e-07,
+      "loss": 0.7354,
+      "step": 77485
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.416855366164123e-07,
+      "loss": 0.6997,
+      "step": 77486
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.416609435662172e-07,
+      "loss": 0.6777,
+      "step": 77487
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.416363508669508e-07,
+      "loss": 0.6382,
+      "step": 77488
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.416117585186324e-07,
+      "loss": 0.6357,
+      "step": 77489
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.41587166521281e-07,
+      "loss": 0.5415,
+      "step": 77490
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.415625748749151e-07,
+      "loss": 0.6348,
+      "step": 77491
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.415379835795534e-07,
+      "loss": 0.5791,
+      "step": 77492
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.415133926352151e-07,
+      "loss": 0.457,
+      "step": 77493
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.414888020419193e-07,
+      "loss": 0.5962,
+      "step": 77494
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.414642117996841e-07,
+      "loss": 0.7637,
+      "step": 77495
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.414396219085287e-07,
+      "loss": 0.7871,
+      "step": 77496
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.414150323684721e-07,
+      "loss": 0.5674,
+      "step": 77497
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.41390443179533e-07,
+      "loss": 0.4844,
+      "step": 77498
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.413658543417299e-07,
+      "loss": 0.572,
+      "step": 77499
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.413412658550818e-07,
+      "loss": 0.5977,
+      "step": 77500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.413166777196076e-07,
+      "loss": 0.6475,
+      "step": 77501
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.412920899353264e-07,
+      "loss": 0.6367,
+      "step": 77502
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.412675025022565e-07,
+      "loss": 0.6494,
+      "step": 77503
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.41242915420417e-07,
+      "loss": 0.5537,
+      "step": 77504
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.412183286898266e-07,
+      "loss": 0.7959,
+      "step": 77505
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.411937423105047e-07,
+      "loss": 0.5447,
+      "step": 77506
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.411691562824696e-07,
+      "loss": 0.5601,
+      "step": 77507
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.411445706057398e-07,
+      "loss": 0.5403,
+      "step": 77508
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.411199852803345e-07,
+      "loss": 0.5043,
+      "step": 77509
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.410954003062725e-07,
+      "loss": 0.4541,
+      "step": 77510
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.410708156835732e-07,
+      "loss": 0.4397,
+      "step": 77511
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.410462314122544e-07,
+      "loss": 0.562,
+      "step": 77512
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.410216474923352e-07,
+      "loss": 0.6943,
+      "step": 77513
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.409970639238348e-07,
+      "loss": 0.6177,
+      "step": 77514
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.409724807067723e-07,
+      "loss": 0.6377,
+      "step": 77515
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.409478978411655e-07,
+      "loss": 0.7017,
+      "step": 77516
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.409233153270344e-07,
+      "loss": 0.6709,
+      "step": 77517
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.408987331643966e-07,
+      "loss": 0.5337,
+      "step": 77518
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40874151353272e-07,
+      "loss": 0.4304,
+      "step": 77519
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.408495698936786e-07,
+      "loss": 0.509,
+      "step": 77520
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.408249887856355e-07,
+      "loss": 0.6294,
+      "step": 77521
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.408004080291617e-07,
+      "loss": 0.647,
+      "step": 77522
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.407758276242762e-07,
+      "loss": 0.6543,
+      "step": 77523
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.407512475709972e-07,
+      "loss": 0.3071,
+      "step": 77524
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.407266678693438e-07,
+      "loss": 0.6348,
+      "step": 77525
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40702088519335e-07,
+      "loss": 0.6172,
+      "step": 77526
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.406775095209898e-07,
+      "loss": 0.554,
+      "step": 77527
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.406529308743266e-07,
+      "loss": 0.5251,
+      "step": 77528
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.406283525793641e-07,
+      "loss": 0.6006,
+      "step": 77529
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.406037746361214e-07,
+      "loss": 0.5483,
+      "step": 77530
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40579197044617e-07,
+      "loss": 0.6953,
+      "step": 77531
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.405546198048706e-07,
+      "loss": 0.6719,
+      "step": 77532
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.405300429169e-07,
+      "loss": 0.6323,
+      "step": 77533
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.405054663807242e-07,
+      "loss": 0.6514,
+      "step": 77534
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.404808901963629e-07,
+      "loss": 0.6118,
+      "step": 77535
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.404563143638337e-07,
+      "loss": 0.5005,
+      "step": 77536
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.404317388831566e-07,
+      "loss": 0.5264,
+      "step": 77537
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40407163754349e-07,
+      "loss": 0.5947,
+      "step": 77538
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.403825889774307e-07,
+      "loss": 0.6934,
+      "step": 77539
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.403580145524207e-07,
+      "loss": 0.8203,
+      "step": 77540
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.403334404793371e-07,
+      "loss": 0.5659,
+      "step": 77541
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40308866758199e-07,
+      "loss": 0.4422,
+      "step": 77542
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.402842933890252e-07,
+      "loss": 0.6235,
+      "step": 77543
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.402597203718351e-07,
+      "loss": 0.707,
+      "step": 77544
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.402351477066466e-07,
+      "loss": 0.5084,
+      "step": 77545
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.402105753934788e-07,
+      "loss": 0.7314,
+      "step": 77546
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.401860034323511e-07,
+      "loss": 0.4905,
+      "step": 77547
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.401614318232814e-07,
+      "loss": 0.6904,
+      "step": 77548
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.401368605662893e-07,
+      "loss": 0.7383,
+      "step": 77549
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40112289661393e-07,
+      "loss": 0.7559,
+      "step": 77550
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.400877191086115e-07,
+      "loss": 0.583,
+      "step": 77551
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.40063148907964e-07,
+      "loss": 0.7002,
+      "step": 77552
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.400385790594686e-07,
+      "loss": 0.7432,
+      "step": 77553
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.400140095631447e-07,
+      "loss": 0.6348,
+      "step": 77554
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.399894404190107e-07,
+      "loss": 0.5957,
+      "step": 77555
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.399648716270862e-07,
+      "loss": 0.5229,
+      "step": 77556
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.399403031873893e-07,
+      "loss": 0.522,
+      "step": 77557
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.399157350999386e-07,
+      "loss": 0.6953,
+      "step": 77558
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.398911673647534e-07,
+      "loss": 0.5742,
+      "step": 77559
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.398665999818523e-07,
+      "loss": 0.7056,
+      "step": 77560
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.398420329512545e-07,
+      "loss": 0.6172,
+      "step": 77561
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.398174662729782e-07,
+      "loss": 0.8154,
+      "step": 77562
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.397928999470424e-07,
+      "loss": 0.6025,
+      "step": 77563
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.397683339734661e-07,
+      "loss": 0.668,
+      "step": 77564
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.397437683522685e-07,
+      "loss": 0.5317,
+      "step": 77565
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.397192030834675e-07,
+      "loss": 0.4374,
+      "step": 77566
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.396946381670826e-07,
+      "loss": 0.5771,
+      "step": 77567
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.39670073603132e-07,
+      "loss": 0.6602,
+      "step": 77568
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.396455093916349e-07,
+      "loss": 0.625,
+      "step": 77569
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.396209455326103e-07,
+      "loss": 0.4783,
+      "step": 77570
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.395963820260767e-07,
+      "loss": 0.5957,
+      "step": 77571
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.395718188720529e-07,
+      "loss": 0.6411,
+      "step": 77572
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.395472560705579e-07,
+      "loss": 0.6475,
+      "step": 77573
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.395226936216102e-07,
+      "loss": 0.6377,
+      "step": 77574
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.394981315252288e-07,
+      "loss": 0.4078,
+      "step": 77575
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.394735697814325e-07,
+      "loss": 0.5981,
+      "step": 77576
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.394490083902406e-07,
+      "loss": 0.6641,
+      "step": 77577
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.394244473516713e-07,
+      "loss": 0.624,
+      "step": 77578
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.393998866657432e-07,
+      "loss": 0.6099,
+      "step": 77579
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.393753263324753e-07,
+      "loss": 0.6787,
+      "step": 77580
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.393507663518866e-07,
+      "loss": 0.5869,
+      "step": 77581
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.393262067239963e-07,
+      "loss": 0.6104,
+      "step": 77582
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.393016474488222e-07,
+      "loss": 0.5713,
+      "step": 77583
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.392770885263837e-07,
+      "loss": 0.6318,
+      "step": 77584
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.392525299566997e-07,
+      "loss": 0.5752,
+      "step": 77585
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.392279717397891e-07,
+      "loss": 0.6777,
+      "step": 77586
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.392034138756704e-07,
+      "loss": 0.7695,
+      "step": 77587
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.391788563643623e-07,
+      "loss": 0.7275,
+      "step": 77588
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.391542992058835e-07,
+      "loss": 0.6626,
+      "step": 77589
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.391297424002536e-07,
+      "loss": 0.6157,
+      "step": 77590
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.391051859474904e-07,
+      "loss": 0.5293,
+      "step": 77591
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.390806298476133e-07,
+      "loss": 0.6689,
+      "step": 77592
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.390560741006408e-07,
+      "loss": 0.563,
+      "step": 77593
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.390315187065924e-07,
+      "loss": 0.52,
+      "step": 77594
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.390069636654859e-07,
+      "loss": 0.6836,
+      "step": 77595
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.389824089773405e-07,
+      "loss": 0.584,
+      "step": 77596
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.389578546421755e-07,
+      "loss": 0.6167,
+      "step": 77597
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.38933300660009e-07,
+      "loss": 0.6494,
+      "step": 77598
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.389087470308604e-07,
+      "loss": 0.6812,
+      "step": 77599
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.388841937547478e-07,
+      "loss": 0.5444,
+      "step": 77600
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.388596408316903e-07,
+      "loss": 0.791,
+      "step": 77601
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.388350882617068e-07,
+      "loss": 0.6553,
+      "step": 77602
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.388105360448165e-07,
+      "loss": 0.5908,
+      "step": 77603
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.387859841810374e-07,
+      "loss": 0.626,
+      "step": 77604
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.387614326703886e-07,
+      "loss": 0.7793,
+      "step": 77605
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.38736881512889e-07,
+      "loss": 0.4219,
+      "step": 77606
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.387123307085581e-07,
+      "loss": 0.549,
+      "step": 77607
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.386877802574132e-07,
+      "loss": 0.5474,
+      "step": 77608
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.386632301594738e-07,
+      "loss": 0.6875,
+      "step": 77609
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.386386804147587e-07,
+      "loss": 0.6396,
+      "step": 77610
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.386141310232872e-07,
+      "loss": 0.6865,
+      "step": 77611
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.385895819850774e-07,
+      "loss": 0.8105,
+      "step": 77612
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.385650333001483e-07,
+      "loss": 0.543,
+      "step": 77613
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.385404849685186e-07,
+      "loss": 0.6567,
+      "step": 77614
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.385159369902077e-07,
+      "loss": 0.6426,
+      "step": 77615
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.384913893652334e-07,
+      "loss": 0.4976,
+      "step": 77616
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.384668420936156e-07,
+      "loss": 0.7783,
+      "step": 77617
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.384422951753721e-07,
+      "loss": 0.6113,
+      "step": 77618
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.384177486105221e-07,
+      "loss": 0.7949,
+      "step": 77619
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.383932023990847e-07,
+      "loss": 0.6143,
+      "step": 77620
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.383686565410781e-07,
+      "loss": 0.4849,
+      "step": 77621
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.383441110365214e-07,
+      "loss": 0.6514,
+      "step": 77622
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.383195658854333e-07,
+      "loss": 0.6641,
+      "step": 77623
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.382950210878332e-07,
+      "loss": 0.5542,
+      "step": 77624
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.382704766437389e-07,
+      "loss": 0.6284,
+      "step": 77625
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.382459325531696e-07,
+      "loss": 0.6299,
+      "step": 77626
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.382213888161447e-07,
+      "loss": 0.5728,
+      "step": 77627
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.381968454326824e-07,
+      "loss": 0.5239,
+      "step": 77628
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.381723024028011e-07,
+      "loss": 0.592,
+      "step": 77629
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.381477597265201e-07,
+      "loss": 0.7051,
+      "step": 77630
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.381232174038581e-07,
+      "loss": 0.5601,
+      "step": 77631
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.380986754348343e-07,
+      "loss": 0.5276,
+      "step": 77632
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.380741338194666e-07,
+      "loss": 0.8984,
+      "step": 77633
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.380495925577745e-07,
+      "loss": 0.7607,
+      "step": 77634
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.380250516497764e-07,
+      "loss": 0.5259,
+      "step": 77635
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.380005110954919e-07,
+      "loss": 0.5864,
+      "step": 77636
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.37975970894939e-07,
+      "loss": 0.6431,
+      "step": 77637
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.379514310481362e-07,
+      "loss": 0.6943,
+      "step": 77638
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.379268915551028e-07,
+      "loss": 0.6719,
+      "step": 77639
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.379023524158575e-07,
+      "loss": 0.5046,
+      "step": 77640
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.378778136304196e-07,
+      "loss": 0.4834,
+      "step": 77641
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.37853275198807e-07,
+      "loss": 0.8223,
+      "step": 77642
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.378287371210389e-07,
+      "loss": 0.606,
+      "step": 77643
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.378041993971345e-07,
+      "loss": 0.624,
+      "step": 77644
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.377796620271118e-07,
+      "loss": 0.5122,
+      "step": 77645
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.377551250109898e-07,
+      "loss": 0.6201,
+      "step": 77646
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.377305883487879e-07,
+      "loss": 0.5781,
+      "step": 77647
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.377060520405241e-07,
+      "loss": 0.7705,
+      "step": 77648
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.376815160862179e-07,
+      "loss": 0.6143,
+      "step": 77649
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.376569804858874e-07,
+      "loss": 0.438,
+      "step": 77650
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.376324452395515e-07,
+      "loss": 0.624,
+      "step": 77651
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.376079103472293e-07,
+      "loss": 0.6738,
+      "step": 77652
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.375833758089399e-07,
+      "loss": 0.5078,
+      "step": 77653
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.375588416247011e-07,
+      "loss": 0.7031,
+      "step": 77654
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.375343077945323e-07,
+      "loss": 0.6396,
+      "step": 77655
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.375097743184522e-07,
+      "loss": 0.6387,
+      "step": 77656
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.374852411964801e-07,
+      "loss": 0.7529,
+      "step": 77657
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.374607084286343e-07,
+      "loss": 0.707,
+      "step": 77658
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.37436176014933e-07,
+      "loss": 0.7393,
+      "step": 77659
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.374116439553956e-07,
+      "loss": 0.6846,
+      "step": 77660
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.373871122500413e-07,
+      "loss": 0.7246,
+      "step": 77661
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.37362580898888e-07,
+      "loss": 0.7109,
+      "step": 77662
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.373380499019549e-07,
+      "loss": 0.6597,
+      "step": 77663
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.373135192592606e-07,
+      "loss": 0.6025,
+      "step": 77664
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.372889889708248e-07,
+      "loss": 0.5171,
+      "step": 77665
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.372644590366648e-07,
+      "loss": 0.7004,
+      "step": 77666
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.372399294568008e-07,
+      "loss": 0.449,
+      "step": 77667
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.372154002312503e-07,
+      "loss": 0.6245,
+      "step": 77668
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.371908713600328e-07,
+      "loss": 0.4824,
+      "step": 77669
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.371663428431674e-07,
+      "loss": 0.4662,
+      "step": 77670
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.371418146806719e-07,
+      "loss": 0.6396,
+      "step": 77671
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.371172868725658e-07,
+      "loss": 0.6299,
+      "step": 77672
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.370927594188676e-07,
+      "loss": 0.5469,
+      "step": 77673
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.370682323195967e-07,
+      "loss": 0.6699,
+      "step": 77674
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.37043705574771e-07,
+      "loss": 0.7949,
+      "step": 77675
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.370191791844095e-07,
+      "loss": 0.5325,
+      "step": 77676
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.369946531485316e-07,
+      "loss": 0.7412,
+      "step": 77677
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.369701274671556e-07,
+      "loss": 0.6543,
+      "step": 77678
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.369456021402998e-07,
+      "loss": 0.533,
+      "step": 77679
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.369210771679835e-07,
+      "loss": 0.5176,
+      "step": 77680
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.368965525502255e-07,
+      "loss": 0.4816,
+      "step": 77681
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.368720282870449e-07,
+      "loss": 0.8022,
+      "step": 77682
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.368475043784598e-07,
+      "loss": 0.7085,
+      "step": 77683
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.368229808244891e-07,
+      "loss": 0.647,
+      "step": 77684
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.367984576251516e-07,
+      "loss": 0.4319,
+      "step": 77685
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.367739347804669e-07,
+      "loss": 0.6074,
+      "step": 77686
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.36749412290453e-07,
+      "loss": 0.5918,
+      "step": 77687
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.367248901551284e-07,
+      "loss": 0.7114,
+      "step": 77688
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.367003683745122e-07,
+      "loss": 0.6924,
+      "step": 77689
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.366758469486233e-07,
+      "loss": 0.7559,
+      "step": 77690
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.366513258774807e-07,
+      "loss": 0.8516,
+      "step": 77691
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.366268051611026e-07,
+      "loss": 0.6475,
+      "step": 77692
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.366022847995079e-07,
+      "loss": 0.582,
+      "step": 77693
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.365777647927157e-07,
+      "loss": 0.707,
+      "step": 77694
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.365532451407449e-07,
+      "loss": 0.6616,
+      "step": 77695
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.365287258436136e-07,
+      "loss": 0.7437,
+      "step": 77696
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.365042069013413e-07,
+      "loss": 0.5894,
+      "step": 77697
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.36479688313946e-07,
+      "loss": 0.6455,
+      "step": 77698
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.364551700814473e-07,
+      "loss": 0.6426,
+      "step": 77699
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.364306522038632e-07,
+      "loss": 0.6035,
+      "step": 77700
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.364061346812128e-07,
+      "loss": 0.5637,
+      "step": 77701
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.36381617513515e-07,
+      "loss": 0.5757,
+      "step": 77702
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.363571007007888e-07,
+      "loss": 0.5801,
+      "step": 77703
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.363325842430522e-07,
+      "loss": 0.6826,
+      "step": 77704
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.363080681403245e-07,
+      "loss": 0.4089,
+      "step": 77705
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.362835523926243e-07,
+      "loss": 0.5908,
+      "step": 77706
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.362590369999709e-07,
+      "loss": 0.7139,
+      "step": 77707
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.362345219623828e-07,
+      "loss": 0.5435,
+      "step": 77708
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.362100072798778e-07,
+      "loss": 0.5439,
+      "step": 77709
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.361854929524759e-07,
+      "loss": 0.479,
+      "step": 77710
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.361609789801953e-07,
+      "loss": 0.6973,
+      "step": 77711
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.361364653630552e-07,
+      "loss": 0.7241,
+      "step": 77712
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.361119521010738e-07,
+      "loss": 0.6157,
+      "step": 77713
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.3608743919427e-07,
+      "loss": 0.4366,
+      "step": 77714
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.360629266426633e-07,
+      "loss": 0.6479,
+      "step": 77715
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.360384144462714e-07,
+      "loss": 0.5645,
+      "step": 77716
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.36013902605114e-07,
+      "loss": 0.5439,
+      "step": 77717
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.359893911192091e-07,
+      "loss": 0.6641,
+      "step": 77718
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.359648799885755e-07,
+      "loss": 0.688,
+      "step": 77719
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.35940369213233e-07,
+      "loss": 0.6914,
+      "step": 77720
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.359158587931989e-07,
+      "loss": 0.6855,
+      "step": 77721
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.358913487284929e-07,
+      "loss": 0.7686,
+      "step": 77722
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.358668390191334e-07,
+      "loss": 0.5527,
+      "step": 77723
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.358423296651398e-07,
+      "loss": 0.5925,
+      "step": 77724
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.3581782066653e-07,
+      "loss": 0.6768,
+      "step": 77725
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.357933120233236e-07,
+      "loss": 0.6211,
+      "step": 77726
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.357688037355385e-07,
+      "loss": 0.564,
+      "step": 77727
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.357442958031938e-07,
+      "loss": 0.562,
+      "step": 77728
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.357197882263087e-07,
+      "loss": 0.6665,
+      "step": 77729
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.356952810049014e-07,
+      "loss": 0.7324,
+      "step": 77730
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.356707741389908e-07,
+      "loss": 0.6885,
+      "step": 77731
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.356462676285957e-07,
+      "loss": 0.5381,
+      "step": 77732
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.356217614737353e-07,
+      "loss": 0.5681,
+      "step": 77733
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.355972556744274e-07,
+      "loss": 0.5967,
+      "step": 77734
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.355727502306917e-07,
+      "loss": 0.6846,
+      "step": 77735
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.355482451425467e-07,
+      "loss": 0.7793,
+      "step": 77736
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.355237404100111e-07,
+      "loss": 0.5879,
+      "step": 77737
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.354992360331033e-07,
+      "loss": 0.4175,
+      "step": 77738
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.354747320118424e-07,
+      "loss": 0.5781,
+      "step": 77739
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.35450228346247e-07,
+      "loss": 0.6157,
+      "step": 77740
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.354257250363366e-07,
+      "loss": 0.6841,
+      "step": 77741
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.354012220821287e-07,
+      "loss": 0.7021,
+      "step": 77742
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.353767194836429e-07,
+      "loss": 0.6846,
+      "step": 77743
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.353522172408977e-07,
+      "loss": 0.6621,
+      "step": 77744
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.353277153539124e-07,
+      "loss": 0.6533,
+      "step": 77745
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.353032138227054e-07,
+      "loss": 0.3726,
+      "step": 77746
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.352787126472947e-07,
+      "loss": 0.4663,
+      "step": 77747
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.352542118277e-07,
+      "loss": 0.6626,
+      "step": 77748
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.352297113639395e-07,
+      "loss": 0.4814,
+      "step": 77749
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.352052112560328e-07,
+      "loss": 0.5896,
+      "step": 77750
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.351807115039977e-07,
+      "loss": 0.4675,
+      "step": 77751
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.351562121078534e-07,
+      "loss": 0.562,
+      "step": 77752
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.35131713067619e-07,
+      "loss": 0.6699,
+      "step": 77753
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.351072143833125e-07,
+      "loss": 0.6631,
+      "step": 77754
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.350827160549528e-07,
+      "loss": 0.6797,
+      "step": 77755
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.350582180825595e-07,
+      "loss": 0.6069,
+      "step": 77756
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.350337204661503e-07,
+      "loss": 0.4756,
+      "step": 77757
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.350092232057447e-07,
+      "loss": 0.6929,
+      "step": 77758
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.349847263013609e-07,
+      "loss": 0.6223,
+      "step": 77759
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.349602297530178e-07,
+      "loss": 0.5869,
+      "step": 77760
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.349357335607343e-07,
+      "loss": 0.7295,
+      "step": 77761
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.349112377245296e-07,
+      "loss": 0.6826,
+      "step": 77762
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.348867422444216e-07,
+      "loss": 0.6094,
+      "step": 77763
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.348622471204294e-07,
+      "loss": 0.5029,
+      "step": 77764
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.348377523525716e-07,
+      "loss": 0.5737,
+      "step": 77765
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.348132579408678e-07,
+      "loss": 0.792,
+      "step": 77766
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.347887638853359e-07,
+      "loss": 0.7168,
+      "step": 77767
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.347642701859946e-07,
+      "loss": 0.6973,
+      "step": 77768
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.347397768428629e-07,
+      "loss": 0.7832,
+      "step": 77769
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.347152838559598e-07,
+      "loss": 0.6333,
+      "step": 77770
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.346907912253036e-07,
+      "loss": 0.6396,
+      "step": 77771
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.346662989509131e-07,
+      "loss": 0.7285,
+      "step": 77772
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.346418070328074e-07,
+      "loss": 0.5414,
+      "step": 77773
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.346173154710054e-07,
+      "loss": 0.7764,
+      "step": 77774
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.345928242655251e-07,
+      "loss": 0.8076,
+      "step": 77775
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.345683334163861e-07,
+      "loss": 0.5205,
+      "step": 77776
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.345438429236062e-07,
+      "loss": 0.5845,
+      "step": 77777
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.345193527872049e-07,
+      "loss": 0.5256,
+      "step": 77778
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.344948630072011e-07,
+      "loss": 0.6208,
+      "step": 77779
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.344703735836126e-07,
+      "loss": 0.5669,
+      "step": 77780
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.344458845164588e-07,
+      "loss": 0.7134,
+      "step": 77781
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.344213958057584e-07,
+      "loss": 0.6543,
+      "step": 77782
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.343969074515305e-07,
+      "loss": 0.6895,
+      "step": 77783
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.343724194537931e-07,
+      "loss": 0.9097,
+      "step": 77784
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.343479318125654e-07,
+      "loss": 0.5679,
+      "step": 77785
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.343234445278664e-07,
+      "loss": 0.5845,
+      "step": 77786
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.342989575997145e-07,
+      "loss": 0.5967,
+      "step": 77787
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.342744710281282e-07,
+      "loss": 0.7021,
+      "step": 77788
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.342499848131264e-07,
+      "loss": 0.6479,
+      "step": 77789
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.342254989547281e-07,
+      "loss": 0.4297,
+      "step": 77790
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.342010134529521e-07,
+      "loss": 0.4536,
+      "step": 77791
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.341765283078167e-07,
+      "loss": 0.6719,
+      "step": 77792
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.34152043519341e-07,
+      "loss": 0.5107,
+      "step": 77793
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.341275590875435e-07,
+      "loss": 0.6548,
+      "step": 77794
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.341030750124436e-07,
+      "loss": 0.6826,
+      "step": 77795
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.340785912940595e-07,
+      "loss": 0.4045,
+      "step": 77796
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.340541079324097e-07,
+      "loss": 0.6836,
+      "step": 77797
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.340296249275131e-07,
+      "loss": 0.5835,
+      "step": 77798
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.340051422793886e-07,
+      "loss": 0.668,
+      "step": 77799
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.339806599880554e-07,
+      "loss": 0.7578,
+      "step": 77800
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.339561780535314e-07,
+      "loss": 0.4971,
+      "step": 77801
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.339316964758357e-07,
+      "loss": 0.3479,
+      "step": 77802
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.339072152549871e-07,
+      "loss": 0.5264,
+      "step": 77803
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.338827343910046e-07,
+      "loss": 0.7139,
+      "step": 77804
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.338582538839065e-07,
+      "loss": 0.666,
+      "step": 77805
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.338337737337119e-07,
+      "loss": 0.6143,
+      "step": 77806
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.33809293940439e-07,
+      "loss": 0.6807,
+      "step": 77807
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.337848145041073e-07,
+      "loss": 0.6523,
+      "step": 77808
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.337603354247346e-07,
+      "loss": 0.6963,
+      "step": 77809
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.337358567023403e-07,
+      "loss": 0.8076,
+      "step": 77810
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.33711378336943e-07,
+      "loss": 0.5547,
+      "step": 77811
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.336869003285617e-07,
+      "loss": 0.668,
+      "step": 77812
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.336624226772148e-07,
+      "loss": 0.5579,
+      "step": 77813
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.336379453829209e-07,
+      "loss": 0.6924,
+      "step": 77814
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.336134684456992e-07,
+      "loss": 0.584,
+      "step": 77815
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.335889918655685e-07,
+      "loss": 0.7041,
+      "step": 77816
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.335645156425472e-07,
+      "loss": 0.6279,
+      "step": 77817
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.335400397766539e-07,
+      "loss": 0.6885,
+      "step": 77818
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.335155642679072e-07,
+      "loss": 0.6387,
+      "step": 77819
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.334910891163265e-07,
+      "loss": 0.5029,
+      "step": 77820
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.334666143219304e-07,
+      "loss": 0.6943,
+      "step": 77821
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.334421398847373e-07,
+      "loss": 0.6357,
+      "step": 77822
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.33417665804766e-07,
+      "loss": 0.5591,
+      "step": 77823
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.333931920820356e-07,
+      "loss": 0.5303,
+      "step": 77824
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.333687187165643e-07,
+      "loss": 0.5498,
+      "step": 77825
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.333442457083715e-07,
+      "loss": 0.6753,
+      "step": 77826
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.333197730574752e-07,
+      "loss": 0.5693,
+      "step": 77827
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.332953007638944e-07,
+      "loss": 0.6855,
+      "step": 77828
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.332708288276484e-07,
+      "loss": 0.4822,
+      "step": 77829
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.332463572487551e-07,
+      "loss": 0.4009,
+      "step": 77830
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.332218860272335e-07,
+      "loss": 0.6382,
+      "step": 77831
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.331974151631024e-07,
+      "loss": 0.6963,
+      "step": 77832
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.331729446563811e-07,
+      "loss": 0.6895,
+      "step": 77833
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.331484745070873e-07,
+      "loss": 0.6616,
+      "step": 77834
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.331240047152404e-07,
+      "loss": 0.666,
+      "step": 77835
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 5.330995352808594e-07,
+      "loss": 0.5493,
+      "step": 77836
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.33075066203962e-07,
+      "loss": 0.5923,
+      "step": 77837
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.33050597484568e-07,
+      "loss": 0.6025,
+      "step": 77838
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.330261291226954e-07,
+      "loss": 0.626,
+      "step": 77839
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.330016611183632e-07,
+      "loss": 0.5688,
+      "step": 77840
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.329771934715906e-07,
+      "loss": 0.7881,
+      "step": 77841
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.329527261823954e-07,
+      "loss": 0.6357,
+      "step": 77842
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.329282592507969e-07,
+      "loss": 0.7026,
+      "step": 77843
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.329037926768136e-07,
+      "loss": 0.4946,
+      "step": 77844
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.32879326460465e-07,
+      "loss": 0.5747,
+      "step": 77845
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.32854860601769e-07,
+      "loss": 0.5571,
+      "step": 77846
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.328303951007442e-07,
+      "loss": 0.7549,
+      "step": 77847
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.328059299574098e-07,
+      "loss": 0.5144,
+      "step": 77848
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.327814651717843e-07,
+      "loss": 0.792,
+      "step": 77849
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.327570007438871e-07,
+      "loss": 0.6138,
+      "step": 77850
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.327325366737359e-07,
+      "loss": 0.5035,
+      "step": 77851
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.327080729613497e-07,
+      "loss": 0.8301,
+      "step": 77852
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.326836096067477e-07,
+      "loss": 0.7725,
+      "step": 77853
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.326591466099488e-07,
+      "loss": 0.6504,
+      "step": 77854
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.326346839709708e-07,
+      "loss": 0.5898,
+      "step": 77855
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.326102216898334e-07,
+      "loss": 0.7393,
+      "step": 77856
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.325857597665544e-07,
+      "loss": 0.5576,
+      "step": 77857
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.325612982011534e-07,
+      "loss": 0.6865,
+      "step": 77858
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.325368369936483e-07,
+      "loss": 0.6208,
+      "step": 77859
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.325123761440583e-07,
+      "loss": 0.6553,
+      "step": 77860
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.32487915652402e-07,
+      "loss": 0.563,
+      "step": 77861
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.324634555186986e-07,
+      "loss": 0.668,
+      "step": 77862
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.324389957429662e-07,
+      "loss": 0.7969,
+      "step": 77863
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.324145363252237e-07,
+      "loss": 0.501,
+      "step": 77864
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.323900772654899e-07,
+      "loss": 0.6748,
+      "step": 77865
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.323656185637837e-07,
+      "loss": 0.582,
+      "step": 77866
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.323411602201238e-07,
+      "loss": 0.7529,
+      "step": 77867
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.323167022345285e-07,
+      "loss": 0.6553,
+      "step": 77868
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.322922446070166e-07,
+      "loss": 0.5688,
+      "step": 77869
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.322677873376071e-07,
+      "loss": 0.5898,
+      "step": 77870
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.322433304263191e-07,
+      "loss": 0.6797,
+      "step": 77871
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.322188738731703e-07,
+      "loss": 0.5088,
+      "step": 77872
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.321944176781801e-07,
+      "loss": 0.6104,
+      "step": 77873
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.321699618413673e-07,
+      "loss": 0.5356,
+      "step": 77874
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.321455063627507e-07,
+      "loss": 0.4834,
+      "step": 77875
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.321210512423486e-07,
+      "loss": 0.5139,
+      "step": 77876
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.320965964801798e-07,
+      "loss": 0.5737,
+      "step": 77877
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.320721420762629e-07,
+      "loss": 0.5469,
+      "step": 77878
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.320476880306173e-07,
+      "loss": 0.5881,
+      "step": 77879
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.320232343432609e-07,
+      "loss": 0.7422,
+      "step": 77880
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.319987810142128e-07,
+      "loss": 0.6777,
+      "step": 77881
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.319743280434917e-07,
+      "loss": 0.5546,
+      "step": 77882
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.319498754311167e-07,
+      "loss": 0.624,
+      "step": 77883
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.319254231771058e-07,
+      "loss": 0.623,
+      "step": 77884
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.319009712814781e-07,
+      "loss": 0.6318,
+      "step": 77885
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.318765197442527e-07,
+      "loss": 0.5859,
+      "step": 77886
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.318520685654475e-07,
+      "loss": 0.5762,
+      "step": 77887
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.31827617745082e-07,
+      "loss": 0.5381,
+      "step": 77888
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.318031672831742e-07,
+      "loss": 0.5127,
+      "step": 77889
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.317787171797432e-07,
+      "loss": 0.5723,
+      "step": 77890
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.317542674348078e-07,
+      "loss": 0.749,
+      "step": 77891
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.31729818048387e-07,
+      "loss": 0.6245,
+      "step": 77892
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.317053690204985e-07,
+      "loss": 0.6758,
+      "step": 77893
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.316809203511618e-07,
+      "loss": 0.5664,
+      "step": 77894
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.316564720403956e-07,
+      "loss": 0.5298,
+      "step": 77895
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.316320240882193e-07,
+      "loss": 0.6416,
+      "step": 77896
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.3160757649465e-07,
+      "loss": 0.437,
+      "step": 77897
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.315831292597072e-07,
+      "loss": 0.707,
+      "step": 77898
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.315586823834096e-07,
+      "loss": 0.5288,
+      "step": 77899
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.315342358657764e-07,
+      "loss": 0.5352,
+      "step": 77900
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.315097897068255e-07,
+      "loss": 0.4922,
+      "step": 77901
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.314853439065761e-07,
+      "loss": 0.6973,
+      "step": 77902
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.314608984650469e-07,
+      "loss": 0.3765,
+      "step": 77903
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.314364533822566e-07,
+      "loss": 0.6943,
+      "step": 77904
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.314120086582237e-07,
+      "loss": 0.5933,
+      "step": 77905
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.313875642929675e-07,
+      "loss": 0.3682,
+      "step": 77906
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.313631202865059e-07,
+      "loss": 0.6025,
+      "step": 77907
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.313386766388579e-07,
+      "loss": 0.6226,
+      "step": 77908
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.313142333500429e-07,
+      "loss": 0.5596,
+      "step": 77909
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.312897904200785e-07,
+      "loss": 0.5249,
+      "step": 77910
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.31265347848984e-07,
+      "loss": 0.4507,
+      "step": 77911
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.31240905636778e-07,
+      "loss": 0.6758,
+      "step": 77912
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.312164637834797e-07,
+      "loss": 0.5771,
+      "step": 77913
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.311920222891071e-07,
+      "loss": 0.6177,
+      "step": 77914
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.311675811536791e-07,
+      "loss": 0.667,
+      "step": 77915
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.311431403772151e-07,
+      "loss": 0.5938,
+      "step": 77916
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.311186999597331e-07,
+      "loss": 0.7441,
+      "step": 77917
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.310942599012515e-07,
+      "loss": 0.5164,
+      "step": 77918
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.310698202017896e-07,
+      "loss": 0.6855,
+      "step": 77919
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.310453808613659e-07,
+      "loss": 0.4844,
+      "step": 77920
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.310209418799996e-07,
+      "loss": 0.7261,
+      "step": 77921
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.309965032577086e-07,
+      "loss": 0.5583,
+      "step": 77922
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.309720649945119e-07,
+      "loss": 0.5669,
+      "step": 77923
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.309476270904285e-07,
+      "loss": 0.4583,
+      "step": 77924
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.309231895454773e-07,
+      "loss": 0.5864,
+      "step": 77925
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.308987523596767e-07,
+      "loss": 0.5811,
+      "step": 77926
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.308743155330449e-07,
+      "loss": 0.7334,
+      "step": 77927
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.308498790656009e-07,
+      "loss": 0.5996,
+      "step": 77928
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.308254429573638e-07,
+      "loss": 0.7246,
+      "step": 77929
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.308010072083523e-07,
+      "loss": 0.3831,
+      "step": 77930
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.307765718185846e-07,
+      "loss": 0.6221,
+      "step": 77931
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.307521367880796e-07,
+      "loss": 0.5891,
+      "step": 77932
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.307277021168567e-07,
+      "loss": 0.6377,
+      "step": 77933
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.307032678049335e-07,
+      "loss": 0.52,
+      "step": 77934
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.306788338523294e-07,
+      "loss": 0.5801,
+      "step": 77935
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.30654400259063e-07,
+      "loss": 0.7056,
+      "step": 77936
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.306299670251529e-07,
+      "loss": 0.5796,
+      "step": 77937
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.306055341506179e-07,
+      "loss": 0.6011,
+      "step": 77938
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.305811016354765e-07,
+      "loss": 0.5596,
+      "step": 77939
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.305566694797476e-07,
+      "loss": 0.644,
+      "step": 77940
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.305322376834498e-07,
+      "loss": 0.7285,
+      "step": 77941
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.305078062466022e-07,
+      "loss": 0.6133,
+      "step": 77942
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.304833751692228e-07,
+      "loss": 0.6147,
+      "step": 77943
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.304589444513308e-07,
+      "loss": 0.5122,
+      "step": 77944
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.304345140929447e-07,
+      "loss": 0.5391,
+      "step": 77945
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.304100840940837e-07,
+      "loss": 0.7666,
+      "step": 77946
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.303856544547663e-07,
+      "loss": 0.7588,
+      "step": 77947
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.303612251750103e-07,
+      "loss": 0.5664,
+      "step": 77948
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.303367962548353e-07,
+      "loss": 0.5894,
+      "step": 77949
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.303123676942603e-07,
+      "loss": 0.502,
+      "step": 77950
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.30287939493303e-07,
+      "loss": 0.7266,
+      "step": 77951
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.302635116519827e-07,
+      "loss": 0.7148,
+      "step": 77952
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.302390841703179e-07,
+      "loss": 0.5723,
+      "step": 77953
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.302146570483279e-07,
+      "loss": 0.6362,
+      "step": 77954
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.301902302860306e-07,
+      "loss": 0.5718,
+      "step": 77955
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.301658038834454e-07,
+      "loss": 0.7017,
+      "step": 77956
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.301413778405901e-07,
+      "loss": 0.7026,
+      "step": 77957
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.301169521574841e-07,
+      "loss": 0.5894,
+      "step": 77958
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.300925268341463e-07,
+      "loss": 0.8467,
+      "step": 77959
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.300681018705946e-07,
+      "loss": 0.7046,
+      "step": 77960
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.300436772668481e-07,
+      "loss": 0.5105,
+      "step": 77961
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.300192530229258e-07,
+      "loss": 0.624,
+      "step": 77962
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.299948291388463e-07,
+      "loss": 0.6455,
+      "step": 77963
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.299704056146279e-07,
+      "loss": 0.7393,
+      "step": 77964
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.299459824502894e-07,
+      "loss": 0.7393,
+      "step": 77965
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.299215596458502e-07,
+      "loss": 0.4507,
+      "step": 77966
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.298971372013283e-07,
+      "loss": 0.582,
+      "step": 77967
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.298727151167423e-07,
+      "loss": 0.5176,
+      "step": 77968
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.298482933921111e-07,
+      "loss": 0.7246,
+      "step": 77969
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.298238720274535e-07,
+      "loss": 0.707,
+      "step": 77970
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.297994510227884e-07,
+      "loss": 0.6196,
+      "step": 77971
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.297750303781339e-07,
+      "loss": 0.4736,
+      "step": 77972
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.297506100935089e-07,
+      "loss": 0.5908,
+      "step": 77973
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.297261901689325e-07,
+      "loss": 0.7344,
+      "step": 77974
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.297017706044234e-07,
+      "loss": 0.5977,
+      "step": 77975
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.296773514000001e-07,
+      "loss": 0.5825,
+      "step": 77976
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.296529325556806e-07,
+      "loss": 0.5464,
+      "step": 77977
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.296285140714845e-07,
+      "loss": 0.6992,
+      "step": 77978
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.296040959474302e-07,
+      "loss": 0.6616,
+      "step": 77979
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.295796781835366e-07,
+      "loss": 0.6694,
+      "step": 77980
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.295552607798219e-07,
+      "loss": 0.6875,
+      "step": 77981
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.29530843736305e-07,
+      "loss": 0.5757,
+      "step": 77982
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.29506427053005e-07,
+      "loss": 0.6685,
+      "step": 77983
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.294820107299405e-07,
+      "loss": 0.604,
+      "step": 77984
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.294575947671296e-07,
+      "loss": 0.6211,
+      "step": 77985
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.294331791645919e-07,
+      "loss": 0.5271,
+      "step": 77986
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.29408763922345e-07,
+      "loss": 0.583,
+      "step": 77987
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.293843490404086e-07,
+      "loss": 0.7217,
+      "step": 77988
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.293599345188007e-07,
+      "loss": 0.5024,
+      "step": 77989
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.293355203575402e-07,
+      "loss": 0.645,
+      "step": 77990
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.293111065566458e-07,
+      "loss": 0.7148,
+      "step": 77991
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.292866931161367e-07,
+      "loss": 0.835,
+      "step": 77992
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.292622800360306e-07,
+      "loss": 0.6465,
+      "step": 77993
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.292378673163469e-07,
+      "loss": 0.709,
+      "step": 77994
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.292134549571039e-07,
+      "loss": 0.7129,
+      "step": 77995
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.291890429583211e-07,
+      "loss": 0.8311,
+      "step": 77996
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.291646313200165e-07,
+      "loss": 0.3918,
+      "step": 77997
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.291402200422086e-07,
+      "loss": 0.6187,
+      "step": 77998
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.291158091249164e-07,
+      "loss": 0.6047,
+      "step": 77999
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.290913985681584e-07,
+      "loss": 0.4841,
+      "step": 78000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.29066988371954e-07,
+      "loss": 0.6318,
+      "step": 78001
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.290425785363209e-07,
+      "loss": 0.6235,
+      "step": 78002
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.290181690612783e-07,
+      "loss": 0.7383,
+      "step": 78003
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.289937599468452e-07,
+      "loss": 0.6357,
+      "step": 78004
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.289693511930396e-07,
+      "loss": 0.6973,
+      "step": 78005
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.289449427998806e-07,
+      "loss": 0.7168,
+      "step": 78006
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.289205347673867e-07,
+      "loss": 0.5439,
+      "step": 78007
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.288961270955766e-07,
+      "loss": 0.6318,
+      "step": 78008
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.288717197844695e-07,
+      "loss": 0.6797,
+      "step": 78009
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.288473128340833e-07,
+      "loss": 0.7441,
+      "step": 78010
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.28822906244437e-07,
+      "loss": 0.4835,
+      "step": 78011
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.287985000155493e-07,
+      "loss": 0.6924,
+      "step": 78012
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.287740941474394e-07,
+      "loss": 0.6274,
+      "step": 78013
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.287496886401251e-07,
+      "loss": 0.6543,
+      "step": 78014
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.287252834936253e-07,
+      "loss": 0.688,
+      "step": 78015
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.287008787079595e-07,
+      "loss": 0.6445,
+      "step": 78016
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.286764742831453e-07,
+      "loss": 0.3557,
+      "step": 78017
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.286520702192022e-07,
+      "loss": 0.5977,
+      "step": 78018
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.286276665161482e-07,
+      "loss": 0.5996,
+      "step": 78019
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.286032631740023e-07,
+      "loss": 0.5674,
+      "step": 78020
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.285788601927836e-07,
+      "loss": 0.6108,
+      "step": 78021
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.285544575725101e-07,
+      "loss": 0.6543,
+      "step": 78022
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.285300553132006e-07,
+      "loss": 0.5542,
+      "step": 78023
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.28505653414874e-07,
+      "loss": 0.7598,
+      "step": 78024
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.284812518775494e-07,
+      "loss": 0.8291,
+      "step": 78025
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.28456850701245e-07,
+      "loss": 0.707,
+      "step": 78026
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.284324498859792e-07,
+      "loss": 0.7422,
+      "step": 78027
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.284080494317709e-07,
+      "loss": 0.7637,
+      "step": 78028
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.283836493386388e-07,
+      "loss": 0.7324,
+      "step": 78029
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.283592496066021e-07,
+      "loss": 0.6748,
+      "step": 78030
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.283348502356785e-07,
+      "loss": 0.6123,
+      "step": 78031
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.283104512258876e-07,
+      "loss": 0.6738,
+      "step": 78032
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.282860525772474e-07,
+      "loss": 0.4316,
+      "step": 78033
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.282616542897773e-07,
+      "loss": 0.5317,
+      "step": 78034
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.282372563634957e-07,
+      "loss": 0.2528,
+      "step": 78035
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.282128587984207e-07,
+      "loss": 0.4692,
+      "step": 78036
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.281884615945713e-07,
+      "loss": 0.4561,
+      "step": 78037
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.281640647519667e-07,
+      "loss": 0.3325,
+      "step": 78038
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.28139668270625e-07,
+      "loss": 0.4868,
+      "step": 78039
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.281152721505649e-07,
+      "loss": 0.491,
+      "step": 78040
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.280908763918053e-07,
+      "loss": 0.3757,
+      "step": 78041
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.280664809943651e-07,
+      "loss": 0.4419,
+      "step": 78042
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.280420859582624e-07,
+      "loss": 0.298,
+      "step": 78043
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.280176912835161e-07,
+      "loss": 0.603,
+      "step": 78044
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.279932969701454e-07,
+      "loss": 0.4521,
+      "step": 78045
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.279689030181681e-07,
+      "loss": 0.4775,
+      "step": 78046
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.279445094276038e-07,
+      "loss": 0.3804,
+      "step": 78047
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.279201161984701e-07,
+      "loss": 0.5171,
+      "step": 78048
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.278957233307864e-07,
+      "loss": 0.4331,
+      "step": 78049
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.278713308245713e-07,
+      "loss": 0.519,
+      "step": 78050
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.278469386798436e-07,
+      "loss": 0.4927,
+      "step": 78051
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.278225468966215e-07,
+      "loss": 0.5303,
+      "step": 78052
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.27798155474924e-07,
+      "loss": 0.5957,
+      "step": 78053
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.277737644147698e-07,
+      "loss": 0.448,
+      "step": 78054
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.277493737161782e-07,
+      "loss": 0.4421,
+      "step": 78055
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.277249833791664e-07,
+      "loss": 0.4062,
+      "step": 78056
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.277005934037539e-07,
+      "loss": 0.4736,
+      "step": 78057
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.276762037899594e-07,
+      "loss": 0.5078,
+      "step": 78058
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.276518145378017e-07,
+      "loss": 0.5005,
+      "step": 78059
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.27627425647299e-07,
+      "loss": 0.4453,
+      "step": 78060
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.276030371184703e-07,
+      "loss": 0.5425,
+      "step": 78061
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.275786489513343e-07,
+      "loss": 0.5029,
+      "step": 78062
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.275542611459098e-07,
+      "loss": 0.4268,
+      "step": 78063
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.275298737022152e-07,
+      "loss": 0.4824,
+      "step": 78064
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.275054866202692e-07,
+      "loss": 0.4678,
+      "step": 78065
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.274810999000905e-07,
+      "loss": 0.4834,
+      "step": 78066
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.274567135416976e-07,
+      "loss": 0.4829,
+      "step": 78067
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.274323275451099e-07,
+      "loss": 0.4746,
+      "step": 78068
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.274079419103451e-07,
+      "loss": 0.3428,
+      "step": 78069
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.273835566374222e-07,
+      "loss": 0.4971,
+      "step": 78070
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.273591717263601e-07,
+      "loss": 0.4116,
+      "step": 78071
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.273347871771777e-07,
+      "loss": 0.4033,
+      "step": 78072
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.273104029898929e-07,
+      "loss": 0.4004,
+      "step": 78073
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.272860191645249e-07,
+      "loss": 0.3368,
+      "step": 78074
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.272616357010926e-07,
+      "loss": 0.4941,
+      "step": 78075
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.272372525996142e-07,
+      "loss": 0.4116,
+      "step": 78076
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.272128698601083e-07,
+      "loss": 0.3885,
+      "step": 78077
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.271884874825936e-07,
+      "loss": 0.4219,
+      "step": 78078
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.27164105467089e-07,
+      "loss": 0.3994,
+      "step": 78079
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.271397238136135e-07,
+      "loss": 0.4431,
+      "step": 78080
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.271153425221851e-07,
+      "loss": 0.3655,
+      "step": 78081
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.270909615928225e-07,
+      "loss": 0.4717,
+      "step": 78082
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.270665810255448e-07,
+      "loss": 0.4053,
+      "step": 78083
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.270422008203708e-07,
+      "loss": 0.2834,
+      "step": 78084
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.270178209773188e-07,
+      "loss": 0.3918,
+      "step": 78085
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.269934414964073e-07,
+      "loss": 0.4204,
+      "step": 78086
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.269690623776548e-07,
+      "loss": 0.4399,
+      "step": 78087
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.269446836210807e-07,
+      "loss": 0.397,
+      "step": 78088
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.269203052267034e-07,
+      "loss": 0.3647,
+      "step": 78089
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.268959271945413e-07,
+      "loss": 0.5059,
+      "step": 78090
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.268715495246133e-07,
+      "loss": 0.481,
+      "step": 78091
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.268471722169377e-07,
+      "loss": 0.3955,
+      "step": 78092
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.268227952715341e-07,
+      "loss": 0.3887,
+      "step": 78093
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.267984186884201e-07,
+      "loss": 0.4668,
+      "step": 78094
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.267740424676151e-07,
+      "loss": 0.4087,
+      "step": 78095
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.267496666091371e-07,
+      "loss": 0.4761,
+      "step": 78096
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.267252911130054e-07,
+      "loss": 0.3792,
+      "step": 78097
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.267009159792382e-07,
+      "loss": 0.4546,
+      "step": 78098
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.266765412078542e-07,
+      "loss": 0.2788,
+      "step": 78099
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.266521667988723e-07,
+      "loss": 0.501,
+      "step": 78100
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.266277927523114e-07,
+      "loss": 0.4751,
+      "step": 78101
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.266034190681894e-07,
+      "loss": 0.355,
+      "step": 78102
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.265790457465256e-07,
+      "loss": 0.3735,
+      "step": 78103
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.265546727873383e-07,
+      "loss": 0.2698,
+      "step": 78104
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.265303001906468e-07,
+      "loss": 0.3696,
+      "step": 78105
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.26505927956469e-07,
+      "loss": 0.3317,
+      "step": 78106
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.264815560848236e-07,
+      "loss": 0.3982,
+      "step": 78107
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.264571845757297e-07,
+      "loss": 0.4165,
+      "step": 78108
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.264328134292055e-07,
+      "loss": 0.4966,
+      "step": 78109
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.264084426452705e-07,
+      "loss": 0.4023,
+      "step": 78110
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.263840722239424e-07,
+      "loss": 0.439,
+      "step": 78111
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.2635970216524e-07,
+      "loss": 0.4658,
+      "step": 78112
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.263353324691829e-07,
+      "loss": 0.4307,
+      "step": 78113
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.263109631357884e-07,
+      "loss": 0.4531,
+      "step": 78114
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.262865941650763e-07,
+      "loss": 0.3076,
+      "step": 78115
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.262622255570644e-07,
+      "loss": 0.3187,
+      "step": 78116
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.262378573117717e-07,
+      "loss": 0.5376,
+      "step": 78117
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.262134894292174e-07,
+      "loss": 0.302,
+      "step": 78118
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.26189121909419e-07,
+      "loss": 0.4541,
+      "step": 78119
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.26164754752396e-07,
+      "loss": 0.3804,
+      "step": 78120
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.261403879581668e-07,
+      "loss": 0.4678,
+      "step": 78121
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.261160215267506e-07,
+      "loss": 0.2551,
+      "step": 78122
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.260916554581651e-07,
+      "loss": 0.4377,
+      "step": 78123
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.260672897524294e-07,
+      "loss": 0.4534,
+      "step": 78124
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.260429244095625e-07,
+      "loss": 0.5771,
+      "step": 78125
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.260185594295825e-07,
+      "loss": 0.4453,
+      "step": 78126
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.259941948125088e-07,
+      "loss": 0.4919,
+      "step": 78127
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.259698305583589e-07,
+      "loss": 0.3728,
+      "step": 78128
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.259454666671523e-07,
+      "loss": 0.5352,
+      "step": 78129
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.259211031389077e-07,
+      "loss": 0.3974,
+      "step": 78130
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.258967399736431e-07,
+      "loss": 0.4629,
+      "step": 78131
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.258723771713778e-07,
+      "loss": 0.3584,
+      "step": 78132
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.258480147321301e-07,
+      "loss": 0.4458,
+      "step": 78133
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.25823652655919e-07,
+      "loss": 0.3635,
+      "step": 78134
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.257992909427631e-07,
+      "loss": 0.4307,
+      "step": 78135
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.257749295926803e-07,
+      "loss": 0.4602,
+      "step": 78136
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.2575056860569e-07,
+      "loss": 0.3545,
+      "step": 78137
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.257262079818108e-07,
+      "loss": 0.4172,
+      "step": 78138
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.257018477210614e-07,
+      "loss": 0.4292,
+      "step": 78139
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.2567748782346e-07,
+      "loss": 0.4927,
+      "step": 78140
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.256531282890256e-07,
+      "loss": 0.3938,
+      "step": 78141
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.256287691177765e-07,
+      "loss": 0.4976,
+      "step": 78142
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.256044103097323e-07,
+      "loss": 0.3514,
+      "step": 78143
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.255800518649105e-07,
+      "loss": 0.3755,
+      "step": 78144
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.255556937833308e-07,
+      "loss": 0.3921,
+      "step": 78145
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.255313360650107e-07,
+      "loss": 0.3545,
+      "step": 78146
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.255069787099699e-07,
+      "loss": 0.4082,
+      "step": 78147
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.254826217182262e-07,
+      "loss": 0.3723,
+      "step": 78148
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.254582650897985e-07,
+      "loss": 0.4307,
+      "step": 78149
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.254339088247057e-07,
+      "loss": 0.5298,
+      "step": 78150
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.254095529229668e-07,
+      "loss": 0.2917,
+      "step": 78151
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.253851973845995e-07,
+      "loss": 0.5732,
+      "step": 78152
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.25360842209623e-07,
+      "loss": 0.4048,
+      "step": 78153
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.253364873980559e-07,
+      "loss": 0.4658,
+      "step": 78154
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.253121329499171e-07,
+      "loss": 0.5146,
+      "step": 78155
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.25287778865225e-07,
+      "loss": 0.5723,
+      "step": 78156
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.252634251439979e-07,
+      "loss": 0.4355,
+      "step": 78157
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.252390717862547e-07,
+      "loss": 0.3872,
+      "step": 78158
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.252147187920141e-07,
+      "loss": 0.5107,
+      "step": 78159
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.251903661612953e-07,
+      "loss": 0.2463,
+      "step": 78160
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.251660138941158e-07,
+      "loss": 0.5508,
+      "step": 78161
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.25141661990495e-07,
+      "loss": 0.4663,
+      "step": 78162
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.251173104504513e-07,
+      "loss": 0.4751,
+      "step": 78163
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.250929592740039e-07,
+      "loss": 0.3884,
+      "step": 78164
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.250686084611711e-07,
+      "loss": 0.3831,
+      "step": 78165
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.250442580119708e-07,
+      "loss": 0.4253,
+      "step": 78166
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.250199079264223e-07,
+      "loss": 0.4326,
+      "step": 78167
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.249955582045447e-07,
+      "loss": 0.2772,
+      "step": 78168
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.249712088463559e-07,
+      "loss": 0.4365,
+      "step": 78169
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.249468598518746e-07,
+      "loss": 0.4316,
+      "step": 78170
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.249225112211197e-07,
+      "loss": 0.4365,
+      "step": 78171
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.248981629541101e-07,
+      "loss": 0.377,
+      "step": 78172
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.248738150508638e-07,
+      "loss": 0.4399,
+      "step": 78173
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.248494675113999e-07,
+      "loss": 0.3857,
+      "step": 78174
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.248251203357372e-07,
+      "loss": 0.3704,
+      "step": 78175
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.248007735238937e-07,
+      "loss": 0.2454,
+      "step": 78176
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.247764270758886e-07,
+      "loss": 0.4395,
+      "step": 78177
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.247520809917402e-07,
+      "loss": 0.5127,
+      "step": 78178
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.247277352714672e-07,
+      "loss": 0.4127,
+      "step": 78179
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.247033899150882e-07,
+      "loss": 0.395,
+      "step": 78180
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.246790449226225e-07,
+      "loss": 0.2844,
+      "step": 78181
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.246547002940878e-07,
+      "loss": 0.5435,
+      "step": 78182
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24630356029503e-07,
+      "loss": 0.1517,
+      "step": 78183
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.246060121288873e-07,
+      "loss": 0.4287,
+      "step": 78184
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24581668592259e-07,
+      "loss": 0.3926,
+      "step": 78185
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.245573254196362e-07,
+      "loss": 0.4087,
+      "step": 78186
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24532982611038e-07,
+      "loss": 0.3584,
+      "step": 78187
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.245086401664832e-07,
+      "loss": 0.4497,
+      "step": 78188
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.244842980859905e-07,
+      "loss": 0.4495,
+      "step": 78189
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24459956369578e-07,
+      "loss": 0.4722,
+      "step": 78190
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.244356150172647e-07,
+      "loss": 0.394,
+      "step": 78191
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.244112740290689e-07,
+      "loss": 0.3804,
+      "step": 78192
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.243869334050101e-07,
+      "loss": 0.4238,
+      "step": 78193
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24362593145106e-07,
+      "loss": 0.3882,
+      "step": 78194
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.24338253249376e-07,
+      "loss": 0.4092,
+      "step": 78195
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.243139137178378e-07,
+      "loss": 0.5234,
+      "step": 78196
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.242895745505107e-07,
+      "loss": 0.4238,
+      "step": 78197
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.242652357474135e-07,
+      "loss": 0.4067,
+      "step": 78198
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.242408973085642e-07,
+      "loss": 0.3896,
+      "step": 78199
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.242165592339817e-07,
+      "loss": 0.4438,
+      "step": 78200
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.241922215236852e-07,
+      "loss": 0.3044,
+      "step": 78201
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.241678841776923e-07,
+      "loss": 0.3755,
+      "step": 78202
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.241435471960223e-07,
+      "loss": 0.4459,
+      "step": 78203
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.241192105786935e-07,
+      "loss": 0.3599,
+      "step": 78204
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.240948743257254e-07,
+      "loss": 0.48,
+      "step": 78205
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.240705384371359e-07,
+      "loss": 0.3394,
+      "step": 78206
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.240462029129433e-07,
+      "loss": 0.4226,
+      "step": 78207
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.240218677531664e-07,
+      "loss": 0.2361,
+      "step": 78208
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.239975329578245e-07,
+      "loss": 0.4756,
+      "step": 78209
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.239731985269359e-07,
+      "loss": 0.4443,
+      "step": 78210
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.239488644605188e-07,
+      "loss": 0.3853,
+      "step": 78211
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.239245307585922e-07,
+      "loss": 0.3284,
+      "step": 78212
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.239001974211748e-07,
+      "loss": 0.4146,
+      "step": 78213
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.238758644482853e-07,
+      "loss": 0.302,
+      "step": 78214
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.238515318399422e-07,
+      "loss": 0.4492,
+      "step": 78215
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.238271995961639e-07,
+      "loss": 0.3931,
+      "step": 78216
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.23802867716969e-07,
+      "loss": 0.3782,
+      "step": 78217
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.237785362023767e-07,
+      "loss": 0.3159,
+      "step": 78218
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.23754205052405e-07,
+      "loss": 0.4565,
+      "step": 78219
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.237298742670728e-07,
+      "loss": 0.3599,
+      "step": 78220
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.237055438463989e-07,
+      "loss": 0.4351,
+      "step": 78221
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.236812137904019e-07,
+      "loss": 0.4226,
+      "step": 78222
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.236568840991e-07,
+      "loss": 0.415,
+      "step": 78223
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.236325547725121e-07,
+      "loss": 0.3713,
+      "step": 78224
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.236082258106574e-07,
+      "loss": 0.3457,
+      "step": 78225
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.235838972135533e-07,
+      "loss": 0.4497,
+      "step": 78226
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 5.235595689812195e-07,
+      "loss": 0.3303,
+      "step": 78227
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.235352411136741e-07,
+      "loss": 0.5103,
+      "step": 78228
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.235109136109357e-07,
+      "loss": 0.3788,
+      "step": 78229
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.234865864730231e-07,
+      "loss": 0.3574,
+      "step": 78230
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.234622596999554e-07,
+      "loss": 0.438,
+      "step": 78231
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.234379332917502e-07,
+      "loss": 0.3907,
+      "step": 78232
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.234136072484268e-07,
+      "loss": 0.4336,
+      "step": 78233
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.233892815700036e-07,
+      "loss": 0.4243,
+      "step": 78234
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.233649562564997e-07,
+      "loss": 0.625,
+      "step": 78235
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.233406313079334e-07,
+      "loss": 0.3159,
+      "step": 78236
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.233163067243226e-07,
+      "loss": 0.4414,
+      "step": 78237
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.23291982505687e-07,
+      "loss": 0.3918,
+      "step": 78238
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.23267658652045e-07,
+      "loss": 0.4395,
+      "step": 78239
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.232433351634147e-07,
+      "loss": 0.4565,
+      "step": 78240
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.23219012039815e-07,
+      "loss": 0.3557,
+      "step": 78241
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.231946892812645e-07,
+      "loss": 0.3911,
+      "step": 78242
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.231703668877824e-07,
+      "loss": 0.3875,
+      "step": 78243
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.231460448593864e-07,
+      "loss": 0.3394,
+      "step": 78244
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.23121723196096e-07,
+      "loss": 0.361,
+      "step": 78245
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.230974018979288e-07,
+      "loss": 0.4106,
+      "step": 78246
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.230730809649042e-07,
+      "loss": 0.3354,
+      "step": 78247
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.230487603970409e-07,
+      "loss": 0.4111,
+      "step": 78248
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.230244401943568e-07,
+      "loss": 0.5405,
+      "step": 78249
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.230001203568711e-07,
+      "loss": 0.4121,
+      "step": 78250
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.229758008846021e-07,
+      "loss": 0.4883,
+      "step": 78251
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.229514817775692e-07,
+      "loss": 0.406,
+      "step": 78252
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.229271630357899e-07,
+      "loss": 0.4976,
+      "step": 78253
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.229028446592834e-07,
+      "loss": 0.4482,
+      "step": 78254
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.228785266480685e-07,
+      "loss": 0.5229,
+      "step": 78255
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.228542090021637e-07,
+      "loss": 0.4502,
+      "step": 78256
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22829891721587e-07,
+      "loss": 0.3556,
+      "step": 78257
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.228055748063576e-07,
+      "loss": 0.5015,
+      "step": 78258
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22781258256494e-07,
+      "loss": 0.3311,
+      "step": 78259
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.227569420720152e-07,
+      "loss": 0.4963,
+      "step": 78260
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.227326262529393e-07,
+      "loss": 0.4893,
+      "step": 78261
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.227083107992847e-07,
+      "loss": 0.4653,
+      "step": 78262
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.226839957110707e-07,
+      "loss": 0.4688,
+      "step": 78263
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.226596809883159e-07,
+      "loss": 0.439,
+      "step": 78264
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.226353666310385e-07,
+      "loss": 0.3685,
+      "step": 78265
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.226110526392571e-07,
+      "loss": 0.3721,
+      "step": 78266
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.225867390129904e-07,
+      "loss": 0.5068,
+      "step": 78267
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22562425752257e-07,
+      "loss": 0.3914,
+      "step": 78268
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.225381128570761e-07,
+      "loss": 0.3552,
+      "step": 78269
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.225138003274653e-07,
+      "loss": 0.4023,
+      "step": 78270
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.224894881634438e-07,
+      "loss": 0.3076,
+      "step": 78271
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.224651763650301e-07,
+      "loss": 0.4596,
+      "step": 78272
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.224408649322433e-07,
+      "loss": 0.4731,
+      "step": 78273
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.224165538651011e-07,
+      "loss": 0.52,
+      "step": 78274
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.223922431636229e-07,
+      "loss": 0.5659,
+      "step": 78275
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.223679328278269e-07,
+      "loss": 0.4146,
+      "step": 78276
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22343622857732e-07,
+      "loss": 0.3315,
+      "step": 78277
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.223193132533563e-07,
+      "loss": 0.5737,
+      "step": 78278
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.222950040147187e-07,
+      "loss": 0.3867,
+      "step": 78279
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.222706951418378e-07,
+      "loss": 0.4194,
+      "step": 78280
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.222463866347328e-07,
+      "loss": 0.5786,
+      "step": 78281
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.222220784934213e-07,
+      "loss": 0.5259,
+      "step": 78282
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.221977707179224e-07,
+      "loss": 0.355,
+      "step": 78283
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.221734633082546e-07,
+      "loss": 0.3086,
+      "step": 78284
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.221491562644371e-07,
+      "loss": 0.3877,
+      "step": 78285
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22124849586488e-07,
+      "loss": 0.4259,
+      "step": 78286
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.221005432744256e-07,
+      "loss": 0.3643,
+      "step": 78287
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.220762373282688e-07,
+      "loss": 0.3862,
+      "step": 78288
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.22051931748036e-07,
+      "loss": 0.4277,
+      "step": 78289
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.220276265337469e-07,
+      "loss": 0.4707,
+      "step": 78290
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.220033216854184e-07,
+      "loss": 0.3848,
+      "step": 78291
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.219790172030704e-07,
+      "loss": 0.4722,
+      "step": 78292
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.219547130867212e-07,
+      "loss": 0.4409,
+      "step": 78293
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.219304093363891e-07,
+      "loss": 0.5356,
+      "step": 78294
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.219061059520932e-07,
+      "loss": 0.3457,
+      "step": 78295
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.218818029338514e-07,
+      "loss": 0.4697,
+      "step": 78296
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.218575002816827e-07,
+      "loss": 0.3975,
+      "step": 78297
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.218331979956062e-07,
+      "loss": 0.4248,
+      "step": 78298
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.218088960756397e-07,
+      "loss": 0.3457,
+      "step": 78299
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.21784594521802e-07,
+      "loss": 0.5015,
+      "step": 78300
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.217602933341119e-07,
+      "loss": 0.5278,
+      "step": 78301
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.217359925125884e-07,
+      "loss": 0.4556,
+      "step": 78302
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.21711692057249e-07,
+      "loss": 0.4375,
+      "step": 78303
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.216873919681133e-07,
+      "loss": 0.5195,
+      "step": 78304
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.216630922451999e-07,
+      "loss": 0.4531,
+      "step": 78305
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.216387928885266e-07,
+      "loss": 0.4712,
+      "step": 78306
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.216144938981128e-07,
+      "loss": 0.459,
+      "step": 78307
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.215901952739765e-07,
+      "loss": 0.2781,
+      "step": 78308
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.215658970161365e-07,
+      "loss": 0.4282,
+      "step": 78309
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.21541599124612e-07,
+      "loss": 0.375,
+      "step": 78310
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.215173015994207e-07,
+      "loss": 0.4541,
+      "step": 78311
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.214930044405817e-07,
+      "loss": 0.4385,
+      "step": 78312
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.214687076481133e-07,
+      "loss": 0.3127,
+      "step": 78313
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.214444112220346e-07,
+      "loss": 0.4001,
+      "step": 78314
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.214201151623642e-07,
+      "loss": 0.5186,
+      "step": 78315
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.2139581946912e-07,
+      "loss": 0.3094,
+      "step": 78316
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.213715241423207e-07,
+      "loss": 0.3958,
+      "step": 78317
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.213472291819855e-07,
+      "loss": 0.417,
+      "step": 78318
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.213229345881331e-07,
+      "loss": 0.4495,
+      "step": 78319
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.212986403607812e-07,
+      "loss": 0.259,
+      "step": 78320
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.212743464999489e-07,
+      "loss": 0.406,
+      "step": 78321
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.21250053005655e-07,
+      "loss": 0.4111,
+      "step": 78322
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.212257598779181e-07,
+      "loss": 0.5415,
+      "step": 78323
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.212014671167568e-07,
+      "loss": 0.3224,
+      "step": 78324
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.211771747221892e-07,
+      "loss": 0.2822,
+      "step": 78325
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.21152882694234e-07,
+      "loss": 0.376,
+      "step": 78326
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.211285910329105e-07,
+      "loss": 0.3643,
+      "step": 78327
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.211042997382364e-07,
+      "loss": 0.3276,
+      "step": 78328
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.210800088102307e-07,
+      "loss": 0.5005,
+      "step": 78329
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.210557182489121e-07,
+      "loss": 0.5151,
+      "step": 78330
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.210314280542995e-07,
+      "loss": 0.3347,
+      "step": 78331
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.210071382264108e-07,
+      "loss": 0.4585,
+      "step": 78332
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.209828487652646e-07,
+      "loss": 0.4683,
+      "step": 78333
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.209585596708806e-07,
+      "loss": 0.4521,
+      "step": 78334
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.209342709432758e-07,
+      "loss": 0.3255,
+      "step": 78335
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.209099825824703e-07,
+      "loss": 0.5591,
+      "step": 78336
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.208856945884814e-07,
+      "loss": 0.478,
+      "step": 78337
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.208614069613284e-07,
+      "loss": 0.3513,
+      "step": 78338
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.208371197010298e-07,
+      "loss": 0.48,
+      "step": 78339
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.208128328076046e-07,
+      "loss": 0.4009,
+      "step": 78340
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.207885462810706e-07,
+      "loss": 0.4883,
+      "step": 78341
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.207642601214467e-07,
+      "loss": 0.5176,
+      "step": 78342
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.207399743287516e-07,
+      "loss": 0.3972,
+      "step": 78343
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.207156889030047e-07,
+      "loss": 0.5098,
+      "step": 78344
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.206914038442229e-07,
+      "loss": 0.4985,
+      "step": 78345
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.206671191524257e-07,
+      "loss": 0.3117,
+      "step": 78346
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.206428348276315e-07,
+      "loss": 0.3259,
+      "step": 78347
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.206185508698596e-07,
+      "loss": 0.479,
+      "step": 78348
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.205942672791276e-07,
+      "loss": 0.3542,
+      "step": 78349
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.205699840554545e-07,
+      "loss": 0.3572,
+      "step": 78350
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.20545701198859e-07,
+      "loss": 0.3229,
+      "step": 78351
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.205214187093598e-07,
+      "loss": 0.4795,
+      "step": 78352
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.204971365869751e-07,
+      "loss": 0.4055,
+      "step": 78353
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.204728548317241e-07,
+      "loss": 0.4194,
+      "step": 78354
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.204485734436245e-07,
+      "loss": 0.3347,
+      "step": 78355
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.204242924226953e-07,
+      "loss": 0.4824,
+      "step": 78356
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.204000117689557e-07,
+      "loss": 0.4829,
+      "step": 78357
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.203757314824235e-07,
+      "loss": 0.4302,
+      "step": 78358
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.203514515631173e-07,
+      "loss": 0.386,
+      "step": 78359
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.20327172011056e-07,
+      "loss": 0.3938,
+      "step": 78360
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.203028928262585e-07,
+      "loss": 0.3851,
+      "step": 78361
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.202786140087427e-07,
+      "loss": 0.4199,
+      "step": 78362
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.202543355585276e-07,
+      "loss": 0.3608,
+      "step": 78363
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.202300574756319e-07,
+      "loss": 0.437,
+      "step": 78364
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.202057797600741e-07,
+      "loss": 0.3745,
+      "step": 78365
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.201815024118721e-07,
+      "loss": 0.3384,
+      "step": 78366
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.201572254310452e-07,
+      "loss": 0.4189,
+      "step": 78367
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.201329488176119e-07,
+      "loss": 0.4014,
+      "step": 78368
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.201086725715911e-07,
+      "loss": 0.4712,
+      "step": 78369
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.200843966930005e-07,
+      "loss": 0.5366,
+      "step": 78370
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.200601211818594e-07,
+      "loss": 0.395,
+      "step": 78371
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.200358460381861e-07,
+      "loss": 0.3218,
+      "step": 78372
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.200115712619997e-07,
+      "loss": 0.2944,
+      "step": 78373
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.199872968533183e-07,
+      "loss": 0.3787,
+      "step": 78374
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.199630228121602e-07,
+      "loss": 0.3656,
+      "step": 78375
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.199387491385443e-07,
+      "loss": 0.4077,
+      "step": 78376
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.199144758324894e-07,
+      "loss": 0.3892,
+      "step": 78377
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.198902028940142e-07,
+      "loss": 0.4746,
+      "step": 78378
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.198659303231366e-07,
+      "loss": 0.3911,
+      "step": 78379
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.198416581198755e-07,
+      "loss": 0.4805,
+      "step": 78380
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.1981738628425e-07,
+      "loss": 0.3213,
+      "step": 78381
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.197931148162778e-07,
+      "loss": 0.4565,
+      "step": 78382
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.19768843715978e-07,
+      "loss": 0.2996,
+      "step": 78383
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.197445729833696e-07,
+      "loss": 0.5024,
+      "step": 78384
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.197203026184702e-07,
+      "loss": 0.4717,
+      "step": 78385
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.196960326212994e-07,
+      "loss": 0.4277,
+      "step": 78386
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.196717629918748e-07,
+      "loss": 0.4177,
+      "step": 78387
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.196474937302153e-07,
+      "loss": 0.4248,
+      "step": 78388
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.196232248363397e-07,
+      "loss": 0.2826,
+      "step": 78389
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.19598956310267e-07,
+      "loss": 0.5049,
+      "step": 78390
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.195746881520148e-07,
+      "loss": 0.3979,
+      "step": 78391
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.195504203616023e-07,
+      "loss": 0.3676,
+      "step": 78392
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.195261529390478e-07,
+      "loss": 0.4375,
+      "step": 78393
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.195018858843705e-07,
+      "loss": 0.4736,
+      "step": 78394
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.194776191975885e-07,
+      "loss": 0.4224,
+      "step": 78395
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.1945335287872e-07,
+      "loss": 0.4182,
+      "step": 78396
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.19429086927784e-07,
+      "loss": 0.4028,
+      "step": 78397
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.19404821344799e-07,
+      "loss": 0.4697,
+      "step": 78398
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.193805561297839e-07,
+      "loss": 0.4094,
+      "step": 78399
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.193562912827569e-07,
+      "loss": 0.4258,
+      "step": 78400
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.193320268037364e-07,
+      "loss": 0.3555,
+      "step": 78401
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.193077626927418e-07,
+      "loss": 0.5117,
+      "step": 78402
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.192834989497909e-07,
+      "loss": 0.5576,
+      "step": 78403
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.192592355749027e-07,
+      "loss": 0.5127,
+      "step": 78404
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.192349725680952e-07,
+      "loss": 0.4497,
+      "step": 78405
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.192107099293875e-07,
+      "loss": 0.46,
+      "step": 78406
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.191864476587984e-07,
+      "loss": 0.5381,
+      "step": 78407
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.191621857563458e-07,
+      "loss": 0.4966,
+      "step": 78408
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.191379242220484e-07,
+      "loss": 0.5332,
+      "step": 78409
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.191136630559252e-07,
+      "loss": 0.3877,
+      "step": 78410
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.190894022579948e-07,
+      "loss": 0.498,
+      "step": 78411
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.190651418282752e-07,
+      "loss": 0.4072,
+      "step": 78412
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.190408817667852e-07,
+      "loss": 0.3403,
+      "step": 78413
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.190166220735441e-07,
+      "loss": 0.4663,
+      "step": 78414
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.189923627485692e-07,
+      "loss": 0.4429,
+      "step": 78415
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.189681037918805e-07,
+      "loss": 0.4507,
+      "step": 78416
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.189438452034952e-07,
+      "loss": 0.5024,
+      "step": 78417
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.189195869834324e-07,
+      "loss": 0.4883,
+      "step": 78418
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.188953291317112e-07,
+      "loss": 0.4878,
+      "step": 78419
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.188710716483493e-07,
+      "loss": 0.4824,
+      "step": 78420
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.188468145333657e-07,
+      "loss": 0.5835,
+      "step": 78421
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.18822557786779e-07,
+      "loss": 0.397,
+      "step": 78422
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.187983014086082e-07,
+      "loss": 0.459,
+      "step": 78423
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.187740453988714e-07,
+      "loss": 0.4468,
+      "step": 78424
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.187497897575867e-07,
+      "loss": 0.4861,
+      "step": 78425
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.187255344847731e-07,
+      "loss": 0.4756,
+      "step": 78426
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.187012795804494e-07,
+      "loss": 0.4951,
+      "step": 78427
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.186770250446344e-07,
+      "loss": 0.4849,
+      "step": 78428
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.186527708773457e-07,
+      "loss": 0.4194,
+      "step": 78429
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.186285170786025e-07,
+      "loss": 0.3706,
+      "step": 78430
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.186042636484233e-07,
+      "loss": 0.4546,
+      "step": 78431
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.185800105868272e-07,
+      "loss": 0.4922,
+      "step": 78432
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.185557578938317e-07,
+      "loss": 0.3339,
+      "step": 78433
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.185315055694566e-07,
+      "loss": 0.3643,
+      "step": 78434
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.185072536137191e-07,
+      "loss": 0.4599,
+      "step": 78435
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.184830020266389e-07,
+      "loss": 0.3054,
+      "step": 78436
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.184587508082338e-07,
+      "loss": 0.4219,
+      "step": 78437
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.184344999585226e-07,
+      "loss": 0.2299,
+      "step": 78438
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.184102494775241e-07,
+      "loss": 0.5078,
+      "step": 78439
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.18385999365257e-07,
+      "loss": 0.4985,
+      "step": 78440
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.183617496217393e-07,
+      "loss": 0.4319,
+      "step": 78441
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.183375002469899e-07,
+      "loss": 0.4062,
+      "step": 78442
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.183132512410271e-07,
+      "loss": 0.373,
+      "step": 78443
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.182890026038702e-07,
+      "loss": 0.4717,
+      "step": 78444
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.182647543355374e-07,
+      "loss": 0.2192,
+      "step": 78445
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.182405064360466e-07,
+      "loss": 0.4229,
+      "step": 78446
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.182162589054169e-07,
+      "loss": 0.4038,
+      "step": 78447
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.181920117436669e-07,
+      "loss": 0.4434,
+      "step": 78448
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.181677649508154e-07,
+      "loss": 0.4939,
+      "step": 78449
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.181435185268803e-07,
+      "loss": 0.458,
+      "step": 78450
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.181192724718808e-07,
+      "loss": 0.4546,
+      "step": 78451
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.18095026785835e-07,
+      "loss": 0.4817,
+      "step": 78452
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.180707814687622e-07,
+      "loss": 0.3473,
+      "step": 78453
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.180465365206804e-07,
+      "loss": 0.4985,
+      "step": 78454
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.180222919416076e-07,
+      "loss": 0.322,
+      "step": 78455
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.179980477315632e-07,
+      "loss": 0.4124,
+      "step": 78456
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.179738038905659e-07,
+      "loss": 0.5259,
+      "step": 78457
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.179495604186335e-07,
+      "loss": 0.3767,
+      "step": 78458
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.179253173157848e-07,
+      "loss": 0.3936,
+      "step": 78459
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.179010745820387e-07,
+      "loss": 0.4751,
+      "step": 78460
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.178768322174138e-07,
+      "loss": 0.4165,
+      "step": 78461
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.178525902219282e-07,
+      "loss": 0.3887,
+      "step": 78462
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.178283485956008e-07,
+      "loss": 0.4287,
+      "step": 78463
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.178041073384502e-07,
+      "loss": 0.3582,
+      "step": 78464
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.177798664504946e-07,
+      "loss": 0.2697,
+      "step": 78465
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.177556259317531e-07,
+      "loss": 0.3721,
+      "step": 78466
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.177313857822437e-07,
+      "loss": 0.3696,
+      "step": 78467
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.17707146001985e-07,
+      "loss": 0.4487,
+      "step": 78468
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.176829065909958e-07,
+      "loss": 0.4062,
+      "step": 78469
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.176586675492951e-07,
+      "loss": 0.2988,
+      "step": 78470
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.176344288769006e-07,
+      "loss": 0.2781,
+      "step": 78471
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.176101905738311e-07,
+      "loss": 0.4165,
+      "step": 78472
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.175859526401059e-07,
+      "loss": 0.4756,
+      "step": 78473
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.175617150757427e-07,
+      "loss": 0.4834,
+      "step": 78474
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.175374778807599e-07,
+      "loss": 0.3411,
+      "step": 78475
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.175132410551768e-07,
+      "loss": 0.4424,
+      "step": 78476
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.174890045990115e-07,
+      "loss": 0.4062,
+      "step": 78477
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.174647685122828e-07,
+      "loss": 0.4121,
+      "step": 78478
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.174405327950091e-07,
+      "loss": 0.4736,
+      "step": 78479
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.174162974472088e-07,
+      "loss": 0.3989,
+      "step": 78480
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.173920624689007e-07,
+      "loss": 0.3669,
+      "step": 78481
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.173678278601037e-07,
+      "loss": 0.3014,
+      "step": 78482
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.173435936208357e-07,
+      "loss": 0.4014,
+      "step": 78483
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.173193597511156e-07,
+      "loss": 0.4167,
+      "step": 78484
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.172951262509617e-07,
+      "loss": 0.5088,
+      "step": 78485
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.172708931203927e-07,
+      "loss": 0.5156,
+      "step": 78486
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.172466603594277e-07,
+      "loss": 0.4492,
+      "step": 78487
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.172224279680842e-07,
+      "loss": 0.4069,
+      "step": 78488
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.171981959463811e-07,
+      "loss": 0.3989,
+      "step": 78489
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.171739642943378e-07,
+      "loss": 0.415,
+      "step": 78490
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.171497330119718e-07,
+      "loss": 0.3573,
+      "step": 78491
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.17125502099302e-07,
+      "loss": 0.3462,
+      "step": 78492
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.17101271556347e-07,
+      "loss": 0.3118,
+      "step": 78493
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.170770413831255e-07,
+      "loss": 0.3899,
+      "step": 78494
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.170528115796563e-07,
+      "loss": 0.3794,
+      "step": 78495
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.17028582145957e-07,
+      "loss": 0.4736,
+      "step": 78496
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.170043530820467e-07,
+      "loss": 0.4072,
+      "step": 78497
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.169801243879439e-07,
+      "loss": 0.3979,
+      "step": 78498
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.169558960636677e-07,
+      "loss": 0.4138,
+      "step": 78499
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.169316681092357e-07,
+      "loss": 0.3862,
+      "step": 78500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.16907440524667e-07,
+      "loss": 0.4111,
+      "step": 78501
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.168832133099798e-07,
+      "loss": 0.4458,
+      "step": 78502
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.168589864651935e-07,
+      "loss": 0.5,
+      "step": 78503
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.168347599903262e-07,
+      "loss": 0.4204,
+      "step": 78504
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.168105338853958e-07,
+      "loss": 0.4819,
+      "step": 78505
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.167863081504214e-07,
+      "loss": 0.2122,
+      "step": 78506
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.167620827854217e-07,
+      "loss": 0.4575,
+      "step": 78507
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.167378577904148e-07,
+      "loss": 0.4292,
+      "step": 78508
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.167136331654194e-07,
+      "loss": 0.394,
+      "step": 78509
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.166894089104543e-07,
+      "loss": 0.3682,
+      "step": 78510
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.166651850255382e-07,
+      "loss": 0.4988,
+      "step": 78511
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.166409615106891e-07,
+      "loss": 0.3689,
+      "step": 78512
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.166167383659255e-07,
+      "loss": 0.4268,
+      "step": 78513
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.165925155912669e-07,
+      "loss": 0.5088,
+      "step": 78514
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.165682931867306e-07,
+      "loss": 0.2888,
+      "step": 78515
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.165440711523363e-07,
+      "loss": 0.5664,
+      "step": 78516
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.165198494881013e-07,
+      "loss": 0.4888,
+      "step": 78517
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.164956281940452e-07,
+      "loss": 0.2608,
+      "step": 78518
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.164714072701859e-07,
+      "loss": 0.4946,
+      "step": 78519
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.164471867165428e-07,
+      "loss": 0.5596,
+      "step": 78520
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.164229665331333e-07,
+      "loss": 0.4773,
+      "step": 78521
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.163987467199764e-07,
+      "loss": 0.332,
+      "step": 78522
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.163745272770909e-07,
+      "loss": 0.4453,
+      "step": 78523
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.16350308204496e-07,
+      "loss": 0.3879,
+      "step": 78524
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.163260895022086e-07,
+      "loss": 0.283,
+      "step": 78525
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.163018711702481e-07,
+      "loss": 0.3926,
+      "step": 78526
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.162776532086331e-07,
+      "loss": 0.394,
+      "step": 78527
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.162534356173822e-07,
+      "loss": 0.363,
+      "step": 78528
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.162292183965135e-07,
+      "loss": 0.4561,
+      "step": 78529
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.16205001546046e-07,
+      "loss": 0.4351,
+      "step": 78530
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.161807850659978e-07,
+      "loss": 0.2691,
+      "step": 78531
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.161565689563884e-07,
+      "loss": 0.3257,
+      "step": 78532
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.161323532172352e-07,
+      "loss": 0.3409,
+      "step": 78533
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.161081378485576e-07,
+      "loss": 0.4316,
+      "step": 78534
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.160839228503733e-07,
+      "loss": 0.4209,
+      "step": 78535
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.160597082227012e-07,
+      "loss": 0.2959,
+      "step": 78536
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.160354939655604e-07,
+      "loss": 0.3508,
+      "step": 78537
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.160112800789686e-07,
+      "loss": 0.3276,
+      "step": 78538
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.159870665629449e-07,
+      "loss": 0.4136,
+      "step": 78539
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.159628534175075e-07,
+      "loss": 0.4636,
+      "step": 78540
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.159386406426753e-07,
+      "loss": 0.4517,
+      "step": 78541
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.159144282384664e-07,
+      "loss": 0.3843,
+      "step": 78542
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.158902162048996e-07,
+      "loss": 0.2074,
+      "step": 78543
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.158660045419938e-07,
+      "loss": 0.5093,
+      "step": 78544
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.158417932497669e-07,
+      "loss": 0.428,
+      "step": 78545
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.158175823282376e-07,
+      "loss": 0.4746,
+      "step": 78546
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.157933717774243e-07,
+      "loss": 0.3682,
+      "step": 78547
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.157691615973459e-07,
+      "loss": 0.4893,
+      "step": 78548
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.15744951788021e-07,
+      "loss": 0.2278,
+      "step": 78549
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.157207423494677e-07,
+      "loss": 0.4609,
+      "step": 78550
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.156965332817048e-07,
+      "loss": 0.4099,
+      "step": 78551
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.156723245847506e-07,
+      "loss": 0.4043,
+      "step": 78552
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.156481162586244e-07,
+      "loss": 0.3447,
+      "step": 78553
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.15623908303344e-07,
+      "loss": 0.3508,
+      "step": 78554
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.155997007189279e-07,
+      "loss": 0.5229,
+      "step": 78555
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.155754935053946e-07,
+      "loss": 0.3821,
+      "step": 78556
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.15551286662763e-07,
+      "loss": 0.4043,
+      "step": 78557
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.155270801910519e-07,
+      "loss": 0.3872,
+      "step": 78558
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.155028740902791e-07,
+      "loss": 0.4561,
+      "step": 78559
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.154786683604633e-07,
+      "loss": 0.3933,
+      "step": 78560
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.154544630016234e-07,
+      "loss": 0.3643,
+      "step": 78561
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.154302580137781e-07,
+      "loss": 0.4387,
+      "step": 78562
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.154060533969451e-07,
+      "loss": 0.254,
+      "step": 78563
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.153818491511438e-07,
+      "loss": 0.4277,
+      "step": 78564
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.153576452763919e-07,
+      "loss": 0.4375,
+      "step": 78565
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.153334417727089e-07,
+      "loss": 0.386,
+      "step": 78566
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.153092386401123e-07,
+      "loss": 0.3025,
+      "step": 78567
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.152850358786211e-07,
+      "loss": 0.4204,
+      "step": 78568
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.15260833488254e-07,
+      "loss": 0.4614,
+      "step": 78569
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.152366314690297e-07,
+      "loss": 0.5425,
+      "step": 78570
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.15212429820966e-07,
+      "loss": 0.603,
+      "step": 78571
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.151882285440818e-07,
+      "loss": 0.304,
+      "step": 78572
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.151640276383958e-07,
+      "loss": 0.4917,
+      "step": 78573
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.151398271039269e-07,
+      "loss": 0.4482,
+      "step": 78574
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.151156269406929e-07,
+      "loss": 0.3727,
+      "step": 78575
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.150914271487124e-07,
+      "loss": 0.4717,
+      "step": 78576
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.150672277280039e-07,
+      "loss": 0.345,
+      "step": 78577
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.150430286785862e-07,
+      "loss": 0.4756,
+      "step": 78578
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.150188300004781e-07,
+      "loss": 0.3235,
+      "step": 78579
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.149946316936975e-07,
+      "loss": 0.4438,
+      "step": 78580
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.149704337582631e-07,
+      "loss": 0.5303,
+      "step": 78581
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.149462361941941e-07,
+      "loss": 0.4941,
+      "step": 78582
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.149220390015079e-07,
+      "loss": 0.4819,
+      "step": 78583
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.148978421802239e-07,
+      "loss": 0.4653,
+      "step": 78584
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.148736457303601e-07,
+      "loss": 0.6069,
+      "step": 78585
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.148494496519353e-07,
+      "loss": 0.3257,
+      "step": 78586
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.148252539449682e-07,
+      "loss": 0.3374,
+      "step": 78587
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.148010586094768e-07,
+      "loss": 0.2526,
+      "step": 78588
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.147768636454799e-07,
+      "loss": 0.3347,
+      "step": 78589
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.14752669052996e-07,
+      "loss": 0.3048,
+      "step": 78590
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.14728474832044e-07,
+      "loss": 0.438,
+      "step": 78591
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.147042809826418e-07,
+      "loss": 0.4099,
+      "step": 78592
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.146800875048083e-07,
+      "loss": 0.4702,
+      "step": 78593
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.146558943985622e-07,
+      "loss": 0.415,
+      "step": 78594
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.146317016639215e-07,
+      "loss": 0.3595,
+      "step": 78595
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.146075093009052e-07,
+      "loss": 0.4878,
+      "step": 78596
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.145833173095313e-07,
+      "loss": 0.3767,
+      "step": 78597
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.145591256898187e-07,
+      "loss": 0.2612,
+      "step": 78598
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.145349344417862e-07,
+      "loss": 0.4951,
+      "step": 78599
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.145107435654516e-07,
+      "loss": 0.4282,
+      "step": 78600
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.144865530608337e-07,
+      "loss": 0.4956,
+      "step": 78601
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.144623629279512e-07,
+      "loss": 0.4697,
+      "step": 78602
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.14438173166823e-07,
+      "loss": 0.417,
+      "step": 78603
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.14413983777467e-07,
+      "loss": 0.4692,
+      "step": 78604
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.143897947599017e-07,
+      "loss": 0.3602,
+      "step": 78605
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.143656061141457e-07,
+      "loss": 0.5024,
+      "step": 78606
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.143414178402176e-07,
+      "loss": 0.3263,
+      "step": 78607
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.143172299381364e-07,
+      "loss": 0.3843,
+      "step": 78608
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.142930424079197e-07,
+      "loss": 0.2759,
+      "step": 78609
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.142688552495865e-07,
+      "loss": 0.4199,
+      "step": 78610
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.142446684631554e-07,
+      "loss": 0.3435,
+      "step": 78611
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.142204820486451e-07,
+      "loss": 0.3486,
+      "step": 78612
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.141962960060739e-07,
+      "loss": 0.3682,
+      "step": 78613
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.141721103354597e-07,
+      "loss": 0.4309,
+      "step": 78614
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.141479250368218e-07,
+      "loss": 0.4854,
+      "step": 78615
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.141237401101788e-07,
+      "loss": 0.4478,
+      "step": 78616
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 5.140995555555485e-07,
+      "loss": 0.4619,
+      "step": 78617
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.140753713729499e-07,
+      "loss": 0.4614,
+      "step": 78618
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.140511875624014e-07,
+      "loss": 0.4287,
+      "step": 78619
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.140270041239219e-07,
+      "loss": 0.3721,
+      "step": 78620
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.140028210575292e-07,
+      "loss": 0.5088,
+      "step": 78621
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.139786383632423e-07,
+      "loss": 0.3442,
+      "step": 78622
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.139544560410799e-07,
+      "loss": 0.4316,
+      "step": 78623
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.139302740910598e-07,
+      "loss": 0.52,
+      "step": 78624
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.139060925132015e-07,
+      "loss": 0.2688,
+      "step": 78625
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.138819113075223e-07,
+      "loss": 0.3862,
+      "step": 78626
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.138577304740417e-07,
+      "loss": 0.4937,
+      "step": 78627
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.138335500127776e-07,
+      "loss": 0.3457,
+      "step": 78628
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.138093699237493e-07,
+      "loss": 0.3499,
+      "step": 78629
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.137851902069744e-07,
+      "loss": 0.459,
+      "step": 78630
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.137610108624719e-07,
+      "loss": 0.3333,
+      "step": 78631
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.137368318902601e-07,
+      "loss": 0.5024,
+      "step": 78632
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.137126532903587e-07,
+      "loss": 0.4531,
+      "step": 78633
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.136884750627841e-07,
+      "loss": 0.439,
+      "step": 78634
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.136642972075559e-07,
+      "loss": 0.4985,
+      "step": 78635
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.136401197246927e-07,
+      "loss": 0.397,
+      "step": 78636
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.136159426142132e-07,
+      "loss": 0.3462,
+      "step": 78637
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.135917658761351e-07,
+      "loss": 0.3052,
+      "step": 78638
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.135675895104776e-07,
+      "loss": 0.4736,
+      "step": 78639
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.135434135172591e-07,
+      "loss": 0.4385,
+      "step": 78640
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.135192378964983e-07,
+      "loss": 0.4419,
+      "step": 78641
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.13495062648213e-07,
+      "loss": 0.4199,
+      "step": 78642
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.134708877724228e-07,
+      "loss": 0.4482,
+      "step": 78643
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.134467132691449e-07,
+      "loss": 0.4917,
+      "step": 78644
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.134225391383986e-07,
+      "loss": 0.3989,
+      "step": 78645
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.133983653802026e-07,
+      "loss": 0.3423,
+      "step": 78646
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.133741919945749e-07,
+      "loss": 0.5015,
+      "step": 78647
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.13350018981534e-07,
+      "loss": 0.4385,
+      "step": 78648
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.133258463410988e-07,
+      "loss": 0.4165,
+      "step": 78649
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.133016740732878e-07,
+      "loss": 0.4834,
+      "step": 78650
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.13277502178119e-07,
+      "loss": 0.4053,
+      "step": 78651
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.132533306556114e-07,
+      "loss": 0.5059,
+      "step": 78652
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.132291595057836e-07,
+      "loss": 0.3955,
+      "step": 78653
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.132049887286539e-07,
+      "loss": 0.4282,
+      "step": 78654
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.131808183242403e-07,
+      "loss": 0.2925,
+      "step": 78655
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.131566482925618e-07,
+      "loss": 0.342,
+      "step": 78656
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.131324786336368e-07,
+      "loss": 0.3667,
+      "step": 78657
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.131083093474844e-07,
+      "loss": 0.4287,
+      "step": 78658
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.130841404341221e-07,
+      "loss": 0.3589,
+      "step": 78659
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.130599718935689e-07,
+      "loss": 0.4141,
+      "step": 78660
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.130358037258434e-07,
+      "loss": 0.3657,
+      "step": 78661
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.130116359309642e-07,
+      "loss": 0.4629,
+      "step": 78662
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.129874685089497e-07,
+      "loss": 0.3833,
+      "step": 78663
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.129633014598178e-07,
+      "loss": 0.3691,
+      "step": 78664
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.129391347835876e-07,
+      "loss": 0.3706,
+      "step": 78665
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.129149684802776e-07,
+      "loss": 0.4521,
+      "step": 78666
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.128908025499065e-07,
+      "loss": 0.3497,
+      "step": 78667
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.128666369924921e-07,
+      "loss": 0.3372,
+      "step": 78668
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.128424718080533e-07,
+      "loss": 0.5107,
+      "step": 78669
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.128183069966091e-07,
+      "loss": 0.4536,
+      "step": 78670
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.12794142558177e-07,
+      "loss": 0.4155,
+      "step": 78671
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.127699784927762e-07,
+      "loss": 0.5718,
+      "step": 78672
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.127458148004253e-07,
+      "loss": 0.4248,
+      "step": 78673
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.127216514811422e-07,
+      "loss": 0.406,
+      "step": 78674
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.126974885349461e-07,
+      "loss": 0.3746,
+      "step": 78675
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.126733259618547e-07,
+      "loss": 0.4204,
+      "step": 78676
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.12649163761887e-07,
+      "loss": 0.3596,
+      "step": 78677
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.126250019350613e-07,
+      "loss": 0.4507,
+      "step": 78678
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.126008404813966e-07,
+      "loss": 0.2878,
+      "step": 78679
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.125766794009107e-07,
+      "loss": 0.5249,
+      "step": 78680
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.125525186936226e-07,
+      "loss": 0.3894,
+      "step": 78681
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.125283583595504e-07,
+      "loss": 0.3418,
+      "step": 78682
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.125041983987133e-07,
+      "loss": 0.3989,
+      "step": 78683
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.124800388111292e-07,
+      "loss": 0.3623,
+      "step": 78684
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.124558795968165e-07,
+      "loss": 0.4015,
+      "step": 78685
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.124317207557938e-07,
+      "loss": 0.312,
+      "step": 78686
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.124075622880801e-07,
+      "loss": 0.3103,
+      "step": 78687
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.123834041936931e-07,
+      "loss": 0.4565,
+      "step": 78688
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.123592464726518e-07,
+      "loss": 0.3923,
+      "step": 78689
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.123350891249747e-07,
+      "loss": 0.4399,
+      "step": 78690
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.123109321506804e-07,
+      "loss": 0.502,
+      "step": 78691
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.122867755497868e-07,
+      "loss": 0.3677,
+      "step": 78692
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.122626193223133e-07,
+      "loss": 0.5039,
+      "step": 78693
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.122384634682775e-07,
+      "loss": 0.4092,
+      "step": 78694
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.122143079876982e-07,
+      "loss": 0.4092,
+      "step": 78695
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.121901528805943e-07,
+      "loss": 0.4028,
+      "step": 78696
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.121659981469837e-07,
+      "loss": 0.3535,
+      "step": 78697
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.121418437868852e-07,
+      "loss": 0.3965,
+      "step": 78698
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.121176898003171e-07,
+      "loss": 0.3828,
+      "step": 78699
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.120935361872986e-07,
+      "loss": 0.519,
+      "step": 78700
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.120693829478471e-07,
+      "loss": 0.4272,
+      "step": 78701
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.120452300819818e-07,
+      "loss": 0.3315,
+      "step": 78702
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.120210775897213e-07,
+      "loss": 0.4663,
+      "step": 78703
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.119969254710838e-07,
+      "loss": 0.3298,
+      "step": 78704
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.119727737260875e-07,
+      "loss": 0.2684,
+      "step": 78705
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.119486223547513e-07,
+      "loss": 0.4644,
+      "step": 78706
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.119244713570934e-07,
+      "loss": 0.4224,
+      "step": 78707
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.11900320733133e-07,
+      "loss": 0.3899,
+      "step": 78708
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.118761704828876e-07,
+      "loss": 0.3794,
+      "step": 78709
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.118520206063763e-07,
+      "loss": 0.4673,
+      "step": 78710
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.118278711036174e-07,
+      "loss": 0.2765,
+      "step": 78711
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.118037219746298e-07,
+      "loss": 0.3813,
+      "step": 78712
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.117795732194316e-07,
+      "loss": 0.3481,
+      "step": 78713
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.117554248380412e-07,
+      "loss": 0.4282,
+      "step": 78714
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.11731276830477e-07,
+      "loss": 0.3789,
+      "step": 78715
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.117071291967578e-07,
+      "loss": 0.3623,
+      "step": 78716
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.116829819369023e-07,
+      "loss": 0.3073,
+      "step": 78717
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.116588350509284e-07,
+      "loss": 0.4214,
+      "step": 78718
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.116346885388547e-07,
+      "loss": 0.26,
+      "step": 78719
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.116105424007e-07,
+      "loss": 0.3879,
+      "step": 78720
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.115863966364831e-07,
+      "loss": 0.4089,
+      "step": 78721
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.115622512462215e-07,
+      "loss": 0.3378,
+      "step": 78722
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.115381062299347e-07,
+      "loss": 0.5015,
+      "step": 78723
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.115139615876403e-07,
+      "loss": 0.4062,
+      "step": 78724
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.114898173193576e-07,
+      "loss": 0.478,
+      "step": 78725
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.114656734251043e-07,
+      "loss": 0.3691,
+      "step": 78726
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.114415299048993e-07,
+      "loss": 0.4316,
+      "step": 78727
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.114173867587611e-07,
+      "loss": 0.376,
+      "step": 78728
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.113932439867084e-07,
+      "loss": 0.4007,
+      "step": 78729
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.113691015887591e-07,
+      "loss": 0.3838,
+      "step": 78730
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.113449595649319e-07,
+      "loss": 0.4146,
+      "step": 78731
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.113208179152455e-07,
+      "loss": 0.4302,
+      "step": 78732
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.112966766397187e-07,
+      "loss": 0.4807,
+      "step": 78733
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.112725357383696e-07,
+      "loss": 0.3211,
+      "step": 78734
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.112483952112161e-07,
+      "loss": 0.4053,
+      "step": 78735
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.112242550582772e-07,
+      "loss": 0.2922,
+      "step": 78736
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.112001152795717e-07,
+      "loss": 0.4075,
+      "step": 78737
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.11175975875118e-07,
+      "loss": 0.4077,
+      "step": 78738
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.111518368449339e-07,
+      "loss": 0.3853,
+      "step": 78739
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.111276981890384e-07,
+      "loss": 0.3555,
+      "step": 78740
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.111035599074502e-07,
+      "loss": 0.3379,
+      "step": 78741
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.110794220001876e-07,
+      "loss": 0.4625,
+      "step": 78742
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.110552844672691e-07,
+      "loss": 0.479,
+      "step": 78743
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.110311473087128e-07,
+      "loss": 0.4463,
+      "step": 78744
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.110070105245374e-07,
+      "loss": 0.385,
+      "step": 78745
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.109828741147617e-07,
+      "loss": 0.3652,
+      "step": 78746
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.109587380794038e-07,
+      "loss": 0.4946,
+      "step": 78747
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.109346024184821e-07,
+      "loss": 0.5786,
+      "step": 78748
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.109104671320154e-07,
+      "loss": 0.4741,
+      "step": 78749
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.108863322200223e-07,
+      "loss": 0.4746,
+      "step": 78750
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.10862197682521e-07,
+      "loss": 0.2917,
+      "step": 78751
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.108380635195296e-07,
+      "loss": 0.3755,
+      "step": 78752
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.108139297310676e-07,
+      "loss": 0.3733,
+      "step": 78753
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.107897963171526e-07,
+      "loss": 0.4685,
+      "step": 78754
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.107656632778034e-07,
+      "loss": 0.406,
+      "step": 78755
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.107415306130385e-07,
+      "loss": 0.46,
+      "step": 78756
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.10717398322876e-07,
+      "loss": 0.4124,
+      "step": 78757
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.106932664073348e-07,
+      "loss": 0.3623,
+      "step": 78758
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.106691348664336e-07,
+      "loss": 0.5234,
+      "step": 78759
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.106450037001902e-07,
+      "loss": 0.3569,
+      "step": 78760
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.106208729086234e-07,
+      "loss": 0.3523,
+      "step": 78761
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.105967424917521e-07,
+      "loss": 0.3298,
+      "step": 78762
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.105726124495944e-07,
+      "loss": 0.5,
+      "step": 78763
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.105484827821683e-07,
+      "loss": 0.4214,
+      "step": 78764
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.105243534894928e-07,
+      "loss": 0.3815,
+      "step": 78765
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.105002245715862e-07,
+      "loss": 0.3162,
+      "step": 78766
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.104760960284674e-07,
+      "loss": 0.3333,
+      "step": 78767
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.104519678601542e-07,
+      "loss": 0.3551,
+      "step": 78768
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.104278400666655e-07,
+      "loss": 0.5093,
+      "step": 78769
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.104037126480196e-07,
+      "loss": 0.481,
+      "step": 78770
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.103795856042357e-07,
+      "loss": 0.4443,
+      "step": 78771
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.10355458935331e-07,
+      "loss": 0.5005,
+      "step": 78772
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.10331332641325e-07,
+      "loss": 0.4473,
+      "step": 78773
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.103072067222355e-07,
+      "loss": 0.3251,
+      "step": 78774
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.102830811780812e-07,
+      "loss": 0.3721,
+      "step": 78775
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.10258956008881e-07,
+      "loss": 0.5043,
+      "step": 78776
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.102348312146527e-07,
+      "loss": 0.263,
+      "step": 78777
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.102107067954148e-07,
+      "loss": 0.4771,
+      "step": 78778
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.101865827511869e-07,
+      "loss": 0.4316,
+      "step": 78779
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.101624590819859e-07,
+      "loss": 0.4211,
+      "step": 78780
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.101383357878311e-07,
+      "loss": 0.3596,
+      "step": 78781
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.101142128687407e-07,
+      "loss": 0.3513,
+      "step": 78782
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.100900903247338e-07,
+      "loss": 0.4006,
+      "step": 78783
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.100659681558283e-07,
+      "loss": 0.3853,
+      "step": 78784
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.100418463620426e-07,
+      "loss": 0.3898,
+      "step": 78785
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.100177249433951e-07,
+      "loss": 0.4868,
+      "step": 78786
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.099936038999046e-07,
+      "loss": 0.4736,
+      "step": 78787
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.0996948323159e-07,
+      "loss": 0.4281,
+      "step": 78788
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.099453629384686e-07,
+      "loss": 0.3356,
+      "step": 78789
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.099212430205595e-07,
+      "loss": 0.4233,
+      "step": 78790
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.098971234778813e-07,
+      "loss": 0.3494,
+      "step": 78791
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.098730043104527e-07,
+      "loss": 0.3166,
+      "step": 78792
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.098488855182919e-07,
+      "loss": 0.4042,
+      "step": 78793
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.098247671014167e-07,
+      "loss": 0.3809,
+      "step": 78794
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.098006490598463e-07,
+      "loss": 0.3633,
+      "step": 78795
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.097765313935992e-07,
+      "loss": 0.3703,
+      "step": 78796
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.097524141026935e-07,
+      "loss": 0.4229,
+      "step": 78797
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.097282971871476e-07,
+      "loss": 0.4224,
+      "step": 78798
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.097041806469804e-07,
+      "loss": 0.3037,
+      "step": 78799
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.096800644822103e-07,
+      "loss": 0.4795,
+      "step": 78800
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.096559486928555e-07,
+      "loss": 0.2552,
+      "step": 78801
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.096318332789344e-07,
+      "loss": 0.3965,
+      "step": 78802
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.096077182404663e-07,
+      "loss": 0.3772,
+      "step": 78803
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.095836035774684e-07,
+      "loss": 0.4697,
+      "step": 78804
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.095594892899603e-07,
+      "loss": 0.3673,
+      "step": 78805
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.095353753779595e-07,
+      "loss": 0.4263,
+      "step": 78806
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.095112618414849e-07,
+      "loss": 0.375,
+      "step": 78807
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.09487148680555e-07,
+      "loss": 0.4065,
+      "step": 78808
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.094630358951887e-07,
+      "loss": 0.3752,
+      "step": 78809
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.094389234854036e-07,
+      "loss": 0.5034,
+      "step": 78810
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.094148114512185e-07,
+      "loss": 0.5215,
+      "step": 78811
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.093906997926518e-07,
+      "loss": 0.4961,
+      "step": 78812
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.09366588509723e-07,
+      "loss": 0.4077,
+      "step": 78813
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.093424776024489e-07,
+      "loss": 0.4185,
+      "step": 78814
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.093183670708486e-07,
+      "loss": 0.3772,
+      "step": 78815
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.092942569149408e-07,
+      "loss": 0.5166,
+      "step": 78816
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.092701471347441e-07,
+      "loss": 0.3999,
+      "step": 78817
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.092460377302763e-07,
+      "loss": 0.4414,
+      "step": 78818
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.092219287015563e-07,
+      "loss": 0.4526,
+      "step": 78819
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.091978200486025e-07,
+      "loss": 0.4556,
+      "step": 78820
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.091737117714336e-07,
+      "loss": 0.519,
+      "step": 78821
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.091496038700676e-07,
+      "loss": 0.3682,
+      "step": 78822
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.091254963445235e-07,
+      "loss": 0.4673,
+      "step": 78823
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.09101389194819e-07,
+      "loss": 0.4507,
+      "step": 78824
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.09077282420973e-07,
+      "loss": 0.2232,
+      "step": 78825
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.090531760230045e-07,
+      "loss": 0.3772,
+      "step": 78826
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.090290700009309e-07,
+      "loss": 0.2639,
+      "step": 78827
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.090049643547711e-07,
+      "loss": 0.417,
+      "step": 78828
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.089808590845438e-07,
+      "loss": 0.5361,
+      "step": 78829
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.089567541902675e-07,
+      "loss": 0.522,
+      "step": 78830
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.089326496719601e-07,
+      "loss": 0.3091,
+      "step": 78831
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.089085455296405e-07,
+      "loss": 0.4448,
+      "step": 78832
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.088844417633274e-07,
+      "loss": 0.4893,
+      "step": 78833
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.088603383730387e-07,
+      "loss": 0.5332,
+      "step": 78834
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.088362353587929e-07,
+      "loss": 0.5054,
+      "step": 78835
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.088121327206085e-07,
+      "loss": 0.3767,
+      "step": 78836
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.08788030458504e-07,
+      "loss": 0.3184,
+      "step": 78837
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.087639285724985e-07,
+      "loss": 0.3984,
+      "step": 78838
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.087398270626093e-07,
+      "loss": 0.3645,
+      "step": 78839
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.087157259288555e-07,
+      "loss": 0.3809,
+      "step": 78840
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.086916251712556e-07,
+      "loss": 0.2244,
+      "step": 78841
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.086675247898281e-07,
+      "loss": 0.2891,
+      "step": 78842
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.086434247845914e-07,
+      "loss": 0.3711,
+      "step": 78843
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.086193251555635e-07,
+      "loss": 0.3826,
+      "step": 78844
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.085952259027631e-07,
+      "loss": 0.4565,
+      "step": 78845
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.085711270262088e-07,
+      "loss": 0.3176,
+      "step": 78846
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.085470285259193e-07,
+      "loss": 0.3423,
+      "step": 78847
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.085229304019123e-07,
+      "loss": 0.3625,
+      "step": 78848
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.084988326542068e-07,
+      "loss": 0.4575,
+      "step": 78849
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.084747352828216e-07,
+      "loss": 0.3513,
+      "step": 78850
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.084506382877742e-07,
+      "loss": 0.4244,
+      "step": 78851
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.084265416690836e-07,
+      "loss": 0.3672,
+      "step": 78852
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.084024454267686e-07,
+      "loss": 0.1902,
+      "step": 78853
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.083783495608467e-07,
+      "loss": 0.4092,
+      "step": 78854
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.083542540713375e-07,
+      "loss": 0.4624,
+      "step": 78855
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.083301589582582e-07,
+      "loss": 0.3463,
+      "step": 78856
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.083060642216281e-07,
+      "loss": 0.4365,
+      "step": 78857
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.082819698614653e-07,
+      "loss": 0.5063,
+      "step": 78858
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.082578758777889e-07,
+      "loss": 0.4492,
+      "step": 78859
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.082337822706164e-07,
+      "loss": 0.3477,
+      "step": 78860
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.082096890399667e-07,
+      "loss": 0.4194,
+      "step": 78861
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.081855961858581e-07,
+      "loss": 0.3235,
+      "step": 78862
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.081615037083098e-07,
+      "loss": 0.3772,
+      "step": 78863
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.081374116073394e-07,
+      "loss": 0.623,
+      "step": 78864
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.081133198829652e-07,
+      "loss": 0.335,
+      "step": 78865
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.080892285352061e-07,
+      "loss": 0.4785,
+      "step": 78866
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.080651375640809e-07,
+      "loss": 0.4395,
+      "step": 78867
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.080410469696071e-07,
+      "loss": 0.3821,
+      "step": 78868
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.080169567518037e-07,
+      "loss": 0.2823,
+      "step": 78869
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07992866910689e-07,
+      "loss": 0.3213,
+      "step": 78870
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07968777446282e-07,
+      "loss": 0.2996,
+      "step": 78871
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.079446883586003e-07,
+      "loss": 0.3726,
+      "step": 78872
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.079205996476632e-07,
+      "loss": 0.2815,
+      "step": 78873
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.078965113134883e-07,
+      "loss": 0.4175,
+      "step": 78874
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.078724233560943e-07,
+      "loss": 0.2892,
+      "step": 78875
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.078483357755001e-07,
+      "loss": 0.3577,
+      "step": 78876
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.078242485717234e-07,
+      "loss": 0.2507,
+      "step": 78877
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.078001617447831e-07,
+      "loss": 0.3133,
+      "step": 78878
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.077760752946975e-07,
+      "loss": 0.3179,
+      "step": 78879
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.077519892214857e-07,
+      "loss": 0.4702,
+      "step": 78880
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.077279035251651e-07,
+      "loss": 0.4155,
+      "step": 78881
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.077038182057547e-07,
+      "loss": 0.3638,
+      "step": 78882
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.076797332632731e-07,
+      "loss": 0.5107,
+      "step": 78883
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.076556486977386e-07,
+      "loss": 0.363,
+      "step": 78884
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07631564509169e-07,
+      "loss": 0.4351,
+      "step": 78885
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.076074806975833e-07,
+      "loss": 0.3932,
+      "step": 78886
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07583397263e-07,
+      "loss": 0.437,
+      "step": 78887
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.075593142054377e-07,
+      "loss": 0.4458,
+      "step": 78888
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.075352315249144e-07,
+      "loss": 0.3481,
+      "step": 78889
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.075111492214485e-07,
+      "loss": 0.46,
+      "step": 78890
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.074870672950588e-07,
+      "loss": 0.3882,
+      "step": 78891
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.074629857457641e-07,
+      "loss": 0.3379,
+      "step": 78892
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.074389045735822e-07,
+      "loss": 0.4453,
+      "step": 78893
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.074148237785314e-07,
+      "loss": 0.3887,
+      "step": 78894
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.073907433606303e-07,
+      "loss": 0.3813,
+      "step": 78895
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.073666633198974e-07,
+      "loss": 0.4287,
+      "step": 78896
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.073425836563517e-07,
+      "loss": 0.3322,
+      "step": 78897
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.073185043700108e-07,
+      "loss": 0.5239,
+      "step": 78898
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.072944254608934e-07,
+      "loss": 0.5508,
+      "step": 78899
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07270346929018e-07,
+      "loss": 0.4409,
+      "step": 78900
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.072462687744035e-07,
+      "loss": 0.4365,
+      "step": 78901
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.072221909970673e-07,
+      "loss": 0.5244,
+      "step": 78902
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.071981135970289e-07,
+      "loss": 0.2012,
+      "step": 78903
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07174036574306e-07,
+      "loss": 0.3721,
+      "step": 78904
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.071499599289174e-07,
+      "loss": 0.3168,
+      "step": 78905
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.071258836608811e-07,
+      "loss": 0.3306,
+      "step": 78906
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.07101807770216e-07,
+      "loss": 0.5698,
+      "step": 78907
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.070777322569402e-07,
+      "loss": 0.3492,
+      "step": 78908
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.070536571210729e-07,
+      "loss": 0.3898,
+      "step": 78909
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.070295823626313e-07,
+      "loss": 0.2341,
+      "step": 78910
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.070055079816348e-07,
+      "loss": 0.4355,
+      "step": 78911
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.069814339781016e-07,
+      "loss": 0.376,
+      "step": 78912
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.069573603520499e-07,
+      "loss": 0.4229,
+      "step": 78913
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.069332871034985e-07,
+      "loss": 0.3999,
+      "step": 78914
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.069092142324654e-07,
+      "loss": 0.4849,
+      "step": 78915
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.06885141738969e-07,
+      "loss": 0.3456,
+      "step": 78916
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.068610696230282e-07,
+      "loss": 0.4375,
+      "step": 78917
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.068369978846613e-07,
+      "loss": 0.3689,
+      "step": 78918
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.068129265238865e-07,
+      "loss": 0.4668,
+      "step": 78919
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.067888555407224e-07,
+      "loss": 0.3696,
+      "step": 78920
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.067647849351873e-07,
+      "loss": 0.3425,
+      "step": 78921
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.067407147073004e-07,
+      "loss": 0.1973,
+      "step": 78922
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.067166448570787e-07,
+      "loss": 0.4653,
+      "step": 78923
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.066925753845415e-07,
+      "loss": 0.3696,
+      "step": 78924
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.066685062897071e-07,
+      "loss": 0.4639,
+      "step": 78925
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.066444375725944e-07,
+      "loss": 0.4512,
+      "step": 78926
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.066203692332208e-07,
+      "loss": 0.2832,
+      "step": 78927
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.065963012716054e-07,
+      "loss": 0.4443,
+      "step": 78928
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.065722336877665e-07,
+      "loss": 0.457,
+      "step": 78929
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.06548166481723e-07,
+      "loss": 0.4673,
+      "step": 78930
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.065240996534924e-07,
+      "loss": 0.4404,
+      "step": 78931
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.06500033203094e-07,
+      "loss": 0.4766,
+      "step": 78932
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.064759671305455e-07,
+      "loss": 0.4717,
+      "step": 78933
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.064519014358657e-07,
+      "loss": 0.5659,
+      "step": 78934
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.064278361190732e-07,
+      "loss": 0.4055,
+      "step": 78935
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.06403771180186e-07,
+      "loss": 0.3042,
+      "step": 78936
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.063797066192226e-07,
+      "loss": 0.4663,
+      "step": 78937
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.063556424362017e-07,
+      "loss": 0.6147,
+      "step": 78938
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.063315786311418e-07,
+      "loss": 0.3376,
+      "step": 78939
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.063075152040608e-07,
+      "loss": 0.4375,
+      "step": 78940
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.062834521549775e-07,
+      "loss": 0.4058,
+      "step": 78941
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.062593894839107e-07,
+      "loss": 0.3818,
+      "step": 78942
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.062353271908782e-07,
+      "loss": 0.582,
+      "step": 78943
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.062112652758983e-07,
+      "loss": 0.4609,
+      "step": 78944
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.061872037389898e-07,
+      "loss": 0.2859,
+      "step": 78945
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.061631425801709e-07,
+      "loss": 0.4873,
+      "step": 78946
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.061390817994606e-07,
+      "loss": 0.4277,
+      "step": 78947
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.061150213968765e-07,
+      "loss": 0.304,
+      "step": 78948
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.060909613724376e-07,
+      "loss": 0.4697,
+      "step": 78949
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.060669017261621e-07,
+      "loss": 0.4414,
+      "step": 78950
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.060428424580687e-07,
+      "loss": 0.4941,
+      "step": 78951
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.060187835681756e-07,
+      "loss": 0.3302,
+      "step": 78952
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.059947250565008e-07,
+      "loss": 0.3652,
+      "step": 78953
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.059706669230633e-07,
+      "loss": 0.415,
+      "step": 78954
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.059466091678813e-07,
+      "loss": 0.458,
+      "step": 78955
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.059225517909735e-07,
+      "loss": 0.4507,
+      "step": 78956
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.058984947923578e-07,
+      "loss": 0.4438,
+      "step": 78957
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.058744381720528e-07,
+      "loss": 0.3712,
+      "step": 78958
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.058503819300775e-07,
+      "loss": 0.5024,
+      "step": 78959
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.058263260664495e-07,
+      "loss": 0.4299,
+      "step": 78960
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.058022705811875e-07,
+      "loss": 0.5039,
+      "step": 78961
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.057782154743104e-07,
+      "loss": 0.3682,
+      "step": 78962
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.057541607458357e-07,
+      "loss": 0.4558,
+      "step": 78963
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.057301063957828e-07,
+      "loss": 0.3867,
+      "step": 78964
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.057060524241693e-07,
+      "loss": 0.343,
+      "step": 78965
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05681998831014e-07,
+      "loss": 0.3582,
+      "step": 78966
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.056579456163351e-07,
+      "loss": 0.332,
+      "step": 78967
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.056338927801515e-07,
+      "loss": 0.3464,
+      "step": 78968
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05609840322481e-07,
+      "loss": 0.478,
+      "step": 78969
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.055857882433423e-07,
+      "loss": 0.457,
+      "step": 78970
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.055617365427538e-07,
+      "loss": 0.271,
+      "step": 78971
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.055376852207344e-07,
+      "loss": 0.3596,
+      "step": 78972
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05513634277302e-07,
+      "loss": 0.4907,
+      "step": 78973
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.054895837124749e-07,
+      "loss": 0.498,
+      "step": 78974
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.054655335262714e-07,
+      "loss": 0.3304,
+      "step": 78975
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.054414837187107e-07,
+      "loss": 0.3896,
+      "step": 78976
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.054174342898103e-07,
+      "loss": 0.3298,
+      "step": 78977
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.053933852395891e-07,
+      "loss": 0.4609,
+      "step": 78978
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.053693365680655e-07,
+      "loss": 0.4536,
+      "step": 78979
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05345288275258e-07,
+      "loss": 0.4386,
+      "step": 78980
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.053212403611847e-07,
+      "loss": 0.4707,
+      "step": 78981
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.052971928258645e-07,
+      "loss": 0.3291,
+      "step": 78982
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.052731456693151e-07,
+      "loss": 0.4375,
+      "step": 78983
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.052490988915553e-07,
+      "loss": 0.4438,
+      "step": 78984
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05225052492604e-07,
+      "loss": 0.3818,
+      "step": 78985
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.052010064724786e-07,
+      "loss": 0.3734,
+      "step": 78986
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05176960831198e-07,
+      "loss": 0.4712,
+      "step": 78987
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.051529155687809e-07,
+      "loss": 0.4756,
+      "step": 78988
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.051288706852456e-07,
+      "loss": 0.4849,
+      "step": 78989
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.051048261806101e-07,
+      "loss": 0.4468,
+      "step": 78990
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.05080782054893e-07,
+      "loss": 0.3303,
+      "step": 78991
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.050567383081132e-07,
+      "loss": 0.3616,
+      "step": 78992
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.050326949402888e-07,
+      "loss": 0.3477,
+      "step": 78993
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.050086519514376e-07,
+      "loss": 0.374,
+      "step": 78994
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.049846093415785e-07,
+      "loss": 0.2576,
+      "step": 78995
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.0496056711073e-07,
+      "loss": 0.3284,
+      "step": 78996
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.049365252589107e-07,
+      "loss": 0.3496,
+      "step": 78997
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.049124837861385e-07,
+      "loss": 0.4077,
+      "step": 78998
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.04888442692432e-07,
+      "loss": 0.5166,
+      "step": 78999
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.048644019778097e-07,
+      "loss": 0.3438,
+      "step": 79000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.048403616422903e-07,
+      "loss": 0.3669,
+      "step": 79001
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.048163216858917e-07,
+      "loss": 0.3625,
+      "step": 79002
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.047922821086324e-07,
+      "loss": 0.2671,
+      "step": 79003
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.047682429105307e-07,
+      "loss": 0.4878,
+      "step": 79004
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.047442040916051e-07,
+      "loss": 0.4346,
+      "step": 79005
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.047201656518745e-07,
+      "loss": 0.4517,
+      "step": 79006
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 5.046961275913566e-07,
+      "loss": 0.4282,
+      "step": 79007
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.046720899100702e-07,
+      "loss": 0.4263,
+      "step": 79008
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.046480526080334e-07,
+      "loss": 0.4771,
+      "step": 79009
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.046240156852654e-07,
+      "loss": 0.4856,
+      "step": 79010
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.045999791417834e-07,
+      "loss": 0.4297,
+      "step": 79011
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.045759429776071e-07,
+      "loss": 0.418,
+      "step": 79012
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.045519071927535e-07,
+      "loss": 0.3604,
+      "step": 79013
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.045278717872424e-07,
+      "loss": 0.2759,
+      "step": 79014
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.04503836761091e-07,
+      "loss": 0.4966,
+      "step": 79015
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.044798021143182e-07,
+      "loss": 0.4038,
+      "step": 79016
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.044557678469425e-07,
+      "loss": 0.4248,
+      "step": 79017
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.044317339589825e-07,
+      "loss": 0.4478,
+      "step": 79018
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.044077004504561e-07,
+      "loss": 0.3789,
+      "step": 79019
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.043836673213818e-07,
+      "loss": 0.3716,
+      "step": 79020
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.043596345717783e-07,
+      "loss": 0.2028,
+      "step": 79021
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.04335602201664e-07,
+      "loss": 0.3611,
+      "step": 79022
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.043115702110573e-07,
+      "loss": 0.4893,
+      "step": 79023
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.042875385999761e-07,
+      "loss": 0.5254,
+      "step": 79024
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.04263507368439e-07,
+      "loss": 0.2472,
+      "step": 79025
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.042394765164646e-07,
+      "loss": 0.4561,
+      "step": 79026
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.042154460440716e-07,
+      "loss": 0.3608,
+      "step": 79027
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.041914159512777e-07,
+      "loss": 0.436,
+      "step": 79028
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.041673862381015e-07,
+      "loss": 0.4047,
+      "step": 79029
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.041433569045621e-07,
+      "loss": 0.4951,
+      "step": 79030
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.041193279506769e-07,
+      "loss": 0.3745,
+      "step": 79031
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.040952993764651e-07,
+      "loss": 0.3303,
+      "step": 79032
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.040712711819444e-07,
+      "loss": 0.4448,
+      "step": 79033
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.040472433671333e-07,
+      "loss": 0.3965,
+      "step": 79034
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.04023215932051e-07,
+      "loss": 0.3069,
+      "step": 79035
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.039991888767149e-07,
+      "loss": 0.457,
+      "step": 79036
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.039751622011438e-07,
+      "loss": 0.498,
+      "step": 79037
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.039511359053561e-07,
+      "loss": 0.4135,
+      "step": 79038
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.039271099893706e-07,
+      "loss": 0.2627,
+      "step": 79039
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.03903084453205e-07,
+      "loss": 0.4775,
+      "step": 79040
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.038790592968778e-07,
+      "loss": 0.3792,
+      "step": 79041
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.038550345204083e-07,
+      "loss": 0.3804,
+      "step": 79042
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.038310101238134e-07,
+      "loss": 0.3402,
+      "step": 79043
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.038069861071129e-07,
+      "loss": 0.3284,
+      "step": 79044
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.037829624703241e-07,
+      "loss": 0.4028,
+      "step": 79045
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.03758939213466e-07,
+      "loss": 0.3549,
+      "step": 79046
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.037349163365572e-07,
+      "loss": 0.5415,
+      "step": 79047
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.037108938396153e-07,
+      "loss": 0.4185,
+      "step": 79048
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.036868717226592e-07,
+      "loss": 0.321,
+      "step": 79049
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.036628499857072e-07,
+      "loss": 0.2151,
+      "step": 79050
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.036388286287781e-07,
+      "loss": 0.2374,
+      "step": 79051
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.0361480765189e-07,
+      "loss": 0.4141,
+      "step": 79052
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.035907870550608e-07,
+      "loss": 0.3271,
+      "step": 79053
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.035667668383094e-07,
+      "loss": 0.5183,
+      "step": 79054
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.03542747001654e-07,
+      "loss": 0.4272,
+      "step": 79055
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.035187275451134e-07,
+      "loss": 0.2955,
+      "step": 79056
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.034947084687055e-07,
+      "loss": 0.4487,
+      "step": 79057
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.034706897724487e-07,
+      "loss": 0.3737,
+      "step": 79058
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.034466714563617e-07,
+      "loss": 0.3623,
+      "step": 79059
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.034226535204631e-07,
+      "loss": 0.5151,
+      "step": 79060
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.033986359647704e-07,
+      "loss": 0.4712,
+      "step": 79061
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.033746187893031e-07,
+      "loss": 0.4595,
+      "step": 79062
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.033506019940785e-07,
+      "loss": 0.374,
+      "step": 79063
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.03326585579116e-07,
+      "loss": 0.5049,
+      "step": 79064
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.033025695444331e-07,
+      "loss": 0.402,
+      "step": 79065
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.032785538900486e-07,
+      "loss": 0.3306,
+      "step": 79066
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.032545386159807e-07,
+      "loss": 0.3538,
+      "step": 79067
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.032305237222485e-07,
+      "loss": 0.4128,
+      "step": 79068
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.032065092088694e-07,
+      "loss": 0.3789,
+      "step": 79069
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.031824950758623e-07,
+      "loss": 0.3569,
+      "step": 79070
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.031584813232454e-07,
+      "loss": 0.4541,
+      "step": 79071
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.031344679510378e-07,
+      "loss": 0.5137,
+      "step": 79072
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.031104549592571e-07,
+      "loss": 0.3862,
+      "step": 79073
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.030864423479215e-07,
+      "loss": 0.4658,
+      "step": 79074
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.030624301170499e-07,
+      "loss": 0.2529,
+      "step": 79075
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.030384182666605e-07,
+      "loss": 0.4187,
+      "step": 79076
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.03014406796772e-07,
+      "loss": 0.3853,
+      "step": 79077
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.029903957074022e-07,
+      "loss": 0.3457,
+      "step": 79078
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.029663849985699e-07,
+      "loss": 0.3654,
+      "step": 79079
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.029423746702932e-07,
+      "loss": 0.4136,
+      "step": 79080
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.029183647225912e-07,
+      "loss": 0.4873,
+      "step": 79081
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.028943551554816e-07,
+      "loss": 0.4551,
+      "step": 79082
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.028703459689827e-07,
+      "loss": 0.3755,
+      "step": 79083
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.028463371631131e-07,
+      "loss": 0.5171,
+      "step": 79084
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.028223287378916e-07,
+      "loss": 0.5884,
+      "step": 79085
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.027983206933358e-07,
+      "loss": 0.4453,
+      "step": 79086
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.027743130294643e-07,
+      "loss": 0.4946,
+      "step": 79087
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.027503057462959e-07,
+      "loss": 0.5732,
+      "step": 79088
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.027262988438489e-07,
+      "loss": 0.4272,
+      "step": 79089
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.027022923221412e-07,
+      "loss": 0.3983,
+      "step": 79090
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.026782861811914e-07,
+      "loss": 0.3595,
+      "step": 79091
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.026542804210185e-07,
+      "loss": 0.4224,
+      "step": 79092
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.026302750416399e-07,
+      "loss": 0.5059,
+      "step": 79093
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.026062700430748e-07,
+      "loss": 0.4849,
+      "step": 79094
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.025822654253409e-07,
+      "loss": 0.3867,
+      "step": 79095
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.025582611884567e-07,
+      "loss": 0.4064,
+      "step": 79096
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.02534257332441e-07,
+      "loss": 0.355,
+      "step": 79097
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.025102538573123e-07,
+      "loss": 0.4854,
+      "step": 79098
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.02486250763088e-07,
+      "loss": 0.4692,
+      "step": 79099
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.024622480497873e-07,
+      "loss": 0.438,
+      "step": 79100
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.024382457174283e-07,
+      "loss": 0.4653,
+      "step": 79101
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.024142437660304e-07,
+      "loss": 0.4001,
+      "step": 79102
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.023902421956101e-07,
+      "loss": 0.4214,
+      "step": 79103
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.023662410061866e-07,
+      "loss": 0.4143,
+      "step": 79104
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.023422401977785e-07,
+      "loss": 0.3359,
+      "step": 79105
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.023182397704047e-07,
+      "loss": 0.5044,
+      "step": 79106
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.022942397240822e-07,
+      "loss": 0.3821,
+      "step": 79107
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.022702400588303e-07,
+      "loss": 0.3687,
+      "step": 79108
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.022462407746672e-07,
+      "loss": 0.4868,
+      "step": 79109
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.022222418716116e-07,
+      "loss": 0.4521,
+      "step": 79110
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.02198243349681e-07,
+      "loss": 0.6328,
+      "step": 79111
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.021742452088949e-07,
+      "loss": 0.3298,
+      "step": 79112
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.021502474492707e-07,
+      "loss": 0.4658,
+      "step": 79113
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.021262500708271e-07,
+      "loss": 0.3872,
+      "step": 79114
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.02102253073583e-07,
+      "loss": 0.455,
+      "step": 79115
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.020782564575559e-07,
+      "loss": 0.4697,
+      "step": 79116
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.020542602227644e-07,
+      "loss": 0.3115,
+      "step": 79117
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.020302643692274e-07,
+      "loss": 0.4878,
+      "step": 79118
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.020062688969631e-07,
+      "loss": 0.4458,
+      "step": 79119
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.019822738059892e-07,
+      "loss": 0.4146,
+      "step": 79120
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.019582790963248e-07,
+      "loss": 0.4521,
+      "step": 79121
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.019342847679882e-07,
+      "loss": 0.4663,
+      "step": 79122
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.019102908209978e-07,
+      "loss": 0.2958,
+      "step": 79123
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.018862972553715e-07,
+      "loss": 0.4561,
+      "step": 79124
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.018623040711278e-07,
+      "loss": 0.4473,
+      "step": 79125
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.018383112682852e-07,
+      "loss": 0.4326,
+      "step": 79126
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.018143188468625e-07,
+      "loss": 0.6157,
+      "step": 79127
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.017903268068773e-07,
+      "loss": 0.2894,
+      "step": 79128
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.017663351483484e-07,
+      "loss": 0.4036,
+      "step": 79129
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.01742343871294e-07,
+      "loss": 0.3962,
+      "step": 79130
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.01718352975733e-07,
+      "loss": 0.3538,
+      "step": 79131
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.016943624616834e-07,
+      "loss": 0.4863,
+      "step": 79132
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.016703723291631e-07,
+      "loss": 0.4468,
+      "step": 79133
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.016463825781909e-07,
+      "loss": 0.4478,
+      "step": 79134
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.016223932087851e-07,
+      "loss": 0.4287,
+      "step": 79135
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.015984042209646e-07,
+      "loss": 0.3638,
+      "step": 79136
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.015744156147468e-07,
+      "loss": 0.3286,
+      "step": 79137
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.015504273901505e-07,
+      "loss": 0.5166,
+      "step": 79138
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.015264395471948e-07,
+      "loss": 0.3276,
+      "step": 79139
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.015024520858967e-07,
+      "loss": 0.3345,
+      "step": 79140
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.014784650062754e-07,
+      "loss": 0.522,
+      "step": 79141
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.014544783083495e-07,
+      "loss": 0.4268,
+      "step": 79142
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.014304919921364e-07,
+      "loss": 0.5234,
+      "step": 79143
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.014065060576556e-07,
+      "loss": 0.4434,
+      "step": 79144
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.013825205049245e-07,
+      "loss": 0.5146,
+      "step": 79145
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.013585353339618e-07,
+      "loss": 0.4668,
+      "step": 79146
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.013345505447861e-07,
+      "loss": 0.4653,
+      "step": 79147
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.013105661374158e-07,
+      "loss": 0.3975,
+      "step": 79148
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.012865821118688e-07,
+      "loss": 0.4414,
+      "step": 79149
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.012625984681637e-07,
+      "loss": 0.3687,
+      "step": 79150
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.012386152063191e-07,
+      "loss": 0.4148,
+      "step": 79151
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.012146323263533e-07,
+      "loss": 0.4526,
+      "step": 79152
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.011906498282845e-07,
+      "loss": 0.3459,
+      "step": 79153
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.011666677121309e-07,
+      "loss": 0.4026,
+      "step": 79154
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.011426859779109e-07,
+      "loss": 0.4116,
+      "step": 79155
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.011187046256434e-07,
+      "loss": 0.302,
+      "step": 79156
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.01094723655346e-07,
+      "loss": 0.3964,
+      "step": 79157
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.010707430670374e-07,
+      "loss": 0.5586,
+      "step": 79158
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.01046762860736e-07,
+      "loss": 0.4214,
+      "step": 79159
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.010227830364606e-07,
+      "loss": 0.3301,
+      "step": 79160
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.009988035942287e-07,
+      "loss": 0.4521,
+      "step": 79161
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.009748245340595e-07,
+      "loss": 0.4506,
+      "step": 79162
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.009508458559704e-07,
+      "loss": 0.418,
+      "step": 79163
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.009268675599804e-07,
+      "loss": 0.3616,
+      "step": 79164
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.009028896461081e-07,
+      "loss": 0.4138,
+      "step": 79165
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.00878912114371e-07,
+      "loss": 0.4424,
+      "step": 79166
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.008549349647883e-07,
+      "loss": 0.498,
+      "step": 79167
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.008309581973777e-07,
+      "loss": 0.4473,
+      "step": 79168
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.008069818121585e-07,
+      "loss": 0.4922,
+      "step": 79169
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.007830058091479e-07,
+      "loss": 0.4976,
+      "step": 79170
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.007590301883648e-07,
+      "loss": 0.5117,
+      "step": 79171
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.00735054949828e-07,
+      "loss": 0.415,
+      "step": 79172
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.007110800935554e-07,
+      "loss": 0.5625,
+      "step": 79173
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.006871056195651e-07,
+      "loss": 0.4272,
+      "step": 79174
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.006631315278755e-07,
+      "loss": 0.2542,
+      "step": 79175
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.006391578185054e-07,
+      "loss": 0.4712,
+      "step": 79176
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.006151844914732e-07,
+      "loss": 0.3774,
+      "step": 79177
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.005912115467965e-07,
+      "loss": 0.4165,
+      "step": 79178
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.005672389844944e-07,
+      "loss": 0.3528,
+      "step": 79179
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.005432668045848e-07,
+      "loss": 0.3735,
+      "step": 79180
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.005192950070867e-07,
+      "loss": 0.5107,
+      "step": 79181
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.00495323592018e-07,
+      "loss": 0.2869,
+      "step": 79182
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.004713525593966e-07,
+      "loss": 0.4355,
+      "step": 79183
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.004473819092415e-07,
+      "loss": 0.3921,
+      "step": 79184
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.004234116415708e-07,
+      "loss": 0.3604,
+      "step": 79185
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.003994417564033e-07,
+      "loss": 0.4033,
+      "step": 79186
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.003754722537565e-07,
+      "loss": 0.3154,
+      "step": 79187
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.003515031336494e-07,
+      "loss": 0.3562,
+      "step": 79188
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.003275343961e-07,
+      "loss": 0.3195,
+      "step": 79189
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.003035660411274e-07,
+      "loss": 0.2976,
+      "step": 79190
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.002795980687489e-07,
+      "loss": 0.395,
+      "step": 79191
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.002556304789836e-07,
+      "loss": 0.5532,
+      "step": 79192
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.002316632718493e-07,
+      "loss": 0.4424,
+      "step": 79193
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.002076964473652e-07,
+      "loss": 0.3875,
+      "step": 79194
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.001837300055484e-07,
+      "loss": 0.314,
+      "step": 79195
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.001597639464181e-07,
+      "loss": 0.5049,
+      "step": 79196
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.001357982699925e-07,
+      "loss": 0.4302,
+      "step": 79197
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.001118329762902e-07,
+      "loss": 0.4258,
+      "step": 79198
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.00087868065329e-07,
+      "loss": 0.4404,
+      "step": 79199
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.000639035371274e-07,
+      "loss": 0.4424,
+      "step": 79200
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.00039939391704e-07,
+      "loss": 0.3699,
+      "step": 79201
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 5.000159756290775e-07,
+      "loss": 0.4722,
+      "step": 79202
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.999920122492657e-07,
+      "loss": 0.478,
+      "step": 79203
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.999680492522867e-07,
+      "loss": 0.2145,
+      "step": 79204
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.99944086638159e-07,
+      "loss": 0.5137,
+      "step": 79205
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.999201244069012e-07,
+      "loss": 0.4175,
+      "step": 79206
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.99896162558532e-07,
+      "loss": 0.2743,
+      "step": 79207
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.998722010930688e-07,
+      "loss": 0.3722,
+      "step": 79208
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.998482400105306e-07,
+      "loss": 0.3262,
+      "step": 79209
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.99824279310936e-07,
+      "loss": 0.4075,
+      "step": 79210
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.998003189943028e-07,
+      "loss": 0.3931,
+      "step": 79211
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.997763590606493e-07,
+      "loss": 0.4312,
+      "step": 79212
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.99752399509994e-07,
+      "loss": 0.2446,
+      "step": 79213
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.997284403423551e-07,
+      "loss": 0.4111,
+      "step": 79214
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.997044815577517e-07,
+      "loss": 0.4424,
+      "step": 79215
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.996805231562012e-07,
+      "loss": 0.4238,
+      "step": 79216
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.996565651377221e-07,
+      "loss": 0.3296,
+      "step": 79217
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.996326075023332e-07,
+      "loss": 0.4199,
+      "step": 79218
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.996086502500529e-07,
+      "loss": 0.3354,
+      "step": 79219
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.995846933808989e-07,
+      "loss": 0.3339,
+      "step": 79220
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.995607368948902e-07,
+      "loss": 0.279,
+      "step": 79221
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.995367807920446e-07,
+      "loss": 0.4673,
+      "step": 79222
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.995128250723804e-07,
+      "loss": 0.4097,
+      "step": 79223
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.994888697359168e-07,
+      "loss": 0.4795,
+      "step": 79224
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.994649147826712e-07,
+      "loss": 0.4893,
+      "step": 79225
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.994409602126622e-07,
+      "loss": 0.4434,
+      "step": 79226
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.994170060259086e-07,
+      "loss": 0.3901,
+      "step": 79227
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.993930522224281e-07,
+      "loss": 0.5073,
+      "step": 79228
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.993690988022394e-07,
+      "loss": 0.3535,
+      "step": 79229
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.993451457653605e-07,
+      "loss": 0.2915,
+      "step": 79230
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.993211931118106e-07,
+      "loss": 0.4878,
+      "step": 79231
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.992972408416074e-07,
+      "loss": 0.4561,
+      "step": 79232
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.992732889547689e-07,
+      "loss": 0.5068,
+      "step": 79233
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.992493374513137e-07,
+      "loss": 0.4883,
+      "step": 79234
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.992253863312604e-07,
+      "loss": 0.4814,
+      "step": 79235
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.992014355946274e-07,
+      "loss": 0.4819,
+      "step": 79236
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.991774852414326e-07,
+      "loss": 0.4122,
+      "step": 79237
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.991535352716945e-07,
+      "loss": 0.3584,
+      "step": 79238
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.991295856854316e-07,
+      "loss": 0.4766,
+      "step": 79239
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.991056364826626e-07,
+      "loss": 0.564,
+      "step": 79240
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.990816876634052e-07,
+      "loss": 0.4985,
+      "step": 79241
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.990577392276776e-07,
+      "loss": 0.3413,
+      "step": 79242
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.990337911754985e-07,
+      "loss": 0.4194,
+      "step": 79243
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.990098435068861e-07,
+      "loss": 0.3704,
+      "step": 79244
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.989858962218593e-07,
+      "loss": 0.5371,
+      "step": 79245
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.989619493204355e-07,
+      "loss": 0.395,
+      "step": 79246
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.989380028026335e-07,
+      "loss": 0.3823,
+      "step": 79247
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.989140566684721e-07,
+      "loss": 0.2991,
+      "step": 79248
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.988901109179687e-07,
+      "loss": 0.4409,
+      "step": 79249
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.988661655511422e-07,
+      "loss": 0.3828,
+      "step": 79250
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.98842220568011e-07,
+      "loss": 0.4902,
+      "step": 79251
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.988182759685932e-07,
+      "loss": 0.3936,
+      "step": 79252
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.987943317529073e-07,
+      "loss": 0.3853,
+      "step": 79253
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.987703879209712e-07,
+      "loss": 0.2297,
+      "step": 79254
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.987464444728037e-07,
+      "loss": 0.4536,
+      "step": 79255
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.98722501408423e-07,
+      "loss": 0.4269,
+      "step": 79256
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.986985587278477e-07,
+      "loss": 0.4233,
+      "step": 79257
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.986746164310955e-07,
+      "loss": 0.458,
+      "step": 79258
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.986506745181852e-07,
+      "loss": 0.54,
+      "step": 79259
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.98626732989135e-07,
+      "loss": 0.4937,
+      "step": 79260
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.986027918439634e-07,
+      "loss": 0.4976,
+      "step": 79261
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.98578851082689e-07,
+      "loss": 0.4404,
+      "step": 79262
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.98554910705329e-07,
+      "loss": 0.3605,
+      "step": 79263
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.985309707119026e-07,
+      "loss": 0.2917,
+      "step": 79264
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.985070311024283e-07,
+      "loss": 0.3784,
+      "step": 79265
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.984830918769237e-07,
+      "loss": 0.418,
+      "step": 79266
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.984591530354076e-07,
+      "loss": 0.4404,
+      "step": 79267
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.984352145778983e-07,
+      "loss": 0.481,
+      "step": 79268
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.984112765044144e-07,
+      "loss": 0.3794,
+      "step": 79269
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.983873388149735e-07,
+      "loss": 0.3821,
+      "step": 79270
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.983634015095948e-07,
+      "loss": 0.4136,
+      "step": 79271
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.983394645882958e-07,
+      "loss": 0.421,
+      "step": 79272
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.983155280510952e-07,
+      "loss": 0.3586,
+      "step": 79273
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.982915918980118e-07,
+      "loss": 0.4033,
+      "step": 79274
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.982676561290629e-07,
+      "loss": 0.4526,
+      "step": 79275
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.982437207442675e-07,
+      "loss": 0.4708,
+      "step": 79276
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.982197857436437e-07,
+      "loss": 0.4414,
+      "step": 79277
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.981958511272104e-07,
+      "loss": 0.3428,
+      "step": 79278
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.981719168949851e-07,
+      "loss": 0.4966,
+      "step": 79279
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.981479830469864e-07,
+      "loss": 0.3438,
+      "step": 79280
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.981240495832333e-07,
+      "loss": 0.3818,
+      "step": 79281
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.981001165037433e-07,
+      "loss": 0.4709,
+      "step": 79282
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.980761838085348e-07,
+      "loss": 0.4409,
+      "step": 79283
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.980522514976261e-07,
+      "loss": 0.3965,
+      "step": 79284
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.980283195710358e-07,
+      "loss": 0.4263,
+      "step": 79285
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.980043880287826e-07,
+      "loss": 0.4399,
+      "step": 79286
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.979804568708838e-07,
+      "loss": 0.4199,
+      "step": 79287
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.979565260973583e-07,
+      "loss": 0.3557,
+      "step": 79288
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.979325957082245e-07,
+      "loss": 0.458,
+      "step": 79289
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.97908665703501e-07,
+      "loss": 0.2637,
+      "step": 79290
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.978847360832057e-07,
+      "loss": 0.4756,
+      "step": 79291
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.978608068473566e-07,
+      "loss": 0.4614,
+      "step": 79292
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.978368779959724e-07,
+      "loss": 0.206,
+      "step": 79293
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.978129495290714e-07,
+      "loss": 0.3862,
+      "step": 79294
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.977890214466722e-07,
+      "loss": 0.3696,
+      "step": 79295
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.977650937487927e-07,
+      "loss": 0.4138,
+      "step": 79296
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.977411664354512e-07,
+      "loss": 0.3521,
+      "step": 79297
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.977172395066663e-07,
+      "loss": 0.4165,
+      "step": 79298
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.976933129624565e-07,
+      "loss": 0.3799,
+      "step": 79299
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.976693868028396e-07,
+      "loss": 0.3347,
+      "step": 79300
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.976454610278344e-07,
+      "loss": 0.375,
+      "step": 79301
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.976215356374587e-07,
+      "loss": 0.4746,
+      "step": 79302
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.975976106317314e-07,
+      "loss": 0.3401,
+      "step": 79303
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.9757368601067e-07,
+      "loss": 0.4038,
+      "step": 79304
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.975497617742936e-07,
+      "loss": 0.3777,
+      "step": 79305
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.975258379226201e-07,
+      "loss": 0.3906,
+      "step": 79306
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.975019144556682e-07,
+      "loss": 0.4644,
+      "step": 79307
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.974779913734558e-07,
+      "loss": 0.3887,
+      "step": 79308
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.974540686760014e-07,
+      "loss": 0.3073,
+      "step": 79309
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.974301463633232e-07,
+      "loss": 0.4326,
+      "step": 79310
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.974062244354401e-07,
+      "loss": 0.3242,
+      "step": 79311
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.973823028923699e-07,
+      "loss": 0.3555,
+      "step": 79312
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.973583817341306e-07,
+      "loss": 0.4316,
+      "step": 79313
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.973344609607408e-07,
+      "loss": 0.3943,
+      "step": 79314
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.97310540572219e-07,
+      "loss": 0.4893,
+      "step": 79315
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.972866205685837e-07,
+      "loss": 0.314,
+      "step": 79316
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.972627009498526e-07,
+      "loss": 0.4795,
+      "step": 79317
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.972387817160443e-07,
+      "loss": 0.5,
+      "step": 79318
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.972148628671775e-07,
+      "loss": 0.3784,
+      "step": 79319
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.971909444032698e-07,
+      "loss": 0.46,
+      "step": 79320
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.971670263243403e-07,
+      "loss": 0.2705,
+      "step": 79321
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.971431086304066e-07,
+      "loss": 0.364,
+      "step": 79322
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.97119191321487e-07,
+      "loss": 0.4028,
+      "step": 79323
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.970952743976008e-07,
+      "loss": 0.4565,
+      "step": 79324
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.97071357858765e-07,
+      "loss": 0.5225,
+      "step": 79325
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.970474417049987e-07,
+      "loss": 0.4192,
+      "step": 79326
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.9702352593632e-07,
+      "loss": 0.3655,
+      "step": 79327
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.969996105527477e-07,
+      "loss": 0.3843,
+      "step": 79328
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.969756955542992e-07,
+      "loss": 0.3252,
+      "step": 79329
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.969517809409932e-07,
+      "loss": 0.562,
+      "step": 79330
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.969278667128486e-07,
+      "loss": 0.4365,
+      "step": 79331
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.969039528698828e-07,
+      "loss": 0.4502,
+      "step": 79332
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.968800394121148e-07,
+      "loss": 0.3181,
+      "step": 79333
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.968561263395624e-07,
+      "loss": 0.4407,
+      "step": 79334
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.968322136522441e-07,
+      "loss": 0.3535,
+      "step": 79335
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.968083013501786e-07,
+      "loss": 0.4277,
+      "step": 79336
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.967843894333834e-07,
+      "loss": 0.5371,
+      "step": 79337
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.967604779018772e-07,
+      "loss": 0.4771,
+      "step": 79338
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.967365667556786e-07,
+      "loss": 0.4395,
+      "step": 79339
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.967126559948058e-07,
+      "loss": 0.3794,
+      "step": 79340
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.966887456192771e-07,
+      "loss": 0.3464,
+      "step": 79341
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.966648356291103e-07,
+      "loss": 0.3755,
+      "step": 79342
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.966409260243241e-07,
+      "loss": 0.2068,
+      "step": 79343
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.966170168049367e-07,
+      "loss": 0.2556,
+      "step": 79344
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.96593107970967e-07,
+      "loss": 0.4041,
+      "step": 79345
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.965691995224325e-07,
+      "loss": 0.4229,
+      "step": 79346
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.965452914593517e-07,
+      "loss": 0.3582,
+      "step": 79347
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.965213837817429e-07,
+      "loss": 0.4121,
+      "step": 79348
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.964974764896252e-07,
+      "loss": 0.3511,
+      "step": 79349
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.964735695830156e-07,
+      "loss": 0.4507,
+      "step": 79350
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.964496630619334e-07,
+      "loss": 0.4033,
+      "step": 79351
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.964257569263963e-07,
+      "loss": 0.2864,
+      "step": 79352
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.964018511764232e-07,
+      "loss": 0.3652,
+      "step": 79353
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.963779458120316e-07,
+      "loss": 0.4031,
+      "step": 79354
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.963540408332404e-07,
+      "loss": 0.4932,
+      "step": 79355
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.963301362400676e-07,
+      "loss": 0.481,
+      "step": 79356
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.963062320325321e-07,
+      "loss": 0.4087,
+      "step": 79357
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.962823282106513e-07,
+      "loss": 0.23,
+      "step": 79358
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.962584247744441e-07,
+      "loss": 0.5635,
+      "step": 79359
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.962345217239285e-07,
+      "loss": 0.3905,
+      "step": 79360
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.962106190591235e-07,
+      "loss": 0.2839,
+      "step": 79361
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.961867167800467e-07,
+      "loss": 0.3682,
+      "step": 79362
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.961628148867164e-07,
+      "loss": 0.3267,
+      "step": 79363
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.961389133791509e-07,
+      "loss": 0.3682,
+      "step": 79364
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.961150122573688e-07,
+      "loss": 0.5068,
+      "step": 79365
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.960911115213885e-07,
+      "loss": 0.4487,
+      "step": 79366
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.960672111712278e-07,
+      "loss": 0.3857,
+      "step": 79367
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.960433112069051e-07,
+      "loss": 0.5044,
+      "step": 79368
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.960194116284389e-07,
+      "loss": 0.5786,
+      "step": 79369
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.95995512435848e-07,
+      "loss": 0.3027,
+      "step": 79370
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.959716136291501e-07,
+      "loss": 0.4692,
+      "step": 79371
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.959477152083631e-07,
+      "loss": 0.3972,
+      "step": 79372
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.959238171735058e-07,
+      "loss": 0.3085,
+      "step": 79373
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.958999195245968e-07,
+      "loss": 0.4878,
+      "step": 79374
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.958760222616538e-07,
+      "loss": 0.4121,
+      "step": 79375
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.958521253846952e-07,
+      "loss": 0.4048,
+      "step": 79376
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.958282288937395e-07,
+      "loss": 0.3848,
+      "step": 79377
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.958043327888054e-07,
+      "loss": 0.5508,
+      "step": 79378
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.957804370699102e-07,
+      "loss": 0.439,
+      "step": 79379
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.957565417370729e-07,
+      "loss": 0.4565,
+      "step": 79380
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.957326467903119e-07,
+      "loss": 0.4316,
+      "step": 79381
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.957087522296449e-07,
+      "loss": 0.4517,
+      "step": 79382
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.95684858055091e-07,
+      "loss": 0.4456,
+      "step": 79383
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.956609642666675e-07,
+      "loss": 0.5371,
+      "step": 79384
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.956370708643932e-07,
+      "loss": 0.3523,
+      "step": 79385
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.956131778482864e-07,
+      "loss": 0.3502,
+      "step": 79386
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.955892852183659e-07,
+      "loss": 0.3015,
+      "step": 79387
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.955653929746489e-07,
+      "loss": 0.4971,
+      "step": 79388
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.955415011171544e-07,
+      "loss": 0.3207,
+      "step": 79389
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.95517609645901e-07,
+      "loss": 0.3926,
+      "step": 79390
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.954937185609065e-07,
+      "loss": 0.4604,
+      "step": 79391
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.954698278621889e-07,
+      "loss": 0.5571,
+      "step": 79392
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.95445937549767e-07,
+      "loss": 0.2381,
+      "step": 79393
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.954220476236587e-07,
+      "loss": 0.345,
+      "step": 79394
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.953981580838831e-07,
+      "loss": 0.4595,
+      "step": 79395
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.953742689304574e-07,
+      "loss": 0.4141,
+      "step": 79396
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 4.953503801634005e-07,
+      "loss": 0.3691,
+      "step": 79397
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.953264917827307e-07,
+      "loss": 0.2675,
+      "step": 79398
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.953026037884664e-07,
+      "loss": 0.457,
+      "step": 79399
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.952787161806255e-07,
+      "loss": 0.4429,
+      "step": 79400
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.952548289592267e-07,
+      "loss": 0.5244,
+      "step": 79401
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.952309421242877e-07,
+      "loss": 0.4829,
+      "step": 79402
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.952070556758272e-07,
+      "loss": 0.3447,
+      "step": 79403
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.951831696138638e-07,
+      "loss": 0.3168,
+      "step": 79404
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.951592839384151e-07,
+      "loss": 0.5513,
+      "step": 79405
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.951353986494997e-07,
+      "loss": 0.4702,
+      "step": 79406
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.95111513747136e-07,
+      "loss": 0.4053,
+      "step": 79407
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.950876292313425e-07,
+      "loss": 0.4502,
+      "step": 79408
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.950637451021367e-07,
+      "loss": 0.3899,
+      "step": 79409
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.950398613595375e-07,
+      "loss": 0.376,
+      "step": 79410
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.950159780035634e-07,
+      "loss": 0.3843,
+      "step": 79411
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.949920950342325e-07,
+      "loss": 0.3293,
+      "step": 79412
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.949682124515624e-07,
+      "loss": 0.4551,
+      "step": 79413
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.94944330255572e-07,
+      "loss": 0.3389,
+      "step": 79414
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.949204484462793e-07,
+      "loss": 0.3174,
+      "step": 79415
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.948965670237033e-07,
+      "loss": 0.4966,
+      "step": 79416
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.948726859878613e-07,
+      "loss": 0.5581,
+      "step": 79417
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.948488053387722e-07,
+      "loss": 0.3721,
+      "step": 79418
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.948249250764542e-07,
+      "loss": 0.3232,
+      "step": 79419
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.948010452009258e-07,
+      "loss": 0.6028,
+      "step": 79420
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.94777165712205e-07,
+      "loss": 0.4521,
+      "step": 79421
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.947532866103096e-07,
+      "loss": 0.3718,
+      "step": 79422
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.947294078952586e-07,
+      "loss": 0.245,
+      "step": 79423
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.9470552956707e-07,
+      "loss": 0.4736,
+      "step": 79424
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.946816516257625e-07,
+      "loss": 0.3513,
+      "step": 79425
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.946577740713536e-07,
+      "loss": 0.3912,
+      "step": 79426
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.94633896903862e-07,
+      "loss": 0.4897,
+      "step": 79427
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.946100201233064e-07,
+      "loss": 0.4155,
+      "step": 79428
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.945861437297043e-07,
+      "loss": 0.4287,
+      "step": 79429
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.945622677230742e-07,
+      "loss": 0.311,
+      "step": 79430
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.945383921034351e-07,
+      "loss": 0.3926,
+      "step": 79431
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.945145168708043e-07,
+      "loss": 0.5488,
+      "step": 79432
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.94490642025201e-07,
+      "loss": 0.4536,
+      "step": 79433
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.944667675666424e-07,
+      "loss": 0.5254,
+      "step": 79434
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.944428934951473e-07,
+      "loss": 0.4082,
+      "step": 79435
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.944190198107342e-07,
+      "loss": 0.3524,
+      "step": 79436
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.943951465134216e-07,
+      "loss": 0.52,
+      "step": 79437
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.943712736032269e-07,
+      "loss": 0.4243,
+      "step": 79438
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.943474010801688e-07,
+      "loss": 0.3708,
+      "step": 79439
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.943235289442657e-07,
+      "loss": 0.4565,
+      "step": 79440
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.942996571955364e-07,
+      "loss": 0.4722,
+      "step": 79441
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.942757858339983e-07,
+      "loss": 0.3533,
+      "step": 79442
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.942519148596698e-07,
+      "loss": 0.5801,
+      "step": 79443
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.942280442725692e-07,
+      "loss": 0.4531,
+      "step": 79444
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.942041740727154e-07,
+      "loss": 0.3157,
+      "step": 79445
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.941803042601259e-07,
+      "loss": 0.4658,
+      "step": 79446
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.941564348348192e-07,
+      "loss": 0.4365,
+      "step": 79447
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.941325657968138e-07,
+      "loss": 0.4448,
+      "step": 79448
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.941086971461279e-07,
+      "loss": 0.4451,
+      "step": 79449
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.940848288827796e-07,
+      "loss": 0.2884,
+      "step": 79450
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.940609610067875e-07,
+      "loss": 0.3097,
+      "step": 79451
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.940370935181694e-07,
+      "loss": 0.3396,
+      "step": 79452
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.940132264169438e-07,
+      "loss": 0.5122,
+      "step": 79453
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.939893597031294e-07,
+      "loss": 0.4663,
+      "step": 79454
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.939654933767437e-07,
+      "loss": 0.5486,
+      "step": 79455
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.939416274378052e-07,
+      "loss": 0.4761,
+      "step": 79456
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.939177618863325e-07,
+      "loss": 0.5088,
+      "step": 79457
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93893896722344e-07,
+      "loss": 0.4673,
+      "step": 79458
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.938700319458574e-07,
+      "loss": 0.3958,
+      "step": 79459
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.938461675568911e-07,
+      "loss": 0.2576,
+      "step": 79460
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.938223035554639e-07,
+      "loss": 0.4873,
+      "step": 79461
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.937984399415939e-07,
+      "loss": 0.3787,
+      "step": 79462
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.937745767152985e-07,
+      "loss": 0.4634,
+      "step": 79463
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.937507138765968e-07,
+      "loss": 0.564,
+      "step": 79464
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.937268514255068e-07,
+      "loss": 0.4922,
+      "step": 79465
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.937029893620475e-07,
+      "loss": 0.4199,
+      "step": 79466
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.936791276862358e-07,
+      "loss": 0.4265,
+      "step": 79467
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93655266398091e-07,
+      "loss": 0.3398,
+      "step": 79468
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.936314054976309e-07,
+      "loss": 0.406,
+      "step": 79469
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.936075449848744e-07,
+      "loss": 0.3894,
+      "step": 79470
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.935836848598394e-07,
+      "loss": 0.4597,
+      "step": 79471
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.935598251225435e-07,
+      "loss": 0.4097,
+      "step": 79472
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.935359657730057e-07,
+      "loss": 0.4014,
+      "step": 79473
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.935121068112441e-07,
+      "loss": 0.4297,
+      "step": 79474
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.934882482372774e-07,
+      "loss": 0.5454,
+      "step": 79475
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93464390051123e-07,
+      "loss": 0.3933,
+      "step": 79476
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.934405322527997e-07,
+      "loss": 0.3945,
+      "step": 79477
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.934166748423256e-07,
+      "loss": 0.4419,
+      "step": 79478
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.933928178197195e-07,
+      "loss": 0.2797,
+      "step": 79479
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93368961184999e-07,
+      "loss": 0.3782,
+      "step": 79480
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.933451049381827e-07,
+      "loss": 0.3718,
+      "step": 79481
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.933212490792885e-07,
+      "loss": 0.4922,
+      "step": 79482
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.932973936083353e-07,
+      "loss": 0.3381,
+      "step": 79483
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.932735385253406e-07,
+      "loss": 0.4438,
+      "step": 79484
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93249683830323e-07,
+      "loss": 0.4819,
+      "step": 79485
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93225829523301e-07,
+      "loss": 0.4248,
+      "step": 79486
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.932019756042929e-07,
+      "loss": 0.3604,
+      "step": 79487
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.931781220733164e-07,
+      "loss": 0.5249,
+      "step": 79488
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.9315426893039e-07,
+      "loss": 0.3329,
+      "step": 79489
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.931304161755323e-07,
+      "loss": 0.4688,
+      "step": 79490
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.931065638087615e-07,
+      "loss": 0.436,
+      "step": 79491
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.930827118300958e-07,
+      "loss": 0.4707,
+      "step": 79492
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.93058860239553e-07,
+      "loss": 0.4346,
+      "step": 79493
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.930350090371516e-07,
+      "loss": 0.3633,
+      "step": 79494
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.930111582229102e-07,
+      "loss": 0.6035,
+      "step": 79495
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.929873077968471e-07,
+      "loss": 0.3816,
+      "step": 79496
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.929634577589798e-07,
+      "loss": 0.4399,
+      "step": 79497
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.929396081093273e-07,
+      "loss": 0.4062,
+      "step": 79498
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.929157588479078e-07,
+      "loss": 0.3752,
+      "step": 79499
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.928919099747395e-07,
+      "loss": 0.4507,
+      "step": 79500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.928680614898402e-07,
+      "loss": 0.2661,
+      "step": 79501
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.928442133932283e-07,
+      "loss": 0.45,
+      "step": 79502
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.928203656849224e-07,
+      "loss": 0.3602,
+      "step": 79503
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.927965183649412e-07,
+      "loss": 0.52,
+      "step": 79504
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.927726714333016e-07,
+      "loss": 0.3542,
+      "step": 79505
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.92748824890023e-07,
+      "loss": 0.3542,
+      "step": 79506
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.927249787351232e-07,
+      "loss": 0.3661,
+      "step": 79507
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.92701132968621e-07,
+      "loss": 0.4453,
+      "step": 79508
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.926772875905336e-07,
+      "loss": 0.3373,
+      "step": 79509
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.926534426008805e-07,
+      "loss": 0.4697,
+      "step": 79510
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.926295979996787e-07,
+      "loss": 0.4165,
+      "step": 79511
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.926057537869474e-07,
+      "loss": 0.5171,
+      "step": 79512
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.925819099627047e-07,
+      "loss": 0.4316,
+      "step": 79513
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.925580665269684e-07,
+      "loss": 0.3118,
+      "step": 79514
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.92534223479757e-07,
+      "loss": 0.3448,
+      "step": 79515
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.925103808210892e-07,
+      "loss": 0.3127,
+      "step": 79516
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.924865385509826e-07,
+      "loss": 0.415,
+      "step": 79517
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.924626966694556e-07,
+      "loss": 0.3618,
+      "step": 79518
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.924388551765266e-07,
+      "loss": 0.4746,
+      "step": 79519
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.924150140722142e-07,
+      "loss": 0.4346,
+      "step": 79520
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.923911733565364e-07,
+      "loss": 0.3628,
+      "step": 79521
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.923673330295108e-07,
+      "loss": 0.4248,
+      "step": 79522
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.923434930911561e-07,
+      "loss": 0.3481,
+      "step": 79523
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.92319653541491e-07,
+      "loss": 0.4497,
+      "step": 79524
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.922958143805335e-07,
+      "loss": 0.4675,
+      "step": 79525
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.922719756083015e-07,
+      "loss": 0.3623,
+      "step": 79526
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.922481372248134e-07,
+      "loss": 0.2687,
+      "step": 79527
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.922242992300876e-07,
+      "loss": 0.3953,
+      "step": 79528
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.922004616241427e-07,
+      "loss": 0.3955,
+      "step": 79529
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.921766244069965e-07,
+      "loss": 0.2925,
+      "step": 79530
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.921527875786669e-07,
+      "loss": 0.3394,
+      "step": 79531
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.921289511391727e-07,
+      "loss": 0.4214,
+      "step": 79532
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.921051150885324e-07,
+      "loss": 0.4536,
+      "step": 79533
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.920812794267633e-07,
+      "loss": 0.2861,
+      "step": 79534
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.920574441538844e-07,
+      "loss": 0.2394,
+      "step": 79535
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.920336092699137e-07,
+      "loss": 0.3516,
+      "step": 79536
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.920097747748699e-07,
+      "loss": 0.3381,
+      "step": 79537
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.919859406687704e-07,
+      "loss": 0.4219,
+      "step": 79538
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.919621069516339e-07,
+      "loss": 0.4182,
+      "step": 79539
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.919382736234791e-07,
+      "loss": 0.3025,
+      "step": 79540
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.919144406843234e-07,
+      "loss": 0.377,
+      "step": 79541
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.918906081341858e-07,
+      "loss": 0.3752,
+      "step": 79542
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.918667759730838e-07,
+      "loss": 0.3486,
+      "step": 79543
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.918429442010361e-07,
+      "loss": 0.5542,
+      "step": 79544
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.918191128180609e-07,
+      "loss": 0.3599,
+      "step": 79545
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.917952818241769e-07,
+      "loss": 0.481,
+      "step": 79546
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.917714512194013e-07,
+      "loss": 0.4165,
+      "step": 79547
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.917476210037531e-07,
+      "loss": 0.5503,
+      "step": 79548
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.917237911772502e-07,
+      "loss": 0.498,
+      "step": 79549
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.91699961739912e-07,
+      "loss": 0.3799,
+      "step": 79550
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.916761326917549e-07,
+      "loss": 0.3892,
+      "step": 79551
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.91652304032798e-07,
+      "loss": 0.4421,
+      "step": 79552
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.916284757630595e-07,
+      "loss": 0.4761,
+      "step": 79553
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.916046478825583e-07,
+      "loss": 0.4673,
+      "step": 79554
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.915808203913114e-07,
+      "loss": 0.4941,
+      "step": 79555
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.915569932893378e-07,
+      "loss": 0.3926,
+      "step": 79556
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.915331665766557e-07,
+      "loss": 0.4233,
+      "step": 79557
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.915093402532836e-07,
+      "loss": 0.4229,
+      "step": 79558
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.91485514319239e-07,
+      "loss": 0.3726,
+      "step": 79559
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.914616887745409e-07,
+      "loss": 0.3557,
+      "step": 79560
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.91437863619207e-07,
+      "loss": 0.4033,
+      "step": 79561
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.914140388532556e-07,
+      "loss": 0.418,
+      "step": 79562
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.913902144767055e-07,
+      "loss": 0.4912,
+      "step": 79563
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.913663904895742e-07,
+      "loss": 0.4434,
+      "step": 79564
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.913425668918802e-07,
+      "loss": 0.3,
+      "step": 79565
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.913187436836417e-07,
+      "loss": 0.4336,
+      "step": 79566
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.912949208648777e-07,
+      "loss": 0.3121,
+      "step": 79567
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.912710984356053e-07,
+      "loss": 0.4253,
+      "step": 79568
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.912472763958431e-07,
+      "loss": 0.4722,
+      "step": 79569
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.912234547456099e-07,
+      "loss": 0.3373,
+      "step": 79570
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.911996334849235e-07,
+      "loss": 0.3716,
+      "step": 79571
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.911758126138019e-07,
+      "loss": 0.324,
+      "step": 79572
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.911519921322634e-07,
+      "loss": 0.3545,
+      "step": 79573
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.911281720403267e-07,
+      "loss": 0.4883,
+      "step": 79574
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.911043523380098e-07,
+      "loss": 0.3237,
+      "step": 79575
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.910805330253307e-07,
+      "loss": 0.4092,
+      "step": 79576
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.910567141023078e-07,
+      "loss": 0.5005,
+      "step": 79577
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.910328955689595e-07,
+      "loss": 0.4858,
+      "step": 79578
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.910090774253041e-07,
+      "loss": 0.5317,
+      "step": 79579
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.909852596713596e-07,
+      "loss": 0.5479,
+      "step": 79580
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.909614423071441e-07,
+      "loss": 0.4531,
+      "step": 79581
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.909376253326759e-07,
+      "loss": 0.3828,
+      "step": 79582
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.909138087479734e-07,
+      "loss": 0.48,
+      "step": 79583
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.908899925530552e-07,
+      "loss": 0.4028,
+      "step": 79584
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.908661767479386e-07,
+      "loss": 0.4321,
+      "step": 79585
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.908423613326425e-07,
+      "loss": 0.5117,
+      "step": 79586
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.908185463071849e-07,
+      "loss": 0.5542,
+      "step": 79587
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.907947316715845e-07,
+      "loss": 0.522,
+      "step": 79588
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.907709174258588e-07,
+      "loss": 0.4644,
+      "step": 79589
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.907471035700269e-07,
+      "loss": 0.4243,
+      "step": 79590
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.907232901041061e-07,
+      "loss": 0.3909,
+      "step": 79591
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.906994770281153e-07,
+      "loss": 0.4614,
+      "step": 79592
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.906756643420723e-07,
+      "loss": 0.5122,
+      "step": 79593
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.906518520459953e-07,
+      "loss": 0.3591,
+      "step": 79594
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.90628040139903e-07,
+      "loss": 0.2665,
+      "step": 79595
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.906042286238136e-07,
+      "loss": 0.3735,
+      "step": 79596
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.905804174977448e-07,
+      "loss": 0.4736,
+      "step": 79597
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.905566067617152e-07,
+      "loss": 0.4304,
+      "step": 79598
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.905327964157429e-07,
+      "loss": 0.3252,
+      "step": 79599
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.905089864598467e-07,
+      "loss": 0.3525,
+      "step": 79600
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.904851768940441e-07,
+      "loss": 0.3516,
+      "step": 79601
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.904613677183534e-07,
+      "loss": 0.3804,
+      "step": 79602
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.904375589327929e-07,
+      "loss": 0.4606,
+      "step": 79603
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.90413750537381e-07,
+      "loss": 0.3254,
+      "step": 79604
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.903899425321363e-07,
+      "loss": 0.4158,
+      "step": 79605
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.903661349170761e-07,
+      "loss": 0.3477,
+      "step": 79606
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.903423276922192e-07,
+      "loss": 0.5186,
+      "step": 79607
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.90318520857584e-07,
+      "loss": 0.4883,
+      "step": 79608
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.902947144131882e-07,
+      "loss": 0.4717,
+      "step": 79609
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.902709083590506e-07,
+      "loss": 0.4917,
+      "step": 79610
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.902471026951888e-07,
+      "loss": 0.4399,
+      "step": 79611
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.902232974216213e-07,
+      "loss": 0.5098,
+      "step": 79612
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.901994925383669e-07,
+      "loss": 0.4268,
+      "step": 79613
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.901756880454428e-07,
+      "loss": 0.5684,
+      "step": 79614
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.901518839428678e-07,
+      "loss": 0.4189,
+      "step": 79615
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.9012808023066e-07,
+      "loss": 0.468,
+      "step": 79616
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.90104276908838e-07,
+      "loss": 0.4127,
+      "step": 79617
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.900804739774195e-07,
+      "loss": 0.4683,
+      "step": 79618
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.900566714364229e-07,
+      "loss": 0.4609,
+      "step": 79619
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.900328692858667e-07,
+      "loss": 0.2561,
+      "step": 79620
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.900090675257686e-07,
+      "loss": 0.4194,
+      "step": 79621
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.899852661561475e-07,
+      "loss": 0.4785,
+      "step": 79622
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.899614651770208e-07,
+      "loss": 0.4004,
+      "step": 79623
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.899376645884072e-07,
+      "loss": 0.4453,
+      "step": 79624
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.899138643903253e-07,
+      "loss": 0.3639,
+      "step": 79625
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.898900645827925e-07,
+      "loss": 0.5063,
+      "step": 79626
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.898662651658274e-07,
+      "loss": 0.3625,
+      "step": 79627
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.898424661394481e-07,
+      "loss": 0.416,
+      "step": 79628
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.898186675036735e-07,
+      "loss": 0.5259,
+      "step": 79629
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.897948692585213e-07,
+      "loss": 0.4277,
+      "step": 79630
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.897710714040093e-07,
+      "loss": 0.359,
+      "step": 79631
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.897472739401562e-07,
+      "loss": 0.3308,
+      "step": 79632
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.8972347686698e-07,
+      "loss": 0.3716,
+      "step": 79633
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.896996801844996e-07,
+      "loss": 0.1527,
+      "step": 79634
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.896758838927322e-07,
+      "loss": 0.4639,
+      "step": 79635
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.896520879916965e-07,
+      "loss": 0.4395,
+      "step": 79636
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.896282924814108e-07,
+      "loss": 0.3206,
+      "step": 79637
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.896044973618936e-07,
+      "loss": 0.4436,
+      "step": 79638
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.895807026331624e-07,
+      "loss": 0.4136,
+      "step": 79639
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.895569082952362e-07,
+      "loss": 0.458,
+      "step": 79640
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.895331143481324e-07,
+      "loss": 0.4155,
+      "step": 79641
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.895093207918699e-07,
+      "loss": 0.4219,
+      "step": 79642
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.894855276264664e-07,
+      "loss": 0.35,
+      "step": 79643
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.894617348519403e-07,
+      "loss": 0.2891,
+      "step": 79644
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.894379424683098e-07,
+      "loss": 0.5137,
+      "step": 79645
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.894141504755938e-07,
+      "loss": 0.3407,
+      "step": 79646
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.893903588738093e-07,
+      "loss": 0.4756,
+      "step": 79647
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.893665676629751e-07,
+      "loss": 0.5107,
+      "step": 79648
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.893427768431097e-07,
+      "loss": 0.4131,
+      "step": 79649
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.893189864142312e-07,
+      "loss": 0.4448,
+      "step": 79650
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.892951963763578e-07,
+      "loss": 0.3799,
+      "step": 79651
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.892714067295071e-07,
+      "loss": 0.4941,
+      "step": 79652
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.892476174736978e-07,
+      "loss": 0.3085,
+      "step": 79653
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.892238286089481e-07,
+      "loss": 0.3079,
+      "step": 79654
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.892000401352768e-07,
+      "loss": 0.4067,
+      "step": 79655
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.89176252052701e-07,
+      "loss": 0.4316,
+      "step": 79656
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.891524643612395e-07,
+      "loss": 0.4287,
+      "step": 79657
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.891286770609104e-07,
+      "loss": 0.3906,
+      "step": 79658
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.891048901517325e-07,
+      "loss": 0.4102,
+      "step": 79659
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.890811036337234e-07,
+      "loss": 0.521,
+      "step": 79660
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.89057317506901e-07,
+      "loss": 0.4941,
+      "step": 79661
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.89033531771284e-07,
+      "loss": 0.5073,
+      "step": 79662
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.890097464268908e-07,
+      "loss": 0.5908,
+      "step": 79663
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.889859614737391e-07,
+      "loss": 0.4043,
+      "step": 79664
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.889621769118473e-07,
+      "loss": 0.4082,
+      "step": 79665
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.889383927412336e-07,
+      "loss": 0.4761,
+      "step": 79666
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.889146089619167e-07,
+      "loss": 0.4976,
+      "step": 79667
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.88890825573914e-07,
+      "loss": 0.4365,
+      "step": 79668
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.888670425772441e-07,
+      "loss": 0.3171,
+      "step": 79669
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.888432599719256e-07,
+      "loss": 0.3091,
+      "step": 79670
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.888194777579759e-07,
+      "loss": 0.4092,
+      "step": 79671
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.88795695935414e-07,
+      "loss": 0.3794,
+      "step": 79672
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.887719145042574e-07,
+      "loss": 0.4048,
+      "step": 79673
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.887481334645246e-07,
+      "loss": 0.4688,
+      "step": 79674
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.887243528162338e-07,
+      "loss": 0.4546,
+      "step": 79675
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.887005725594038e-07,
+      "loss": 0.5425,
+      "step": 79676
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.886767926940517e-07,
+      "loss": 0.2775,
+      "step": 79677
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.886530132201964e-07,
+      "loss": 0.3247,
+      "step": 79678
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.886292341378561e-07,
+      "loss": 0.4167,
+      "step": 79679
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.886054554470492e-07,
+      "loss": 0.185,
+      "step": 79680
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.885816771477931e-07,
+      "loss": 0.4561,
+      "step": 79681
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.885578992401063e-07,
+      "loss": 0.4272,
+      "step": 79682
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.885341217240075e-07,
+      "loss": 0.4736,
+      "step": 79683
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.885103445995149e-07,
+      "loss": 0.4253,
+      "step": 79684
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.884865678666459e-07,
+      "loss": 0.4238,
+      "step": 79685
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.884627915254194e-07,
+      "loss": 0.4204,
+      "step": 79686
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.884390155758533e-07,
+      "loss": 0.2114,
+      "step": 79687
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.884152400179663e-07,
+      "loss": 0.5347,
+      "step": 79688
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.883914648517758e-07,
+      "loss": 0.3599,
+      "step": 79689
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.883676900773009e-07,
+      "loss": 0.3037,
+      "step": 79690
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.88343915694559e-07,
+      "loss": 0.3513,
+      "step": 79691
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.883201417035685e-07,
+      "loss": 0.2686,
+      "step": 79692
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.882963681043482e-07,
+      "loss": 0.4531,
+      "step": 79693
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.882725948969155e-07,
+      "loss": 0.5132,
+      "step": 79694
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.882488220812889e-07,
+      "loss": 0.4051,
+      "step": 79695
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.88225049657487e-07,
+      "loss": 0.4062,
+      "step": 79696
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.882012776255273e-07,
+      "loss": 0.3367,
+      "step": 79697
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.881775059854284e-07,
+      "loss": 0.499,
+      "step": 79698
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.881537347372084e-07,
+      "loss": 0.3916,
+      "step": 79699
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.881299638808859e-07,
+      "loss": 0.3215,
+      "step": 79700
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.881061934164787e-07,
+      "loss": 0.3032,
+      "step": 79701
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.880824233440048e-07,
+      "loss": 0.349,
+      "step": 79702
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.880586536634826e-07,
+      "loss": 0.4182,
+      "step": 79703
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.880348843749304e-07,
+      "loss": 0.5332,
+      "step": 79704
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.880111154783666e-07,
+      "loss": 0.4365,
+      "step": 79705
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.879873469738089e-07,
+      "loss": 0.4639,
+      "step": 79706
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.879635788612756e-07,
+      "loss": 0.4341,
+      "step": 79707
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.87939811140785e-07,
+      "loss": 0.4604,
+      "step": 79708
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.879160438123559e-07,
+      "loss": 0.4436,
+      "step": 79709
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.87892276876006e-07,
+      "loss": 0.459,
+      "step": 79710
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.878685103317529e-07,
+      "loss": 0.4001,
+      "step": 79711
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.878447441796154e-07,
+      "loss": 0.4175,
+      "step": 79712
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.87820978419612e-07,
+      "loss": 0.374,
+      "step": 79713
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.877972130517601e-07,
+      "loss": 0.5747,
+      "step": 79714
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.877734480760783e-07,
+      "loss": 0.4167,
+      "step": 79715
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.877496834925849e-07,
+      "loss": 0.5034,
+      "step": 79716
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.877259193012983e-07,
+      "loss": 0.4717,
+      "step": 79717
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.877021555022361e-07,
+      "loss": 0.4888,
+      "step": 79718
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.876783920954169e-07,
+      "loss": 0.375,
+      "step": 79719
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.876546290808591e-07,
+      "loss": 0.4639,
+      "step": 79720
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.876308664585801e-07,
+      "loss": 0.478,
+      "step": 79721
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.876071042285991e-07,
+      "loss": 0.4707,
+      "step": 79722
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.875833423909334e-07,
+      "loss": 0.5356,
+      "step": 79723
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.875595809456014e-07,
+      "loss": 0.4155,
+      "step": 79724
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.875358198926218e-07,
+      "loss": 0.3975,
+      "step": 79725
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.875120592320126e-07,
+      "loss": 0.6172,
+      "step": 79726
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.874882989637914e-07,
+      "loss": 0.4028,
+      "step": 79727
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.874645390879769e-07,
+      "loss": 0.439,
+      "step": 79728
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.874407796045873e-07,
+      "loss": 0.2925,
+      "step": 79729
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.874170205136415e-07,
+      "loss": 0.5352,
+      "step": 79730
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.873932618151561e-07,
+      "loss": 0.3418,
+      "step": 79731
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.873695035091502e-07,
+      "loss": 0.4722,
+      "step": 79732
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.873457455956418e-07,
+      "loss": 0.4673,
+      "step": 79733
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.873219880746497e-07,
+      "loss": 0.3787,
+      "step": 79734
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.872982309461911e-07,
+      "loss": 0.3042,
+      "step": 79735
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.872744742102845e-07,
+      "loss": 0.5117,
+      "step": 79736
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.872507178669485e-07,
+      "loss": 0.3127,
+      "step": 79737
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.872269619162014e-07,
+      "loss": 0.4233,
+      "step": 79738
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.872032063580607e-07,
+      "loss": 0.3766,
+      "step": 79739
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.871794511925451e-07,
+      "loss": 0.3586,
+      "step": 79740
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.871556964196724e-07,
+      "loss": 0.4014,
+      "step": 79741
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.871319420394609e-07,
+      "loss": 0.5493,
+      "step": 79742
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.871081880519294e-07,
+      "loss": 0.4717,
+      "step": 79743
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.870844344570951e-07,
+      "loss": 0.4053,
+      "step": 79744
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.870606812549767e-07,
+      "loss": 0.3286,
+      "step": 79745
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.870369284455925e-07,
+      "loss": 0.3928,
+      "step": 79746
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.870131760289607e-07,
+      "loss": 0.48,
+      "step": 79747
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.86989424005099e-07,
+      "loss": 0.3706,
+      "step": 79748
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.869656723740259e-07,
+      "loss": 0.4434,
+      "step": 79749
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.8694192113576e-07,
+      "loss": 0.4575,
+      "step": 79750
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.86918170290319e-07,
+      "loss": 0.5039,
+      "step": 79751
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.868944198377208e-07,
+      "loss": 0.2842,
+      "step": 79752
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.86870669777984e-07,
+      "loss": 0.4644,
+      "step": 79753
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.868469201111268e-07,
+      "loss": 0.5132,
+      "step": 79754
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.868231708371676e-07,
+      "loss": 0.4658,
+      "step": 79755
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.86799421956124e-07,
+      "loss": 0.4436,
+      "step": 79756
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.867756734680144e-07,
+      "loss": 0.6001,
+      "step": 79757
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.867519253728572e-07,
+      "loss": 0.3306,
+      "step": 79758
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.867281776706708e-07,
+      "loss": 0.328,
+      "step": 79759
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.867044303614729e-07,
+      "loss": 0.499,
+      "step": 79760
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.866806834452815e-07,
+      "loss": 0.4436,
+      "step": 79761
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.866569369221151e-07,
+      "loss": 0.5127,
+      "step": 79762
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.866331907919918e-07,
+      "loss": 0.4189,
+      "step": 79763
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.866094450549304e-07,
+      "loss": 0.3916,
+      "step": 79764
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.865856997109481e-07,
+      "loss": 0.3892,
+      "step": 79765
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.865619547600636e-07,
+      "loss": 0.3682,
+      "step": 79766
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.865382102022948e-07,
+      "loss": 0.394,
+      "step": 79767
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.865144660376607e-07,
+      "loss": 0.2832,
+      "step": 79768
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.864907222661782e-07,
+      "loss": 0.3228,
+      "step": 79769
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.864669788878668e-07,
+      "loss": 0.3499,
+      "step": 79770
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.864432359027435e-07,
+      "loss": 0.4072,
+      "step": 79771
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.864194933108274e-07,
+      "loss": 0.5293,
+      "step": 79772
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.863957511121359e-07,
+      "loss": 0.481,
+      "step": 79773
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.863720093066876e-07,
+      "loss": 0.3118,
+      "step": 79774
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.863482678945005e-07,
+      "loss": 0.4663,
+      "step": 79775
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.863245268755935e-07,
+      "loss": 0.3896,
+      "step": 79776
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.863007862499836e-07,
+      "loss": 0.3506,
+      "step": 79777
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.862770460176897e-07,
+      "loss": 0.3599,
+      "step": 79778
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.862533061787298e-07,
+      "loss": 0.4187,
+      "step": 79779
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.862295667331226e-07,
+      "loss": 0.4314,
+      "step": 79780
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.862058276808856e-07,
+      "loss": 0.4214,
+      "step": 79781
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.86182089022037e-07,
+      "loss": 0.4229,
+      "step": 79782
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.861583507565949e-07,
+      "loss": 0.3242,
+      "step": 79783
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.861346128845779e-07,
+      "loss": 0.3557,
+      "step": 79784
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.861108754060043e-07,
+      "loss": 0.479,
+      "step": 79785
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.860871383208918e-07,
+      "loss": 0.4717,
+      "step": 79786
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 4.860634016292586e-07,
+      "loss": 0.4707,
+      "step": 79787
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.860396653311234e-07,
+      "loss": 0.3567,
+      "step": 79788
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.860159294265036e-07,
+      "loss": 0.5137,
+      "step": 79789
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.859921939154181e-07,
+      "loss": 0.5098,
+      "step": 79790
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.859684587978844e-07,
+      "loss": 0.5049,
+      "step": 79791
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.859447240739211e-07,
+      "loss": 0.4956,
+      "step": 79792
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.859209897435466e-07,
+      "loss": 0.4468,
+      "step": 79793
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.858972558067783e-07,
+      "loss": 0.5356,
+      "step": 79794
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.858735222636349e-07,
+      "loss": 0.437,
+      "step": 79795
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.858497891141347e-07,
+      "loss": 0.3267,
+      "step": 79796
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.858260563582957e-07,
+      "loss": 0.4902,
+      "step": 79797
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.85802323996136e-07,
+      "loss": 0.237,
+      "step": 79798
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.85778592027674e-07,
+      "loss": 0.4343,
+      "step": 79799
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.857548604529274e-07,
+      "loss": 0.2676,
+      "step": 79800
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.857311292719145e-07,
+      "loss": 0.4531,
+      "step": 79801
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.857073984846542e-07,
+      "loss": 0.4055,
+      "step": 79802
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.856836680911635e-07,
+      "loss": 0.4048,
+      "step": 79803
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.856599380914614e-07,
+      "loss": 0.4551,
+      "step": 79804
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.856362084855661e-07,
+      "loss": 0.46,
+      "step": 79805
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.856124792734951e-07,
+      "loss": 0.4438,
+      "step": 79806
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.85588750455267e-07,
+      "loss": 0.4165,
+      "step": 79807
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.855650220309e-07,
+      "loss": 0.4517,
+      "step": 79808
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.855412940004125e-07,
+      "loss": 0.3223,
+      "step": 79809
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.855175663638225e-07,
+      "loss": 0.377,
+      "step": 79810
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.854938391211475e-07,
+      "loss": 0.3975,
+      "step": 79811
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.854701122724063e-07,
+      "loss": 0.2061,
+      "step": 79812
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.85446385817617e-07,
+      "loss": 0.4189,
+      "step": 79813
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.854226597567981e-07,
+      "loss": 0.5015,
+      "step": 79814
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.853989340899671e-07,
+      "loss": 0.3643,
+      "step": 79815
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.853752088171422e-07,
+      "loss": 0.3346,
+      "step": 79816
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.853514839383422e-07,
+      "loss": 0.3643,
+      "step": 79817
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.853277594535849e-07,
+      "loss": 0.3982,
+      "step": 79818
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.853040353628888e-07,
+      "loss": 0.4668,
+      "step": 79819
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.852803116662712e-07,
+      "loss": 0.4043,
+      "step": 79820
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.852565883637508e-07,
+      "loss": 0.3069,
+      "step": 79821
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.852328654553462e-07,
+      "loss": 0.5425,
+      "step": 79822
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.852091429410747e-07,
+      "loss": 0.3896,
+      "step": 79823
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.851854208209549e-07,
+      "loss": 0.4946,
+      "step": 79824
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.85161699095005e-07,
+      "loss": 0.3542,
+      "step": 79825
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.851379777632435e-07,
+      "loss": 0.5552,
+      "step": 79826
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.851142568256878e-07,
+      "loss": 0.3892,
+      "step": 79827
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.850905362823563e-07,
+      "loss": 0.2218,
+      "step": 79828
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.850668161332679e-07,
+      "loss": 0.4136,
+      "step": 79829
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.850430963784396e-07,
+      "loss": 0.3217,
+      "step": 79830
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.850193770178905e-07,
+      "loss": 0.4407,
+      "step": 79831
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.84995658051638e-07,
+      "loss": 0.4478,
+      "step": 79832
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.849719394797007e-07,
+      "loss": 0.5654,
+      "step": 79833
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.849482213020968e-07,
+      "loss": 0.4614,
+      "step": 79834
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.849245035188446e-07,
+      "loss": 0.4531,
+      "step": 79835
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.849007861299616e-07,
+      "loss": 0.4907,
+      "step": 79836
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.848770691354666e-07,
+      "loss": 0.4194,
+      "step": 79837
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.848533525353774e-07,
+      "loss": 0.4873,
+      "step": 79838
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.848296363297131e-07,
+      "loss": 0.3784,
+      "step": 79839
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.848059205184902e-07,
+      "loss": 0.3579,
+      "step": 79840
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.847822051017278e-07,
+      "loss": 0.3696,
+      "step": 79841
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.84758490079444e-07,
+      "loss": 0.4104,
+      "step": 79842
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.847347754516573e-07,
+      "loss": 0.4648,
+      "step": 79843
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.847110612183851e-07,
+      "loss": 0.3838,
+      "step": 79844
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.846873473796459e-07,
+      "loss": 0.3811,
+      "step": 79845
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.84663633935458e-07,
+      "loss": 0.356,
+      "step": 79846
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.846399208858398e-07,
+      "loss": 0.4082,
+      "step": 79847
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.846162082308088e-07,
+      "loss": 0.3806,
+      "step": 79848
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.845924959703837e-07,
+      "loss": 0.4033,
+      "step": 79849
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.845687841045821e-07,
+      "loss": 0.4036,
+      "step": 79850
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.845450726334225e-07,
+      "loss": 0.5581,
+      "step": 79851
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.845213615569234e-07,
+      "loss": 0.4536,
+      "step": 79852
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.844976508751022e-07,
+      "loss": 0.2458,
+      "step": 79853
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.844739405879776e-07,
+      "loss": 0.4526,
+      "step": 79854
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.844502306955673e-07,
+      "loss": 0.3564,
+      "step": 79855
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.844265211978904e-07,
+      "loss": 0.3051,
+      "step": 79856
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.844028120949639e-07,
+      "loss": 0.5732,
+      "step": 79857
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.843791033868066e-07,
+      "loss": 0.3945,
+      "step": 79858
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.843553950734368e-07,
+      "loss": 0.3579,
+      "step": 79859
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.843316871548723e-07,
+      "loss": 0.2959,
+      "step": 79860
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.843079796311309e-07,
+      "loss": 0.3752,
+      "step": 79861
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.842842725022312e-07,
+      "loss": 0.4697,
+      "step": 79862
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.842605657681914e-07,
+      "loss": 0.5005,
+      "step": 79863
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.8423685942903e-07,
+      "loss": 0.3915,
+      "step": 79864
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.842131534847642e-07,
+      "loss": 0.5,
+      "step": 79865
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.841894479354127e-07,
+      "loss": 0.2767,
+      "step": 79866
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.841657427809938e-07,
+      "loss": 0.4956,
+      "step": 79867
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.841420380215256e-07,
+      "loss": 0.4209,
+      "step": 79868
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.841183336570263e-07,
+      "loss": 0.4404,
+      "step": 79869
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.840946296875134e-07,
+      "loss": 0.4878,
+      "step": 79870
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.840709261130055e-07,
+      "loss": 0.3748,
+      "step": 79871
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.840472229335209e-07,
+      "loss": 0.366,
+      "step": 79872
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.840235201490779e-07,
+      "loss": 0.3994,
+      "step": 79873
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.839998177596939e-07,
+      "loss": 0.2949,
+      "step": 79874
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.839761157653877e-07,
+      "loss": 0.4502,
+      "step": 79875
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.839524141661775e-07,
+      "loss": 0.415,
+      "step": 79876
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.839287129620808e-07,
+      "loss": 0.4927,
+      "step": 79877
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.839050121531162e-07,
+      "loss": 0.3391,
+      "step": 79878
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.838813117393022e-07,
+      "loss": 0.3398,
+      "step": 79879
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.838576117206563e-07,
+      "loss": 0.4221,
+      "step": 79880
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.838339120971972e-07,
+      "loss": 0.3984,
+      "step": 79881
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.838102128689422e-07,
+      "loss": 0.3687,
+      "step": 79882
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.837865140359101e-07,
+      "loss": 0.3999,
+      "step": 79883
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.837628155981189e-07,
+      "loss": 0.3621,
+      "step": 79884
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.837391175555873e-07,
+      "loss": 0.4727,
+      "step": 79885
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.837154199083323e-07,
+      "loss": 0.3538,
+      "step": 79886
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.836917226563728e-07,
+      "loss": 0.3689,
+      "step": 79887
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.836680257997269e-07,
+      "loss": 0.4849,
+      "step": 79888
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.836443293384128e-07,
+      "loss": 0.3916,
+      "step": 79889
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.836206332724487e-07,
+      "loss": 0.4766,
+      "step": 79890
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.835969376018521e-07,
+      "loss": 0.3424,
+      "step": 79891
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.835732423266416e-07,
+      "loss": 0.4039,
+      "step": 79892
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.835495474468356e-07,
+      "loss": 0.4214,
+      "step": 79893
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.835258529624516e-07,
+      "loss": 0.4326,
+      "step": 79894
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.835021588735082e-07,
+      "loss": 0.3818,
+      "step": 79895
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.834784651800233e-07,
+      "loss": 0.4243,
+      "step": 79896
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.834547718820157e-07,
+      "loss": 0.4468,
+      "step": 79897
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.834310789795026e-07,
+      "loss": 0.4419,
+      "step": 79898
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.83407386472503e-07,
+      "loss": 0.4023,
+      "step": 79899
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.833836943610342e-07,
+      "loss": 0.4854,
+      "step": 79900
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.833600026451148e-07,
+      "loss": 0.3462,
+      "step": 79901
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.833363113247631e-07,
+      "loss": 0.2418,
+      "step": 79902
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.833126203999968e-07,
+      "loss": 0.4246,
+      "step": 79903
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.832889298708342e-07,
+      "loss": 0.3193,
+      "step": 79904
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.832652397372936e-07,
+      "loss": 0.4185,
+      "step": 79905
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.832415499993934e-07,
+      "loss": 0.4517,
+      "step": 79906
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.83217860657151e-07,
+      "loss": 0.4141,
+      "step": 79907
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.831941717105849e-07,
+      "loss": 0.4346,
+      "step": 79908
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.831704831597135e-07,
+      "loss": 0.4351,
+      "step": 79909
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.831467950045545e-07,
+      "loss": 0.3402,
+      "step": 79910
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.831231072451266e-07,
+      "loss": 0.5142,
+      "step": 79911
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.83099419881447e-07,
+      "loss": 0.5415,
+      "step": 79912
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.830757329135346e-07,
+      "loss": 0.5542,
+      "step": 79913
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.830520463414077e-07,
+      "loss": 0.3496,
+      "step": 79914
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.830283601650838e-07,
+      "loss": 0.3735,
+      "step": 79915
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.830046743845811e-07,
+      "loss": 0.3975,
+      "step": 79916
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.829809889999181e-07,
+      "loss": 0.5156,
+      "step": 79917
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.829573040111132e-07,
+      "loss": 0.4465,
+      "step": 79918
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.829336194181839e-07,
+      "loss": 0.4126,
+      "step": 79919
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.829099352211483e-07,
+      "loss": 0.3245,
+      "step": 79920
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.828862514200249e-07,
+      "loss": 0.396,
+      "step": 79921
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.828625680148315e-07,
+      "loss": 0.3105,
+      "step": 79922
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.828388850055869e-07,
+      "loss": 0.4526,
+      "step": 79923
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.828152023923084e-07,
+      "loss": 0.4268,
+      "step": 79924
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.827915201750147e-07,
+      "loss": 0.3818,
+      "step": 79925
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.827678383537236e-07,
+      "loss": 0.3823,
+      "step": 79926
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.827441569284537e-07,
+      "loss": 0.2917,
+      "step": 79927
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.827204758992225e-07,
+      "loss": 0.3882,
+      "step": 79928
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.826967952660489e-07,
+      "loss": 0.2955,
+      "step": 79929
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.826731150289501e-07,
+      "loss": 0.3923,
+      "step": 79930
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.826494351879451e-07,
+      "loss": 0.3752,
+      "step": 79931
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.826257557430513e-07,
+      "loss": 0.4746,
+      "step": 79932
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.826020766942872e-07,
+      "loss": 0.3347,
+      "step": 79933
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.825783980416708e-07,
+      "loss": 0.2912,
+      "step": 79934
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.825547197852207e-07,
+      "loss": 0.4526,
+      "step": 79935
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.825310419249545e-07,
+      "loss": 0.4697,
+      "step": 79936
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.825073644608903e-07,
+      "loss": 0.481,
+      "step": 79937
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.824836873930463e-07,
+      "loss": 0.4558,
+      "step": 79938
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.824600107214414e-07,
+      "loss": 0.4399,
+      "step": 79939
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.824363344460929e-07,
+      "loss": 0.4902,
+      "step": 79940
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.824126585670188e-07,
+      "loss": 0.4116,
+      "step": 79941
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.823889830842375e-07,
+      "loss": 0.4497,
+      "step": 79942
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.823653079977671e-07,
+      "loss": 0.3672,
+      "step": 79943
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.823416333076262e-07,
+      "loss": 0.4678,
+      "step": 79944
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.823179590138321e-07,
+      "loss": 0.415,
+      "step": 79945
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.822942851164033e-07,
+      "loss": 0.4785,
+      "step": 79946
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.822706116153581e-07,
+      "loss": 0.2102,
+      "step": 79947
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.822469385107148e-07,
+      "loss": 0.3667,
+      "step": 79948
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.822232658024912e-07,
+      "loss": 0.3724,
+      "step": 79949
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.821995934907051e-07,
+      "loss": 0.4067,
+      "step": 79950
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.821759215753749e-07,
+      "loss": 0.2727,
+      "step": 79951
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.821522500565191e-07,
+      "loss": 0.3127,
+      "step": 79952
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.821285789341554e-07,
+      "loss": 0.3257,
+      "step": 79953
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.821049082083017e-07,
+      "loss": 0.5283,
+      "step": 79954
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.820812378789767e-07,
+      "loss": 0.3691,
+      "step": 79955
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.820575679461985e-07,
+      "loss": 0.4351,
+      "step": 79956
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.820338984099847e-07,
+      "loss": 0.353,
+      "step": 79957
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.820102292703538e-07,
+      "loss": 0.5107,
+      "step": 79958
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.819865605273242e-07,
+      "loss": 0.2261,
+      "step": 79959
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.819628921809133e-07,
+      "loss": 0.5256,
+      "step": 79960
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.819392242311398e-07,
+      "loss": 0.3765,
+      "step": 79961
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.819155566780214e-07,
+      "loss": 0.5815,
+      "step": 79962
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.818918895215764e-07,
+      "loss": 0.3735,
+      "step": 79963
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.818682227618229e-07,
+      "loss": 0.5073,
+      "step": 79964
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.818445563987797e-07,
+      "loss": 0.519,
+      "step": 79965
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.818208904324636e-07,
+      "loss": 0.4321,
+      "step": 79966
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.817972248628935e-07,
+      "loss": 0.4668,
+      "step": 79967
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.817735596900879e-07,
+      "loss": 0.4624,
+      "step": 79968
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.817498949140645e-07,
+      "loss": 0.2599,
+      "step": 79969
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.817262305348409e-07,
+      "loss": 0.5098,
+      "step": 79970
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.817025665524357e-07,
+      "loss": 0.5605,
+      "step": 79971
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.816789029668669e-07,
+      "loss": 0.2808,
+      "step": 79972
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.816552397781533e-07,
+      "loss": 0.4282,
+      "step": 79973
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.816315769863121e-07,
+      "loss": 0.4062,
+      "step": 79974
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.816079145913616e-07,
+      "loss": 0.4429,
+      "step": 79975
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.815842525933201e-07,
+      "loss": 0.4131,
+      "step": 79976
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.815605909922063e-07,
+      "loss": 0.4766,
+      "step": 79977
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.815369297880371e-07,
+      "loss": 0.2898,
+      "step": 79978
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.815132689808317e-07,
+      "loss": 0.4771,
+      "step": 79979
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.814896085706073e-07,
+      "loss": 0.3019,
+      "step": 79980
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.814659485573825e-07,
+      "loss": 0.478,
+      "step": 79981
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.814422889411757e-07,
+      "loss": 0.4014,
+      "step": 79982
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.814186297220045e-07,
+      "loss": 0.3464,
+      "step": 79983
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.813949708998871e-07,
+      "loss": 0.4478,
+      "step": 79984
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.813713124748422e-07,
+      "loss": 0.2253,
+      "step": 79985
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.81347654446887e-07,
+      "loss": 0.458,
+      "step": 79986
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.8132399681604e-07,
+      "loss": 0.4058,
+      "step": 79987
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.813003395823193e-07,
+      "loss": 0.4341,
+      "step": 79988
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.812766827457435e-07,
+      "loss": 0.5054,
+      "step": 79989
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.812530263063303e-07,
+      "loss": 0.417,
+      "step": 79990
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.812293702640976e-07,
+      "loss": 0.3774,
+      "step": 79991
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.812057146190636e-07,
+      "loss": 0.4565,
+      "step": 79992
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.811820593712465e-07,
+      "loss": 0.5327,
+      "step": 79993
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.811584045206648e-07,
+      "loss": 0.3313,
+      "step": 79994
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.811347500673359e-07,
+      "loss": 0.3779,
+      "step": 79995
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.811110960112783e-07,
+      "loss": 0.3782,
+      "step": 79996
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.810874423525099e-07,
+      "loss": 0.3647,
+      "step": 79997
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.810637890910496e-07,
+      "loss": 0.4595,
+      "step": 79998
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.810401362269149e-07,
+      "loss": 0.4429,
+      "step": 79999
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.810164837601233e-07,
+      "loss": 0.2094,
+      "step": 80000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.809928316906937e-07,
+      "loss": 0.478,
+      "step": 80001
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.809691800186444e-07,
+      "loss": 0.3979,
+      "step": 80002
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.809455287439927e-07,
+      "loss": 0.4854,
+      "step": 80003
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.80921877866757e-07,
+      "loss": 0.3679,
+      "step": 80004
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.808982273869558e-07,
+      "loss": 0.3752,
+      "step": 80005
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.808745773046073e-07,
+      "loss": 0.3575,
+      "step": 80006
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.808509276197288e-07,
+      "loss": 0.3848,
+      "step": 80007
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.808272783323389e-07,
+      "loss": 0.5474,
+      "step": 80008
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.80803629442456e-07,
+      "loss": 0.3726,
+      "step": 80009
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.807799809500975e-07,
+      "loss": 0.4082,
+      "step": 80010
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.807563328552824e-07,
+      "loss": 0.394,
+      "step": 80011
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.807326851580279e-07,
+      "loss": 0.3645,
+      "step": 80012
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.807090378583523e-07,
+      "loss": 0.3838,
+      "step": 80013
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.806853909562742e-07,
+      "loss": 0.4473,
+      "step": 80014
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.806617444518117e-07,
+      "loss": 0.2947,
+      "step": 80015
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.806380983449822e-07,
+      "loss": 0.4424,
+      "step": 80016
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.806144526358042e-07,
+      "loss": 0.4253,
+      "step": 80017
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.805908073242959e-07,
+      "loss": 0.2925,
+      "step": 80018
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.805671624104761e-07,
+      "loss": 0.4191,
+      "step": 80019
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.805435178943613e-07,
+      "loss": 0.2361,
+      "step": 80020
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.805198737759705e-07,
+      "loss": 0.3389,
+      "step": 80021
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.804962300553217e-07,
+      "loss": 0.5381,
+      "step": 80022
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.804725867324335e-07,
+      "loss": 0.3413,
+      "step": 80023
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.804489438073232e-07,
+      "loss": 0.3723,
+      "step": 80024
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.804253012800092e-07,
+      "loss": 0.3683,
+      "step": 80025
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.804016591505096e-07,
+      "loss": 0.2881,
+      "step": 80026
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.803780174188429e-07,
+      "loss": 0.4253,
+      "step": 80027
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.803543760850266e-07,
+      "loss": 0.417,
+      "step": 80028
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.803307351490795e-07,
+      "loss": 0.302,
+      "step": 80029
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.803070946110187e-07,
+      "loss": 0.5156,
+      "step": 80030
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.802834544708631e-07,
+      "loss": 0.3745,
+      "step": 80031
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.802598147286307e-07,
+      "loss": 0.4135,
+      "step": 80032
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.802361753843391e-07,
+      "loss": 0.4531,
+      "step": 80033
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.802125364380069e-07,
+      "loss": 0.4663,
+      "step": 80034
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.80188897889652e-07,
+      "loss": 0.3912,
+      "step": 80035
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.801652597392928e-07,
+      "loss": 0.459,
+      "step": 80036
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.801416219869471e-07,
+      "loss": 0.4771,
+      "step": 80037
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.801179846326328e-07,
+      "loss": 0.3088,
+      "step": 80038
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.800943476763687e-07,
+      "loss": 0.3079,
+      "step": 80039
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.800707111181724e-07,
+      "loss": 0.354,
+      "step": 80040
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.800470749580617e-07,
+      "loss": 0.3778,
+      "step": 80041
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.800234391960552e-07,
+      "loss": 0.4673,
+      "step": 80042
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.799998038321707e-07,
+      "loss": 0.3024,
+      "step": 80043
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.799761688664267e-07,
+      "loss": 0.364,
+      "step": 80044
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.799525342988409e-07,
+      "loss": 0.4585,
+      "step": 80045
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.799289001294315e-07,
+      "loss": 0.4849,
+      "step": 80046
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.799052663582165e-07,
+      "loss": 0.4038,
+      "step": 80047
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.798816329852145e-07,
+      "loss": 0.3892,
+      "step": 80048
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.798580000104433e-07,
+      "loss": 0.3818,
+      "step": 80049
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.798343674339205e-07,
+      "loss": 0.5532,
+      "step": 80050
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.798107352556646e-07,
+      "loss": 0.3816,
+      "step": 80051
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.797871034756938e-07,
+      "loss": 0.5444,
+      "step": 80052
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.797634720940264e-07,
+      "loss": 0.4199,
+      "step": 80053
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.797398411106797e-07,
+      "loss": 0.3523,
+      "step": 80054
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.797162105256723e-07,
+      "loss": 0.3094,
+      "step": 80055
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.796925803390228e-07,
+      "loss": 0.4492,
+      "step": 80056
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.796689505507483e-07,
+      "loss": 0.5449,
+      "step": 80057
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.796453211608674e-07,
+      "loss": 0.4302,
+      "step": 80058
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.796216921693985e-07,
+      "loss": 0.3945,
+      "step": 80059
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79598063576359e-07,
+      "loss": 0.3857,
+      "step": 80060
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.795744353817676e-07,
+      "loss": 0.5581,
+      "step": 80061
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79550807585642e-07,
+      "loss": 0.4536,
+      "step": 80062
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.795271801880001e-07,
+      "loss": 0.3765,
+      "step": 80063
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.795035531888605e-07,
+      "loss": 0.4146,
+      "step": 80064
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.794799265882415e-07,
+      "loss": 0.4507,
+      "step": 80065
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.794563003861602e-07,
+      "loss": 0.5078,
+      "step": 80066
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.794326745826355e-07,
+      "loss": 0.3967,
+      "step": 80067
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.794090491776851e-07,
+      "loss": 0.356,
+      "step": 80068
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.793854241713277e-07,
+      "loss": 0.2376,
+      "step": 80069
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79361799563581e-07,
+      "loss": 0.3418,
+      "step": 80070
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.793381753544626e-07,
+      "loss": 0.5186,
+      "step": 80071
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79314551543991e-07,
+      "loss": 0.3401,
+      "step": 80072
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.792909281321843e-07,
+      "loss": 0.302,
+      "step": 80073
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79267305119061e-07,
+      "loss": 0.4336,
+      "step": 80074
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.792436825046385e-07,
+      "loss": 0.3428,
+      "step": 80075
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79220060288935e-07,
+      "loss": 0.48,
+      "step": 80076
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.791964384719692e-07,
+      "loss": 0.4268,
+      "step": 80077
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.791728170537583e-07,
+      "loss": 0.4648,
+      "step": 80078
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.791491960343214e-07,
+      "loss": 0.4951,
+      "step": 80079
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.791255754136753e-07,
+      "loss": 0.3103,
+      "step": 80080
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.791019551918392e-07,
+      "loss": 0.4893,
+      "step": 80081
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.790783353688309e-07,
+      "loss": 0.3643,
+      "step": 80082
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79054715944668e-07,
+      "loss": 0.3325,
+      "step": 80083
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.790310969193691e-07,
+      "loss": 0.3306,
+      "step": 80084
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.79007478292952e-07,
+      "loss": 0.4189,
+      "step": 80085
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.789838600654352e-07,
+      "loss": 0.2318,
+      "step": 80086
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.789602422368363e-07,
+      "loss": 0.4419,
+      "step": 80087
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.789366248071735e-07,
+      "loss": 0.3582,
+      "step": 80088
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.789130077764654e-07,
+      "loss": 0.4907,
+      "step": 80089
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.788893911447293e-07,
+      "loss": 0.2634,
+      "step": 80090
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.788657749119839e-07,
+      "loss": 0.4492,
+      "step": 80091
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.788421590782467e-07,
+      "loss": 0.2969,
+      "step": 80092
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.788185436435361e-07,
+      "loss": 0.3945,
+      "step": 80093
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.787949286078705e-07,
+      "loss": 0.3868,
+      "step": 80094
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.787713139712674e-07,
+      "loss": 0.4021,
+      "step": 80095
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.787476997337449e-07,
+      "loss": 0.4985,
+      "step": 80096
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.787240858953215e-07,
+      "loss": 0.4341,
+      "step": 80097
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.787004724560155e-07,
+      "loss": 0.3706,
+      "step": 80098
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.786768594158446e-07,
+      "loss": 0.5161,
+      "step": 80099
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.786532467748264e-07,
+      "loss": 0.2827,
+      "step": 80100
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.786296345329795e-07,
+      "loss": 0.5195,
+      "step": 80101
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.786060226903218e-07,
+      "loss": 0.4146,
+      "step": 80102
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.785824112468719e-07,
+      "loss": 0.3562,
+      "step": 80103
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.785588002026471e-07,
+      "loss": 0.4536,
+      "step": 80104
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.78535189557666e-07,
+      "loss": 0.3647,
+      "step": 80105
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.785115793119464e-07,
+      "loss": 0.3726,
+      "step": 80106
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.784879694655069e-07,
+      "loss": 0.4346,
+      "step": 80107
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.784643600183651e-07,
+      "loss": 0.4011,
+      "step": 80108
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.784407509705389e-07,
+      "loss": 0.4277,
+      "step": 80109
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.784171423220466e-07,
+      "loss": 0.3547,
+      "step": 80110
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.783935340729067e-07,
+      "loss": 0.4397,
+      "step": 80111
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.783699262231364e-07,
+      "loss": 0.4044,
+      "step": 80112
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.783463187727544e-07,
+      "loss": 0.478,
+      "step": 80113
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.783227117217786e-07,
+      "loss": 0.4097,
+      "step": 80114
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.782991050702274e-07,
+      "loss": 0.2642,
+      "step": 80115
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.782754988181183e-07,
+      "loss": 0.5137,
+      "step": 80116
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.782518929654698e-07,
+      "loss": 0.4199,
+      "step": 80117
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.782282875123e-07,
+      "loss": 0.3872,
+      "step": 80118
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.782046824586265e-07,
+      "loss": 0.4927,
+      "step": 80119
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.781810778044681e-07,
+      "loss": 0.3596,
+      "step": 80120
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.781574735498419e-07,
+      "loss": 0.293,
+      "step": 80121
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.781338696947667e-07,
+      "loss": 0.365,
+      "step": 80122
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.781102662392603e-07,
+      "loss": 0.2646,
+      "step": 80123
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.780866631833414e-07,
+      "loss": 0.5059,
+      "step": 80124
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.78063060527027e-07,
+      "loss": 0.4106,
+      "step": 80125
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.780394582703358e-07,
+      "loss": 0.3108,
+      "step": 80126
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.780158564132858e-07,
+      "loss": 0.4185,
+      "step": 80127
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.779922549558958e-07,
+      "loss": 0.375,
+      "step": 80128
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.779686538981823e-07,
+      "loss": 0.46,
+      "step": 80129
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.779450532401643e-07,
+      "loss": 0.4371,
+      "step": 80130
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.779214529818597e-07,
+      "loss": 0.4629,
+      "step": 80131
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.77897853123287e-07,
+      "loss": 0.5186,
+      "step": 80132
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.778742536644637e-07,
+      "loss": 0.5005,
+      "step": 80133
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.778506546054078e-07,
+      "loss": 0.4419,
+      "step": 80134
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.778270559461378e-07,
+      "loss": 0.3047,
+      "step": 80135
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.778034576866721e-07,
+      "loss": 0.3857,
+      "step": 80136
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.777798598270277e-07,
+      "loss": 0.3215,
+      "step": 80137
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.777562623672238e-07,
+      "loss": 0.4526,
+      "step": 80138
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.777326653072774e-07,
+      "loss": 0.501,
+      "step": 80139
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.777090686472071e-07,
+      "loss": 0.4614,
+      "step": 80140
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.776854723870314e-07,
+      "loss": 0.4482,
+      "step": 80141
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.776618765267674e-07,
+      "loss": 0.3604,
+      "step": 80142
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.776382810664339e-07,
+      "loss": 0.4099,
+      "step": 80143
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.776146860060486e-07,
+      "loss": 0.3979,
+      "step": 80144
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.775910913456302e-07,
+      "loss": 0.3364,
+      "step": 80145
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.775674970851958e-07,
+      "loss": 0.4951,
+      "step": 80146
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.775439032247639e-07,
+      "loss": 0.4414,
+      "step": 80147
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.775203097643531e-07,
+      "loss": 0.5229,
+      "step": 80148
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.77496716703981e-07,
+      "loss": 0.4194,
+      "step": 80149
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.774731240436651e-07,
+      "loss": 0.4248,
+      "step": 80150
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.774495317834241e-07,
+      "loss": 0.2455,
+      "step": 80151
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.774259399232759e-07,
+      "loss": 0.4951,
+      "step": 80152
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.774023484632393e-07,
+      "loss": 0.3459,
+      "step": 80153
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.77378757403331e-07,
+      "loss": 0.3735,
+      "step": 80154
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.773551667435697e-07,
+      "loss": 0.3853,
+      "step": 80155
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.773315764839738e-07,
+      "loss": 0.311,
+      "step": 80156
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.773079866245613e-07,
+      "loss": 0.3423,
+      "step": 80157
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.772843971653501e-07,
+      "loss": 0.2499,
+      "step": 80158
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.772608081063577e-07,
+      "loss": 0.4019,
+      "step": 80159
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.772372194476028e-07,
+      "loss": 0.373,
+      "step": 80160
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.772136311891032e-07,
+      "loss": 0.4214,
+      "step": 80161
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.771900433308775e-07,
+      "loss": 0.3262,
+      "step": 80162
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.77166455872943e-07,
+      "loss": 0.4226,
+      "step": 80163
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.771428688153182e-07,
+      "loss": 0.4099,
+      "step": 80164
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.771192821580213e-07,
+      "loss": 0.4419,
+      "step": 80165
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.770956959010698e-07,
+      "loss": 0.2192,
+      "step": 80166
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.770721100444821e-07,
+      "loss": 0.4839,
+      "step": 80167
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.770485245882766e-07,
+      "loss": 0.3728,
+      "step": 80168
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.770249395324706e-07,
+      "loss": 0.4019,
+      "step": 80169
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.770013548770829e-07,
+      "loss": 0.3574,
+      "step": 80170
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.769777706221309e-07,
+      "loss": 0.2683,
+      "step": 80171
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.76954186767633e-07,
+      "loss": 0.3575,
+      "step": 80172
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.769306033136072e-07,
+      "loss": 0.3289,
+      "step": 80173
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.769070202600719e-07,
+      "loss": 0.4668,
+      "step": 80174
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.768834376070446e-07,
+      "loss": 0.5957,
+      "step": 80175
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.768598553545434e-07,
+      "loss": 0.4255,
+      "step": 80176
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 4.768362735025866e-07,
+      "loss": 0.3986,
+      "step": 80177
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.768126920511927e-07,
+      "loss": 0.4287,
+      "step": 80178
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7678911100037923e-07,
+      "loss": 0.355,
+      "step": 80179
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.767655303501639e-07,
+      "loss": 0.5361,
+      "step": 80180
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7674195010056504e-07,
+      "loss": 0.397,
+      "step": 80181
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7671837025160136e-07,
+      "loss": 0.4019,
+      "step": 80182
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.766947908032899e-07,
+      "loss": 0.394,
+      "step": 80183
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7667121175564915e-07,
+      "loss": 0.4468,
+      "step": 80184
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7664763310869726e-07,
+      "loss": 0.502,
+      "step": 80185
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7662405486245247e-07,
+      "loss": 0.4419,
+      "step": 80186
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.766004770169323e-07,
+      "loss": 0.3311,
+      "step": 80187
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7657689957215543e-07,
+      "loss": 0.3149,
+      "step": 80188
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7655332252813917e-07,
+      "loss": 0.5168,
+      "step": 80189
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.76529745884902e-07,
+      "loss": 0.2542,
+      "step": 80190
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7650616964246226e-07,
+      "loss": 0.3669,
+      "step": 80191
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7648259380083733e-07,
+      "loss": 0.4656,
+      "step": 80192
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7645901836004565e-07,
+      "loss": 0.4629,
+      "step": 80193
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.764354433201052e-07,
+      "loss": 0.4597,
+      "step": 80194
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7641186868103436e-07,
+      "loss": 0.4482,
+      "step": 80195
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.763882944428506e-07,
+      "loss": 0.3779,
+      "step": 80196
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.763647206055722e-07,
+      "loss": 0.5029,
+      "step": 80197
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7634114716921756e-07,
+      "loss": 0.3911,
+      "step": 80198
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.763175741338046e-07,
+      "loss": 0.4375,
+      "step": 80199
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7629400149935064e-07,
+      "loss": 0.4048,
+      "step": 80200
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7627042926587436e-07,
+      "loss": 0.3374,
+      "step": 80201
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7624685743339375e-07,
+      "loss": 0.3098,
+      "step": 80202
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.762232860019272e-07,
+      "loss": 0.2904,
+      "step": 80203
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.76199714971492e-07,
+      "loss": 0.3193,
+      "step": 80204
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.761761443421066e-07,
+      "loss": 0.4253,
+      "step": 80205
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.761525741137891e-07,
+      "loss": 0.3275,
+      "step": 80206
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7612900428655776e-07,
+      "loss": 0.4277,
+      "step": 80207
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.761054348604304e-07,
+      "loss": 0.4414,
+      "step": 80208
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.760818658354246e-07,
+      "loss": 0.4098,
+      "step": 80209
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7605829721155887e-07,
+      "loss": 0.4785,
+      "step": 80210
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.760347289888512e-07,
+      "loss": 0.3868,
+      "step": 80211
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.760111611673201e-07,
+      "loss": 0.2522,
+      "step": 80212
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.759875937469827e-07,
+      "loss": 0.5254,
+      "step": 80213
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.759640267278575e-07,
+      "loss": 0.4028,
+      "step": 80214
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.759404601099626e-07,
+      "loss": 0.4854,
+      "step": 80215
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7591689389331636e-07,
+      "loss": 0.4751,
+      "step": 80216
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7589332807793604e-07,
+      "loss": 0.4824,
+      "step": 80217
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.758697626638405e-07,
+      "loss": 0.3042,
+      "step": 80218
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.758461976510469e-07,
+      "loss": 0.3916,
+      "step": 80219
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.758226330395744e-07,
+      "loss": 0.481,
+      "step": 80220
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.757990688294398e-07,
+      "loss": 0.606,
+      "step": 80221
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.757755050206619e-07,
+      "loss": 0.4658,
+      "step": 80222
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.757519416132585e-07,
+      "loss": 0.4043,
+      "step": 80223
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.757283786072481e-07,
+      "loss": 0.3833,
+      "step": 80224
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.75704816002648e-07,
+      "loss": 0.4062,
+      "step": 80225
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.756812537994767e-07,
+      "loss": 0.4185,
+      "step": 80226
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.756576919977521e-07,
+      "loss": 0.4614,
+      "step": 80227
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7563413059749257e-07,
+      "loss": 0.5098,
+      "step": 80228
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7561056959871604e-07,
+      "loss": 0.4204,
+      "step": 80229
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.755870090014399e-07,
+      "loss": 0.5205,
+      "step": 80230
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7556344880568267e-07,
+      "loss": 0.416,
+      "step": 80231
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.755398890114625e-07,
+      "loss": 0.3735,
+      "step": 80232
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7551632961879763e-07,
+      "loss": 0.3716,
+      "step": 80233
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7549277062770543e-07,
+      "loss": 0.5229,
+      "step": 80234
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.754692120382042e-07,
+      "loss": 0.3818,
+      "step": 80235
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7544565385031223e-07,
+      "loss": 0.4104,
+      "step": 80236
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.754220960640477e-07,
+      "loss": 0.2992,
+      "step": 80237
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7539853867942834e-07,
+      "loss": 0.3342,
+      "step": 80238
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7537498169647185e-07,
+      "loss": 0.4185,
+      "step": 80239
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7535142511519655e-07,
+      "loss": 0.5151,
+      "step": 80240
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7532786893562106e-07,
+      "loss": 0.4658,
+      "step": 80241
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7530431315776243e-07,
+      "loss": 0.4443,
+      "step": 80242
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7528075778163914e-07,
+      "loss": 0.4102,
+      "step": 80243
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.752572028072692e-07,
+      "loss": 0.3901,
+      "step": 80244
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7523364823467117e-07,
+      "loss": 0.4888,
+      "step": 80245
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7521009406386216e-07,
+      "loss": 0.4629,
+      "step": 80246
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.751865402948607e-07,
+      "loss": 0.4756,
+      "step": 80247
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.751629869276852e-07,
+      "loss": 0.4507,
+      "step": 80248
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.751394339623528e-07,
+      "loss": 0.3386,
+      "step": 80249
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7511588139888234e-07,
+      "loss": 0.4565,
+      "step": 80250
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7509232923729123e-07,
+      "loss": 0.5078,
+      "step": 80251
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7506877747759777e-07,
+      "loss": 0.313,
+      "step": 80252
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7504522611982e-07,
+      "loss": 0.2368,
+      "step": 80253
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.750216751639764e-07,
+      "loss": 0.5381,
+      "step": 80254
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7499812461008403e-07,
+      "loss": 0.4678,
+      "step": 80255
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.749745744581616e-07,
+      "loss": 0.3357,
+      "step": 80256
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7495102470822734e-07,
+      "loss": 0.28,
+      "step": 80257
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.74927475360299e-07,
+      "loss": 0.2958,
+      "step": 80258
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7490392641439403e-07,
+      "loss": 0.1877,
+      "step": 80259
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7488037787053104e-07,
+      "loss": 0.3604,
+      "step": 80260
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.74856829728728e-07,
+      "loss": 0.4482,
+      "step": 80261
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.748332819890033e-07,
+      "loss": 0.4453,
+      "step": 80262
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.748097346513743e-07,
+      "loss": 0.5068,
+      "step": 80263
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.747861877158593e-07,
+      "loss": 0.5337,
+      "step": 80264
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7476264118247636e-07,
+      "loss": 0.3755,
+      "step": 80265
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.747390950512439e-07,
+      "loss": 0.4241,
+      "step": 80266
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7471554932217915e-07,
+      "loss": 0.4404,
+      "step": 80267
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7469200399530097e-07,
+      "loss": 0.4189,
+      "step": 80268
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.746684590706266e-07,
+      "loss": 0.3662,
+      "step": 80269
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7464491454817437e-07,
+      "loss": 0.4966,
+      "step": 80270
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.746213704279628e-07,
+      "loss": 0.3196,
+      "step": 80271
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7459782671000904e-07,
+      "loss": 0.3066,
+      "step": 80272
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.745742833943316e-07,
+      "loss": 0.4849,
+      "step": 80273
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.745507404809489e-07,
+      "loss": 0.3438,
+      "step": 80274
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.745271979698782e-07,
+      "loss": 0.449,
+      "step": 80275
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7450365586113773e-07,
+      "loss": 0.4663,
+      "step": 80276
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7448011415474576e-07,
+      "loss": 0.3755,
+      "step": 80277
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.744565728507205e-07,
+      "loss": 0.3191,
+      "step": 80278
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.744330319490797e-07,
+      "loss": 0.3424,
+      "step": 80279
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.74409491449841e-07,
+      "loss": 0.3442,
+      "step": 80280
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.743859513530228e-07,
+      "loss": 0.4556,
+      "step": 80281
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.743624116586431e-07,
+      "loss": 0.3761,
+      "step": 80282
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.743388723667203e-07,
+      "loss": 0.4421,
+      "step": 80283
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.743153334772717e-07,
+      "loss": 0.4336,
+      "step": 80284
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7429179499031555e-07,
+      "loss": 0.4021,
+      "step": 80285
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7426825690587013e-07,
+      "loss": 0.4956,
+      "step": 80286
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.742447192239537e-07,
+      "loss": 0.3994,
+      "step": 80287
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7422118194458395e-07,
+      "loss": 0.4268,
+      "step": 80288
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.741976450677784e-07,
+      "loss": 0.436,
+      "step": 80289
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.741741085935556e-07,
+      "loss": 0.3447,
+      "step": 80290
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.741505725219338e-07,
+      "loss": 0.3643,
+      "step": 80291
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7412703685293043e-07,
+      "loss": 0.4673,
+      "step": 80292
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7410350158656374e-07,
+      "loss": 0.3918,
+      "step": 80293
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.740799667228519e-07,
+      "loss": 0.4133,
+      "step": 80294
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7405643226181314e-07,
+      "loss": 0.4446,
+      "step": 80295
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7403289820346493e-07,
+      "loss": 0.4478,
+      "step": 80296
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7400936454782546e-07,
+      "loss": 0.4158,
+      "step": 80297
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.739858312949132e-07,
+      "loss": 0.423,
+      "step": 80298
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7396229844474546e-07,
+      "loss": 0.5928,
+      "step": 80299
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7393876599734094e-07,
+      "loss": 0.4038,
+      "step": 80300
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7391523395271695e-07,
+      "loss": 0.2812,
+      "step": 80301
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7389170231089195e-07,
+      "loss": 0.3796,
+      "step": 80302
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.738681710718837e-07,
+      "loss": 0.4756,
+      "step": 80303
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.738446402357109e-07,
+      "loss": 0.3287,
+      "step": 80304
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7382110980239075e-07,
+      "loss": 0.436,
+      "step": 80305
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7379757977194144e-07,
+      "loss": 0.4395,
+      "step": 80306
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7377405014438133e-07,
+      "loss": 0.4092,
+      "step": 80307
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.737505209197288e-07,
+      "loss": 0.3999,
+      "step": 80308
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.737269920980006e-07,
+      "loss": 0.4214,
+      "step": 80309
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.737034636792154e-07,
+      "loss": 0.4041,
+      "step": 80310
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7367993566339135e-07,
+      "loss": 0.3923,
+      "step": 80311
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7365640805054676e-07,
+      "loss": 0.4214,
+      "step": 80312
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7363288084069873e-07,
+      "loss": 0.3541,
+      "step": 80313
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.736093540338659e-07,
+      "loss": 0.5264,
+      "step": 80314
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7358582763006625e-07,
+      "loss": 0.267,
+      "step": 80315
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.73562301629318e-07,
+      "loss": 0.4771,
+      "step": 80316
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.735387760316386e-07,
+      "loss": 0.3402,
+      "step": 80317
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.735152508370467e-07,
+      "loss": 0.4893,
+      "step": 80318
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.734917260455595e-07,
+      "loss": 0.4287,
+      "step": 80319
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.734682016571956e-07,
+      "loss": 0.5054,
+      "step": 80320
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7344467767197316e-07,
+      "loss": 0.3735,
+      "step": 80321
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7342115408990955e-07,
+      "loss": 0.3667,
+      "step": 80322
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7339763091102324e-07,
+      "loss": 0.3496,
+      "step": 80323
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.733741081353321e-07,
+      "loss": 0.4482,
+      "step": 80324
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.733505857628545e-07,
+      "loss": 0.4341,
+      "step": 80325
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.733270637936079e-07,
+      "loss": 0.4021,
+      "step": 80326
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7330354222761037e-07,
+      "loss": 0.417,
+      "step": 80327
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7328002106488064e-07,
+      "loss": 0.48,
+      "step": 80328
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7325650030543606e-07,
+      "loss": 0.4233,
+      "step": 80329
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7323297994929446e-07,
+      "loss": 0.3828,
+      "step": 80330
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.732094599964742e-07,
+      "loss": 0.2794,
+      "step": 80331
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.731859404469931e-07,
+      "loss": 0.3293,
+      "step": 80332
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.731624213008698e-07,
+      "loss": 0.3716,
+      "step": 80333
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.731389025581213e-07,
+      "loss": 0.3745,
+      "step": 80334
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.731153842187662e-07,
+      "loss": 0.3914,
+      "step": 80335
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.730918662828224e-07,
+      "loss": 0.4551,
+      "step": 80336
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.730683487503083e-07,
+      "loss": 0.3625,
+      "step": 80337
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7304483162124166e-07,
+      "loss": 0.3638,
+      "step": 80338
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.730213148956398e-07,
+      "loss": 0.3635,
+      "step": 80339
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.729977985735214e-07,
+      "loss": 0.4355,
+      "step": 80340
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7297428265490433e-07,
+      "loss": 0.4253,
+      "step": 80341
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7295076713980696e-07,
+      "loss": 0.4165,
+      "step": 80342
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7292725202824656e-07,
+      "loss": 0.3462,
+      "step": 80343
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7290373732024157e-07,
+      "loss": 0.418,
+      "step": 80344
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7288022301581045e-07,
+      "loss": 0.4365,
+      "step": 80345
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.728567091149701e-07,
+      "loss": 0.4556,
+      "step": 80346
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7283319561773925e-07,
+      "loss": 0.319,
+      "step": 80347
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7280968252413623e-07,
+      "loss": 0.4517,
+      "step": 80348
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7278616983417807e-07,
+      "loss": 0.5044,
+      "step": 80349
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.727626575478838e-07,
+      "loss": 0.3696,
+      "step": 80350
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.727391456652704e-07,
+      "loss": 0.4702,
+      "step": 80351
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.727156341863565e-07,
+      "loss": 0.3506,
+      "step": 80352
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7269212311116e-07,
+      "loss": 0.4673,
+      "step": 80353
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7266861243969925e-07,
+      "loss": 0.4294,
+      "step": 80354
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.726451021719915e-07,
+      "loss": 0.3323,
+      "step": 80355
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.726215923080552e-07,
+      "loss": 0.3459,
+      "step": 80356
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7259808284790826e-07,
+      "loss": 0.254,
+      "step": 80357
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7257457379156904e-07,
+      "loss": 0.4019,
+      "step": 80358
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.725510651390553e-07,
+      "loss": 0.3945,
+      "step": 80359
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.725275568903845e-07,
+      "loss": 0.4441,
+      "step": 80360
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.725040490455752e-07,
+      "loss": 0.5366,
+      "step": 80361
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.724805416046456e-07,
+      "loss": 0.277,
+      "step": 80362
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.72457034567613e-07,
+      "loss": 0.4902,
+      "step": 80363
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7243352793449586e-07,
+      "loss": 0.3479,
+      "step": 80364
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7241002170531206e-07,
+      "loss": 0.417,
+      "step": 80365
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7238651588008006e-07,
+      "loss": 0.2025,
+      "step": 80366
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.72363010458817e-07,
+      "loss": 0.4858,
+      "step": 80367
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7233950544154177e-07,
+      "loss": 0.3989,
+      "step": 80368
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.723160008282715e-07,
+      "loss": 0.3579,
+      "step": 80369
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.722924966190246e-07,
+      "loss": 0.5205,
+      "step": 80370
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.722689928138195e-07,
+      "loss": 0.4485,
+      "step": 80371
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.722454894126733e-07,
+      "loss": 0.4429,
+      "step": 80372
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7222198641560453e-07,
+      "loss": 0.3838,
+      "step": 80373
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7219848382263105e-07,
+      "loss": 0.35,
+      "step": 80374
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7217498163377143e-07,
+      "loss": 0.4214,
+      "step": 80375
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.721514798490427e-07,
+      "loss": 0.4756,
+      "step": 80376
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7212797846846333e-07,
+      "loss": 0.582,
+      "step": 80377
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7210447749205164e-07,
+      "loss": 0.4219,
+      "step": 80378
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.720809769198253e-07,
+      "loss": 0.4771,
+      "step": 80379
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7205747675180184e-07,
+      "loss": 0.3098,
+      "step": 80380
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.720339769879997e-07,
+      "loss": 0.5288,
+      "step": 80381
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.72010477628437e-07,
+      "loss": 0.4619,
+      "step": 80382
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7198697867313187e-07,
+      "loss": 0.4204,
+      "step": 80383
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7196348012210173e-07,
+      "loss": 0.457,
+      "step": 80384
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7193998197536475e-07,
+      "loss": 0.3643,
+      "step": 80385
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7191648423293915e-07,
+      "loss": 0.3167,
+      "step": 80386
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7189298689484314e-07,
+      "loss": 0.3396,
+      "step": 80387
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7186948996109446e-07,
+      "loss": 0.4714,
+      "step": 80388
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.718459934317106e-07,
+      "loss": 0.418,
+      "step": 80389
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7182249730671e-07,
+      "loss": 0.4756,
+      "step": 80390
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7179900158611054e-07,
+      "loss": 0.54,
+      "step": 80391
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.717755062699308e-07,
+      "loss": 0.3877,
+      "step": 80392
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7175201135818775e-07,
+      "loss": 0.4126,
+      "step": 80393
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.717285168509e-07,
+      "loss": 0.502,
+      "step": 80394
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7170502274808544e-07,
+      "loss": 0.3877,
+      "step": 80395
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7168152904976243e-07,
+      "loss": 0.4683,
+      "step": 80396
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.716580357559485e-07,
+      "loss": 0.4331,
+      "step": 80397
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7163454286666135e-07,
+      "loss": 0.3176,
+      "step": 80398
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.716110503819194e-07,
+      "loss": 0.5029,
+      "step": 80399
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.715875583017409e-07,
+      "loss": 0.2347,
+      "step": 80400
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.715640666261432e-07,
+      "loss": 0.3994,
+      "step": 80401
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.715405753551446e-07,
+      "loss": 0.3955,
+      "step": 80402
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7151708448876304e-07,
+      "loss": 0.4666,
+      "step": 80403
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.714935940270168e-07,
+      "loss": 0.4463,
+      "step": 80404
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.714701039699235e-07,
+      "loss": 0.4272,
+      "step": 80405
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.71446614317501e-07,
+      "loss": 0.3694,
+      "step": 80406
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.71423125069768e-07,
+      "loss": 0.4248,
+      "step": 80407
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.713996362267416e-07,
+      "loss": 0.4648,
+      "step": 80408
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.713761477884407e-07,
+      "loss": 0.2157,
+      "step": 80409
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7135265975488223e-07,
+      "loss": 0.4517,
+      "step": 80410
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.713291721260848e-07,
+      "loss": 0.4014,
+      "step": 80411
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7130568490206634e-07,
+      "loss": 0.4893,
+      "step": 80412
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7128219808284506e-07,
+      "loss": 0.3275,
+      "step": 80413
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7125871166843847e-07,
+      "loss": 0.4355,
+      "step": 80414
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7123522565886476e-07,
+      "loss": 0.3041,
+      "step": 80415
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7121174005414177e-07,
+      "loss": 0.4004,
+      "step": 80416
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7118825485428846e-07,
+      "loss": 0.4019,
+      "step": 80417
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.711647700593213e-07,
+      "loss": 0.3081,
+      "step": 80418
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.711412856692589e-07,
+      "loss": 0.3931,
+      "step": 80419
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.711178016841194e-07,
+      "loss": 0.6548,
+      "step": 80420
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7109431810392095e-07,
+      "loss": 0.542,
+      "step": 80421
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7107083492868094e-07,
+      "loss": 0.4958,
+      "step": 80422
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.710473521584176e-07,
+      "loss": 0.4497,
+      "step": 80423
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.71023869793149e-07,
+      "loss": 0.4028,
+      "step": 80424
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7100038783289353e-07,
+      "loss": 0.417,
+      "step": 80425
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.709769062776683e-07,
+      "loss": 0.5435,
+      "step": 80426
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.709534251274921e-07,
+      "loss": 0.3552,
+      "step": 80427
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7092994438238233e-07,
+      "loss": 0.4629,
+      "step": 80428
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.70906464042357e-07,
+      "loss": 0.4463,
+      "step": 80429
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7088298410743466e-07,
+      "loss": 0.3906,
+      "step": 80430
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7085950457763255e-07,
+      "loss": 0.4814,
+      "step": 80431
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7083602545296895e-07,
+      "loss": 0.355,
+      "step": 80432
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.708125467334619e-07,
+      "loss": 0.4722,
+      "step": 80433
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.707890684191297e-07,
+      "loss": 0.4771,
+      "step": 80434
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7076559050998966e-07,
+      "loss": 0.4326,
+      "step": 80435
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7074211300606e-07,
+      "loss": 0.541,
+      "step": 80436
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7071863590735916e-07,
+      "loss": 0.5435,
+      "step": 80437
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.706951592139048e-07,
+      "loss": 0.51,
+      "step": 80438
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7067168292571435e-07,
+      "loss": 0.2495,
+      "step": 80439
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.706482070428063e-07,
+      "loss": 0.4434,
+      "step": 80440
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7062473156519853e-07,
+      "loss": 0.5259,
+      "step": 80441
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.706012564929095e-07,
+      "loss": 0.4253,
+      "step": 80442
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.705777818259563e-07,
+      "loss": 0.3523,
+      "step": 80443
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7055430756435745e-07,
+      "loss": 0.355,
+      "step": 80444
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7053083370813074e-07,
+      "loss": 0.4526,
+      "step": 80445
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.705073602572946e-07,
+      "loss": 0.457,
+      "step": 80446
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.704838872118666e-07,
+      "loss": 0.3628,
+      "step": 80447
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7046041457186446e-07,
+      "loss": 0.4077,
+      "step": 80448
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7043694233730635e-07,
+      "loss": 0.4209,
+      "step": 80449
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.704134705082103e-07,
+      "loss": 0.5015,
+      "step": 80450
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.703899990845949e-07,
+      "loss": 0.5322,
+      "step": 80451
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.703665280664769e-07,
+      "loss": 0.356,
+      "step": 80452
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.70343057453875e-07,
+      "loss": 0.4036,
+      "step": 80453
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7031958724680743e-07,
+      "loss": 0.343,
+      "step": 80454
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7029611744529143e-07,
+      "loss": 0.3672,
+      "step": 80455
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.702726480493453e-07,
+      "loss": 0.293,
+      "step": 80456
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7024917905898743e-07,
+      "loss": 0.4515,
+      "step": 80457
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.702257104742351e-07,
+      "loss": 0.2966,
+      "step": 80458
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7020224229510686e-07,
+      "loss": 0.3381,
+      "step": 80459
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7017877452162014e-07,
+      "loss": 0.4844,
+      "step": 80460
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.70155307153793e-07,
+      "loss": 0.3989,
+      "step": 80461
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7013184019164374e-07,
+      "loss": 0.2706,
+      "step": 80462
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7010837363519053e-07,
+      "loss": 0.3347,
+      "step": 80463
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.700849074844505e-07,
+      "loss": 0.437,
+      "step": 80464
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.700614417394422e-07,
+      "loss": 0.4595,
+      "step": 80465
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7003797640018337e-07,
+      "loss": 0.3106,
+      "step": 80466
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.7001451146669257e-07,
+      "loss": 0.3523,
+      "step": 80467
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.699910469389872e-07,
+      "loss": 0.332,
+      "step": 80468
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.69967582817085e-07,
+      "loss": 0.5073,
+      "step": 80469
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.699441191010042e-07,
+      "loss": 0.4565,
+      "step": 80470
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6992065579076323e-07,
+      "loss": 0.5483,
+      "step": 80471
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.698971928863793e-07,
+      "loss": 0.4375,
+      "step": 80472
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.698737303878707e-07,
+      "loss": 0.2845,
+      "step": 80473
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.698502682952553e-07,
+      "loss": 0.3923,
+      "step": 80474
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.698268066085517e-07,
+      "loss": 0.2846,
+      "step": 80475
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6980334532777686e-07,
+      "loss": 0.3479,
+      "step": 80476
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6977988445294967e-07,
+      "loss": 0.3416,
+      "step": 80477
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.697564239840872e-07,
+      "loss": 0.499,
+      "step": 80478
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6973296392120785e-07,
+      "loss": 0.6392,
+      "step": 80479
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6970950426432996e-07,
+      "loss": 0.4084,
+      "step": 80480
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6968604501347086e-07,
+      "loss": 0.4678,
+      "step": 80481
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.696625861686486e-07,
+      "loss": 0.4316,
+      "step": 80482
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6963912772988143e-07,
+      "loss": 0.3696,
+      "step": 80483
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.696156696971876e-07,
+      "loss": 0.3916,
+      "step": 80484
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.695922120705842e-07,
+      "loss": 0.4434,
+      "step": 80485
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6956875485008965e-07,
+      "loss": 0.4263,
+      "step": 80486
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.695452980357223e-07,
+      "loss": 0.5264,
+      "step": 80487
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6952184162749973e-07,
+      "loss": 0.4678,
+      "step": 80488
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6949838562543963e-07,
+      "loss": 0.3901,
+      "step": 80489
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6947493002956004e-07,
+      "loss": 0.4189,
+      "step": 80490
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6945147483987924e-07,
+      "loss": 0.418,
+      "step": 80491
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6942802005641537e-07,
+      "loss": 0.2683,
+      "step": 80492
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6940456567918574e-07,
+      "loss": 0.3092,
+      "step": 80493
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.693811117082086e-07,
+      "loss": 0.392,
+      "step": 80494
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6935765814350193e-07,
+      "loss": 0.4563,
+      "step": 80495
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.69334204985084e-07,
+      "loss": 0.4062,
+      "step": 80496
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.693107522329726e-07,
+      "loss": 0.2913,
+      "step": 80497
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.692872998871852e-07,
+      "loss": 0.3079,
+      "step": 80498
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6926384794773996e-07,
+      "loss": 0.4326,
+      "step": 80499
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6924039641465507e-07,
+      "loss": 0.373,
+      "step": 80500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6921694528794874e-07,
+      "loss": 0.4209,
+      "step": 80501
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6919349456763834e-07,
+      "loss": 0.376,
+      "step": 80502
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6917004425374197e-07,
+      "loss": 0.4011,
+      "step": 80503
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.691465943462777e-07,
+      "loss": 0.4556,
+      "step": 80504
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6912314484526384e-07,
+      "loss": 0.2974,
+      "step": 80505
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.690996957507176e-07,
+      "loss": 0.2841,
+      "step": 80506
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.690762470626578e-07,
+      "loss": 0.3005,
+      "step": 80507
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6905279878110135e-07,
+      "loss": 0.2868,
+      "step": 80508
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.690293509060672e-07,
+      "loss": 0.5352,
+      "step": 80509
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.690059034375725e-07,
+      "loss": 0.3318,
+      "step": 80510
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6898245637563563e-07,
+      "loss": 0.382,
+      "step": 80511
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6895900972027446e-07,
+      "loss": 0.4258,
+      "step": 80512
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.689355634715072e-07,
+      "loss": 0.3972,
+      "step": 80513
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6891211762935134e-07,
+      "loss": 0.4302,
+      "step": 80514
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6888867219382497e-07,
+      "loss": 0.3828,
+      "step": 80515
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6886522716494615e-07,
+      "loss": 0.408,
+      "step": 80516
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.688417825427331e-07,
+      "loss": 0.4653,
+      "step": 80517
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6881833832720363e-07,
+      "loss": 0.2723,
+      "step": 80518
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.687948945183751e-07,
+      "loss": 0.5146,
+      "step": 80519
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.687714511162658e-07,
+      "loss": 0.376,
+      "step": 80520
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.687480081208939e-07,
+      "loss": 0.4771,
+      "step": 80521
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.687245655322775e-07,
+      "loss": 0.4072,
+      "step": 80522
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.687011233504339e-07,
+      "loss": 0.4153,
+      "step": 80523
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.686776815753815e-07,
+      "loss": 0.3346,
+      "step": 80524
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.686542402071384e-07,
+      "loss": 0.3843,
+      "step": 80525
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.686307992457221e-07,
+      "loss": 0.4258,
+      "step": 80526
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.686073586911511e-07,
+      "loss": 0.4526,
+      "step": 80527
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.685839185434426e-07,
+      "loss": 0.4873,
+      "step": 80528
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6856047880261496e-07,
+      "loss": 0.4653,
+      "step": 80529
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.685370394686865e-07,
+      "loss": 0.4058,
+      "step": 80530
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6851360054167443e-07,
+      "loss": 0.429,
+      "step": 80531
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.684901620215971e-07,
+      "loss": 0.3804,
+      "step": 80532
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.684667239084723e-07,
+      "loss": 0.4189,
+      "step": 80533
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.684432862023185e-07,
+      "loss": 0.4402,
+      "step": 80534
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.684198489031529e-07,
+      "loss": 0.426,
+      "step": 80535
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.683964120109938e-07,
+      "loss": 0.3142,
+      "step": 80536
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6837297552585954e-07,
+      "loss": 0.4985,
+      "step": 80537
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6834953944776714e-07,
+      "loss": 0.4487,
+      "step": 80538
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.683261037767354e-07,
+      "loss": 0.3567,
+      "step": 80539
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6830266851278164e-07,
+      "loss": 0.5342,
+      "step": 80540
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6827923365592413e-07,
+      "loss": 0.3447,
+      "step": 80541
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.682557992061811e-07,
+      "loss": 0.3879,
+      "step": 80542
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.682323651635698e-07,
+      "loss": 0.3772,
+      "step": 80543
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.682089315281084e-07,
+      "loss": 0.2927,
+      "step": 80544
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6818549829981503e-07,
+      "loss": 0.3882,
+      "step": 80545
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.68162065478708e-07,
+      "loss": 0.5327,
+      "step": 80546
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.681386330648048e-07,
+      "loss": 0.418,
+      "step": 80547
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6811520105812297e-07,
+      "loss": 0.4233,
+      "step": 80548
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.680917694586809e-07,
+      "loss": 0.4697,
+      "step": 80549
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.680683382664965e-07,
+      "loss": 0.4683,
+      "step": 80550
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.680449074815881e-07,
+      "loss": 0.4507,
+      "step": 80551
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6802147710397286e-07,
+      "loss": 0.3691,
+      "step": 80552
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.679980471336692e-07,
+      "loss": 0.4443,
+      "step": 80553
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.679746175706949e-07,
+      "loss": 0.5732,
+      "step": 80554
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.679511884150683e-07,
+      "loss": 0.3762,
+      "step": 80555
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6792775966680674e-07,
+      "loss": 0.3867,
+      "step": 80556
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.679043313259288e-07,
+      "loss": 0.3286,
+      "step": 80557
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6788090339245156e-07,
+      "loss": 0.4242,
+      "step": 80558
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.678574758663939e-07,
+      "loss": 0.5059,
+      "step": 80559
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6783404874777286e-07,
+      "loss": 0.2932,
+      "step": 80560
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6781062203660695e-07,
+      "loss": 0.4634,
+      "step": 80561
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.677871957329139e-07,
+      "loss": 0.3843,
+      "step": 80562
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.677637698367121e-07,
+      "loss": 0.4658,
+      "step": 80563
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.677403443480187e-07,
+      "loss": 0.4556,
+      "step": 80564
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6771691926685207e-07,
+      "loss": 0.4082,
+      "step": 80565
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.6769349459323016e-07,
+      "loss": 0.4154,
+      "step": 80566
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.676700703271712e-07,
+      "loss": 0.356,
+      "step": 80567
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 4.676466464686928e-07,
+      "loss": 0.4624,
+      "step": 80568
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6762322301781257e-07,
+      "loss": 0.3276,
+      "step": 80569
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6759979997454856e-07,
+      "loss": 0.41,
+      "step": 80570
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.675763773389191e-07,
+      "loss": 0.3401,
+      "step": 80571
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6755295511094215e-07,
+      "loss": 0.3887,
+      "step": 80572
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6752953329063517e-07,
+      "loss": 0.3756,
+      "step": 80573
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6750611187801636e-07,
+      "loss": 0.324,
+      "step": 80574
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6748269087310353e-07,
+      "loss": 0.447,
+      "step": 80575
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6745927027591514e-07,
+      "loss": 0.3464,
+      "step": 80576
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6743585008646866e-07,
+      "loss": 0.4248,
+      "step": 80577
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6741243030478183e-07,
+      "loss": 0.4897,
+      "step": 80578
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.673890109308727e-07,
+      "loss": 0.4346,
+      "step": 80579
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6736559196475965e-07,
+      "loss": 0.5088,
+      "step": 80580
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6734217340646e-07,
+      "loss": 0.5288,
+      "step": 80581
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6731875525599185e-07,
+      "loss": 0.5278,
+      "step": 80582
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6729533751337336e-07,
+      "loss": 0.3762,
+      "step": 80583
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.672719201786226e-07,
+      "loss": 0.436,
+      "step": 80584
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.672485032517569e-07,
+      "loss": 0.4463,
+      "step": 80585
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.672250867327946e-07,
+      "loss": 0.2716,
+      "step": 80586
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6720167062175387e-07,
+      "loss": 0.3486,
+      "step": 80587
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.671782549186519e-07,
+      "loss": 0.3511,
+      "step": 80588
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6715483962350744e-07,
+      "loss": 0.4109,
+      "step": 80589
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.671314247363376e-07,
+      "loss": 0.4221,
+      "step": 80590
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6710801025716083e-07,
+      "loss": 0.4692,
+      "step": 80591
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6708459618599496e-07,
+      "loss": 0.3616,
+      "step": 80592
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.670611825228582e-07,
+      "loss": 0.5137,
+      "step": 80593
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6703776926776785e-07,
+      "loss": 0.3918,
+      "step": 80594
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6701435642074215e-07,
+      "loss": 0.3948,
+      "step": 80595
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6699094398179907e-07,
+      "loss": 0.3687,
+      "step": 80596
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6696753195095716e-07,
+      "loss": 0.3135,
+      "step": 80597
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6694412032823304e-07,
+      "loss": 0.4646,
+      "step": 80598
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6692070911364533e-07,
+      "loss": 0.3003,
+      "step": 80599
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6689729830721194e-07,
+      "loss": 0.499,
+      "step": 80600
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.668738879089511e-07,
+      "loss": 0.4434,
+      "step": 80601
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.668504779188801e-07,
+      "loss": 0.3584,
+      "step": 80602
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.668270683370171e-07,
+      "loss": 0.3315,
+      "step": 80603
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6680365916338007e-07,
+      "loss": 0.415,
+      "step": 80604
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.667802503979874e-07,
+      "loss": 0.3853,
+      "step": 80605
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6675684204085617e-07,
+      "loss": 0.344,
+      "step": 80606
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6673343409200516e-07,
+      "loss": 0.2678,
+      "step": 80607
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.667100265514513e-07,
+      "loss": 0.4746,
+      "step": 80608
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6668661941921315e-07,
+      "loss": 0.376,
+      "step": 80609
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6666321269530894e-07,
+      "loss": 0.3442,
+      "step": 80610
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6663980637975577e-07,
+      "loss": 0.415,
+      "step": 80611
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.66616400472572e-07,
+      "loss": 0.4126,
+      "step": 80612
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6659299497377555e-07,
+      "loss": 0.4795,
+      "step": 80613
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6656958988338455e-07,
+      "loss": 0.2474,
+      "step": 80614
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6654618520141653e-07,
+      "loss": 0.3984,
+      "step": 80615
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6652278092788944e-07,
+      "loss": 0.4309,
+      "step": 80616
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.664993770628217e-07,
+      "loss": 0.4224,
+      "step": 80617
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6647597360623083e-07,
+      "loss": 0.4336,
+      "step": 80618
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6645257055813447e-07,
+      "loss": 0.4233,
+      "step": 80619
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.664291679185508e-07,
+      "loss": 0.4397,
+      "step": 80620
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.664057656874978e-07,
+      "loss": 0.272,
+      "step": 80621
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.663823638649937e-07,
+      "loss": 0.3748,
+      "step": 80622
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.663589624510558e-07,
+      "loss": 0.4443,
+      "step": 80623
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.663355614457023e-07,
+      "loss": 0.4248,
+      "step": 80624
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.66312160848951e-07,
+      "loss": 0.3757,
+      "step": 80625
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6628876066082046e-07,
+      "loss": 0.4712,
+      "step": 80626
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6626536088132794e-07,
+      "loss": 0.3354,
+      "step": 80627
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6624196151049124e-07,
+      "loss": 0.4106,
+      "step": 80628
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.662185625483285e-07,
+      "loss": 0.5088,
+      "step": 80629
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6619516399485757e-07,
+      "loss": 0.2471,
+      "step": 80630
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6617176585009687e-07,
+      "loss": 0.3997,
+      "step": 80631
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6614836811406356e-07,
+      "loss": 0.396,
+      "step": 80632
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6612497078677595e-07,
+      "loss": 0.3589,
+      "step": 80633
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.661015738682522e-07,
+      "loss": 0.3318,
+      "step": 80634
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.660781773585095e-07,
+      "loss": 0.3618,
+      "step": 80635
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6605478125756636e-07,
+      "loss": 0.4194,
+      "step": 80636
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6603138556544075e-07,
+      "loss": 0.5244,
+      "step": 80637
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6600799028215e-07,
+      "loss": 0.4072,
+      "step": 80638
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6598459540771284e-07,
+      "loss": 0.4702,
+      "step": 80639
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.659612009421463e-07,
+      "loss": 0.3528,
+      "step": 80640
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.659378068854687e-07,
+      "loss": 0.4673,
+      "step": 80641
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.65914413237698e-07,
+      "loss": 0.4004,
+      "step": 80642
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6589101999885237e-07,
+      "loss": 0.3755,
+      "step": 80643
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6586762716894914e-07,
+      "loss": 0.2965,
+      "step": 80644
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.658442347480064e-07,
+      "loss": 0.3486,
+      "step": 80645
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6582084273604226e-07,
+      "loss": 0.4302,
+      "step": 80646
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.657974511330749e-07,
+      "loss": 0.3289,
+      "step": 80647
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6577405993912176e-07,
+      "loss": 0.4526,
+      "step": 80648
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.657506691542006e-07,
+      "loss": 0.4692,
+      "step": 80649
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6572727877832963e-07,
+      "loss": 0.5186,
+      "step": 80650
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.65703888811527e-07,
+      "loss": 0.5312,
+      "step": 80651
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.656804992538099e-07,
+      "loss": 0.3261,
+      "step": 80652
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6565711010519683e-07,
+      "loss": 0.353,
+      "step": 80653
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6563372136570544e-07,
+      "loss": 0.542,
+      "step": 80654
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6561033303535414e-07,
+      "loss": 0.4761,
+      "step": 80655
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.655869451141601e-07,
+      "loss": 0.438,
+      "step": 80656
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6556355760214194e-07,
+      "loss": 0.3494,
+      "step": 80657
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.655401704993167e-07,
+      "loss": 0.3691,
+      "step": 80658
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.655167838057029e-07,
+      "loss": 0.4971,
+      "step": 80659
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6549339752131867e-07,
+      "loss": 0.3667,
+      "step": 80660
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6547001164618114e-07,
+      "loss": 0.3503,
+      "step": 80661
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.654466261803087e-07,
+      "loss": 0.2986,
+      "step": 80662
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.654232411237191e-07,
+      "loss": 0.4824,
+      "step": 80663
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6539985647643087e-07,
+      "loss": 0.3887,
+      "step": 80664
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6537647223846086e-07,
+      "loss": 0.4346,
+      "step": 80665
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.653530884098277e-07,
+      "loss": 0.4131,
+      "step": 80666
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.653297049905494e-07,
+      "loss": 0.5298,
+      "step": 80667
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6530632198064347e-07,
+      "loss": 0.5049,
+      "step": 80668
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6528293938012754e-07,
+      "loss": 0.4294,
+      "step": 80669
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6525955718901986e-07,
+      "loss": 0.2806,
+      "step": 80670
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.652361754073384e-07,
+      "loss": 0.4385,
+      "step": 80671
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.652127940351014e-07,
+      "loss": 0.2031,
+      "step": 80672
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6518941307232596e-07,
+      "loss": 0.3877,
+      "step": 80673
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6516603251903043e-07,
+      "loss": 0.4551,
+      "step": 80674
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6514265237523266e-07,
+      "loss": 0.3467,
+      "step": 80675
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.65119272640951e-07,
+      "loss": 0.3833,
+      "step": 80676
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.650958933162028e-07,
+      "loss": 0.3313,
+      "step": 80677
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.650725144010058e-07,
+      "loss": 0.3032,
+      "step": 80678
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.650491358953782e-07,
+      "loss": 0.519,
+      "step": 80679
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6502575779933773e-07,
+      "loss": 0.4192,
+      "step": 80680
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6500238011290293e-07,
+      "loss": 0.4236,
+      "step": 80681
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.649790028360908e-07,
+      "loss": 0.4414,
+      "step": 80682
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.649556259689197e-07,
+      "loss": 0.5122,
+      "step": 80683
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.649322495114074e-07,
+      "loss": 0.521,
+      "step": 80684
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6490887346357233e-07,
+      "loss": 0.3597,
+      "step": 80685
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6488549782543186e-07,
+      "loss": 0.3801,
+      "step": 80686
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6486212259700366e-07,
+      "loss": 0.4175,
+      "step": 80687
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.648387477783058e-07,
+      "loss": 0.4531,
+      "step": 80688
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.648153733693568e-07,
+      "loss": 0.2231,
+      "step": 80689
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6479199937017364e-07,
+      "loss": 0.3516,
+      "step": 80690
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.647686257807746e-07,
+      "loss": 0.4517,
+      "step": 80691
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.647452526011777e-07,
+      "loss": 0.4219,
+      "step": 80692
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.647218798314011e-07,
+      "loss": 0.306,
+      "step": 80693
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.646985074714619e-07,
+      "loss": 0.374,
+      "step": 80694
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.646751355213785e-07,
+      "loss": 0.3781,
+      "step": 80695
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.64651763981169e-07,
+      "loss": 0.4854,
+      "step": 80696
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.646283928508508e-07,
+      "loss": 0.3572,
+      "step": 80697
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.646050221304423e-07,
+      "loss": 0.5146,
+      "step": 80698
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6458165181996077e-07,
+      "loss": 0.3651,
+      "step": 80699
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.645582819194245e-07,
+      "loss": 0.4395,
+      "step": 80700
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.645349124288513e-07,
+      "loss": 0.4824,
+      "step": 80701
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6451154334825946e-07,
+      "loss": 0.4702,
+      "step": 80702
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.644881746776661e-07,
+      "loss": 0.4546,
+      "step": 80703
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6446480641708965e-07,
+      "loss": 0.5107,
+      "step": 80704
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.644414385665482e-07,
+      "loss": 0.28,
+      "step": 80705
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.644180711260592e-07,
+      "loss": 0.3843,
+      "step": 80706
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.643947040956403e-07,
+      "loss": 0.5649,
+      "step": 80707
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.643713374753099e-07,
+      "loss": 0.418,
+      "step": 80708
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6434797126508563e-07,
+      "loss": 0.4043,
+      "step": 80709
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.643246054649859e-07,
+      "loss": 0.5298,
+      "step": 80710
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.643012400750277e-07,
+      "loss": 0.4458,
+      "step": 80711
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.642778750952295e-07,
+      "loss": 0.3994,
+      "step": 80712
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6425451052560914e-07,
+      "loss": 0.3606,
+      "step": 80713
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.642311463661848e-07,
+      "loss": 0.4883,
+      "step": 80714
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6420778261697356e-07,
+      "loss": 0.3994,
+      "step": 80715
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6418441927799437e-07,
+      "loss": 0.3623,
+      "step": 80716
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6416105634926394e-07,
+      "loss": 0.2247,
+      "step": 80717
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.641376938308009e-07,
+      "loss": 0.4409,
+      "step": 80718
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6411433172262327e-07,
+      "loss": 0.4736,
+      "step": 80719
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.640909700247484e-07,
+      "loss": 0.257,
+      "step": 80720
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.640676087371943e-07,
+      "loss": 0.2482,
+      "step": 80721
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6404424785997944e-07,
+      "loss": 0.2606,
+      "step": 80722
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6402088739312084e-07,
+      "loss": 0.4817,
+      "step": 80723
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.639975273366369e-07,
+      "loss": 0.3417,
+      "step": 80724
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.639741676905453e-07,
+      "loss": 0.3174,
+      "step": 80725
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6395080845486446e-07,
+      "loss": 0.3706,
+      "step": 80726
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6392744962961183e-07,
+      "loss": 0.4175,
+      "step": 80727
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.639040912148049e-07,
+      "loss": 0.3975,
+      "step": 80728
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6388073321046197e-07,
+      "loss": 0.4614,
+      "step": 80729
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.638573756166009e-07,
+      "loss": 0.3364,
+      "step": 80730
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6383401843324e-07,
+      "loss": 0.3884,
+      "step": 80731
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6381066166039626e-07,
+      "loss": 0.4766,
+      "step": 80732
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6378730529808817e-07,
+      "loss": 0.3218,
+      "step": 80733
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6376394934633345e-07,
+      "loss": 0.4121,
+      "step": 80734
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.637405938051504e-07,
+      "loss": 0.3412,
+      "step": 80735
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6371723867455646e-07,
+      "loss": 0.5586,
+      "step": 80736
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6369388395456923e-07,
+      "loss": 0.4351,
+      "step": 80737
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.636705296452069e-07,
+      "loss": 0.4312,
+      "step": 80738
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6364717574648746e-07,
+      "loss": 0.4204,
+      "step": 80739
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6362382225842913e-07,
+      "loss": 0.4575,
+      "step": 80740
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.636004691810489e-07,
+      "loss": 0.5269,
+      "step": 80741
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6357711651436514e-07,
+      "loss": 0.4453,
+      "step": 80742
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6355376425839607e-07,
+      "loss": 0.4492,
+      "step": 80743
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6353041241315883e-07,
+      "loss": 0.4553,
+      "step": 80744
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6350706097867174e-07,
+      "loss": 0.2825,
+      "step": 80745
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.63483709954953e-07,
+      "loss": 0.5254,
+      "step": 80746
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6346035934201976e-07,
+      "loss": 0.4238,
+      "step": 80747
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.634370091398906e-07,
+      "loss": 0.3601,
+      "step": 80748
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.634136593485827e-07,
+      "loss": 0.2351,
+      "step": 80749
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.633903099681142e-07,
+      "loss": 0.4065,
+      "step": 80750
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.633669609985031e-07,
+      "loss": 0.3481,
+      "step": 80751
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.633436124397676e-07,
+      "loss": 0.4436,
+      "step": 80752
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6332026429192493e-07,
+      "loss": 0.3896,
+      "step": 80753
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6329691655499325e-07,
+      "loss": 0.4468,
+      "step": 80754
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.632735692289904e-07,
+      "loss": 0.3218,
+      "step": 80755
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6325022231393464e-07,
+      "loss": 0.3779,
+      "step": 80756
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6322687580984353e-07,
+      "loss": 0.395,
+      "step": 80757
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.632035297167346e-07,
+      "loss": 0.438,
+      "step": 80758
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.63180184034626e-07,
+      "loss": 0.4194,
+      "step": 80759
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.63156838763536e-07,
+      "loss": 0.2635,
+      "step": 80760
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.631334939034818e-07,
+      "loss": 0.4214,
+      "step": 80761
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6311014945448154e-07,
+      "loss": 0.479,
+      "step": 80762
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6308680541655323e-07,
+      "loss": 0.3833,
+      "step": 80763
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.63063461789715e-07,
+      "loss": 0.5195,
+      "step": 80764
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6304011857398394e-07,
+      "loss": 0.2659,
+      "step": 80765
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.630167757693788e-07,
+      "loss": 0.5066,
+      "step": 80766
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6299343337591667e-07,
+      "loss": 0.4053,
+      "step": 80767
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6297009139361577e-07,
+      "loss": 0.3593,
+      "step": 80768
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6294674982249436e-07,
+      "loss": 0.4207,
+      "step": 80769
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.629234086625695e-07,
+      "loss": 0.3367,
+      "step": 80770
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.629000679138596e-07,
+      "loss": 0.1906,
+      "step": 80771
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6287672757638243e-07,
+      "loss": 0.3358,
+      "step": 80772
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.628533876501561e-07,
+      "loss": 0.5586,
+      "step": 80773
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.628300481351979e-07,
+      "loss": 0.3108,
+      "step": 80774
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6280670903152606e-07,
+      "loss": 0.2964,
+      "step": 80775
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6278337033915884e-07,
+      "loss": 0.4136,
+      "step": 80776
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6276003205811356e-07,
+      "loss": 0.4746,
+      "step": 80777
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6273669418840787e-07,
+      "loss": 0.5283,
+      "step": 80778
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6271335673006007e-07,
+      "loss": 0.4189,
+      "step": 80779
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6269001968308796e-07,
+      "loss": 0.4497,
+      "step": 80780
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6266668304750964e-07,
+      "loss": 0.4692,
+      "step": 80781
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.626433468233425e-07,
+      "loss": 0.4971,
+      "step": 80782
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6262001101060457e-07,
+      "loss": 0.3867,
+      "step": 80783
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.625966756093138e-07,
+      "loss": 0.3268,
+      "step": 80784
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.625733406194885e-07,
+      "loss": 0.3441,
+      "step": 80785
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.625500060411459e-07,
+      "loss": 0.5679,
+      "step": 80786
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6252667187430386e-07,
+      "loss": 0.4321,
+      "step": 80787
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6250333811898034e-07,
+      "loss": 0.4302,
+      "step": 80788
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6248000477519334e-07,
+      "loss": 0.4463,
+      "step": 80789
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.62456671842961e-07,
+      "loss": 0.3936,
+      "step": 80790
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6243333932230057e-07,
+      "loss": 0.4331,
+      "step": 80791
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.624100072132301e-07,
+      "loss": 0.3784,
+      "step": 80792
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.623866755157676e-07,
+      "loss": 0.2679,
+      "step": 80793
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.623633442299314e-07,
+      "loss": 0.4868,
+      "step": 80794
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6234001335573837e-07,
+      "loss": 0.5508,
+      "step": 80795
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6231668289320726e-07,
+      "loss": 0.3494,
+      "step": 80796
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6229335284235513e-07,
+      "loss": 0.4985,
+      "step": 80797
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6227002320320077e-07,
+      "loss": 0.4272,
+      "step": 80798
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.622466939757611e-07,
+      "loss": 0.3052,
+      "step": 80799
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6222336516005434e-07,
+      "loss": 0.377,
+      "step": 80800
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.622000367560984e-07,
+      "loss": 0.4438,
+      "step": 80801
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.621767087639117e-07,
+      "loss": 0.3521,
+      "step": 80802
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.621533811835111e-07,
+      "loss": 0.4194,
+      "step": 80803
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.621300540149149e-07,
+      "loss": 0.4414,
+      "step": 80804
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6210672725814095e-07,
+      "loss": 0.4067,
+      "step": 80805
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.620834009132076e-07,
+      "loss": 0.4465,
+      "step": 80806
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.620600749801322e-07,
+      "loss": 0.3618,
+      "step": 80807
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6203674945893234e-07,
+      "loss": 0.3979,
+      "step": 80808
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.620134243496262e-07,
+      "loss": 0.418,
+      "step": 80809
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.619900996522316e-07,
+      "loss": 0.4326,
+      "step": 80810
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6196677536676686e-07,
+      "loss": 0.4668,
+      "step": 80811
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.61943451493249e-07,
+      "loss": 0.4644,
+      "step": 80812
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6192012803169633e-07,
+      "loss": 0.3978,
+      "step": 80813
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.61896804982127e-07,
+      "loss": 0.3838,
+      "step": 80814
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6187348234455826e-07,
+      "loss": 0.3794,
+      "step": 80815
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.618501601190086e-07,
+      "loss": 0.5713,
+      "step": 80816
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6182683830549516e-07,
+      "loss": 0.4038,
+      "step": 80817
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6180351690403616e-07,
+      "loss": 0.4373,
+      "step": 80818
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.617801959146498e-07,
+      "loss": 0.4214,
+      "step": 80819
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.617568753373532e-07,
+      "loss": 0.4031,
+      "step": 80820
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.617335551721647e-07,
+      "loss": 0.4238,
+      "step": 80821
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6171023541910203e-07,
+      "loss": 0.2961,
+      "step": 80822
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6168691607818345e-07,
+      "loss": 0.429,
+      "step": 80823
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.61663597149426e-07,
+      "loss": 0.2859,
+      "step": 80824
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6164027863284807e-07,
+      "loss": 0.354,
+      "step": 80825
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.616169605284678e-07,
+      "loss": 0.3584,
+      "step": 80826
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.615936428363022e-07,
+      "loss": 0.3611,
+      "step": 80827
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.615703255563701e-07,
+      "loss": 0.2996,
+      "step": 80828
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6154700868868834e-07,
+      "loss": 0.3523,
+      "step": 80829
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.615236922332754e-07,
+      "loss": 0.5054,
+      "step": 80830
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6150037619014925e-07,
+      "loss": 0.226,
+      "step": 80831
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6147706055932725e-07,
+      "loss": 0.5015,
+      "step": 80832
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.614537453408275e-07,
+      "loss": 0.4216,
+      "step": 80833
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.614304305346678e-07,
+      "loss": 0.4058,
+      "step": 80834
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6140711614086636e-07,
+      "loss": 0.4712,
+      "step": 80835
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.613838021594408e-07,
+      "loss": 0.3433,
+      "step": 80836
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6136048859040856e-07,
+      "loss": 0.3823,
+      "step": 80837
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.613371754337878e-07,
+      "loss": 0.4045,
+      "step": 80838
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.613138626895964e-07,
+      "loss": 0.3506,
+      "step": 80839
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.612905503578525e-07,
+      "loss": 0.4976,
+      "step": 80840
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.612672384385734e-07,
+      "loss": 0.3965,
+      "step": 80841
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.612439269317772e-07,
+      "loss": 0.3616,
+      "step": 80842
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.612206158374816e-07,
+      "loss": 0.4233,
+      "step": 80843
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.611973051557051e-07,
+      "loss": 0.5635,
+      "step": 80844
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.611739948864647e-07,
+      "loss": 0.396,
+      "step": 80845
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6115068502977884e-07,
+      "loss": 0.3163,
+      "step": 80846
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6112737558566485e-07,
+      "loss": 0.5137,
+      "step": 80847
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6110406655414123e-07,
+      "loss": 0.4604,
+      "step": 80848
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6108075793522505e-07,
+      "loss": 0.415,
+      "step": 80849
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6105744972893457e-07,
+      "loss": 0.4561,
+      "step": 80850
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6103414193528757e-07,
+      "loss": 0.5278,
+      "step": 80851
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6101083455430236e-07,
+      "loss": 0.3506,
+      "step": 80852
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.60987527585996e-07,
+      "loss": 0.3918,
+      "step": 80853
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.609642210303867e-07,
+      "loss": 0.3303,
+      "step": 80854
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.609409148874923e-07,
+      "loss": 0.3491,
+      "step": 80855
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.60917609157331e-07,
+      "loss": 0.4551,
+      "step": 80856
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6089430383992034e-07,
+      "loss": 0.5986,
+      "step": 80857
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.608709989352777e-07,
+      "loss": 0.4448,
+      "step": 80858
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6084769444342143e-07,
+      "loss": 0.3413,
+      "step": 80859
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.608243903643693e-07,
+      "loss": 0.3478,
+      "step": 80860
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6080108669813945e-07,
+      "loss": 0.5654,
+      "step": 80861
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6077778344474903e-07,
+      "loss": 0.2997,
+      "step": 80862
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.607544806042163e-07,
+      "loss": 0.4116,
+      "step": 80863
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6073117817655905e-07,
+      "loss": 0.3999,
+      "step": 80864
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.607078761617955e-07,
+      "loss": 0.4351,
+      "step": 80865
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6068457455994314e-07,
+      "loss": 0.4253,
+      "step": 80866
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6066127337101945e-07,
+      "loss": 0.4038,
+      "step": 80867
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.606379725950427e-07,
+      "loss": 0.396,
+      "step": 80868
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6061467223203086e-07,
+      "loss": 0.4424,
+      "step": 80869
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.605913722820013e-07,
+      "loss": 0.4609,
+      "step": 80870
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.605680727449721e-07,
+      "loss": 0.4956,
+      "step": 80871
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.605447736209611e-07,
+      "loss": 0.2132,
+      "step": 80872
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.605214749099866e-07,
+      "loss": 0.4443,
+      "step": 80873
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6049817661206556e-07,
+      "loss": 0.4287,
+      "step": 80874
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6047487872721633e-07,
+      "loss": 0.5801,
+      "step": 80875
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.60451581255457e-07,
+      "loss": 0.3868,
+      "step": 80876
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.604282841968047e-07,
+      "loss": 0.3496,
+      "step": 80877
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.604049875512781e-07,
+      "loss": 0.2908,
+      "step": 80878
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6038169131889413e-07,
+      "loss": 0.5059,
+      "step": 80879
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6035839549967117e-07,
+      "loss": 0.3616,
+      "step": 80880
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6033510009362686e-07,
+      "loss": 0.4937,
+      "step": 80881
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.603118051007796e-07,
+      "loss": 0.4839,
+      "step": 80882
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.602885105211464e-07,
+      "loss": 0.5151,
+      "step": 80883
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.602652163547455e-07,
+      "loss": 0.4729,
+      "step": 80884
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.60241922601595e-07,
+      "loss": 0.4121,
+      "step": 80885
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.602186292617124e-07,
+      "loss": 0.3458,
+      "step": 80886
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.601953363351153e-07,
+      "loss": 0.4233,
+      "step": 80887
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6017204382182183e-07,
+      "loss": 0.4023,
+      "step": 80888
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6014875172184976e-07,
+      "loss": 0.5137,
+      "step": 80889
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6012546003521733e-07,
+      "loss": 0.4648,
+      "step": 80890
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6010216876194163e-07,
+      "loss": 0.519,
+      "step": 80891
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6007887790204094e-07,
+      "loss": 0.4155,
+      "step": 80892
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.60055587455533e-07,
+      "loss": 0.3979,
+      "step": 80893
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6003229742243597e-07,
+      "loss": 0.5044,
+      "step": 80894
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.6000900780276706e-07,
+      "loss": 0.3779,
+      "step": 80895
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.599857185965448e-07,
+      "loss": 0.4521,
+      "step": 80896
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.599624298037863e-07,
+      "loss": 0.5024,
+      "step": 80897
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5993914142450976e-07,
+      "loss": 0.3845,
+      "step": 80898
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.599158534587333e-07,
+      "loss": 0.4038,
+      "step": 80899
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.598925659064741e-07,
+      "loss": 0.4438,
+      "step": 80900
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5986927876775027e-07,
+      "loss": 0.4097,
+      "step": 80901
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5984599204257966e-07,
+      "loss": 0.3225,
+      "step": 80902
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.598227057309806e-07,
+      "loss": 0.5103,
+      "step": 80903
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.597994198329701e-07,
+      "loss": 0.5234,
+      "step": 80904
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5977613434856635e-07,
+      "loss": 0.3235,
+      "step": 80905
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5975284927778746e-07,
+      "loss": 0.5059,
+      "step": 80906
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5972956462065096e-07,
+      "loss": 0.5474,
+      "step": 80907
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.597062803771744e-07,
+      "loss": 0.4023,
+      "step": 80908
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5968299654737584e-07,
+      "loss": 0.2737,
+      "step": 80909
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5965971313127326e-07,
+      "loss": 0.5342,
+      "step": 80910
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5963643012888466e-07,
+      "loss": 0.3239,
+      "step": 80911
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.596131475402273e-07,
+      "loss": 0.2686,
+      "step": 80912
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.595898653653193e-07,
+      "loss": 0.3743,
+      "step": 80913
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5956658360417846e-07,
+      "loss": 0.3608,
+      "step": 80914
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5954330225682294e-07,
+      "loss": 0.5435,
+      "step": 80915
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.595200213232704e-07,
+      "loss": 0.4297,
+      "step": 80916
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5949674080353806e-07,
+      "loss": 0.438,
+      "step": 80917
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5947346069764426e-07,
+      "loss": 0.5293,
+      "step": 80918
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5945018100560675e-07,
+      "loss": 0.459,
+      "step": 80919
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.594269017274437e-07,
+      "loss": 0.3491,
+      "step": 80920
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5940362286317237e-07,
+      "loss": 0.3215,
+      "step": 80921
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.593803444128107e-07,
+      "loss": 0.4023,
+      "step": 80922
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.593570663763772e-07,
+      "loss": 0.459,
+      "step": 80923
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.593337887538886e-07,
+      "loss": 0.4312,
+      "step": 80924
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.593105115453633e-07,
+      "loss": 0.4277,
+      "step": 80925
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5928723475081943e-07,
+      "loss": 0.3357,
+      "step": 80926
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.592639583702741e-07,
+      "loss": 0.406,
+      "step": 80927
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.592406824037459e-07,
+      "loss": 0.5366,
+      "step": 80928
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.592174068512519e-07,
+      "loss": 0.4028,
+      "step": 80929
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.591941317128102e-07,
+      "loss": 0.5957,
+      "step": 80930
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.591708569884386e-07,
+      "loss": 0.3818,
+      "step": 80931
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.591475826781556e-07,
+      "loss": 0.4238,
+      "step": 80932
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.591243087819778e-07,
+      "loss": 0.49,
+      "step": 80933
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5910103529992386e-07,
+      "loss": 0.2736,
+      "step": 80934
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.590777622320113e-07,
+      "loss": 0.4438,
+      "step": 80935
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5905448957825844e-07,
+      "loss": 0.3677,
+      "step": 80936
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5903121733868255e-07,
+      "loss": 0.397,
+      "step": 80937
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5900794551330137e-07,
+      "loss": 0.4302,
+      "step": 80938
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.589846741021328e-07,
+      "loss": 0.4302,
+      "step": 80939
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.589614031051953e-07,
+      "loss": 0.384,
+      "step": 80940
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5893813252250567e-07,
+      "loss": 0.4033,
+      "step": 80941
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.589148623540824e-07,
+      "loss": 0.5205,
+      "step": 80942
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.58891592599943e-07,
+      "loss": 0.3425,
+      "step": 80943
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.588683232601058e-07,
+      "loss": 0.348,
+      "step": 80944
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.58845054334588e-07,
+      "loss": 0.373,
+      "step": 80945
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.588217858234079e-07,
+      "loss": 0.6011,
+      "step": 80946
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.587985177265827e-07,
+      "loss": 0.3679,
+      "step": 80947
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5877525004413064e-07,
+      "loss": 0.3629,
+      "step": 80948
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5875198277606984e-07,
+      "loss": 0.3107,
+      "step": 80949
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.587287159224175e-07,
+      "loss": 0.5347,
+      "step": 80950
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5870544948319144e-07,
+      "loss": 0.22,
+      "step": 80951
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.586821834584099e-07,
+      "loss": 0.4033,
+      "step": 80952
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5865891784809083e-07,
+      "loss": 0.3657,
+      "step": 80953
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5863565265225146e-07,
+      "loss": 0.3245,
+      "step": 80954
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5861238787090983e-07,
+      "loss": 0.4209,
+      "step": 80955
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.585891235040842e-07,
+      "loss": 0.3989,
+      "step": 80956
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5856585955179196e-07,
+      "loss": 0.415,
+      "step": 80957
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 4.5854259601405055e-07,
+      "loss": 0.4126,
+      "step": 80958
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.585193328908782e-07,
+      "loss": 0.3833,
+      "step": 80959
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5849607018229263e-07,
+      "loss": 0.4639,
+      "step": 80960
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.584728078883122e-07,
+      "loss": 0.5322,
+      "step": 80961
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.584495460089538e-07,
+      "loss": 0.4641,
+      "step": 80962
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.584262845442357e-07,
+      "loss": 0.4868,
+      "step": 80963
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.584030234941757e-07,
+      "loss": 0.4077,
+      "step": 80964
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.58379762858792e-07,
+      "loss": 0.4209,
+      "step": 80965
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.583565026381019e-07,
+      "loss": 0.4214,
+      "step": 80966
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5833324283212306e-07,
+      "loss": 0.502,
+      "step": 80967
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.583099834408736e-07,
+      "loss": 0.3513,
+      "step": 80968
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5828672446437113e-07,
+      "loss": 0.4109,
+      "step": 80969
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.58263465902634e-07,
+      "loss": 0.4414,
+      "step": 80970
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.582402077556793e-07,
+      "loss": 0.4753,
+      "step": 80971
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.582169500235252e-07,
+      "loss": 0.3408,
+      "step": 80972
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5819369270618937e-07,
+      "loss": 0.4417,
+      "step": 80973
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.581704358036902e-07,
+      "loss": 0.4163,
+      "step": 80974
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.581471793160445e-07,
+      "loss": 0.3525,
+      "step": 80975
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.581239232432711e-07,
+      "loss": 0.3153,
+      "step": 80976
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.581006675853869e-07,
+      "loss": 0.5103,
+      "step": 80977
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.580774123424104e-07,
+      "loss": 0.4155,
+      "step": 80978
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5805415751435885e-07,
+      "loss": 0.4902,
+      "step": 80979
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5803090310125027e-07,
+      "loss": 0.4292,
+      "step": 80980
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.580076491031024e-07,
+      "loss": 0.3821,
+      "step": 80981
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5798439551993364e-07,
+      "loss": 0.4482,
+      "step": 80982
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.579611423517609e-07,
+      "loss": 0.3353,
+      "step": 80983
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5793788959860237e-07,
+      "loss": 0.3965,
+      "step": 80984
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.579146372604763e-07,
+      "loss": 0.5205,
+      "step": 80985
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5789138533739956e-07,
+      "loss": 0.4192,
+      "step": 80986
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5786813382939096e-07,
+      "loss": 0.2108,
+      "step": 80987
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5784488273646736e-07,
+      "loss": 0.2762,
+      "step": 80988
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.578216320586471e-07,
+      "loss": 0.4482,
+      "step": 80989
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.577983817959479e-07,
+      "loss": 0.1984,
+      "step": 80990
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.577751319483878e-07,
+      "loss": 0.3457,
+      "step": 80991
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5775188251598397e-07,
+      "loss": 0.5044,
+      "step": 80992
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.577286334987547e-07,
+      "loss": 0.3745,
+      "step": 80993
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5770538489671797e-07,
+      "loss": 0.4868,
+      "step": 80994
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.576821367098913e-07,
+      "loss": 0.3101,
+      "step": 80995
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.576588889382922e-07,
+      "loss": 0.3082,
+      "step": 80996
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5763564158193866e-07,
+      "loss": 0.4463,
+      "step": 80997
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5761239464084854e-07,
+      "loss": 0.3945,
+      "step": 80998
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5758914811504014e-07,
+      "loss": 0.439,
+      "step": 80999
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5756590200453035e-07,
+      "loss": 0.4399,
+      "step": 81000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.575426563093374e-07,
+      "loss": 0.4028,
+      "step": 81001
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.575194110294791e-07,
+      "loss": 0.3462,
+      "step": 81002
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5749616616497367e-07,
+      "loss": 0.4961,
+      "step": 81003
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5747292171583807e-07,
+      "loss": 0.5342,
+      "step": 81004
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5744967768209097e-07,
+      "loss": 0.4072,
+      "step": 81005
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5742643406374915e-07,
+      "loss": 0.5142,
+      "step": 81006
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.574031908608311e-07,
+      "loss": 0.3945,
+      "step": 81007
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.573799480733548e-07,
+      "loss": 0.3188,
+      "step": 81008
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.573567057013373e-07,
+      "loss": 0.3711,
+      "step": 81009
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5733346374479697e-07,
+      "loss": 0.4585,
+      "step": 81010
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.573102222037517e-07,
+      "loss": 0.3738,
+      "step": 81011
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.572869810782187e-07,
+      "loss": 0.4846,
+      "step": 81012
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5726374036821615e-07,
+      "loss": 0.3333,
+      "step": 81013
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.572405000737617e-07,
+      "loss": 0.5039,
+      "step": 81014
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.572172601948737e-07,
+      "loss": 0.4023,
+      "step": 81015
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5719402073156944e-07,
+      "loss": 0.3481,
+      "step": 81016
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.571707816838665e-07,
+      "loss": 0.4653,
+      "step": 81017
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.571475430517828e-07,
+      "loss": 0.437,
+      "step": 81018
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5712430483533627e-07,
+      "loss": 0.3309,
+      "step": 81019
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.571010670345451e-07,
+      "loss": 0.356,
+      "step": 81020
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5707782964942633e-07,
+      "loss": 0.3893,
+      "step": 81021
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.57054592679998e-07,
+      "loss": 0.3689,
+      "step": 81022
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5703135612627817e-07,
+      "loss": 0.4429,
+      "step": 81023
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.570081199882847e-07,
+      "loss": 0.4614,
+      "step": 81024
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.569848842660352e-07,
+      "loss": 0.3315,
+      "step": 81025
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.56961648959547e-07,
+      "loss": 0.4229,
+      "step": 81026
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.569384140688384e-07,
+      "loss": 0.4829,
+      "step": 81027
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5691517959392735e-07,
+      "loss": 0.4022,
+      "step": 81028
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5689194553483103e-07,
+      "loss": 0.4285,
+      "step": 81029
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5686871189156763e-07,
+      "loss": 0.3672,
+      "step": 81030
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.568454786641548e-07,
+      "loss": 0.2462,
+      "step": 81031
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5682224585261085e-07,
+      "loss": 0.2977,
+      "step": 81032
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.567990134569527e-07,
+      "loss": 0.4434,
+      "step": 81033
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5677578147719866e-07,
+      "loss": 0.3784,
+      "step": 81034
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5675254991336663e-07,
+      "loss": 0.4878,
+      "step": 81035
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5672931876547395e-07,
+      "loss": 0.4844,
+      "step": 81036
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.56706088033539e-07,
+      "loss": 0.4702,
+      "step": 81037
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5668285771757886e-07,
+      "loss": 0.4404,
+      "step": 81038
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5665962781761167e-07,
+      "loss": 0.3389,
+      "step": 81039
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5663639833365517e-07,
+      "loss": 0.4595,
+      "step": 81040
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.566131692657276e-07,
+      "loss": 0.3373,
+      "step": 81041
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5658994061384603e-07,
+      "loss": 0.4893,
+      "step": 81042
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.565667123780285e-07,
+      "loss": 0.3265,
+      "step": 81043
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5654348455829283e-07,
+      "loss": 0.3408,
+      "step": 81044
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5652025715465757e-07,
+      "loss": 0.3152,
+      "step": 81045
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5649703016713905e-07,
+      "loss": 0.0738,
+      "step": 81046
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.564738035957557e-07,
+      "loss": 0.3961,
+      "step": 81047
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.564505774405254e-07,
+      "loss": 0.52,
+      "step": 81048
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.564273517014663e-07,
+      "loss": 0.3274,
+      "step": 81049
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.564041263785954e-07,
+      "loss": 0.5005,
+      "step": 81050
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5638090147193087e-07,
+      "loss": 0.4487,
+      "step": 81051
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.563576769814904e-07,
+      "loss": 0.3787,
+      "step": 81052
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.563344529072923e-07,
+      "loss": 0.4873,
+      "step": 81053
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5631122924935363e-07,
+      "loss": 0.3833,
+      "step": 81054
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.562880060076927e-07,
+      "loss": 0.3179,
+      "step": 81055
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.562647831823268e-07,
+      "loss": 0.4272,
+      "step": 81056
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.562415607732738e-07,
+      "loss": 0.4438,
+      "step": 81057
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5621833878055203e-07,
+      "loss": 0.4419,
+      "step": 81058
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.561951172041786e-07,
+      "loss": 0.4344,
+      "step": 81059
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.561718960441715e-07,
+      "loss": 0.3176,
+      "step": 81060
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5614867530054855e-07,
+      "loss": 0.3455,
+      "step": 81061
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.56125454973328e-07,
+      "loss": 0.4795,
+      "step": 81062
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.561022350625268e-07,
+      "loss": 0.266,
+      "step": 81063
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.560790155681631e-07,
+      "loss": 0.405,
+      "step": 81064
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5605579649025496e-07,
+      "loss": 0.4434,
+      "step": 81065
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5603257782882e-07,
+      "loss": 0.326,
+      "step": 81066
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5600935958387545e-07,
+      "loss": 0.2787,
+      "step": 81067
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.559861417554395e-07,
+      "loss": 0.3594,
+      "step": 81068
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5596292434353e-07,
+      "loss": 0.3599,
+      "step": 81069
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.55939707348165e-07,
+      "loss": 0.333,
+      "step": 81070
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5591649076936166e-07,
+      "loss": 0.3982,
+      "step": 81071
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.55893274607138e-07,
+      "loss": 0.3979,
+      "step": 81072
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.558700588615119e-07,
+      "loss": 0.4375,
+      "step": 81073
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.558468435325014e-07,
+      "loss": 0.4314,
+      "step": 81074
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.558236286201238e-07,
+      "loss": 0.4448,
+      "step": 81075
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5580041412439686e-07,
+      "loss": 0.4785,
+      "step": 81076
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.557772000453385e-07,
+      "loss": 0.3773,
+      "step": 81077
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5575398638296635e-07,
+      "loss": 0.4971,
+      "step": 81078
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.557307731372989e-07,
+      "loss": 0.4023,
+      "step": 81079
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.557075603083529e-07,
+      "loss": 0.5034,
+      "step": 81080
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.556843478961466e-07,
+      "loss": 0.4312,
+      "step": 81081
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5566113590069786e-07,
+      "loss": 0.3787,
+      "step": 81082
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.556379243220246e-07,
+      "loss": 0.4087,
+      "step": 81083
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.556147131601441e-07,
+      "loss": 0.3533,
+      "step": 81084
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.555915024150747e-07,
+      "loss": 0.3555,
+      "step": 81085
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5556829208683347e-07,
+      "loss": 0.3994,
+      "step": 81086
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5554508217543897e-07,
+      "loss": 0.4937,
+      "step": 81087
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5552187268090814e-07,
+      "loss": 0.366,
+      "step": 81088
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.554986636032592e-07,
+      "loss": 0.3989,
+      "step": 81089
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5547545494250994e-07,
+      "loss": 0.3418,
+      "step": 81090
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.554522466986784e-07,
+      "loss": 0.5239,
+      "step": 81091
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.554290388717816e-07,
+      "loss": 0.5151,
+      "step": 81092
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.554058314618379e-07,
+      "loss": 0.3779,
+      "step": 81093
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.553826244688649e-07,
+      "loss": 0.4111,
+      "step": 81094
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.553594178928807e-07,
+      "loss": 0.1779,
+      "step": 81095
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5533621173390266e-07,
+      "loss": 0.3248,
+      "step": 81096
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5531300599194833e-07,
+      "loss": 0.2855,
+      "step": 81097
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.552898006670358e-07,
+      "loss": 0.3406,
+      "step": 81098
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5526659575918293e-07,
+      "loss": 0.4355,
+      "step": 81099
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.552433912684076e-07,
+      "loss": 0.3799,
+      "step": 81100
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.552201871947271e-07,
+      "loss": 0.4272,
+      "step": 81101
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.551969835381594e-07,
+      "loss": 0.4829,
+      "step": 81102
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5517378029872265e-07,
+      "loss": 0.3234,
+      "step": 81103
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5515057747643393e-07,
+      "loss": 0.5107,
+      "step": 81104
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5512737507131174e-07,
+      "loss": 0.3079,
+      "step": 81105
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.551041730833731e-07,
+      "loss": 0.3009,
+      "step": 81106
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5508097151263614e-07,
+      "loss": 0.5435,
+      "step": 81107
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5505777035911895e-07,
+      "loss": 0.3958,
+      "step": 81108
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.550345696228387e-07,
+      "loss": 0.625,
+      "step": 81109
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5501136930381334e-07,
+      "loss": 0.3657,
+      "step": 81110
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5498816940206076e-07,
+      "loss": 0.5117,
+      "step": 81111
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.54964969917599e-07,
+      "loss": 0.4023,
+      "step": 81112
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5494177085044515e-07,
+      "loss": 0.5376,
+      "step": 81113
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.549185722006174e-07,
+      "loss": 0.3667,
+      "step": 81114
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.548953739681337e-07,
+      "loss": 0.4116,
+      "step": 81115
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.548721761530112e-07,
+      "loss": 0.3187,
+      "step": 81116
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.548489787552685e-07,
+      "loss": 0.3555,
+      "step": 81117
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.548257817749224e-07,
+      "loss": 0.3943,
+      "step": 81118
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.548025852119911e-07,
+      "loss": 0.3984,
+      "step": 81119
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5477938906649283e-07,
+      "loss": 0.239,
+      "step": 81120
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.547561933384444e-07,
+      "loss": 0.3975,
+      "step": 81121
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.547329980278642e-07,
+      "loss": 0.2649,
+      "step": 81122
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5470980313476993e-07,
+      "loss": 0.4893,
+      "step": 81123
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.546866086591795e-07,
+      "loss": 0.3909,
+      "step": 81124
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5466341460111056e-07,
+      "loss": 0.3452,
+      "step": 81125
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.546402209605804e-07,
+      "loss": 0.3313,
+      "step": 81126
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5461702773760714e-07,
+      "loss": 0.5122,
+      "step": 81127
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.545938349322085e-07,
+      "loss": 0.4199,
+      "step": 81128
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.545706425444027e-07,
+      "loss": 0.3809,
+      "step": 81129
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.545474505742067e-07,
+      "loss": 0.3672,
+      "step": 81130
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.545242590216386e-07,
+      "loss": 0.4617,
+      "step": 81131
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.545010678867163e-07,
+      "loss": 0.4307,
+      "step": 81132
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5447787716945773e-07,
+      "loss": 0.4348,
+      "step": 81133
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5445468686987997e-07,
+      "loss": 0.4262,
+      "step": 81134
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.544314969880017e-07,
+      "loss": 0.479,
+      "step": 81135
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5440830752383964e-07,
+      "loss": 0.3804,
+      "step": 81136
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5438511847741234e-07,
+      "loss": 0.2686,
+      "step": 81137
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5436192984873703e-07,
+      "loss": 0.324,
+      "step": 81138
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.543387416378317e-07,
+      "loss": 0.4751,
+      "step": 81139
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.543155538447141e-07,
+      "loss": 0.3643,
+      "step": 81140
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.542923664694023e-07,
+      "loss": 0.3207,
+      "step": 81141
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5426917951191335e-07,
+      "loss": 0.4902,
+      "step": 81142
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.542459929722655e-07,
+      "loss": 0.3528,
+      "step": 81143
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.542228068504763e-07,
+      "loss": 0.3887,
+      "step": 81144
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5419962114656405e-07,
+      "loss": 0.4521,
+      "step": 81145
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.541764358605461e-07,
+      "loss": 0.2946,
+      "step": 81146
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.541532509924396e-07,
+      "loss": 0.448,
+      "step": 81147
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5413006654226306e-07,
+      "loss": 0.3921,
+      "step": 81148
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5410688251003395e-07,
+      "loss": 0.5112,
+      "step": 81149
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5408369889577045e-07,
+      "loss": 0.4775,
+      "step": 81150
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5406051569948954e-07,
+      "loss": 0.4805,
+      "step": 81151
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.540373329212095e-07,
+      "loss": 0.3934,
+      "step": 81152
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5401415056094793e-07,
+      "loss": 0.4731,
+      "step": 81153
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5399096861872286e-07,
+      "loss": 0.3423,
+      "step": 81154
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5396778709455196e-07,
+      "loss": 0.4316,
+      "step": 81155
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5394460598845243e-07,
+      "loss": 0.4648,
+      "step": 81156
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5392142530044244e-07,
+      "loss": 0.397,
+      "step": 81157
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5389824503054e-07,
+      "loss": 0.385,
+      "step": 81158
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5387506517876215e-07,
+      "loss": 0.3943,
+      "step": 81159
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.538518857451271e-07,
+      "loss": 0.4219,
+      "step": 81160
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5382870672965255e-07,
+      "loss": 0.3501,
+      "step": 81161
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.538055281323566e-07,
+      "loss": 0.4087,
+      "step": 81162
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5378234995325636e-07,
+      "loss": 0.4717,
+      "step": 81163
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5375917219236983e-07,
+      "loss": 0.4517,
+      "step": 81164
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5373599484971514e-07,
+      "loss": 0.3096,
+      "step": 81165
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.537128179253092e-07,
+      "loss": 0.3406,
+      "step": 81166
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.536896414191708e-07,
+      "loss": 0.4258,
+      "step": 81167
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5366646533131657e-07,
+      "loss": 0.3049,
+      "step": 81168
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.536432896617649e-07,
+      "loss": 0.2534,
+      "step": 81169
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5362011441053336e-07,
+      "loss": 0.3369,
+      "step": 81170
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.535969395776403e-07,
+      "loss": 0.4565,
+      "step": 81171
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.535737651631024e-07,
+      "loss": 0.4717,
+      "step": 81172
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.535505911669381e-07,
+      "loss": 0.4238,
+      "step": 81173
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.535274175891652e-07,
+      "loss": 0.3491,
+      "step": 81174
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5350424442980137e-07,
+      "loss": 0.4429,
+      "step": 81175
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5348107168886387e-07,
+      "loss": 0.5044,
+      "step": 81176
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.534578993663706e-07,
+      "loss": 0.3821,
+      "step": 81177
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.534347274623396e-07,
+      "loss": 0.3945,
+      "step": 81178
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.534115559767888e-07,
+      "loss": 0.2937,
+      "step": 81179
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5338838490973543e-07,
+      "loss": 0.4209,
+      "step": 81180
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.533652142611972e-07,
+      "loss": 0.3181,
+      "step": 81181
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5334204403119226e-07,
+      "loss": 0.3687,
+      "step": 81182
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.533188742197386e-07,
+      "loss": 0.3997,
+      "step": 81183
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5329570482685297e-07,
+      "loss": 0.4431,
+      "step": 81184
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.532725358525542e-07,
+      "loss": 0.4824,
+      "step": 81185
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5324936729685915e-07,
+      "loss": 0.3035,
+      "step": 81186
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.532261991597858e-07,
+      "loss": 0.3311,
+      "step": 81187
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5320303144135244e-07,
+      "loss": 0.4629,
+      "step": 81188
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5317986414157607e-07,
+      "loss": 0.4756,
+      "step": 81189
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.531566972604747e-07,
+      "loss": 0.4893,
+      "step": 81190
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5313353079806637e-07,
+      "loss": 0.4956,
+      "step": 81191
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5311036475436824e-07,
+      "loss": 0.3508,
+      "step": 81192
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.530871991293984e-07,
+      "loss": 0.4722,
+      "step": 81193
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.530640339231745e-07,
+      "loss": 0.5303,
+      "step": 81194
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.530408691357147e-07,
+      "loss": 0.3403,
+      "step": 81195
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.530177047670364e-07,
+      "loss": 0.4844,
+      "step": 81196
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.529945408171569e-07,
+      "loss": 0.3848,
+      "step": 81197
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5297137728609436e-07,
+      "loss": 0.4211,
+      "step": 81198
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.529482141738664e-07,
+      "loss": 0.3257,
+      "step": 81199
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.529250514804912e-07,
+      "loss": 0.2804,
+      "step": 81200
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.529018892059858e-07,
+      "loss": 0.4702,
+      "step": 81201
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5287872735036835e-07,
+      "loss": 0.4648,
+      "step": 81202
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.528555659136565e-07,
+      "loss": 0.3356,
+      "step": 81203
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5283240489586826e-07,
+      "loss": 0.3291,
+      "step": 81204
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5280924429702105e-07,
+      "loss": 0.4268,
+      "step": 81205
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.527860841171324e-07,
+      "loss": 0.3575,
+      "step": 81206
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.527629243562202e-07,
+      "loss": 0.4902,
+      "step": 81207
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.527397650143027e-07,
+      "loss": 0.4727,
+      "step": 81208
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.527166060913967e-07,
+      "loss": 0.4932,
+      "step": 81209
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.526934475875205e-07,
+      "loss": 0.3394,
+      "step": 81210
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.526702895026918e-07,
+      "loss": 0.4614,
+      "step": 81211
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.526471318369286e-07,
+      "loss": 0.4092,
+      "step": 81212
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5262397459024803e-07,
+      "loss": 0.4761,
+      "step": 81213
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.526008177626681e-07,
+      "loss": 0.4473,
+      "step": 81214
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.525776613542068e-07,
+      "loss": 0.5039,
+      "step": 81215
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5255450536488135e-07,
+      "loss": 0.3896,
+      "step": 81216
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.525313497947102e-07,
+      "loss": 0.3064,
+      "step": 81217
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5250819464371013e-07,
+      "loss": 0.3455,
+      "step": 81218
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5248503991189937e-07,
+      "loss": 0.3516,
+      "step": 81219
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5246188559929563e-07,
+      "loss": 0.324,
+      "step": 81220
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5243873170591704e-07,
+      "loss": 0.3624,
+      "step": 81221
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.524155782317807e-07,
+      "loss": 0.3833,
+      "step": 81222
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5239242517690436e-07,
+      "loss": 0.3523,
+      "step": 81223
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5236927254130607e-07,
+      "loss": 0.5186,
+      "step": 81224
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5234612032500416e-07,
+      "loss": 0.4124,
+      "step": 81225
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5232296852801486e-07,
+      "loss": 0.3147,
+      "step": 81226
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5229981715035684e-07,
+      "loss": 0.4497,
+      "step": 81227
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.522766661920475e-07,
+      "loss": 0.2792,
+      "step": 81228
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.522535156531051e-07,
+      "loss": 0.2914,
+      "step": 81229
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5223036553354666e-07,
+      "loss": 0.4146,
+      "step": 81230
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5220721583339027e-07,
+      "loss": 0.5195,
+      "step": 81231
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5218406655265363e-07,
+      "loss": 0.3254,
+      "step": 81232
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.521609176913549e-07,
+      "loss": 0.3103,
+      "step": 81233
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5213776924951084e-07,
+      "loss": 0.3335,
+      "step": 81234
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5211462122714017e-07,
+      "loss": 0.4219,
+      "step": 81235
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.520914736242597e-07,
+      "loss": 0.406,
+      "step": 81236
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5206832644088765e-07,
+      "loss": 0.4038,
+      "step": 81237
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.52045179677042e-07,
+      "loss": 0.4238,
+      "step": 81238
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.520220333327399e-07,
+      "loss": 0.4966,
+      "step": 81239
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.519988874079993e-07,
+      "loss": 0.5918,
+      "step": 81240
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.519757419028379e-07,
+      "loss": 0.3304,
+      "step": 81241
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5195259681727394e-07,
+      "loss": 0.4609,
+      "step": 81242
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5192945215132415e-07,
+      "loss": 0.5195,
+      "step": 81243
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.519063079050069e-07,
+      "loss": 0.5107,
+      "step": 81244
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5188316407834015e-07,
+      "loss": 0.2512,
+      "step": 81245
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.518600206713412e-07,
+      "loss": 0.4526,
+      "step": 81246
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5183687768402744e-07,
+      "loss": 0.3308,
+      "step": 81247
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5181373511641707e-07,
+      "loss": 0.5073,
+      "step": 81248
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5179059296852773e-07,
+      "loss": 0.4165,
+      "step": 81249
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.517674512403774e-07,
+      "loss": 0.4009,
+      "step": 81250
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.517443099319832e-07,
+      "loss": 0.5098,
+      "step": 81251
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.517211690433632e-07,
+      "loss": 0.3621,
+      "step": 81252
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.51698028574535e-07,
+      "loss": 0.5952,
+      "step": 81253
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5167488852551683e-07,
+      "loss": 0.5713,
+      "step": 81254
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5165174889632607e-07,
+      "loss": 0.479,
+      "step": 81255
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.516286096869798e-07,
+      "loss": 0.3732,
+      "step": 81256
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5160547089749645e-07,
+      "loss": 0.4507,
+      "step": 81257
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5158233252789345e-07,
+      "loss": 0.4346,
+      "step": 81258
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.515591945781891e-07,
+      "loss": 0.4666,
+      "step": 81259
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.515360570484003e-07,
+      "loss": 0.5093,
+      "step": 81260
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.515129199385451e-07,
+      "loss": 0.4619,
+      "step": 81261
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5148978324864116e-07,
+      "loss": 0.4329,
+      "step": 81262
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.514666469787067e-07,
+      "loss": 0.4908,
+      "step": 81263
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5144351112875864e-07,
+      "loss": 0.3843,
+      "step": 81264
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.514203756988155e-07,
+      "loss": 0.5415,
+      "step": 81265
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5139724068889404e-07,
+      "loss": 0.4146,
+      "step": 81266
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5137410609901303e-07,
+      "loss": 0.3284,
+      "step": 81267
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5135097192918916e-07,
+      "loss": 0.3691,
+      "step": 81268
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5132783817944065e-07,
+      "loss": 0.4521,
+      "step": 81269
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.513047048497851e-07,
+      "loss": 0.3281,
+      "step": 81270
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5128157194024075e-07,
+      "loss": 0.2812,
+      "step": 81271
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5125843945082453e-07,
+      "loss": 0.5186,
+      "step": 81272
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5123530738155444e-07,
+      "loss": 0.4077,
+      "step": 81273
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5121217573244864e-07,
+      "loss": 0.4966,
+      "step": 81274
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.511890445035241e-07,
+      "loss": 0.3767,
+      "step": 81275
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.511659136947992e-07,
+      "loss": 0.4132,
+      "step": 81276
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5114278330629096e-07,
+      "loss": 0.3499,
+      "step": 81277
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.511196533380175e-07,
+      "loss": 0.4106,
+      "step": 81278
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.510965237899964e-07,
+      "loss": 0.4956,
+      "step": 81279
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5107339466224583e-07,
+      "loss": 0.4712,
+      "step": 81280
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.510502659547828e-07,
+      "loss": 0.3245,
+      "step": 81281
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5102713766762536e-07,
+      "loss": 0.3774,
+      "step": 81282
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5100400980079155e-07,
+      "loss": 0.3973,
+      "step": 81283
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5098088235429864e-07,
+      "loss": 0.3872,
+      "step": 81284
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5095775532816414e-07,
+      "loss": 0.4033,
+      "step": 81285
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5093462872240606e-07,
+      "loss": 0.3848,
+      "step": 81286
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.50911502537042e-07,
+      "loss": 0.4771,
+      "step": 81287
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5088837677209025e-07,
+      "loss": 0.2858,
+      "step": 81288
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.508652514275675e-07,
+      "loss": 0.4814,
+      "step": 81289
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5084212650349207e-07,
+      "loss": 0.3118,
+      "step": 81290
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5081900199988156e-07,
+      "loss": 0.4006,
+      "step": 81291
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.507958779167541e-07,
+      "loss": 0.4111,
+      "step": 81292
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5077275425412655e-07,
+      "loss": 0.3433,
+      "step": 81293
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5074963101201746e-07,
+      "loss": 0.4546,
+      "step": 81294
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5072650819044366e-07,
+      "loss": 0.4351,
+      "step": 81295
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5070338578942346e-07,
+      "loss": 0.4175,
+      "step": 81296
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.506802638089747e-07,
+      "loss": 0.3734,
+      "step": 81297
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5065714224911445e-07,
+      "loss": 0.3735,
+      "step": 81298
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.506340211098607e-07,
+      "loss": 0.4634,
+      "step": 81299
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.506109003912317e-07,
+      "loss": 0.4233,
+      "step": 81300
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5058778009324425e-07,
+      "loss": 0.4331,
+      "step": 81301
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5056466021591653e-07,
+      "loss": 0.4702,
+      "step": 81302
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.505415407592661e-07,
+      "loss": 0.502,
+      "step": 81303
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.505184217233111e-07,
+      "loss": 0.3389,
+      "step": 81304
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5049530310806904e-07,
+      "loss": 0.3994,
+      "step": 81305
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.50472184913557e-07,
+      "loss": 0.4292,
+      "step": 81306
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.50449067139793e-07,
+      "loss": 0.439,
+      "step": 81307
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5042594978679505e-07,
+      "loss": 0.5088,
+      "step": 81308
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5040283285458103e-07,
+      "loss": 0.4531,
+      "step": 81309
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5037971634316784e-07,
+      "loss": 0.3384,
+      "step": 81310
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.503566002525737e-07,
+      "loss": 0.4922,
+      "step": 81311
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.503334845828163e-07,
+      "loss": 0.3564,
+      "step": 81312
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5031036933391355e-07,
+      "loss": 0.3745,
+      "step": 81313
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.502872545058829e-07,
+      "loss": 0.2004,
+      "step": 81314
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.502641400987416e-07,
+      "loss": 0.4004,
+      "step": 81315
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.502410261125078e-07,
+      "loss": 0.3385,
+      "step": 81316
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5021791254719955e-07,
+      "loss": 0.4238,
+      "step": 81317
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5019479940283376e-07,
+      "loss": 0.5112,
+      "step": 81318
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.501716866794286e-07,
+      "loss": 0.3787,
+      "step": 81319
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.501485743770017e-07,
+      "loss": 0.3015,
+      "step": 81320
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5012546249557105e-07,
+      "loss": 0.5796,
+      "step": 81321
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.501023510351537e-07,
+      "loss": 0.3751,
+      "step": 81322
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.500792399957678e-07,
+      "loss": 0.4004,
+      "step": 81323
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.500561293774312e-07,
+      "loss": 0.4448,
+      "step": 81324
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.50033019180161e-07,
+      "loss": 0.4272,
+      "step": 81325
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.5000990940397556e-07,
+      "loss": 0.4453,
+      "step": 81326
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.499868000488919e-07,
+      "loss": 0.3309,
+      "step": 81327
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4996369111492806e-07,
+      "loss": 0.4214,
+      "step": 81328
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.499405826021018e-07,
+      "loss": 0.4751,
+      "step": 81329
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4991747451043094e-07,
+      "loss": 0.3054,
+      "step": 81330
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4989436683993277e-07,
+      "loss": 0.4292,
+      "step": 81331
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.498712595906251e-07,
+      "loss": 0.4438,
+      "step": 81332
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4984815276252573e-07,
+      "loss": 0.4448,
+      "step": 81333
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.498250463556531e-07,
+      "loss": 0.3936,
+      "step": 81334
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.498019403700234e-07,
+      "loss": 0.4211,
+      "step": 81335
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.49778834805655e-07,
+      "loss": 0.522,
+      "step": 81336
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.497557296625657e-07,
+      "loss": 0.5591,
+      "step": 81337
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.497326249407735e-07,
+      "loss": 0.3539,
+      "step": 81338
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.497095206402953e-07,
+      "loss": 0.4358,
+      "step": 81339
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.496864167611493e-07,
+      "loss": 0.4102,
+      "step": 81340
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4966331330335306e-07,
+      "loss": 0.3267,
+      "step": 81341
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4964021026692466e-07,
+      "loss": 0.3748,
+      "step": 81342
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.496171076518811e-07,
+      "loss": 0.3785,
+      "step": 81343
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4959400545824076e-07,
+      "loss": 0.3853,
+      "step": 81344
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.495709036860207e-07,
+      "loss": 0.353,
+      "step": 81345
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.495478023352388e-07,
+      "loss": 0.395,
+      "step": 81346
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.4952470140591325e-07,
+      "loss": 0.3508,
+      "step": 81347
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 4.495016008980609e-07,
+      "loss": 0.4526,
+      "step": 81348
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.494785008116999e-07,
+      "loss": 0.3813,
+      "step": 81349
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.49455401146848e-07,
+      "loss": 0.5049,
+      "step": 81350
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.49432301903523e-07,
+      "loss": 0.5049,
+      "step": 81351
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4940920308174203e-07,
+      "loss": 0.3628,
+      "step": 81352
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4938610468152316e-07,
+      "loss": 0.4402,
+      "step": 81353
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.493630067028844e-07,
+      "loss": 0.4004,
+      "step": 81354
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.493399091458431e-07,
+      "loss": 0.4429,
+      "step": 81355
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4931681201041647e-07,
+      "loss": 0.4037,
+      "step": 81356
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.492937152966226e-07,
+      "loss": 0.3988,
+      "step": 81357
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.492706190044793e-07,
+      "loss": 0.2332,
+      "step": 81358
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4924752313400447e-07,
+      "loss": 0.3516,
+      "step": 81359
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.492244276852152e-07,
+      "loss": 0.4272,
+      "step": 81360
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4920133265812945e-07,
+      "loss": 0.4526,
+      "step": 81361
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4917823805276477e-07,
+      "loss": 0.4385,
+      "step": 81362
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4915514386913954e-07,
+      "loss": 0.4556,
+      "step": 81363
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.491320501072707e-07,
+      "loss": 0.4717,
+      "step": 81364
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4910895676717595e-07,
+      "loss": 0.4443,
+      "step": 81365
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4908586384887305e-07,
+      "loss": 0.4048,
+      "step": 81366
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.490627713523798e-07,
+      "loss": 0.311,
+      "step": 81367
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4903967927771415e-07,
+      "loss": 0.2778,
+      "step": 81368
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4901658762489314e-07,
+      "loss": 0.5142,
+      "step": 81369
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.489934963939348e-07,
+      "loss": 0.2972,
+      "step": 81370
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.489704055848572e-07,
+      "loss": 0.401,
+      "step": 81371
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.489473151976773e-07,
+      "loss": 0.4246,
+      "step": 81372
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4892422523241303e-07,
+      "loss": 0.396,
+      "step": 81373
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.489011356890825e-07,
+      "loss": 0.4141,
+      "step": 81374
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.488780465677027e-07,
+      "loss": 0.3374,
+      "step": 81375
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4885495786829197e-07,
+      "loss": 0.481,
+      "step": 81376
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.488318695908673e-07,
+      "loss": 0.353,
+      "step": 81377
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4880878173544676e-07,
+      "loss": 0.4038,
+      "step": 81378
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4878569430204796e-07,
+      "loss": 0.3562,
+      "step": 81379
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4876260729068893e-07,
+      "loss": 0.4688,
+      "step": 81380
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.487395207013868e-07,
+      "loss": 0.3467,
+      "step": 81381
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.487164345341593e-07,
+      "loss": 0.4409,
+      "step": 81382
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.486933487890244e-07,
+      "loss": 0.4873,
+      "step": 81383
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.48670263466e-07,
+      "loss": 0.436,
+      "step": 81384
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4864717856510334e-07,
+      "loss": 0.4414,
+      "step": 81385
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.486240940863519e-07,
+      "loss": 0.3304,
+      "step": 81386
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4860101002976356e-07,
+      "loss": 0.4492,
+      "step": 81387
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4857792639535666e-07,
+      "loss": 0.4243,
+      "step": 81388
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.485548431831476e-07,
+      "loss": 0.3771,
+      "step": 81389
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.48531760393155e-07,
+      "loss": 0.4844,
+      "step": 81390
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4850867802539617e-07,
+      "loss": 0.387,
+      "step": 81391
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.484855960798892e-07,
+      "loss": 0.2753,
+      "step": 81392
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.484625145566512e-07,
+      "loss": 0.5161,
+      "step": 81393
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4843943345570036e-07,
+      "loss": 0.3867,
+      "step": 81394
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4841635277705383e-07,
+      "loss": 0.3936,
+      "step": 81395
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4839327252072946e-07,
+      "loss": 0.4097,
+      "step": 81396
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4837019268674535e-07,
+      "loss": 0.3335,
+      "step": 81397
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4834711327511844e-07,
+      "loss": 0.3999,
+      "step": 81398
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4832403428586685e-07,
+      "loss": 0.3813,
+      "step": 81399
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.483009557190081e-07,
+      "loss": 0.4482,
+      "step": 81400
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4827787757456036e-07,
+      "loss": 0.4199,
+      "step": 81401
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4825479985254046e-07,
+      "loss": 0.3645,
+      "step": 81402
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.482317225529666e-07,
+      "loss": 0.4204,
+      "step": 81403
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.482086456758566e-07,
+      "loss": 0.3928,
+      "step": 81404
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4818556922122755e-07,
+      "loss": 0.3325,
+      "step": 81405
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4816249318909783e-07,
+      "loss": 0.5127,
+      "step": 81406
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4813941757948424e-07,
+      "loss": 0.3816,
+      "step": 81407
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4811634239240504e-07,
+      "loss": 0.4565,
+      "step": 81408
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.48093267627878e-07,
+      "loss": 0.3542,
+      "step": 81409
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4807019328592034e-07,
+      "loss": 0.4326,
+      "step": 81410
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4804711936654993e-07,
+      "loss": 0.4546,
+      "step": 81411
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.480240458697845e-07,
+      "loss": 0.4609,
+      "step": 81412
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.48000972795642e-07,
+      "loss": 0.4648,
+      "step": 81413
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4797790014413973e-07,
+      "loss": 0.4761,
+      "step": 81414
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4795482791529504e-07,
+      "loss": 0.3789,
+      "step": 81415
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4793175610912604e-07,
+      "loss": 0.3513,
+      "step": 81416
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4790868472565025e-07,
+      "loss": 0.5059,
+      "step": 81417
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.478856137648858e-07,
+      "loss": 0.4751,
+      "step": 81418
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4786254322684945e-07,
+      "loss": 0.3391,
+      "step": 81419
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4783947311155944e-07,
+      "loss": 0.4697,
+      "step": 81420
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4781640341903336e-07,
+      "loss": 0.3735,
+      "step": 81421
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4779333414928933e-07,
+      "loss": 0.3477,
+      "step": 81422
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4777026530234407e-07,
+      "loss": 0.4225,
+      "step": 81423
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4774719687821616e-07,
+      "loss": 0.3307,
+      "step": 81424
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4772412887692234e-07,
+      "loss": 0.4775,
+      "step": 81425
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.477010612984813e-07,
+      "loss": 0.3936,
+      "step": 81426
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.476779941429096e-07,
+      "loss": 0.29,
+      "step": 81427
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4765492741022573e-07,
+      "loss": 0.2347,
+      "step": 81428
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.476318611004468e-07,
+      "loss": 0.501,
+      "step": 81429
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.476087952135913e-07,
+      "loss": 0.4277,
+      "step": 81430
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.475857297496759e-07,
+      "loss": 0.6162,
+      "step": 81431
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.475626647087187e-07,
+      "loss": 0.3004,
+      "step": 81432
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.475396000907374e-07,
+      "loss": 0.5615,
+      "step": 81433
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.475165358957501e-07,
+      "loss": 0.4761,
+      "step": 81434
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4749347212377384e-07,
+      "loss": 0.322,
+      "step": 81435
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.474704087748261e-07,
+      "loss": 0.3264,
+      "step": 81436
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4744734584892485e-07,
+      "loss": 0.3396,
+      "step": 81437
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4742428334608786e-07,
+      "loss": 0.4512,
+      "step": 81438
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4740122126633287e-07,
+      "loss": 0.4014,
+      "step": 81439
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.473781596096771e-07,
+      "loss": 0.3647,
+      "step": 81440
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.473550983761385e-07,
+      "loss": 0.313,
+      "step": 81441
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.473320375657346e-07,
+      "loss": 0.3831,
+      "step": 81442
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.473089771784835e-07,
+      "loss": 0.4565,
+      "step": 81443
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4728591721440256e-07,
+      "loss": 0.3329,
+      "step": 81444
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.47262857673509e-07,
+      "loss": 0.3203,
+      "step": 81445
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4723979855582083e-07,
+      "loss": 0.3018,
+      "step": 81446
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.472167398613562e-07,
+      "loss": 0.3965,
+      "step": 81447
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.471936815901318e-07,
+      "loss": 0.4023,
+      "step": 81448
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4717062374216586e-07,
+      "loss": 0.3212,
+      "step": 81449
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.47147566317476e-07,
+      "loss": 0.2191,
+      "step": 81450
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4712450931608e-07,
+      "loss": 0.3862,
+      "step": 81451
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4710145273799514e-07,
+      "loss": 0.4253,
+      "step": 81452
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.470783965832392e-07,
+      "loss": 0.4429,
+      "step": 81453
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.470553408518303e-07,
+      "loss": 0.2822,
+      "step": 81454
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4703228554378534e-07,
+      "loss": 0.4717,
+      "step": 81455
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4700923065912275e-07,
+      "loss": 0.4077,
+      "step": 81456
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.469861761978593e-07,
+      "loss": 0.37,
+      "step": 81457
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.469631221600132e-07,
+      "loss": 0.3718,
+      "step": 81458
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4694006854560197e-07,
+      "loss": 0.4702,
+      "step": 81459
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4691701535464365e-07,
+      "loss": 0.4805,
+      "step": 81460
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4689396258715516e-07,
+      "loss": 0.4351,
+      "step": 81461
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4687091024315446e-07,
+      "loss": 0.397,
+      "step": 81462
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.468478583226598e-07,
+      "loss": 0.4722,
+      "step": 81463
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4682480682568824e-07,
+      "loss": 0.5327,
+      "step": 81464
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4680175575225707e-07,
+      "loss": 0.4136,
+      "step": 81465
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4677870510238447e-07,
+      "loss": 0.5015,
+      "step": 81466
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.467556548760879e-07,
+      "loss": 0.3149,
+      "step": 81467
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.467326050733854e-07,
+      "loss": 0.3428,
+      "step": 81468
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4670955569429404e-07,
+      "loss": 0.4639,
+      "step": 81469
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.466865067388317e-07,
+      "loss": 0.4609,
+      "step": 81470
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.46663458207016e-07,
+      "loss": 0.5894,
+      "step": 81471
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.466404100988651e-07,
+      "loss": 0.2927,
+      "step": 81472
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4661736241439575e-07,
+      "loss": 0.5791,
+      "step": 81473
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4659431515362643e-07,
+      "loss": 0.2277,
+      "step": 81474
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.465712683165741e-07,
+      "loss": 0.2753,
+      "step": 81475
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.465482219032566e-07,
+      "loss": 0.4761,
+      "step": 81476
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.465251759136921e-07,
+      "loss": 0.5137,
+      "step": 81477
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4650213034789743e-07,
+      "loss": 0.3904,
+      "step": 81478
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.464790852058907e-07,
+      "loss": 0.4658,
+      "step": 81479
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4645604048768983e-07,
+      "loss": 0.4531,
+      "step": 81480
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4643299619331163e-07,
+      "loss": 0.354,
+      "step": 81481
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.464099523227743e-07,
+      "loss": 0.4297,
+      "step": 81482
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.463869088760955e-07,
+      "loss": 0.3225,
+      "step": 81483
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4636386585329313e-07,
+      "loss": 0.3657,
+      "step": 81484
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4634082325438436e-07,
+      "loss": 0.1786,
+      "step": 81485
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4631778107938677e-07,
+      "loss": 0.4231,
+      "step": 81486
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4629473932831806e-07,
+      "loss": 0.5391,
+      "step": 81487
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.46271698001196e-07,
+      "loss": 0.4819,
+      "step": 81488
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4624865709803873e-07,
+      "loss": 0.3735,
+      "step": 81489
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4622561661886294e-07,
+      "loss": 0.334,
+      "step": 81490
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4620257656368675e-07,
+      "loss": 0.417,
+      "step": 81491
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4617953693252785e-07,
+      "loss": 0.3252,
+      "step": 81492
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.461564977254041e-07,
+      "loss": 0.3335,
+      "step": 81493
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4613345894233287e-07,
+      "loss": 0.4224,
+      "step": 81494
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.461104205833315e-07,
+      "loss": 0.4653,
+      "step": 81495
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4608738264841783e-07,
+      "loss": 0.4941,
+      "step": 81496
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.460643451376099e-07,
+      "loss": 0.467,
+      "step": 81497
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4604130805092475e-07,
+      "loss": 0.3901,
+      "step": 81498
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.460182713883802e-07,
+      "loss": 0.4175,
+      "step": 81499
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4599523514999415e-07,
+      "loss": 0.4634,
+      "step": 81500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4597219933578435e-07,
+      "loss": 0.4697,
+      "step": 81501
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4594916394576774e-07,
+      "loss": 0.436,
+      "step": 81502
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.459261289799625e-07,
+      "loss": 0.3926,
+      "step": 81503
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4590309443838634e-07,
+      "loss": 0.4253,
+      "step": 81504
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4588006032105653e-07,
+      "loss": 0.427,
+      "step": 81505
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.458570266279912e-07,
+      "loss": 0.2946,
+      "step": 81506
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4583399335920724e-07,
+      "loss": 0.3418,
+      "step": 81507
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.458109605147227e-07,
+      "loss": 0.318,
+      "step": 81508
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.457879280945552e-07,
+      "loss": 0.4639,
+      "step": 81509
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.457648960987228e-07,
+      "loss": 0.542,
+      "step": 81510
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.457418645272424e-07,
+      "loss": 0.4746,
+      "step": 81511
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.457188333801319e-07,
+      "loss": 0.4434,
+      "step": 81512
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4569580265740916e-07,
+      "loss": 0.5156,
+      "step": 81513
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.456727723590923e-07,
+      "loss": 0.4971,
+      "step": 81514
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.456497424851975e-07,
+      "loss": 0.4189,
+      "step": 81515
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.456267130357433e-07,
+      "loss": 0.4214,
+      "step": 81516
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.456036840107472e-07,
+      "loss": 0.3335,
+      "step": 81517
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4558065541022726e-07,
+      "loss": 0.4829,
+      "step": 81518
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.455576272342003e-07,
+      "loss": 0.4165,
+      "step": 81519
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4553459948268437e-07,
+      "loss": 0.2367,
+      "step": 81520
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.455115721556971e-07,
+      "loss": 0.5952,
+      "step": 81521
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.454885452532565e-07,
+      "loss": 0.4355,
+      "step": 81522
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4546551877537954e-07,
+      "loss": 0.2894,
+      "step": 81523
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4544249272208445e-07,
+      "loss": 0.2942,
+      "step": 81524
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.454194670933882e-07,
+      "loss": 0.5488,
+      "step": 81525
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.453964418893087e-07,
+      "loss": 0.5454,
+      "step": 81526
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4537341710986396e-07,
+      "loss": 0.3343,
+      "step": 81527
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4535039275507105e-07,
+      "loss": 0.3584,
+      "step": 81528
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.453273688249478e-07,
+      "loss": 0.3901,
+      "step": 81529
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.453043453195119e-07,
+      "loss": 0.4565,
+      "step": 81530
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4528132223878124e-07,
+      "loss": 0.3994,
+      "step": 81531
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4525829958277283e-07,
+      "loss": 0.353,
+      "step": 81532
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.452352773515047e-07,
+      "loss": 0.4692,
+      "step": 81533
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4521225554499474e-07,
+      "loss": 0.4091,
+      "step": 81534
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4518923416326026e-07,
+      "loss": 0.4106,
+      "step": 81535
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.451662132063184e-07,
+      "loss": 0.4314,
+      "step": 81536
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.451431926741874e-07,
+      "loss": 0.395,
+      "step": 81537
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4512017256688473e-07,
+      "loss": 0.239,
+      "step": 81538
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4509715288442826e-07,
+      "loss": 0.5,
+      "step": 81539
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.450741336268352e-07,
+      "loss": 0.4678,
+      "step": 81540
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.450511147941233e-07,
+      "loss": 0.417,
+      "step": 81541
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.450280963863102e-07,
+      "loss": 0.3613,
+      "step": 81542
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4500507840341397e-07,
+      "loss": 0.4517,
+      "step": 81543
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4498206084545176e-07,
+      "loss": 0.5303,
+      "step": 81544
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4495904371244095e-07,
+      "loss": 0.3457,
+      "step": 81545
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.449360270043995e-07,
+      "loss": 0.3483,
+      "step": 81546
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.44913010721345e-07,
+      "loss": 0.5684,
+      "step": 81547
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.448899948632955e-07,
+      "loss": 0.249,
+      "step": 81548
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4486697943026774e-07,
+      "loss": 0.321,
+      "step": 81549
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.448439644222798e-07,
+      "loss": 0.3435,
+      "step": 81550
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.448209498393498e-07,
+      "loss": 0.5645,
+      "step": 81551
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4479793568149437e-07,
+      "loss": 0.5029,
+      "step": 81552
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.447749219487317e-07,
+      "loss": 0.4595,
+      "step": 81553
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.447519086410797e-07,
+      "loss": 0.3794,
+      "step": 81554
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4472889575855533e-07,
+      "loss": 0.5117,
+      "step": 81555
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4470588330117676e-07,
+      "loss": 0.3228,
+      "step": 81556
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.446828712689611e-07,
+      "loss": 0.4609,
+      "step": 81557
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.446598596619262e-07,
+      "loss": 0.3779,
+      "step": 81558
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4463684848008964e-07,
+      "loss": 0.4263,
+      "step": 81559
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.446138377234695e-07,
+      "loss": 0.3505,
+      "step": 81560
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4459082739208273e-07,
+      "loss": 0.4189,
+      "step": 81561
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.445678174859472e-07,
+      "loss": 0.4399,
+      "step": 81562
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.445448080050804e-07,
+      "loss": 0.4153,
+      "step": 81563
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.445217989495007e-07,
+      "loss": 0.3848,
+      "step": 81564
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.44498790319225e-07,
+      "loss": 0.3505,
+      "step": 81565
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4447578211427064e-07,
+      "loss": 0.3994,
+      "step": 81566
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4445277433465567e-07,
+      "loss": 0.4536,
+      "step": 81567
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4442976698039803e-07,
+      "loss": 0.3831,
+      "step": 81568
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.444067600515146e-07,
+      "loss": 0.4277,
+      "step": 81569
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.443837535480234e-07,
+      "loss": 0.4678,
+      "step": 81570
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4436074746994187e-07,
+      "loss": 0.3367,
+      "step": 81571
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4433774181728836e-07,
+      "loss": 0.3691,
+      "step": 81572
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.443147365900797e-07,
+      "loss": 0.4204,
+      "step": 81573
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4429173178833335e-07,
+      "loss": 0.4312,
+      "step": 81574
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.442687274120673e-07,
+      "loss": 0.3481,
+      "step": 81575
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.442457234612992e-07,
+      "loss": 0.3989,
+      "step": 81576
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.44222719936047e-07,
+      "loss": 0.4482,
+      "step": 81577
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.441997168363274e-07,
+      "loss": 0.543,
+      "step": 81578
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.441767141621585e-07,
+      "loss": 0.4438,
+      "step": 81579
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.441537119135581e-07,
+      "loss": 0.3489,
+      "step": 81580
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4413071009054393e-07,
+      "loss": 0.479,
+      "step": 81581
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4410770869313285e-07,
+      "loss": 0.3533,
+      "step": 81582
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4408470772134344e-07,
+      "loss": 0.4419,
+      "step": 81583
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.440617071751924e-07,
+      "loss": 0.3149,
+      "step": 81584
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.440387070546977e-07,
+      "loss": 0.4023,
+      "step": 81585
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.440157073598775e-07,
+      "loss": 0.415,
+      "step": 81586
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.439927080907484e-07,
+      "loss": 0.459,
+      "step": 81587
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4396970924732857e-07,
+      "loss": 0.3149,
+      "step": 81588
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.439467108296359e-07,
+      "loss": 0.478,
+      "step": 81589
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4392371283768735e-07,
+      "loss": 0.3521,
+      "step": 81590
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4390071527150074e-07,
+      "loss": 0.2952,
+      "step": 81591
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4387771813109387e-07,
+      "loss": 0.4702,
+      "step": 81592
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4385472141648464e-07,
+      "loss": 0.3232,
+      "step": 81593
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.438317251276903e-07,
+      "loss": 0.3862,
+      "step": 81594
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.43808729264728e-07,
+      "loss": 0.4575,
+      "step": 81595
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4378573382761573e-07,
+      "loss": 0.5381,
+      "step": 81596
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.437627388163713e-07,
+      "loss": 0.4204,
+      "step": 81597
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4373974423101246e-07,
+      "loss": 0.3433,
+      "step": 81598
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.437167500715562e-07,
+      "loss": 0.4316,
+      "step": 81599
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.436937563380203e-07,
+      "loss": 0.3466,
+      "step": 81600
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4367076303042263e-07,
+      "loss": 0.3502,
+      "step": 81601
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4364777014878106e-07,
+      "loss": 0.4746,
+      "step": 81602
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.436247776931128e-07,
+      "loss": 0.3845,
+      "step": 81603
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.43601785663435e-07,
+      "loss": 0.4873,
+      "step": 81604
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.435787940597658e-07,
+      "loss": 0.4648,
+      "step": 81605
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4355580288212314e-07,
+      "loss": 0.374,
+      "step": 81606
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4353281213052374e-07,
+      "loss": 0.3009,
+      "step": 81607
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4350982180498577e-07,
+      "loss": 0.4767,
+      "step": 81608
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.434868319055267e-07,
+      "loss": 0.5103,
+      "step": 81609
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.434638424321646e-07,
+      "loss": 0.4067,
+      "step": 81610
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.434408533849162e-07,
+      "loss": 0.4355,
+      "step": 81611
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.434178647637996e-07,
+      "loss": 0.5449,
+      "step": 81612
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4339487656883266e-07,
+      "loss": 0.2874,
+      "step": 81613
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.433718888000324e-07,
+      "loss": 0.4131,
+      "step": 81614
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.43348901457417e-07,
+      "loss": 0.4009,
+      "step": 81615
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.433259145410033e-07,
+      "loss": 0.5405,
+      "step": 81616
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4330292805080947e-07,
+      "loss": 0.4316,
+      "step": 81617
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.432799419868529e-07,
+      "loss": 0.4355,
+      "step": 81618
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4325695634915173e-07,
+      "loss": 0.4648,
+      "step": 81619
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.432339711377228e-07,
+      "loss": 0.5752,
+      "step": 81620
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.432109863525838e-07,
+      "loss": 0.3794,
+      "step": 81621
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.431880019937527e-07,
+      "loss": 0.4346,
+      "step": 81622
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4316501806124773e-07,
+      "loss": 0.4907,
+      "step": 81623
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4314203455508483e-07,
+      "loss": 0.3351,
+      "step": 81624
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.431190514752825e-07,
+      "loss": 0.4814,
+      "step": 81625
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4309606882185835e-07,
+      "loss": 0.356,
+      "step": 81626
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4307308659483035e-07,
+      "loss": 0.3756,
+      "step": 81627
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4305010479421525e-07,
+      "loss": 0.5332,
+      "step": 81628
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.43027123420031e-07,
+      "loss": 0.4424,
+      "step": 81629
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4300414247229533e-07,
+      "loss": 0.3865,
+      "step": 81630
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.429811619510262e-07,
+      "loss": 0.4453,
+      "step": 81631
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4295818185624036e-07,
+      "loss": 0.376,
+      "step": 81632
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.429352021879562e-07,
+      "loss": 0.4546,
+      "step": 81633
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.429122229461906e-07,
+      "loss": 0.4512,
+      "step": 81634
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4288924413096154e-07,
+      "loss": 0.4211,
+      "step": 81635
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.428662657422868e-07,
+      "loss": 0.3411,
+      "step": 81636
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4284328778018344e-07,
+      "loss": 0.4146,
+      "step": 81637
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.428203102446694e-07,
+      "loss": 0.4001,
+      "step": 81638
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4279733313576216e-07,
+      "loss": 0.4154,
+      "step": 81639
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.427743564534798e-07,
+      "loss": 0.4141,
+      "step": 81640
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.427513801978391e-07,
+      "loss": 0.4888,
+      "step": 81641
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4272840436885805e-07,
+      "loss": 0.4033,
+      "step": 81642
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.427054289665546e-07,
+      "loss": 0.4424,
+      "step": 81643
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4268245399094596e-07,
+      "loss": 0.438,
+      "step": 81644
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.426594794420494e-07,
+      "loss": 0.4009,
+      "step": 81645
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.426365053198828e-07,
+      "loss": 0.4019,
+      "step": 81646
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.426135316244639e-07,
+      "loss": 0.3704,
+      "step": 81647
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.425905583558105e-07,
+      "loss": 0.4829,
+      "step": 81648
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4256758551393946e-07,
+      "loss": 0.3979,
+      "step": 81649
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.425446130988689e-07,
+      "loss": 0.4585,
+      "step": 81650
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.425216411106162e-07,
+      "loss": 0.5098,
+      "step": 81651
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.424986695491995e-07,
+      "loss": 0.2915,
+      "step": 81652
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4247569841463585e-07,
+      "loss": 0.5264,
+      "step": 81653
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4245272770694263e-07,
+      "loss": 0.4023,
+      "step": 81654
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4242975742613766e-07,
+      "loss": 0.4292,
+      "step": 81655
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.424067875722386e-07,
+      "loss": 0.3555,
+      "step": 81656
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.423838181452634e-07,
+      "loss": 0.3916,
+      "step": 81657
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.423608491452289e-07,
+      "loss": 0.3564,
+      "step": 81658
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4233788057215304e-07,
+      "loss": 0.2751,
+      "step": 81659
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.423149124260538e-07,
+      "loss": 0.3892,
+      "step": 81660
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.422919447069481e-07,
+      "loss": 0.4268,
+      "step": 81661
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.422689774148537e-07,
+      "loss": 0.3158,
+      "step": 81662
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.422460105497887e-07,
+      "loss": 0.5557,
+      "step": 81663
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.422230441117699e-07,
+      "loss": 0.4746,
+      "step": 81664
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.422000781008157e-07,
+      "loss": 0.3403,
+      "step": 81665
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.421771125169429e-07,
+      "loss": 0.458,
+      "step": 81666
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.421541473601693e-07,
+      "loss": 0.5435,
+      "step": 81667
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4213118263051275e-07,
+      "loss": 0.4092,
+      "step": 81668
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4210821832799104e-07,
+      "loss": 0.4556,
+      "step": 81669
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.42085254452621e-07,
+      "loss": 0.4946,
+      "step": 81670
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.420622910044206e-07,
+      "loss": 0.3322,
+      "step": 81671
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.420393279834076e-07,
+      "loss": 0.3955,
+      "step": 81672
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.420163653895996e-07,
+      "loss": 0.4241,
+      "step": 81673
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4199340322301415e-07,
+      "loss": 0.4453,
+      "step": 81674
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4197044148366835e-07,
+      "loss": 0.3589,
+      "step": 81675
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4194748017158e-07,
+      "loss": 0.2864,
+      "step": 81676
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.419245192867672e-07,
+      "loss": 0.3843,
+      "step": 81677
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4190155882924684e-07,
+      "loss": 0.3652,
+      "step": 81678
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4187859879903676e-07,
+      "loss": 0.4438,
+      "step": 81679
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4185563919615454e-07,
+      "loss": 0.4458,
+      "step": 81680
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.418326800206182e-07,
+      "loss": 0.2975,
+      "step": 81681
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4180972127244455e-07,
+      "loss": 0.3638,
+      "step": 81682
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4178676295165183e-07,
+      "loss": 0.4922,
+      "step": 81683
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4176380505825707e-07,
+      "loss": 0.4072,
+      "step": 81684
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4174084759227803e-07,
+      "loss": 0.3613,
+      "step": 81685
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4171789055373274e-07,
+      "loss": 0.5557,
+      "step": 81686
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.41694933942638e-07,
+      "loss": 0.3262,
+      "step": 81687
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4167197775901175e-07,
+      "loss": 0.5879,
+      "step": 81688
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.416490220028717e-07,
+      "loss": 0.3721,
+      "step": 81689
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.416260666742355e-07,
+      "loss": 0.2849,
+      "step": 81690
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4160311177312037e-07,
+      "loss": 0.3721,
+      "step": 81691
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.41580157299544e-07,
+      "loss": 0.3833,
+      "step": 81692
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.415572032535243e-07,
+      "loss": 0.5684,
+      "step": 81693
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.415342496350787e-07,
+      "loss": 0.4937,
+      "step": 81694
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4151129644422424e-07,
+      "loss": 0.3428,
+      "step": 81695
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4148834368097885e-07,
+      "loss": 0.3657,
+      "step": 81696
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.414653913453602e-07,
+      "loss": 0.3301,
+      "step": 81697
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4144243943738624e-07,
+      "loss": 0.4194,
+      "step": 81698
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4141948795707363e-07,
+      "loss": 0.4756,
+      "step": 81699
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4139653690444045e-07,
+      "loss": 0.3967,
+      "step": 81700
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.413735862795044e-07,
+      "loss": 0.5781,
+      "step": 81701
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4135063608228316e-07,
+      "loss": 0.3428,
+      "step": 81702
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4132768631279404e-07,
+      "loss": 0.4146,
+      "step": 81703
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.413047369710543e-07,
+      "loss": 0.4316,
+      "step": 81704
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4128178805708183e-07,
+      "loss": 0.324,
+      "step": 81705
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4125883957089416e-07,
+      "loss": 0.4241,
+      "step": 81706
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4123589151250927e-07,
+      "loss": 0.3916,
+      "step": 81707
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.412129438819441e-07,
+      "loss": 0.4277,
+      "step": 81708
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4118999667921643e-07,
+      "loss": 0.3252,
+      "step": 81709
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4116704990434385e-07,
+      "loss": 0.3779,
+      "step": 81710
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4114410355734434e-07,
+      "loss": 0.416,
+      "step": 81711
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4112115763823476e-07,
+      "loss": 0.4795,
+      "step": 81712
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4109821214703335e-07,
+      "loss": 0.4048,
+      "step": 81713
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4107526708375696e-07,
+      "loss": 0.4399,
+      "step": 81714
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4105232244842395e-07,
+      "loss": 0.2355,
+      "step": 81715
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.41029378241051e-07,
+      "loss": 0.4399,
+      "step": 81716
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.410064344616563e-07,
+      "loss": 0.4431,
+      "step": 81717
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4098349111025725e-07,
+      "loss": 0.4639,
+      "step": 81718
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.409605481868717e-07,
+      "loss": 0.3293,
+      "step": 81719
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.409376056915166e-07,
+      "loss": 0.4248,
+      "step": 81720
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4091466362420994e-07,
+      "loss": 0.5913,
+      "step": 81721
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4089172198496927e-07,
+      "loss": 0.3354,
+      "step": 81722
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4086878077381227e-07,
+      "loss": 0.4087,
+      "step": 81723
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.408458399907565e-07,
+      "loss": 0.4395,
+      "step": 81724
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4082289963581896e-07,
+      "loss": 0.3555,
+      "step": 81725
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.407999597090176e-07,
+      "loss": 0.3298,
+      "step": 81726
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4077702021037e-07,
+      "loss": 0.3828,
+      "step": 81727
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.407540811398941e-07,
+      "loss": 0.4185,
+      "step": 81728
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.407311424976067e-07,
+      "loss": 0.4434,
+      "step": 81729
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4070820428352573e-07,
+      "loss": 0.2829,
+      "step": 81730
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.406852664976691e-07,
+      "loss": 0.5283,
+      "step": 81731
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4066232914005373e-07,
+      "loss": 0.4268,
+      "step": 81732
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.406393922106979e-07,
+      "loss": 0.5244,
+      "step": 81733
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4061645570961836e-07,
+      "loss": 0.4128,
+      "step": 81734
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.405935196368331e-07,
+      "loss": 0.2625,
+      "step": 81735
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4057058399236e-07,
+      "loss": 0.3403,
+      "step": 81736
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.4054764877621597e-07,
+      "loss": 0.418,
+      "step": 81737
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 4.405247139884188e-07,
+      "loss": 0.479,
+      "step": 81738
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.405017796289863e-07,
+      "loss": 0.3071,
+      "step": 81739
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.40478845697936e-07,
+      "loss": 0.3611,
+      "step": 81740
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.404559121952851e-07,
+      "loss": 0.4204,
+      "step": 81741
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4043297912105127e-07,
+      "loss": 0.3423,
+      "step": 81742
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.404100464752526e-07,
+      "loss": 0.5151,
+      "step": 81743
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4038711425790586e-07,
+      "loss": 0.4282,
+      "step": 81744
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.403641824690294e-07,
+      "loss": 0.335,
+      "step": 81745
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4034125110863994e-07,
+      "loss": 0.4819,
+      "step": 81746
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.403183201767554e-07,
+      "loss": 0.5112,
+      "step": 81747
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.402953896733934e-07,
+      "loss": 0.4087,
+      "step": 81748
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4027245959857196e-07,
+      "loss": 0.3411,
+      "step": 81749
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.402495299523077e-07,
+      "loss": 0.4346,
+      "step": 81750
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.402266007346187e-07,
+      "loss": 0.2797,
+      "step": 81751
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.402036719455228e-07,
+      "loss": 0.5371,
+      "step": 81752
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.401807435850371e-07,
+      "loss": 0.4214,
+      "step": 81753
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.40157815653179e-07,
+      "loss": 0.4014,
+      "step": 81754
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4013488814996624e-07,
+      "loss": 0.3663,
+      "step": 81755
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.401119610754166e-07,
+      "loss": 0.3352,
+      "step": 81756
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4008903442954773e-07,
+      "loss": 0.5444,
+      "step": 81757
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4006610821237653e-07,
+      "loss": 0.4097,
+      "step": 81758
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.400431824239211e-07,
+      "loss": 0.314,
+      "step": 81759
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.4002025706419877e-07,
+      "loss": 0.5669,
+      "step": 81760
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.399973321332277e-07,
+      "loss": 0.3416,
+      "step": 81761
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3997440763102434e-07,
+      "loss": 0.4072,
+      "step": 81762
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3995148355760724e-07,
+      "loss": 0.4263,
+      "step": 81763
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3992855991299327e-07,
+      "loss": 0.3624,
+      "step": 81764
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3990563669720015e-07,
+      "loss": 0.4175,
+      "step": 81765
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.39882713910246e-07,
+      "loss": 0.5234,
+      "step": 81766
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.398597915521475e-07,
+      "loss": 0.3218,
+      "step": 81767
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3983686962292256e-07,
+      "loss": 0.4551,
+      "step": 81768
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.398139481225891e-07,
+      "loss": 0.3994,
+      "step": 81769
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.397910270511641e-07,
+      "loss": 0.3501,
+      "step": 81770
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.397681064086652e-07,
+      "loss": 0.3203,
+      "step": 81771
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3974518619511024e-07,
+      "loss": 0.4502,
+      "step": 81772
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.397222664105169e-07,
+      "loss": 0.5146,
+      "step": 81773
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.396993470549025e-07,
+      "loss": 0.3071,
+      "step": 81774
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.396764281282842e-07,
+      "loss": 0.2755,
+      "step": 81775
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3965350963067995e-07,
+      "loss": 0.5283,
+      "step": 81776
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3963059156210716e-07,
+      "loss": 0.4873,
+      "step": 81777
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.396076739225838e-07,
+      "loss": 0.395,
+      "step": 81778
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.395847567121267e-07,
+      "loss": 0.4233,
+      "step": 81779
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.395618399307538e-07,
+      "loss": 0.3511,
+      "step": 81780
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.395389235784827e-07,
+      "loss": 0.4697,
+      "step": 81781
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3951600765533116e-07,
+      "loss": 0.2593,
+      "step": 81782
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.394930921613165e-07,
+      "loss": 0.4414,
+      "step": 81783
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3947017709645583e-07,
+      "loss": 0.394,
+      "step": 81784
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.39447262460767e-07,
+      "loss": 0.4316,
+      "step": 81785
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.394243482542681e-07,
+      "loss": 0.3706,
+      "step": 81786
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3940143447697573e-07,
+      "loss": 0.457,
+      "step": 81787
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3937852112890803e-07,
+      "loss": 0.2739,
+      "step": 81788
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.393556082100823e-07,
+      "loss": 0.4214,
+      "step": 81789
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3933269572051665e-07,
+      "loss": 0.436,
+      "step": 81790
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3930978366022776e-07,
+      "loss": 0.4688,
+      "step": 81791
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3928687202923364e-07,
+      "loss": 0.2868,
+      "step": 81792
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3926396082755214e-07,
+      "loss": 0.2986,
+      "step": 81793
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.392410500552001e-07,
+      "loss": 0.2681,
+      "step": 81794
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.392181397121957e-07,
+      "loss": 0.364,
+      "step": 81795
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3919522979855574e-07,
+      "loss": 0.4316,
+      "step": 81796
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.391723203142983e-07,
+      "loss": 0.4907,
+      "step": 81797
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3914941125944084e-07,
+      "loss": 0.3945,
+      "step": 81798
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.391265026340013e-07,
+      "loss": 0.4065,
+      "step": 81799
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3910359443799636e-07,
+      "loss": 0.4697,
+      "step": 81800
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3908068667144406e-07,
+      "loss": 0.373,
+      "step": 81801
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.390577793343618e-07,
+      "loss": 0.3755,
+      "step": 81802
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3903487242676794e-07,
+      "loss": 0.4321,
+      "step": 81803
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3901196594867864e-07,
+      "loss": 0.3989,
+      "step": 81804
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.38989059900112e-07,
+      "loss": 0.3833,
+      "step": 81805
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3896615428108576e-07,
+      "loss": 0.4746,
+      "step": 81806
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3894324909161764e-07,
+      "loss": 0.4395,
+      "step": 81807
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.389203443317245e-07,
+      "loss": 0.4263,
+      "step": 81808
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3889744000142426e-07,
+      "loss": 0.498,
+      "step": 81809
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3887453610073445e-07,
+      "loss": 0.5801,
+      "step": 81810
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3885163262967294e-07,
+      "loss": 0.3821,
+      "step": 81811
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3882872958825656e-07,
+      "loss": 0.4419,
+      "step": 81812
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.388058269765037e-07,
+      "loss": 0.3159,
+      "step": 81813
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.387829247944309e-07,
+      "loss": 0.4678,
+      "step": 81814
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3876002304205627e-07,
+      "loss": 0.4192,
+      "step": 81815
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.387371217193977e-07,
+      "loss": 0.48,
+      "step": 81816
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.387142208264719e-07,
+      "loss": 0.5439,
+      "step": 81817
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.386913203632968e-07,
+      "loss": 0.342,
+      "step": 81818
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3866842032988993e-07,
+      "loss": 0.3169,
+      "step": 81819
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3864552072626926e-07,
+      "loss": 0.345,
+      "step": 81820
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3862262155245154e-07,
+      "loss": 0.3807,
+      "step": 81821
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.385997228084546e-07,
+      "loss": 0.3123,
+      "step": 81822
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3857682449429647e-07,
+      "loss": 0.4258,
+      "step": 81823
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3855392660999424e-07,
+      "loss": 0.3093,
+      "step": 81824
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3853102915556504e-07,
+      "loss": 0.2477,
+      "step": 81825
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3850813213102687e-07,
+      "loss": 0.5315,
+      "step": 81826
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3848523553639714e-07,
+      "loss": 0.3264,
+      "step": 81827
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3846233937169387e-07,
+      "loss": 0.4258,
+      "step": 81828
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.384394436369337e-07,
+      "loss": 0.3511,
+      "step": 81829
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.384165483321348e-07,
+      "loss": 0.373,
+      "step": 81830
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3839365345731437e-07,
+      "loss": 0.4512,
+      "step": 81831
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3837075901249053e-07,
+      "loss": 0.6045,
+      "step": 81832
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.383478649976804e-07,
+      "loss": 0.2977,
+      "step": 81833
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3832497141290104e-07,
+      "loss": 0.4448,
+      "step": 81834
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.383020782581706e-07,
+      "loss": 0.5083,
+      "step": 81835
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.382791855335062e-07,
+      "loss": 0.3293,
+      "step": 81836
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.382562932389261e-07,
+      "loss": 0.3843,
+      "step": 81837
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.382334013744469e-07,
+      "loss": 0.342,
+      "step": 81838
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3821050994008656e-07,
+      "loss": 0.3762,
+      "step": 81839
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3818761893586296e-07,
+      "loss": 0.3945,
+      "step": 81840
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3816472836179297e-07,
+      "loss": 0.4214,
+      "step": 81841
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3814183821789444e-07,
+      "loss": 0.479,
+      "step": 81842
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.381189485041852e-07,
+      "loss": 0.4395,
+      "step": 81843
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.380960592206819e-07,
+      "loss": 0.3232,
+      "step": 81844
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.380731703674032e-07,
+      "loss": 0.4032,
+      "step": 81845
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3805028194436555e-07,
+      "loss": 0.4331,
+      "step": 81846
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3802739395158707e-07,
+      "loss": 0.4673,
+      "step": 81847
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.38004506389085e-07,
+      "loss": 0.4697,
+      "step": 81848
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.379816192568776e-07,
+      "loss": 0.4575,
+      "step": 81849
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.379587325549814e-07,
+      "loss": 0.5161,
+      "step": 81850
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.379358462834143e-07,
+      "loss": 0.5234,
+      "step": 81851
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3791296044219395e-07,
+      "loss": 0.4565,
+      "step": 81852
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.378900750313381e-07,
+      "loss": 0.4312,
+      "step": 81853
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.37867190050864e-07,
+      "loss": 0.3591,
+      "step": 81854
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3784430550078876e-07,
+      "loss": 0.4932,
+      "step": 81855
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.378214213811303e-07,
+      "loss": 0.3955,
+      "step": 81856
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3779853769190654e-07,
+      "loss": 0.2921,
+      "step": 81857
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.377756544331341e-07,
+      "loss": 0.4265,
+      "step": 81858
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3775277160483113e-07,
+      "loss": 0.4297,
+      "step": 81859
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.37729889207015e-07,
+      "loss": 0.4612,
+      "step": 81860
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3770700723970356e-07,
+      "loss": 0.5005,
+      "step": 81861
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.376841257029136e-07,
+      "loss": 0.3412,
+      "step": 81862
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3766124459666354e-07,
+      "loss": 0.3938,
+      "step": 81863
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3763836392096986e-07,
+      "loss": 0.4263,
+      "step": 81864
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3761548367585076e-07,
+      "loss": 0.4951,
+      "step": 81865
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3759260386132404e-07,
+      "loss": 0.4043,
+      "step": 81866
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3756972447740627e-07,
+      "loss": 0.4451,
+      "step": 81867
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.375468455241156e-07,
+      "loss": 0.4341,
+      "step": 81868
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3752396700146945e-07,
+      "loss": 0.2817,
+      "step": 81869
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3750108890948566e-07,
+      "loss": 0.4536,
+      "step": 81870
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3747821124818107e-07,
+      "loss": 0.3799,
+      "step": 81871
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3745533401757394e-07,
+      "loss": 0.3665,
+      "step": 81872
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3743245721768105e-07,
+      "loss": 0.4507,
+      "step": 81873
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3740958084852056e-07,
+      "loss": 0.4238,
+      "step": 81874
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.373867049101093e-07,
+      "loss": 0.4766,
+      "step": 81875
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.373638294024653e-07,
+      "loss": 0.5103,
+      "step": 81876
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.373409543256058e-07,
+      "loss": 0.4336,
+      "step": 81877
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3731807967954893e-07,
+      "loss": 0.4858,
+      "step": 81878
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3729520546431133e-07,
+      "loss": 0.5317,
+      "step": 81879
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3727233167991095e-07,
+      "loss": 0.3926,
+      "step": 81880
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3724945832636515e-07,
+      "loss": 0.4268,
+      "step": 81881
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3722658540369204e-07,
+      "loss": 0.3923,
+      "step": 81882
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.372037129119086e-07,
+      "loss": 0.4912,
+      "step": 81883
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3718084085103214e-07,
+      "loss": 0.4092,
+      "step": 81884
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.371579692210804e-07,
+      "loss": 0.314,
+      "step": 81885
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3713509802207083e-07,
+      "loss": 0.4023,
+      "step": 81886
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3711222725402143e-07,
+      "loss": 0.3891,
+      "step": 81887
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3708935691694895e-07,
+      "loss": 0.3198,
+      "step": 81888
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3706648701087135e-07,
+      "loss": 0.5049,
+      "step": 81889
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.37043617535806e-07,
+      "loss": 0.3833,
+      "step": 81890
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3702074849177075e-07,
+      "loss": 0.4072,
+      "step": 81891
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.369978798787829e-07,
+      "loss": 0.3993,
+      "step": 81892
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3697501169685956e-07,
+      "loss": 0.3945,
+      "step": 81893
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3695214394601854e-07,
+      "loss": 0.3301,
+      "step": 81894
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3692927662627764e-07,
+      "loss": 0.4077,
+      "step": 81895
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.369064097376539e-07,
+      "loss": 0.4025,
+      "step": 81896
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3688354328016495e-07,
+      "loss": 0.4546,
+      "step": 81897
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.368606772538284e-07,
+      "loss": 0.425,
+      "step": 81898
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3683781165866207e-07,
+      "loss": 0.3999,
+      "step": 81899
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3681494649468275e-07,
+      "loss": 0.4268,
+      "step": 81900
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.367920817619084e-07,
+      "loss": 0.3645,
+      "step": 81901
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3676921746035677e-07,
+      "loss": 0.3606,
+      "step": 81902
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3674635359004463e-07,
+      "loss": 0.3967,
+      "step": 81903
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.367234901509904e-07,
+      "loss": 0.5176,
+      "step": 81904
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3670062714321065e-07,
+      "loss": 0.3181,
+      "step": 81905
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3667776456672333e-07,
+      "loss": 0.5127,
+      "step": 81906
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.36654902421546e-07,
+      "loss": 0.412,
+      "step": 81907
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3663204070769645e-07,
+      "loss": 0.5024,
+      "step": 81908
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3660917942519135e-07,
+      "loss": 0.3887,
+      "step": 81909
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.365863185740488e-07,
+      "loss": 0.312,
+      "step": 81910
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.365634581542863e-07,
+      "loss": 0.3118,
+      "step": 81911
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.365405981659218e-07,
+      "loss": 0.343,
+      "step": 81912
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.365177386089717e-07,
+      "loss": 0.3181,
+      "step": 81913
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.36494879483454e-07,
+      "loss": 0.4524,
+      "step": 81914
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3647202078938617e-07,
+      "loss": 0.3643,
+      "step": 81915
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.364491625267862e-07,
+      "loss": 0.3511,
+      "step": 81916
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.364263046956709e-07,
+      "loss": 0.3574,
+      "step": 81917
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3640344729605805e-07,
+      "loss": 0.415,
+      "step": 81918
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3638059032796514e-07,
+      "loss": 0.2694,
+      "step": 81919
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3635773379141005e-07,
+      "loss": 0.4609,
+      "step": 81920
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.363348776864095e-07,
+      "loss": 0.4668,
+      "step": 81921
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.363120220129819e-07,
+      "loss": 0.3755,
+      "step": 81922
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.362891667711438e-07,
+      "loss": 0.4434,
+      "step": 81923
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3626631196091323e-07,
+      "loss": 0.3798,
+      "step": 81924
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.36243457582308e-07,
+      "loss": 0.4009,
+      "step": 81925
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3622060363534484e-07,
+      "loss": 0.3492,
+      "step": 81926
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3619775012004166e-07,
+      "loss": 0.3634,
+      "step": 81927
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3617489703641585e-07,
+      "loss": 0.3507,
+      "step": 81928
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.361520443844855e-07,
+      "loss": 0.3545,
+      "step": 81929
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.361291921642671e-07,
+      "loss": 0.5527,
+      "step": 81930
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3610634037577877e-07,
+      "loss": 0.4019,
+      "step": 81931
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3608348901903826e-07,
+      "loss": 0.3911,
+      "step": 81932
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3606063809406257e-07,
+      "loss": 0.3159,
+      "step": 81933
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.360377876008691e-07,
+      "loss": 0.5005,
+      "step": 81934
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.360149375394754e-07,
+      "loss": 0.4185,
+      "step": 81935
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.359920879098993e-07,
+      "loss": 0.425,
+      "step": 81936
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.359692387121583e-07,
+      "loss": 0.4062,
+      "step": 81937
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3594638994626954e-07,
+      "loss": 0.3031,
+      "step": 81938
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3592354161225054e-07,
+      "loss": 0.4448,
+      "step": 81939
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.359006937101191e-07,
+      "loss": 0.4414,
+      "step": 81940
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.358778462398928e-07,
+      "loss": 0.313,
+      "step": 81941
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3585499920158887e-07,
+      "loss": 0.2969,
+      "step": 81942
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3583215259522453e-07,
+      "loss": 0.4255,
+      "step": 81943
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3580930642081757e-07,
+      "loss": 0.3621,
+      "step": 81944
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3578646067838555e-07,
+      "loss": 0.3906,
+      "step": 81945
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3576361536794615e-07,
+      "loss": 0.4043,
+      "step": 81946
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3574077048951627e-07,
+      "loss": 0.4463,
+      "step": 81947
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3571792604311375e-07,
+      "loss": 0.4653,
+      "step": 81948
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3569508202875647e-07,
+      "loss": 0.519,
+      "step": 81949
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3567223844646106e-07,
+      "loss": 0.2714,
+      "step": 81950
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.356493952962456e-07,
+      "loss": 0.4243,
+      "step": 81951
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3562655257812765e-07,
+      "loss": 0.4219,
+      "step": 81952
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3560371029212427e-07,
+      "loss": 0.5439,
+      "step": 81953
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.355808684382535e-07,
+      "loss": 0.365,
+      "step": 81954
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.355580270165321e-07,
+      "loss": 0.366,
+      "step": 81955
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3553518602697805e-07,
+      "loss": 0.3887,
+      "step": 81956
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3551234546960867e-07,
+      "loss": 0.3979,
+      "step": 81957
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3548950534444197e-07,
+      "loss": 0.458,
+      "step": 81958
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.354666656514945e-07,
+      "loss": 0.375,
+      "step": 81959
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3544382639078436e-07,
+      "loss": 0.4907,
+      "step": 81960
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.354209875623289e-07,
+      "loss": 0.3574,
+      "step": 81961
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3539814916614605e-07,
+      "loss": 0.2478,
+      "step": 81962
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.353753112022528e-07,
+      "loss": 0.3018,
+      "step": 81963
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.353524736706663e-07,
+      "loss": 0.3929,
+      "step": 81964
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3532963657140453e-07,
+      "loss": 0.3774,
+      "step": 81965
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3530679990448527e-07,
+      "loss": 0.4067,
+      "step": 81966
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3528396366992523e-07,
+      "loss": 0.4116,
+      "step": 81967
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.352611278677423e-07,
+      "loss": 0.481,
+      "step": 81968
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3523829249795396e-07,
+      "loss": 0.4695,
+      "step": 81969
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.35215457560578e-07,
+      "loss": 0.4546,
+      "step": 81970
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3519262305563133e-07,
+      "loss": 0.3656,
+      "step": 81971
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3516978898313195e-07,
+      "loss": 0.4897,
+      "step": 81972
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.351469553430969e-07,
+      "loss": 0.3989,
+      "step": 81973
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.351241221355437e-07,
+      "loss": 0.4629,
+      "step": 81974
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.351012893604904e-07,
+      "loss": 0.562,
+      "step": 81975
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.350784570179537e-07,
+      "loss": 0.4702,
+      "step": 81976
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.350556251079515e-07,
+      "loss": 0.3723,
+      "step": 81977
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3503279363050106e-07,
+      "loss": 0.4824,
+      "step": 81978
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.350099625856205e-07,
+      "loss": 0.4465,
+      "step": 81979
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.349871319733266e-07,
+      "loss": 0.4308,
+      "step": 81980
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3496430179363685e-07,
+      "loss": 0.4951,
+      "step": 81981
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.349414720465695e-07,
+      "loss": 0.4102,
+      "step": 81982
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3491864273214143e-07,
+      "loss": 0.3772,
+      "step": 81983
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.348958138503698e-07,
+      "loss": 0.5127,
+      "step": 81984
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.348729854012724e-07,
+      "loss": 0.3254,
+      "step": 81985
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.348501573848667e-07,
+      "loss": 0.459,
+      "step": 81986
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3482732980117086e-07,
+      "loss": 0.395,
+      "step": 81987
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.348045026502012e-07,
+      "loss": 0.5405,
+      "step": 81988
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.347816759319758e-07,
+      "loss": 0.3989,
+      "step": 81989
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.347588496465121e-07,
+      "loss": 0.4307,
+      "step": 81990
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.347360237938279e-07,
+      "loss": 0.5132,
+      "step": 81991
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.347131983739403e-07,
+      "loss": 0.2974,
+      "step": 81992
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3469037338686654e-07,
+      "loss": 0.4402,
+      "step": 81993
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3466754883262424e-07,
+      "loss": 0.3707,
+      "step": 81994
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3464472471123103e-07,
+      "loss": 0.4668,
+      "step": 81995
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.346219010227048e-07,
+      "loss": 0.4287,
+      "step": 81996
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3459907776706216e-07,
+      "loss": 0.4595,
+      "step": 81997
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.34576254944321e-07,
+      "loss": 0.4307,
+      "step": 81998
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3455343255449894e-07,
+      "loss": 0.5122,
+      "step": 81999
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.345306105976135e-07,
+      "loss": 0.4087,
+      "step": 82000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3450778907368167e-07,
+      "loss": 0.4067,
+      "step": 82001
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.344849679827216e-07,
+      "loss": 0.3677,
+      "step": 82002
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3446214732475004e-07,
+      "loss": 0.4243,
+      "step": 82003
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3443932709978515e-07,
+      "loss": 0.4839,
+      "step": 82004
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.344165073078437e-07,
+      "loss": 0.4458,
+      "step": 82005
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.343936879489436e-07,
+      "loss": 0.4072,
+      "step": 82006
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3437086902310216e-07,
+      "loss": 0.2474,
+      "step": 82007
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3434805053033727e-07,
+      "loss": 0.375,
+      "step": 82008
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.343252324706658e-07,
+      "loss": 0.5015,
+      "step": 82009
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.343024148441055e-07,
+      "loss": 0.2089,
+      "step": 82010
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.342795976506737e-07,
+      "loss": 0.4897,
+      "step": 82011
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.342567808903884e-07,
+      "loss": 0.4316,
+      "step": 82012
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3423396456326677e-07,
+      "loss": 0.4785,
+      "step": 82013
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3421114866932576e-07,
+      "loss": 0.4062,
+      "step": 82014
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.341883332085832e-07,
+      "loss": 0.3679,
+      "step": 82015
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.341655181810566e-07,
+      "loss": 0.5864,
+      "step": 82016
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.34142703586764e-07,
+      "loss": 0.3621,
+      "step": 82017
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3411988942572177e-07,
+      "loss": 0.3455,
+      "step": 82018
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.34097075697948e-07,
+      "loss": 0.4795,
+      "step": 82019
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3407426240346035e-07,
+      "loss": 0.4272,
+      "step": 82020
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.340514495422757e-07,
+      "loss": 0.3164,
+      "step": 82021
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.340286371144122e-07,
+      "loss": 0.4868,
+      "step": 82022
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.340058251198866e-07,
+      "loss": 0.5,
+      "step": 82023
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.339830135587167e-07,
+      "loss": 0.3604,
+      "step": 82024
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3396020243092033e-07,
+      "loss": 0.4204,
+      "step": 82025
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3393739173651436e-07,
+      "loss": 0.3672,
+      "step": 82026
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.339145814755163e-07,
+      "loss": 0.4868,
+      "step": 82027
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.33891771647944e-07,
+      "loss": 0.3092,
+      "step": 82028
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3386896225381496e-07,
+      "loss": 0.499,
+      "step": 82029
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.338461532931461e-07,
+      "loss": 0.3621,
+      "step": 82030
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.338233447659553e-07,
+      "loss": 0.417,
+      "step": 82031
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3380053667226026e-07,
+      "loss": 0.4785,
+      "step": 82032
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3377772901207766e-07,
+      "loss": 0.4258,
+      "step": 82033
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3375492178542585e-07,
+      "loss": 0.3726,
+      "step": 82034
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3373211499232154e-07,
+      "loss": 0.4712,
+      "step": 82035
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3370930863278244e-07,
+      "loss": 0.2783,
+      "step": 82036
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3368650270682637e-07,
+      "loss": 0.3645,
+      "step": 82037
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.336636972144703e-07,
+      "loss": 0.3412,
+      "step": 82038
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3364089215573184e-07,
+      "loss": 0.5117,
+      "step": 82039
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3361808753062843e-07,
+      "loss": 0.4038,
+      "step": 82040
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3359528333917805e-07,
+      "loss": 0.4658,
+      "step": 82041
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.335724795813978e-07,
+      "loss": 0.3435,
+      "step": 82042
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.335496762573045e-07,
+      "loss": 0.4307,
+      "step": 82043
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.335268733669163e-07,
+      "loss": 0.3525,
+      "step": 82044
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3350407091025057e-07,
+      "loss": 0.4487,
+      "step": 82045
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3348126888732496e-07,
+      "loss": 0.4746,
+      "step": 82046
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.334584672981564e-07,
+      "loss": 0.564,
+      "step": 82047
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3343566614276264e-07,
+      "loss": 0.3049,
+      "step": 82048
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3341286542116106e-07,
+      "loss": 0.4434,
+      "step": 82049
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3339006513336964e-07,
+      "loss": 0.3294,
+      "step": 82050
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3336726527940504e-07,
+      "loss": 0.3794,
+      "step": 82051
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.333444658592854e-07,
+      "loss": 0.4126,
+      "step": 82052
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.333216668730275e-07,
+      "loss": 0.4678,
+      "step": 82053
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3329886832064953e-07,
+      "loss": 0.438,
+      "step": 82054
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3327607020216814e-07,
+      "loss": 0.521,
+      "step": 82055
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3325327251760124e-07,
+      "loss": 0.3794,
+      "step": 82056
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.332304752669662e-07,
+      "loss": 0.4805,
+      "step": 82057
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3320767845028096e-07,
+      "loss": 0.5708,
+      "step": 82058
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.331848820675621e-07,
+      "loss": 0.3173,
+      "step": 82059
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3316208611882754e-07,
+      "loss": 0.3507,
+      "step": 82060
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3313929060409473e-07,
+      "loss": 0.3901,
+      "step": 82061
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.331164955233815e-07,
+      "loss": 0.4629,
+      "step": 82062
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.330937008767048e-07,
+      "loss": 0.4834,
+      "step": 82063
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.33070906664082e-07,
+      "loss": 0.3613,
+      "step": 82064
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3304811288553055e-07,
+      "loss": 0.3048,
+      "step": 82065
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3302531954106815e-07,
+      "loss": 0.3394,
+      "step": 82066
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3300252663071257e-07,
+      "loss": 0.3481,
+      "step": 82067
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3297973415448054e-07,
+      "loss": 0.3896,
+      "step": 82068
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3295694211238974e-07,
+      "loss": 0.376,
+      "step": 82069
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.329341505044579e-07,
+      "loss": 0.3811,
+      "step": 82070
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3291135933070256e-07,
+      "loss": 0.4717,
+      "step": 82071
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3288856859114096e-07,
+      "loss": 0.4722,
+      "step": 82072
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.328657782857902e-07,
+      "loss": 0.4697,
+      "step": 82073
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3284298841466793e-07,
+      "loss": 0.397,
+      "step": 82074
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.328201989777921e-07,
+      "loss": 0.3763,
+      "step": 82075
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3279740997517943e-07,
+      "loss": 0.4033,
+      "step": 82076
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.327746214068477e-07,
+      "loss": 0.3585,
+      "step": 82077
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3275183327281426e-07,
+      "loss": 0.5527,
+      "step": 82078
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.327290455730972e-07,
+      "loss": 0.3403,
+      "step": 82079
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.327062583077129e-07,
+      "loss": 0.4287,
+      "step": 82080
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.326834714766795e-07,
+      "loss": 0.4404,
+      "step": 82081
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.326606850800145e-07,
+      "loss": 0.3576,
+      "step": 82082
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3263789911773487e-07,
+      "loss": 0.2698,
+      "step": 82083
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.326151135898586e-07,
+      "loss": 0.4072,
+      "step": 82084
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.325923284964025e-07,
+      "loss": 0.3523,
+      "step": 82085
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.325695438373843e-07,
+      "loss": 0.3105,
+      "step": 82086
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.325467596128216e-07,
+      "loss": 0.4792,
+      "step": 82087
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.32523975822732e-07,
+      "loss": 0.4543,
+      "step": 82088
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3250119246713245e-07,
+      "loss": 0.4727,
+      "step": 82089
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3247840954604065e-07,
+      "loss": 0.4272,
+      "step": 82090
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.32455627059474e-07,
+      "loss": 0.4224,
+      "step": 82091
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3243284500745067e-07,
+      "loss": 0.3804,
+      "step": 82092
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3241006338998674e-07,
+      "loss": 0.5371,
+      "step": 82093
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.323872822071004e-07,
+      "loss": 0.5073,
+      "step": 82094
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3236450145880885e-07,
+      "loss": 0.3984,
+      "step": 82095
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.323417211451302e-07,
+      "loss": 0.2361,
+      "step": 82096
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.32318941266081e-07,
+      "loss": 0.2567,
+      "step": 82097
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3229616182167905e-07,
+      "loss": 0.2469,
+      "step": 82098
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3227338281194183e-07,
+      "loss": 0.5024,
+      "step": 82099
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.322506042368872e-07,
+      "loss": 0.2769,
+      "step": 82100
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3222782609653174e-07,
+      "loss": 0.395,
+      "step": 82101
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3220504839089365e-07,
+      "loss": 0.3564,
+      "step": 82102
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.321822711199897e-07,
+      "loss": 0.3584,
+      "step": 82103
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.321594942838378e-07,
+      "loss": 0.3033,
+      "step": 82104
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3213671788245555e-07,
+      "loss": 0.5078,
+      "step": 82105
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.321139419158597e-07,
+      "loss": 0.4067,
+      "step": 82106
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.320911663840682e-07,
+      "loss": 0.3726,
+      "step": 82107
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.320683912870983e-07,
+      "loss": 0.5308,
+      "step": 82108
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3204561662496784e-07,
+      "loss": 0.4307,
+      "step": 82109
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.320228423976936e-07,
+      "loss": 0.48,
+      "step": 82110
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.320000686052934e-07,
+      "loss": 0.2898,
+      "step": 82111
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3197729524778505e-07,
+      "loss": 0.3497,
+      "step": 82112
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3195452232518547e-07,
+      "loss": 0.4805,
+      "step": 82113
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.319317498375118e-07,
+      "loss": 0.4053,
+      "step": 82114
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3190897778478196e-07,
+      "loss": 0.459,
+      "step": 82115
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3188620616701323e-07,
+      "loss": 0.4336,
+      "step": 82116
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.318634349842235e-07,
+      "loss": 0.418,
+      "step": 82117
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.318406642364294e-07,
+      "loss": 0.3164,
+      "step": 82118
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3181789392364887e-07,
+      "loss": 0.4644,
+      "step": 82119
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.317951240458991e-07,
+      "loss": 0.2532,
+      "step": 82120
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.317723546031983e-07,
+      "loss": 0.4619,
+      "step": 82121
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3174958559556307e-07,
+      "loss": 0.3735,
+      "step": 82122
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.317268170230107e-07,
+      "loss": 0.4546,
+      "step": 82123
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.31704048885559e-07,
+      "loss": 0.3057,
+      "step": 82124
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3168128118322546e-07,
+      "loss": 0.5654,
+      "step": 82125
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.316585139160277e-07,
+      "loss": 0.4424,
+      "step": 82126
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.316357470839825e-07,
+      "loss": 0.2769,
+      "step": 82127
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 4.3161298068710783e-07,
+      "loss": 0.5625,
+      "step": 82128
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3159021472542125e-07,
+      "loss": 0.4131,
+      "step": 82129
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.315674491989395e-07,
+      "loss": 0.4541,
+      "step": 82130
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.315446841076804e-07,
+      "loss": 0.4282,
+      "step": 82131
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.315219194516618e-07,
+      "loss": 0.2381,
+      "step": 82132
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.314991552309004e-07,
+      "loss": 0.3429,
+      "step": 82133
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3147639144541427e-07,
+      "loss": 0.4453,
+      "step": 82134
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.314536280952201e-07,
+      "loss": 0.3371,
+      "step": 82135
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.314308651803359e-07,
+      "loss": 0.3572,
+      "step": 82136
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3140810270077886e-07,
+      "loss": 0.2562,
+      "step": 82137
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.313853406565668e-07,
+      "loss": 0.3823,
+      "step": 82138
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3136257904771655e-07,
+      "loss": 0.2991,
+      "step": 82139
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.313398178742459e-07,
+      "loss": 0.5137,
+      "step": 82140
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.313170571361721e-07,
+      "loss": 0.4531,
+      "step": 82141
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3129429683351304e-07,
+      "loss": 0.394,
+      "step": 82142
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.312715369662858e-07,
+      "loss": 0.4937,
+      "step": 82143
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.312487775345074e-07,
+      "loss": 0.2415,
+      "step": 82144
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.312260185381957e-07,
+      "loss": 0.3362,
+      "step": 82145
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.312032599773685e-07,
+      "loss": 0.4634,
+      "step": 82146
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3118050185204234e-07,
+      "loss": 0.4448,
+      "step": 82147
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3115774416223516e-07,
+      "loss": 0.4424,
+      "step": 82148
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3113498690796437e-07,
+      "loss": 0.3174,
+      "step": 82149
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3111223008924767e-07,
+      "loss": 0.5156,
+      "step": 82150
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3108947370610183e-07,
+      "loss": 0.4785,
+      "step": 82151
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3106671775854496e-07,
+      "loss": 0.4644,
+      "step": 82152
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.310439622465939e-07,
+      "loss": 0.429,
+      "step": 82153
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3102120717026626e-07,
+      "loss": 0.4775,
+      "step": 82154
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.309984525295798e-07,
+      "loss": 0.439,
+      "step": 82155
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.309756983245514e-07,
+      "loss": 0.335,
+      "step": 82156
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.309529445551987e-07,
+      "loss": 0.3657,
+      "step": 82157
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3093019122153915e-07,
+      "loss": 0.4861,
+      "step": 82158
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.309074383235906e-07,
+      "loss": 0.499,
+      "step": 82159
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.308846858613696e-07,
+      "loss": 0.38,
+      "step": 82160
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3086193383489444e-07,
+      "loss": 0.3828,
+      "step": 82161
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.308391822441818e-07,
+      "loss": 0.4683,
+      "step": 82162
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.308164310892497e-07,
+      "loss": 0.3556,
+      "step": 82163
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.30793680370115e-07,
+      "loss": 0.2393,
+      "step": 82164
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3077093008679544e-07,
+      "loss": 0.2894,
+      "step": 82165
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.307481802393084e-07,
+      "loss": 0.3464,
+      "step": 82166
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3072543082767156e-07,
+      "loss": 0.4341,
+      "step": 82167
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3070268185190174e-07,
+      "loss": 0.3445,
+      "step": 82168
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3067993331201677e-07,
+      "loss": 0.4497,
+      "step": 82169
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3065718520803396e-07,
+      "loss": 0.3394,
+      "step": 82170
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.306344375399712e-07,
+      "loss": 0.3813,
+      "step": 82171
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.306116903078454e-07,
+      "loss": 0.3016,
+      "step": 82172
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.305889435116736e-07,
+      "loss": 0.6045,
+      "step": 82173
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.305661971514738e-07,
+      "loss": 0.4575,
+      "step": 82174
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.305434512272632e-07,
+      "loss": 0.5288,
+      "step": 82175
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3052070573905973e-07,
+      "loss": 0.5366,
+      "step": 82176
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3049796068687994e-07,
+      "loss": 0.3804,
+      "step": 82177
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.304752160707418e-07,
+      "loss": 0.4443,
+      "step": 82178
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3045247189066246e-07,
+      "loss": 0.5044,
+      "step": 82179
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.304297281466599e-07,
+      "loss": 0.3955,
+      "step": 82180
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3040698483875105e-07,
+      "loss": 0.4307,
+      "step": 82181
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.303842419669531e-07,
+      "loss": 0.3475,
+      "step": 82182
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3036149953128366e-07,
+      "loss": 0.3716,
+      "step": 82183
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3033875753176063e-07,
+      "loss": 0.5103,
+      "step": 82184
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.303160159684007e-07,
+      "loss": 0.4011,
+      "step": 82185
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.302932748412216e-07,
+      "loss": 0.3867,
+      "step": 82186
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.302705341502406e-07,
+      "loss": 0.4341,
+      "step": 82187
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3024779389547573e-07,
+      "loss": 0.4673,
+      "step": 82188
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3022505407694353e-07,
+      "loss": 0.4272,
+      "step": 82189
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3020231469466184e-07,
+      "loss": 0.3938,
+      "step": 82190
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.301795757486485e-07,
+      "loss": 0.4219,
+      "step": 82191
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3015683723891993e-07,
+      "loss": 0.3213,
+      "step": 82192
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3013409916549446e-07,
+      "loss": 0.394,
+      "step": 82193
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.301113615283888e-07,
+      "loss": 0.4829,
+      "step": 82194
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.300886243276206e-07,
+      "loss": 0.4355,
+      "step": 82195
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.300658875632074e-07,
+      "loss": 0.3472,
+      "step": 82196
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.300431512351669e-07,
+      "loss": 0.3687,
+      "step": 82197
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.3002041534351575e-07,
+      "loss": 0.4678,
+      "step": 82198
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2999767988827187e-07,
+      "loss": 0.3406,
+      "step": 82199
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2997494486945276e-07,
+      "loss": 0.2095,
+      "step": 82200
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2995221028707575e-07,
+      "loss": 0.4009,
+      "step": 82201
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.299294761411576e-07,
+      "loss": 0.3246,
+      "step": 82202
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2990674243171644e-07,
+      "loss": 0.3227,
+      "step": 82203
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.298840091587693e-07,
+      "loss": 0.4741,
+      "step": 82204
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2986127632233424e-07,
+      "loss": 0.4648,
+      "step": 82205
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.298385439224278e-07,
+      "loss": 0.4739,
+      "step": 82206
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.298158119590677e-07,
+      "loss": 0.3182,
+      "step": 82207
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2979308043227145e-07,
+      "loss": 0.4165,
+      "step": 82208
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.297703493420568e-07,
+      "loss": 0.355,
+      "step": 82209
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2974761868844046e-07,
+      "loss": 0.4502,
+      "step": 82210
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.297248884714404e-07,
+      "loss": 0.3701,
+      "step": 82211
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.297021586910735e-07,
+      "loss": 0.3657,
+      "step": 82212
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2967942934735737e-07,
+      "loss": 0.4094,
+      "step": 82213
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.296567004403099e-07,
+      "loss": 0.3279,
+      "step": 82214
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2963397196994767e-07,
+      "loss": 0.4282,
+      "step": 82215
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2961124393628846e-07,
+      "loss": 0.375,
+      "step": 82216
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2958851633935013e-07,
+      "loss": 0.4253,
+      "step": 82217
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.295657891791492e-07,
+      "loss": 0.3381,
+      "step": 82218
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.295430624557036e-07,
+      "loss": 0.3972,
+      "step": 82219
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.295203361690305e-07,
+      "loss": 0.4104,
+      "step": 82220
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2949761031914797e-07,
+      "loss": 0.4307,
+      "step": 82221
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2947488490607276e-07,
+      "loss": 0.4937,
+      "step": 82222
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.294521599298222e-07,
+      "loss": 0.4873,
+      "step": 82223
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2942943539041375e-07,
+      "loss": 0.3152,
+      "step": 82224
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.29406711287865e-07,
+      "loss": 0.4071,
+      "step": 82225
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.293839876221936e-07,
+      "loss": 0.4165,
+      "step": 82226
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2936126439341637e-07,
+      "loss": 0.439,
+      "step": 82227
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.293385416015509e-07,
+      "loss": 0.4189,
+      "step": 82228
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2931581924661466e-07,
+      "loss": 0.3992,
+      "step": 82229
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.292930973286255e-07,
+      "loss": 0.3171,
+      "step": 82230
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2927037584760026e-07,
+      "loss": 0.4075,
+      "step": 82231
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.292476548035562e-07,
+      "loss": 0.3381,
+      "step": 82232
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2922493419651093e-07,
+      "loss": 0.4492,
+      "step": 82233
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2920221402648226e-07,
+      "loss": 0.2314,
+      "step": 82234
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2917949429348686e-07,
+      "loss": 0.3567,
+      "step": 82235
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2915677499754245e-07,
+      "loss": 0.4199,
+      "step": 82236
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2913405613866636e-07,
+      "loss": 0.4517,
+      "step": 82237
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.291113377168766e-07,
+      "loss": 0.4485,
+      "step": 82238
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.290886197321896e-07,
+      "loss": 0.502,
+      "step": 82239
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2906590218462314e-07,
+      "loss": 0.4014,
+      "step": 82240
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.290431850741951e-07,
+      "loss": 0.4707,
+      "step": 82241
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.29020468400922e-07,
+      "loss": 0.519,
+      "step": 82242
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2899775216482217e-07,
+      "loss": 0.3723,
+      "step": 82243
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2897503636591205e-07,
+      "loss": 0.4401,
+      "step": 82244
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.289523210042094e-07,
+      "loss": 0.4722,
+      "step": 82245
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.289296060797318e-07,
+      "loss": 0.4004,
+      "step": 82246
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2890689159249694e-07,
+      "loss": 0.4175,
+      "step": 82247
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.288841775425215e-07,
+      "loss": 0.4297,
+      "step": 82248
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.28861463929823e-07,
+      "loss": 0.4497,
+      "step": 82249
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.288387507544192e-07,
+      "loss": 0.314,
+      "step": 82250
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.288160380163276e-07,
+      "loss": 0.3215,
+      "step": 82251
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2879332571556524e-07,
+      "loss": 0.3999,
+      "step": 82252
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.287706138521493e-07,
+      "loss": 0.4292,
+      "step": 82253
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2874790242609737e-07,
+      "loss": 0.3955,
+      "step": 82254
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2872519143742727e-07,
+      "loss": 0.5166,
+      "step": 82255
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2870248088615567e-07,
+      "loss": 0.2712,
+      "step": 82256
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.286797707723002e-07,
+      "loss": 0.4531,
+      "step": 82257
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.286570610958785e-07,
+      "loss": 0.4668,
+      "step": 82258
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2863435185690824e-07,
+      "loss": 0.3813,
+      "step": 82259
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2861164305540586e-07,
+      "loss": 0.4678,
+      "step": 82260
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.285889346913897e-07,
+      "loss": 0.3442,
+      "step": 82261
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.285662267648764e-07,
+      "loss": 0.4575,
+      "step": 82262
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.285435192758836e-07,
+      "loss": 0.5356,
+      "step": 82263
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2852081222442917e-07,
+      "loss": 0.3396,
+      "step": 82264
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2849810561052966e-07,
+      "loss": 0.2955,
+      "step": 82265
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.284753994342029e-07,
+      "loss": 0.46,
+      "step": 82266
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.284526936954662e-07,
+      "loss": 0.2488,
+      "step": 82267
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.284299883943374e-07,
+      "loss": 0.4346,
+      "step": 82268
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.284072835308331e-07,
+      "loss": 0.5625,
+      "step": 82269
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2838457910497117e-07,
+      "loss": 0.4443,
+      "step": 82270
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2836187511676913e-07,
+      "loss": 0.4163,
+      "step": 82271
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.283391715662441e-07,
+      "loss": 0.3682,
+      "step": 82272
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.283164684534132e-07,
+      "loss": 0.438,
+      "step": 82273
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2829376577829403e-07,
+      "loss": 0.3813,
+      "step": 82274
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.28271063540904e-07,
+      "loss": 0.4897,
+      "step": 82275
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.28248361741261e-07,
+      "loss": 0.4458,
+      "step": 82276
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2822566037938145e-07,
+      "loss": 0.3389,
+      "step": 82277
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.282029594552833e-07,
+      "loss": 0.4331,
+      "step": 82278
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2818025896898393e-07,
+      "loss": 0.4519,
+      "step": 82279
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2815755892050085e-07,
+      "loss": 0.2495,
+      "step": 82280
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.281348593098513e-07,
+      "loss": 0.4526,
+      "step": 82281
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.281121601370522e-07,
+      "loss": 0.406,
+      "step": 82282
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2808946140212145e-07,
+      "loss": 0.3585,
+      "step": 82283
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.280667631050762e-07,
+      "loss": 0.4961,
+      "step": 82284
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2804406524593426e-07,
+      "loss": 0.4214,
+      "step": 82285
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2802136782471233e-07,
+      "loss": 0.2615,
+      "step": 82286
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2799867084142814e-07,
+      "loss": 0.3612,
+      "step": 82287
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.279759742960991e-07,
+      "loss": 0.4482,
+      "step": 82288
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.279532781887429e-07,
+      "loss": 0.415,
+      "step": 82289
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.279305825193763e-07,
+      "loss": 0.4771,
+      "step": 82290
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.279078872880172e-07,
+      "loss": 0.3945,
+      "step": 82291
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2788519249468247e-07,
+      "loss": 0.4482,
+      "step": 82292
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2786249813939e-07,
+      "loss": 0.1985,
+      "step": 82293
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.278398042221566e-07,
+      "loss": 0.3542,
+      "step": 82294
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2781711074299986e-07,
+      "loss": 0.324,
+      "step": 82295
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2779441770193734e-07,
+      "loss": 0.3618,
+      "step": 82296
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2777172509898663e-07,
+      "loss": 0.3418,
+      "step": 82297
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.277490329341645e-07,
+      "loss": 0.4258,
+      "step": 82298
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.277263412074885e-07,
+      "loss": 0.3887,
+      "step": 82299
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.277036499189763e-07,
+      "loss": 0.5103,
+      "step": 82300
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.276809590686453e-07,
+      "loss": 0.3403,
+      "step": 82301
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2765826865651276e-07,
+      "loss": 0.3159,
+      "step": 82302
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.276355786825957e-07,
+      "loss": 0.3525,
+      "step": 82303
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.276128891469116e-07,
+      "loss": 0.4927,
+      "step": 82304
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.275902000494781e-07,
+      "loss": 0.4536,
+      "step": 82305
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2756751139031276e-07,
+      "loss": 0.4155,
+      "step": 82306
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.275448231694323e-07,
+      "loss": 0.4399,
+      "step": 82307
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.275221353868544e-07,
+      "loss": 0.3555,
+      "step": 82308
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2749944804259696e-07,
+      "loss": 0.4307,
+      "step": 82309
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.274767611366764e-07,
+      "loss": 0.4355,
+      "step": 82310
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2745407466911096e-07,
+      "loss": 0.3843,
+      "step": 82311
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2743138863991726e-07,
+      "loss": 0.4209,
+      "step": 82312
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2740870304911303e-07,
+      "loss": 0.3867,
+      "step": 82313
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.273860178967159e-07,
+      "loss": 0.4438,
+      "step": 82314
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2736333318274276e-07,
+      "loss": 0.5005,
+      "step": 82315
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.27340648907211e-07,
+      "loss": 0.312,
+      "step": 82316
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2731796507013827e-07,
+      "loss": 0.3936,
+      "step": 82317
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.272952816715423e-07,
+      "loss": 0.4351,
+      "step": 82318
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.272725987114395e-07,
+      "loss": 0.4624,
+      "step": 82319
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2724991618984775e-07,
+      "loss": 0.3962,
+      "step": 82320
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2722723410678483e-07,
+      "loss": 0.6025,
+      "step": 82321
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.272045524622673e-07,
+      "loss": 0.3618,
+      "step": 82322
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2718187125631323e-07,
+      "loss": 0.4673,
+      "step": 82323
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2715919048893933e-07,
+      "loss": 0.4604,
+      "step": 82324
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.271365101601633e-07,
+      "loss": 0.4336,
+      "step": 82325
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.271138302700028e-07,
+      "loss": 0.3345,
+      "step": 82326
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.270911508184746e-07,
+      "loss": 0.3916,
+      "step": 82327
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2706847180559634e-07,
+      "loss": 0.4336,
+      "step": 82328
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2704579323138546e-07,
+      "loss": 0.3467,
+      "step": 82329
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2702311509585956e-07,
+      "loss": 0.4902,
+      "step": 82330
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.270004373990357e-07,
+      "loss": 0.4907,
+      "step": 82331
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.26977760140931e-07,
+      "loss": 0.396,
+      "step": 82332
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2695508332156294e-07,
+      "loss": 0.4751,
+      "step": 82333
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.269324069409491e-07,
+      "loss": 0.4287,
+      "step": 82334
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.269097309991072e-07,
+      "loss": 0.4277,
+      "step": 82335
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2688705549605377e-07,
+      "loss": 0.4529,
+      "step": 82336
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.268643804318065e-07,
+      "loss": 0.4263,
+      "step": 82337
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.268417058063828e-07,
+      "loss": 0.3345,
+      "step": 82338
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2681903161980046e-07,
+      "loss": 0.3936,
+      "step": 82339
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.267963578720761e-07,
+      "loss": 0.4219,
+      "step": 82340
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.267736845632278e-07,
+      "loss": 0.3854,
+      "step": 82341
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.267510116932721e-07,
+      "loss": 0.4102,
+      "step": 82342
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.267283392622272e-07,
+      "loss": 0.6729,
+      "step": 82343
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.267056672701097e-07,
+      "loss": 0.3865,
+      "step": 82344
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2668299571693723e-07,
+      "loss": 0.4277,
+      "step": 82345
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2666032460272725e-07,
+      "loss": 0.4351,
+      "step": 82346
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.266376539274975e-07,
+      "loss": 0.3555,
+      "step": 82347
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.266149836912646e-07,
+      "loss": 0.3401,
+      "step": 82348
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2659231389404617e-07,
+      "loss": 0.4331,
+      "step": 82349
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2656964453585964e-07,
+      "loss": 0.3652,
+      "step": 82350
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2654697561672284e-07,
+      "loss": 0.458,
+      "step": 82351
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2652430713665256e-07,
+      "loss": 0.405,
+      "step": 82352
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2650163909566584e-07,
+      "loss": 0.4678,
+      "step": 82353
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.264789714937805e-07,
+      "loss": 0.4399,
+      "step": 82354
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2645630433101387e-07,
+      "loss": 0.4202,
+      "step": 82355
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.264336376073836e-07,
+      "loss": 0.3535,
+      "step": 82356
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2641097132290637e-07,
+      "loss": 0.2379,
+      "step": 82357
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2638830547759983e-07,
+      "loss": 0.392,
+      "step": 82358
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2636564007148147e-07,
+      "loss": 0.3994,
+      "step": 82359
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2634297510456895e-07,
+      "loss": 0.3877,
+      "step": 82360
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2632031057687913e-07,
+      "loss": 0.385,
+      "step": 82361
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2629764648842914e-07,
+      "loss": 0.3916,
+      "step": 82362
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.262749828392367e-07,
+      "loss": 0.4614,
+      "step": 82363
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2625231962931953e-07,
+      "loss": 0.4102,
+      "step": 82364
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2622965685869415e-07,
+      "loss": 0.3474,
+      "step": 82365
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.262069945273783e-07,
+      "loss": 0.3238,
+      "step": 82366
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2618433263538944e-07,
+      "loss": 0.2572,
+      "step": 82367
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2616167118274516e-07,
+      "loss": 0.4395,
+      "step": 82368
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2613901016946216e-07,
+      "loss": 0.4028,
+      "step": 82369
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.261163495955582e-07,
+      "loss": 0.4736,
+      "step": 82370
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2609368946105085e-07,
+      "loss": 0.4302,
+      "step": 82371
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.260710297659568e-07,
+      "loss": 0.4009,
+      "step": 82372
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.260483705102942e-07,
+      "loss": 0.3955,
+      "step": 82373
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.260257116940795e-07,
+      "loss": 0.3794,
+      "step": 82374
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2600305331733066e-07,
+      "loss": 0.4331,
+      "step": 82375
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.259803953800648e-07,
+      "loss": 0.448,
+      "step": 82376
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2595773788229973e-07,
+      "loss": 0.4268,
+      "step": 82377
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2593508082405205e-07,
+      "loss": 0.3442,
+      "step": 82378
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2591242420533957e-07,
+      "loss": 0.4487,
+      "step": 82379
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.258897680261798e-07,
+      "loss": 0.4169,
+      "step": 82380
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.258671122865899e-07,
+      "loss": 0.4639,
+      "step": 82381
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.258444569865868e-07,
+      "loss": 0.3979,
+      "step": 82382
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.258218021261881e-07,
+      "loss": 0.3135,
+      "step": 82383
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2579914770541136e-07,
+      "loss": 0.3531,
+      "step": 82384
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2577649372427413e-07,
+      "loss": 0.3779,
+      "step": 82385
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2575384018279317e-07,
+      "loss": 0.4834,
+      "step": 82386
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.257311870809859e-07,
+      "loss": 0.363,
+      "step": 82387
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2570853441887e-07,
+      "loss": 0.5117,
+      "step": 82388
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.25685882196463e-07,
+      "loss": 0.4155,
+      "step": 82389
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.256632304137815e-07,
+      "loss": 0.377,
+      "step": 82390
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.256405790708436e-07,
+      "loss": 0.4229,
+      "step": 82391
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2561792816766607e-07,
+      "loss": 0.2994,
+      "step": 82392
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.255952777042664e-07,
+      "loss": 0.2295,
+      "step": 82393
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2557262768066225e-07,
+      "loss": 0.4648,
+      "step": 82394
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.255499780968705e-07,
+      "loss": 0.3738,
+      "step": 82395
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2552732895290875e-07,
+      "loss": 0.376,
+      "step": 82396
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.255046802487946e-07,
+      "loss": 0.3405,
+      "step": 82397
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.254820319845448e-07,
+      "loss": 0.4175,
+      "step": 82398
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.254593841601769e-07,
+      "loss": 0.4636,
+      "step": 82399
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.254367367757085e-07,
+      "loss": 0.4346,
+      "step": 82400
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.25414089831157e-07,
+      "loss": 0.2913,
+      "step": 82401
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2539144332653953e-07,
+      "loss": 0.3589,
+      "step": 82402
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2536879726187303e-07,
+      "loss": 0.5571,
+      "step": 82403
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2534615163717524e-07,
+      "loss": 0.4297,
+      "step": 82404
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.253235064524635e-07,
+      "loss": 0.4033,
+      "step": 82405
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2530086170775546e-07,
+      "loss": 0.4277,
+      "step": 82406
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2527821740306767e-07,
+      "loss": 0.4966,
+      "step": 82407
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.252555735384181e-07,
+      "loss": 0.4375,
+      "step": 82408
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2523293011382375e-07,
+      "loss": 0.4043,
+      "step": 82409
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.252102871293026e-07,
+      "loss": 0.5044,
+      "step": 82410
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.251876445848715e-07,
+      "loss": 0.3638,
+      "step": 82411
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2516500248054743e-07,
+      "loss": 0.3665,
+      "step": 82412
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.25142360816348e-07,
+      "loss": 0.4395,
+      "step": 82413
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2511971959229067e-07,
+      "loss": 0.4297,
+      "step": 82414
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2509707880839316e-07,
+      "loss": 0.302,
+      "step": 82415
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2507443846467194e-07,
+      "loss": 0.417,
+      "step": 82416
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.250517985611448e-07,
+      "loss": 0.4478,
+      "step": 82417
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2502915909782945e-07,
+      "loss": 0.5928,
+      "step": 82418
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2500652007474245e-07,
+      "loss": 0.4917,
+      "step": 82419
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2498388149190155e-07,
+      "loss": 0.4224,
+      "step": 82420
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2496124334932435e-07,
+      "loss": 0.415,
+      "step": 82421
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.249386056470277e-07,
+      "loss": 0.3621,
+      "step": 82422
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.249159683850293e-07,
+      "loss": 0.4639,
+      "step": 82423
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2489333156334594e-07,
+      "loss": 0.4131,
+      "step": 82424
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2487069518199535e-07,
+      "loss": 0.4558,
+      "step": 82425
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2484805924099485e-07,
+      "loss": 0.502,
+      "step": 82426
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.248254237403621e-07,
+      "loss": 0.345,
+      "step": 82427
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2480278868011366e-07,
+      "loss": 0.4795,
+      "step": 82428
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2478015406026735e-07,
+      "loss": 0.3318,
+      "step": 82429
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.247575198808404e-07,
+      "loss": 0.4609,
+      "step": 82430
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.247348861418505e-07,
+      "loss": 0.2961,
+      "step": 82431
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.247122528433147e-07,
+      "loss": 0.4441,
+      "step": 82432
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.246896199852499e-07,
+      "loss": 0.4766,
+      "step": 82433
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2466698756767373e-07,
+      "loss": 0.3584,
+      "step": 82434
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.246443555906041e-07,
+      "loss": 0.3921,
+      "step": 82435
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.246217240540573e-07,
+      "loss": 0.3612,
+      "step": 82436
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.245990929580513e-07,
+      "loss": 0.3955,
+      "step": 82437
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2457646230260325e-07,
+      "loss": 0.3787,
+      "step": 82438
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2455383208773096e-07,
+      "loss": 0.3059,
+      "step": 82439
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2453120231345097e-07,
+      "loss": 0.2856,
+      "step": 82440
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2450857297978137e-07,
+      "loss": 0.3103,
+      "step": 82441
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.244859440867387e-07,
+      "loss": 0.4135,
+      "step": 82442
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.244633156343407e-07,
+      "loss": 0.3385,
+      "step": 82443
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2444068762260496e-07,
+      "loss": 0.4038,
+      "step": 82444
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2441806005154824e-07,
+      "loss": 0.4487,
+      "step": 82445
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2439543292118805e-07,
+      "loss": 0.3982,
+      "step": 82446
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.243728062315419e-07,
+      "loss": 0.3901,
+      "step": 82447
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2435017998262733e-07,
+      "loss": 0.4268,
+      "step": 82448
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.243275541744611e-07,
+      "loss": 0.4883,
+      "step": 82449
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.243049288070607e-07,
+      "loss": 0.3843,
+      "step": 82450
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.242823038804439e-07,
+      "loss": 0.358,
+      "step": 82451
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.242596793946277e-07,
+      "loss": 0.3162,
+      "step": 82452
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2423705534962905e-07,
+      "loss": 0.4238,
+      "step": 82453
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.242144317454656e-07,
+      "loss": 0.4824,
+      "step": 82454
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2419180858215466e-07,
+      "loss": 0.4678,
+      "step": 82455
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2416918585971396e-07,
+      "loss": 0.5044,
+      "step": 82456
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2414656357816004e-07,
+      "loss": 0.3955,
+      "step": 82457
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2412394173751065e-07,
+      "loss": 0.5078,
+      "step": 82458
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.24101320337783e-07,
+      "loss": 0.4097,
+      "step": 82459
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.24078699378995e-07,
+      "loss": 0.4746,
+      "step": 82460
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.240560788611633e-07,
+      "loss": 0.4683,
+      "step": 82461
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.240334587843052e-07,
+      "loss": 0.2338,
+      "step": 82462
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2401083914843805e-07,
+      "loss": 0.3672,
+      "step": 82463
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2398821995357934e-07,
+      "loss": 0.5132,
+      "step": 82464
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2396560119974676e-07,
+      "loss": 0.4685,
+      "step": 82465
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2394298288695683e-07,
+      "loss": 0.4434,
+      "step": 82466
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.239203650152273e-07,
+      "loss": 0.3609,
+      "step": 82467
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.238977475845754e-07,
+      "loss": 0.4001,
+      "step": 82468
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.23875130595019e-07,
+      "loss": 0.4353,
+      "step": 82469
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2385251404657484e-07,
+      "loss": 0.416,
+      "step": 82470
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2382989793925996e-07,
+      "loss": 0.4326,
+      "step": 82471
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.23807282273092e-07,
+      "loss": 0.5464,
+      "step": 82472
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.237846670480887e-07,
+      "loss": 0.4585,
+      "step": 82473
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2376205226426667e-07,
+      "loss": 0.3619,
+      "step": 82474
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2373943792164345e-07,
+      "loss": 0.3767,
+      "step": 82475
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2371682402023647e-07,
+      "loss": 0.4595,
+      "step": 82476
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.236942105600634e-07,
+      "loss": 0.3804,
+      "step": 82477
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2367159754114077e-07,
+      "loss": 0.5088,
+      "step": 82478
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2364898496348646e-07,
+      "loss": 0.3232,
+      "step": 82479
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.236263728271178e-07,
+      "loss": 0.5439,
+      "step": 82480
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2360376113205163e-07,
+      "loss": 0.5037,
+      "step": 82481
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2358114987830596e-07,
+      "loss": 0.5591,
+      "step": 82482
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.235585390658972e-07,
+      "loss": 0.543,
+      "step": 82483
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2353592869484335e-07,
+      "loss": 0.4888,
+      "step": 82484
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2351331876516137e-07,
+      "loss": 0.3436,
+      "step": 82485
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2349070927686915e-07,
+      "loss": 0.4224,
+      "step": 82486
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2346810022998327e-07,
+      "loss": 0.3716,
+      "step": 82487
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2344549162452126e-07,
+      "loss": 0.4644,
+      "step": 82488
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2342288346050105e-07,
+      "loss": 0.3887,
+      "step": 82489
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.234002757379392e-07,
+      "loss": 0.3633,
+      "step": 82490
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.233776684568531e-07,
+      "loss": 0.3521,
+      "step": 82491
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.233550616172601e-07,
+      "loss": 0.4497,
+      "step": 82492
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.233324552191776e-07,
+      "loss": 0.3262,
+      "step": 82493
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2330984926262334e-07,
+      "loss": 0.5342,
+      "step": 82494
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.232872437476138e-07,
+      "loss": 0.4185,
+      "step": 82495
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2326463867416674e-07,
+      "loss": 0.4331,
+      "step": 82496
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.232420340422994e-07,
+      "loss": 0.4351,
+      "step": 82497
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2321942985202954e-07,
+      "loss": 0.3484,
+      "step": 82498
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.231968261033736e-07,
+      "loss": 0.397,
+      "step": 82499
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2317422279634974e-07,
+      "loss": 0.4541,
+      "step": 82500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2315161993097447e-07,
+      "loss": 0.3596,
+      "step": 82501
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2312901750726547e-07,
+      "loss": 0.4424,
+      "step": 82502
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.231064155252404e-07,
+      "loss": 0.2982,
+      "step": 82503
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.230838139849159e-07,
+      "loss": 0.2219,
+      "step": 82504
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.230612128863096e-07,
+      "loss": 0.3444,
+      "step": 82505
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2303861222943915e-07,
+      "loss": 0.4009,
+      "step": 82506
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2301601201432113e-07,
+      "loss": 0.4434,
+      "step": 82507
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.229934122409733e-07,
+      "loss": 0.4351,
+      "step": 82508
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.229708129094127e-07,
+      "loss": 0.4814,
+      "step": 82509
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.229482140196573e-07,
+      "loss": 0.3994,
+      "step": 82510
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.229256155717239e-07,
+      "loss": 0.2307,
+      "step": 82511
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2290301756562954e-07,
+      "loss": 0.458,
+      "step": 82512
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.228804200013917e-07,
+      "loss": 0.3064,
+      "step": 82513
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.228578228790278e-07,
+      "loss": 0.4512,
+      "step": 82514
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.228352261985555e-07,
+      "loss": 0.3716,
+      "step": 82515
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2281262995999143e-07,
+      "loss": 0.3789,
+      "step": 82516
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.227900341633531e-07,
+      "loss": 0.3257,
+      "step": 82517
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 4.2276743880865795e-07,
+      "loss": 0.4104,
+      "step": 82518
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2274484389592356e-07,
+      "loss": 0.3188,
+      "step": 82519
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.22722249425167e-07,
+      "loss": 0.4751,
+      "step": 82520
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2269965539640497e-07,
+      "loss": 0.4346,
+      "step": 82521
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2267706180965535e-07,
+      "loss": 0.2236,
+      "step": 82522
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.226544686649357e-07,
+      "loss": 0.4893,
+      "step": 82523
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2263187596226256e-07,
+      "loss": 0.438,
+      "step": 82524
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.226092837016537e-07,
+      "loss": 0.3857,
+      "step": 82525
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2258669188312633e-07,
+      "loss": 0.5234,
+      "step": 82526
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.225641005066982e-07,
+      "loss": 0.4526,
+      "step": 82527
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.225415095723858e-07,
+      "loss": 0.3342,
+      "step": 82528
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2251891908020676e-07,
+      "loss": 0.3956,
+      "step": 82529
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2249632903017883e-07,
+      "loss": 0.3711,
+      "step": 82530
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2247373942231857e-07,
+      "loss": 0.4246,
+      "step": 82531
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.22451150256644e-07,
+      "loss": 0.457,
+      "step": 82532
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2242856153317164e-07,
+      "loss": 0.4053,
+      "step": 82533
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2240597325191907e-07,
+      "loss": 0.3503,
+      "step": 82534
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2238338541290384e-07,
+      "loss": 0.3645,
+      "step": 82535
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2236079801614335e-07,
+      "loss": 0.4292,
+      "step": 82536
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2233821106165434e-07,
+      "loss": 0.3848,
+      "step": 82537
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2231562454945433e-07,
+      "loss": 0.3156,
+      "step": 82538
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.222930384795608e-07,
+      "loss": 0.4399,
+      "step": 82539
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2227045285199146e-07,
+      "loss": 0.3418,
+      "step": 82540
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.222478676667626e-07,
+      "loss": 0.5522,
+      "step": 82541
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.222252829238918e-07,
+      "loss": 0.4678,
+      "step": 82542
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.222026986233965e-07,
+      "loss": 0.5889,
+      "step": 82543
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2218011476529446e-07,
+      "loss": 0.214,
+      "step": 82544
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.221575313496022e-07,
+      "loss": 0.3885,
+      "step": 82545
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2213494837633736e-07,
+      "loss": 0.2524,
+      "step": 82546
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2211236584551725e-07,
+      "loss": 0.4055,
+      "step": 82547
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.220897837571594e-07,
+      "loss": 0.3539,
+      "step": 82548
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.220672021112804e-07,
+      "loss": 0.4512,
+      "step": 82549
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.220446209078985e-07,
+      "loss": 0.4316,
+      "step": 82550
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2202204014703005e-07,
+      "loss": 0.2832,
+      "step": 82551
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2199945982869266e-07,
+      "loss": 0.3761,
+      "step": 82552
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.219768799529042e-07,
+      "loss": 0.4438,
+      "step": 82553
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2195430051968094e-07,
+      "loss": 0.3273,
+      "step": 82554
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2193172152904086e-07,
+      "loss": 0.3936,
+      "step": 82555
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.21909142981001e-07,
+      "loss": 0.3618,
+      "step": 82556
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.218865648755791e-07,
+      "loss": 0.4673,
+      "step": 82557
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2186398721279173e-07,
+      "loss": 0.4302,
+      "step": 82558
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.218414099926565e-07,
+      "loss": 0.4062,
+      "step": 82559
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2181883321519105e-07,
+      "loss": 0.4521,
+      "step": 82560
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.217962568804124e-07,
+      "loss": 0.4307,
+      "step": 82561
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.217736809883373e-07,
+      "loss": 0.3745,
+      "step": 82562
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2175110553898365e-07,
+      "loss": 0.4541,
+      "step": 82563
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2172853053236843e-07,
+      "loss": 0.428,
+      "step": 82564
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2170595596850957e-07,
+      "loss": 0.3882,
+      "step": 82565
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2168338184742347e-07,
+      "loss": 0.4478,
+      "step": 82566
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2166080816912786e-07,
+      "loss": 0.4321,
+      "step": 82567
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.216382349336399e-07,
+      "loss": 0.3508,
+      "step": 82568
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2161566214097733e-07,
+      "loss": 0.417,
+      "step": 82569
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2159308979115714e-07,
+      "loss": 0.2996,
+      "step": 82570
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2157051788419605e-07,
+      "loss": 0.3213,
+      "step": 82571
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.215479464201118e-07,
+      "loss": 0.4243,
+      "step": 82572
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2152537539892176e-07,
+      "loss": 0.213,
+      "step": 82573
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.215028048206435e-07,
+      "loss": 0.356,
+      "step": 82574
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.214802346852936e-07,
+      "loss": 0.46,
+      "step": 82575
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.214576649928897e-07,
+      "loss": 0.5615,
+      "step": 82576
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.214350957434489e-07,
+      "loss": 0.303,
+      "step": 82577
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2141252693698913e-07,
+      "loss": 0.3389,
+      "step": 82578
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.213899585735269e-07,
+      "loss": 0.4102,
+      "step": 82579
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2136739065308e-07,
+      "loss": 0.499,
+      "step": 82580
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.213448231756652e-07,
+      "loss": 0.3623,
+      "step": 82581
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2132225614130034e-07,
+      "loss": 0.246,
+      "step": 82582
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.212996895500022e-07,
+      "loss": 0.4644,
+      "step": 82583
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2127712340178814e-07,
+      "loss": 0.364,
+      "step": 82584
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.212545576966756e-07,
+      "loss": 0.2921,
+      "step": 82585
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.212319924346822e-07,
+      "loss": 0.3506,
+      "step": 82586
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2120942761582454e-07,
+      "loss": 0.4316,
+      "step": 82587
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.211868632401202e-07,
+      "loss": 0.3386,
+      "step": 82588
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2116429930758643e-07,
+      "loss": 0.3967,
+      "step": 82589
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.21141735818241e-07,
+      "loss": 0.5249,
+      "step": 82590
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.211191727721006e-07,
+      "loss": 0.3345,
+      "step": 82591
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.210966101691823e-07,
+      "loss": 0.397,
+      "step": 82592
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2107404800950373e-07,
+      "loss": 0.4536,
+      "step": 82593
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2105148629308205e-07,
+      "loss": 0.3462,
+      "step": 82594
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2102892501993516e-07,
+      "loss": 0.3171,
+      "step": 82595
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.210063641900793e-07,
+      "loss": 0.4448,
+      "step": 82596
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.209838038035323e-07,
+      "loss": 0.4902,
+      "step": 82597
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.209612438603117e-07,
+      "loss": 0.3755,
+      "step": 82598
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.209386843604342e-07,
+      "loss": 0.3837,
+      "step": 82599
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.209161253039176e-07,
+      "loss": 0.478,
+      "step": 82600
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.208935666907785e-07,
+      "loss": 0.3967,
+      "step": 82601
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.208710085210346e-07,
+      "loss": 0.4058,
+      "step": 82602
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2084845079470357e-07,
+      "loss": 0.3732,
+      "step": 82603
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2082589351180173e-07,
+      "loss": 0.4175,
+      "step": 82604
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.20803336672347e-07,
+      "loss": 0.3943,
+      "step": 82605
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2078078027635653e-07,
+      "loss": 0.3765,
+      "step": 82606
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.207582243238479e-07,
+      "loss": 0.4536,
+      "step": 82607
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.207356688148377e-07,
+      "loss": 0.3647,
+      "step": 82608
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.207131137493436e-07,
+      "loss": 0.3469,
+      "step": 82609
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.206905591273832e-07,
+      "loss": 0.4062,
+      "step": 82610
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.206680049489729e-07,
+      "loss": 0.4834,
+      "step": 82611
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.20645451214131e-07,
+      "loss": 0.3082,
+      "step": 82612
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.206228979228739e-07,
+      "loss": 0.5039,
+      "step": 82613
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2060034507521904e-07,
+      "loss": 0.4072,
+      "step": 82614
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.205777926711843e-07,
+      "loss": 0.3823,
+      "step": 82615
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.205552407107862e-07,
+      "loss": 0.4653,
+      "step": 82616
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2053268919404215e-07,
+      "loss": 0.2732,
+      "step": 82617
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2051013812096967e-07,
+      "loss": 0.395,
+      "step": 82618
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2048758749158633e-07,
+      "loss": 0.4604,
+      "step": 82619
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2046503730590897e-07,
+      "loss": 0.3584,
+      "step": 82620
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2044248756395453e-07,
+      "loss": 0.4722,
+      "step": 82621
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2041993826574064e-07,
+      "loss": 0.3889,
+      "step": 82622
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2039738941128445e-07,
+      "loss": 0.3801,
+      "step": 82623
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.203748410006038e-07,
+      "loss": 0.5107,
+      "step": 82624
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.203522930337151e-07,
+      "loss": 0.3877,
+      "step": 82625
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.203297455106359e-07,
+      "loss": 0.224,
+      "step": 82626
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2030719843138375e-07,
+      "loss": 0.4316,
+      "step": 82627
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2028465179597596e-07,
+      "loss": 0.4668,
+      "step": 82628
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.202621056044292e-07,
+      "loss": 0.4434,
+      "step": 82629
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.202395598567614e-07,
+      "loss": 0.4016,
+      "step": 82630
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.202170145529893e-07,
+      "loss": 0.2842,
+      "step": 82631
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.201944696931307e-07,
+      "loss": 0.3557,
+      "step": 82632
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2017192527720205e-07,
+      "loss": 0.3459,
+      "step": 82633
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2014938130522117e-07,
+      "loss": 0.5405,
+      "step": 82634
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2012683777720525e-07,
+      "loss": 0.3887,
+      "step": 82635
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2010429469317187e-07,
+      "loss": 0.406,
+      "step": 82636
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2008175205313765e-07,
+      "loss": 0.4688,
+      "step": 82637
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.2005920985712016e-07,
+      "loss": 0.2112,
+      "step": 82638
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.200366681051366e-07,
+      "loss": 0.418,
+      "step": 82639
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.200141267972047e-07,
+      "loss": 0.4419,
+      "step": 82640
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.199915859333414e-07,
+      "loss": 0.3616,
+      "step": 82641
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.199690455135635e-07,
+      "loss": 0.2949,
+      "step": 82642
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1994650553788845e-07,
+      "loss": 0.4763,
+      "step": 82643
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1992396600633385e-07,
+      "loss": 0.3931,
+      "step": 82644
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1990142691891717e-07,
+      "loss": 0.4717,
+      "step": 82645
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1987888827565487e-07,
+      "loss": 0.5439,
+      "step": 82646
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1985635007656463e-07,
+      "loss": 0.5239,
+      "step": 82647
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.198338123216637e-07,
+      "loss": 0.4307,
+      "step": 82648
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.198112750109698e-07,
+      "loss": 0.4141,
+      "step": 82649
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1978873814449964e-07,
+      "loss": 0.3623,
+      "step": 82650
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1976620172227017e-07,
+      "loss": 0.5,
+      "step": 82651
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.197436657442991e-07,
+      "loss": 0.3068,
+      "step": 82652
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1972113021060397e-07,
+      "loss": 0.3616,
+      "step": 82653
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1969859512120133e-07,
+      "loss": 0.2964,
+      "step": 82654
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.196760604761088e-07,
+      "loss": 0.3719,
+      "step": 82655
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.196535262753437e-07,
+      "loss": 0.4712,
+      "step": 82656
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.196309925189234e-07,
+      "loss": 0.2635,
+      "step": 82657
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.196084592068647e-07,
+      "loss": 0.4697,
+      "step": 82658
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.195859263391851e-07,
+      "loss": 0.4492,
+      "step": 82659
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1956339391590224e-07,
+      "loss": 0.356,
+      "step": 82660
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1954086193703264e-07,
+      "loss": 0.519,
+      "step": 82661
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.195183304025943e-07,
+      "loss": 0.2437,
+      "step": 82662
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1949579931260373e-07,
+      "loss": 0.4805,
+      "step": 82663
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1947326866707845e-07,
+      "loss": 0.3569,
+      "step": 82664
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.194507384660358e-07,
+      "loss": 0.3838,
+      "step": 82665
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.194282087094935e-07,
+      "loss": 0.5352,
+      "step": 82666
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1940567939746785e-07,
+      "loss": 0.4233,
+      "step": 82667
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1938315052997664e-07,
+      "loss": 0.4819,
+      "step": 82668
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.193606221070374e-07,
+      "loss": 0.3668,
+      "step": 82669
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1933809412866697e-07,
+      "loss": 0.4331,
+      "step": 82670
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1931556659488244e-07,
+      "loss": 0.4448,
+      "step": 82671
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1929303950570117e-07,
+      "loss": 0.5693,
+      "step": 82672
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1927051286114047e-07,
+      "loss": 0.3572,
+      "step": 82673
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1924798666121806e-07,
+      "loss": 0.4136,
+      "step": 82674
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1922546090595047e-07,
+      "loss": 0.4736,
+      "step": 82675
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1920293559535516e-07,
+      "loss": 0.3136,
+      "step": 82676
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1918041072944956e-07,
+      "loss": 0.2109,
+      "step": 82677
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1915788630825097e-07,
+      "loss": 0.3084,
+      "step": 82678
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.191353623317763e-07,
+      "loss": 0.2535,
+      "step": 82679
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1911283880004324e-07,
+      "loss": 0.3701,
+      "step": 82680
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.190903157130685e-07,
+      "loss": 0.457,
+      "step": 82681
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.190677930708695e-07,
+      "loss": 0.4365,
+      "step": 82682
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1904527087346395e-07,
+      "loss": 0.3896,
+      "step": 82683
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1902274912086845e-07,
+      "loss": 0.6333,
+      "step": 82684
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.190002278131004e-07,
+      "loss": 0.4232,
+      "step": 82685
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1897770695017763e-07,
+      "loss": 0.4561,
+      "step": 82686
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1895518653211647e-07,
+      "loss": 0.3843,
+      "step": 82687
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.189326665589347e-07,
+      "loss": 0.3477,
+      "step": 82688
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.189101470306493e-07,
+      "loss": 0.3003,
+      "step": 82689
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1888762794727817e-07,
+      "loss": 0.5327,
+      "step": 82690
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.188651093088381e-07,
+      "loss": 0.3892,
+      "step": 82691
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.188425911153458e-07,
+      "loss": 0.4634,
+      "step": 82692
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1882007336681915e-07,
+      "loss": 0.4888,
+      "step": 82693
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1879755606327515e-07,
+      "loss": 0.4492,
+      "step": 82694
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1877503920473157e-07,
+      "loss": 0.4116,
+      "step": 82695
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1875252279120475e-07,
+      "loss": 0.366,
+      "step": 82696
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1873000682271253e-07,
+      "loss": 0.4976,
+      "step": 82697
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1870749129927185e-07,
+      "loss": 0.3777,
+      "step": 82698
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.186849762209006e-07,
+      "loss": 0.3359,
+      "step": 82699
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1866246158761555e-07,
+      "loss": 0.3918,
+      "step": 82700
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1863994739943354e-07,
+      "loss": 0.3457,
+      "step": 82701
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1861743365637216e-07,
+      "loss": 0.3281,
+      "step": 82702
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.18594920358449e-07,
+      "loss": 0.4111,
+      "step": 82703
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.185724075056807e-07,
+      "loss": 0.342,
+      "step": 82704
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.185498950980847e-07,
+      "loss": 0.4961,
+      "step": 82705
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.185273831356784e-07,
+      "loss": 0.3665,
+      "step": 82706
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1850487161847925e-07,
+      "loss": 0.4941,
+      "step": 82707
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1848236054650376e-07,
+      "loss": 0.4189,
+      "step": 82708
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1845984991976966e-07,
+      "loss": 0.418,
+      "step": 82709
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.184373397382944e-07,
+      "loss": 0.3555,
+      "step": 82710
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1841483000209466e-07,
+      "loss": 0.3841,
+      "step": 82711
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1839232071118837e-07,
+      "loss": 0.4961,
+      "step": 82712
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.183698118655918e-07,
+      "loss": 0.3951,
+      "step": 82713
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.183473034653228e-07,
+      "loss": 0.479,
+      "step": 82714
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.183247955103986e-07,
+      "loss": 0.4141,
+      "step": 82715
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1830228800083657e-07,
+      "loss": 0.2834,
+      "step": 82716
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.182797809366535e-07,
+      "loss": 0.3994,
+      "step": 82717
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1825727431786674e-07,
+      "loss": 0.4761,
+      "step": 82718
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1823476814449365e-07,
+      "loss": 0.457,
+      "step": 82719
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1821226241655226e-07,
+      "loss": 0.4189,
+      "step": 82720
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1818975713405824e-07,
+      "loss": 0.3533,
+      "step": 82721
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1816725229702955e-07,
+      "loss": 0.3308,
+      "step": 82722
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.181447479054835e-07,
+      "loss": 0.394,
+      "step": 82723
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.181222439594375e-07,
+      "loss": 0.3923,
+      "step": 82724
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.180997404589083e-07,
+      "loss": 0.4299,
+      "step": 82725
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.180772374039133e-07,
+      "loss": 0.4036,
+      "step": 82726
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.180547347944698e-07,
+      "loss": 0.4111,
+      "step": 82727
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1803223263059537e-07,
+      "loss": 0.5005,
+      "step": 82728
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.180097309123066e-07,
+      "loss": 0.4546,
+      "step": 82729
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.179872296396214e-07,
+      "loss": 0.3779,
+      "step": 82730
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.179647288125562e-07,
+      "loss": 0.4062,
+      "step": 82731
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.179422284311287e-07,
+      "loss": 0.3447,
+      "step": 82732
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.179197284953564e-07,
+      "loss": 0.373,
+      "step": 82733
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.178972290052558e-07,
+      "loss": 0.4189,
+      "step": 82734
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.178747299608445e-07,
+      "loss": 0.4243,
+      "step": 82735
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.178522313621399e-07,
+      "loss": 0.5938,
+      "step": 82736
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.178297332091593e-07,
+      "loss": 0.4204,
+      "step": 82737
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.178072355019194e-07,
+      "loss": 0.4138,
+      "step": 82738
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.177847382404377e-07,
+      "loss": 0.3682,
+      "step": 82739
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1776224142473184e-07,
+      "loss": 0.4253,
+      "step": 82740
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1773974505481866e-07,
+      "loss": 0.4214,
+      "step": 82741
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1771724913071506e-07,
+      "loss": 0.4338,
+      "step": 82742
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.176947536524386e-07,
+      "loss": 0.4902,
+      "step": 82743
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1767225862000646e-07,
+      "loss": 0.4253,
+      "step": 82744
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1764976403343623e-07,
+      "loss": 0.3813,
+      "step": 82745
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.176272698927443e-07,
+      "loss": 0.3247,
+      "step": 82746
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1760477619794863e-07,
+      "loss": 0.3989,
+      "step": 82747
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1758228294906604e-07,
+      "loss": 0.1436,
+      "step": 82748
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.175597901461144e-07,
+      "loss": 0.4263,
+      "step": 82749
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.175372977891104e-07,
+      "loss": 0.3298,
+      "step": 82750
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.17514805878071e-07,
+      "loss": 0.2788,
+      "step": 82751
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.174923144130136e-07,
+      "loss": 0.4163,
+      "step": 82752
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1746982339395564e-07,
+      "loss": 0.4006,
+      "step": 82753
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.174473328209146e-07,
+      "loss": 0.3382,
+      "step": 82754
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.17424842693907e-07,
+      "loss": 0.3911,
+      "step": 82755
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1740235301295037e-07,
+      "loss": 0.3402,
+      "step": 82756
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.17379863778062e-07,
+      "loss": 0.3337,
+      "step": 82757
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.173573749892595e-07,
+      "loss": 0.3152,
+      "step": 82758
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.173348866465596e-07,
+      "loss": 0.3259,
+      "step": 82759
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1731239874997916e-07,
+      "loss": 0.3779,
+      "step": 82760
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1728991129953594e-07,
+      "loss": 0.3904,
+      "step": 82761
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.172674242952473e-07,
+      "loss": 0.4312,
+      "step": 82762
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1724493773712985e-07,
+      "loss": 0.2676,
+      "step": 82763
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.172224516252013e-07,
+      "loss": 0.4438,
+      "step": 82764
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.171999659594786e-07,
+      "loss": 0.3815,
+      "step": 82765
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1717748073997947e-07,
+      "loss": 0.4741,
+      "step": 82766
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.171549959667203e-07,
+      "loss": 0.4141,
+      "step": 82767
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1713251163971886e-07,
+      "loss": 0.3677,
+      "step": 82768
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.171100277589926e-07,
+      "loss": 0.4658,
+      "step": 82769
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1708754432455806e-07,
+      "loss": 0.4453,
+      "step": 82770
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1706506133643306e-07,
+      "loss": 0.4736,
+      "step": 82771
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1704257879463423e-07,
+      "loss": 0.3296,
+      "step": 82772
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1702009669917905e-07,
+      "loss": 0.3657,
+      "step": 82773
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.169976150500848e-07,
+      "loss": 0.3979,
+      "step": 82774
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1697513384736905e-07,
+      "loss": 0.4526,
+      "step": 82775
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1695265309104834e-07,
+      "loss": 0.4478,
+      "step": 82776
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.169301727811401e-07,
+      "loss": 0.4155,
+      "step": 82777
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1690769291766193e-07,
+      "loss": 0.4583,
+      "step": 82778
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.168852135006308e-07,
+      "loss": 0.4629,
+      "step": 82779
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.168627345300634e-07,
+      "loss": 0.3276,
+      "step": 82780
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.168402560059774e-07,
+      "loss": 0.4253,
+      "step": 82781
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.168177779283901e-07,
+      "loss": 0.4922,
+      "step": 82782
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.167953002973189e-07,
+      "loss": 0.3975,
+      "step": 82783
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.167728231127804e-07,
+      "loss": 0.4082,
+      "step": 82784
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1675034637479213e-07,
+      "loss": 0.3335,
+      "step": 82785
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.167278700833713e-07,
+      "loss": 0.4111,
+      "step": 82786
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1670539423853543e-07,
+      "loss": 0.3877,
+      "step": 82787
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1668291884030104e-07,
+      "loss": 0.4116,
+      "step": 82788
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1666044388868616e-07,
+      "loss": 0.3489,
+      "step": 82789
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.166379693837072e-07,
+      "loss": 0.4546,
+      "step": 82790
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1661549532538167e-07,
+      "loss": 0.4189,
+      "step": 82791
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.165930217137272e-07,
+      "loss": 0.3459,
+      "step": 82792
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.165705485487603e-07,
+      "loss": 0.4287,
+      "step": 82793
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1654807583049845e-07,
+      "loss": 0.2396,
+      "step": 82794
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.165256035589594e-07,
+      "loss": 0.3682,
+      "step": 82795
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.165031317341593e-07,
+      "loss": 0.4844,
+      "step": 82796
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1648066035611604e-07,
+      "loss": 0.374,
+      "step": 82797
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.164581894248468e-07,
+      "loss": 0.4458,
+      "step": 82798
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.164357189403689e-07,
+      "loss": 0.4321,
+      "step": 82799
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1641324890269935e-07,
+      "loss": 0.2822,
+      "step": 82800
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.163907793118551e-07,
+      "loss": 0.3342,
+      "step": 82801
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.163683101678535e-07,
+      "loss": 0.2921,
+      "step": 82802
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.163458414707118e-07,
+      "loss": 0.3314,
+      "step": 82803
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1632337322044775e-07,
+      "loss": 0.3582,
+      "step": 82804
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1630090541707753e-07,
+      "loss": 0.3638,
+      "step": 82805
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1627843806061904e-07,
+      "loss": 0.3028,
+      "step": 82806
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1625597115108923e-07,
+      "loss": 0.52,
+      "step": 82807
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.162335046885057e-07,
+      "loss": 0.4897,
+      "step": 82808
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.162110386728854e-07,
+      "loss": 0.5034,
+      "step": 82809
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1618857310424507e-07,
+      "loss": 0.4927,
+      "step": 82810
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1616610798260233e-07,
+      "loss": 0.2697,
+      "step": 82811
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1614364330797467e-07,
+      "loss": 0.3853,
+      "step": 82812
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.161211790803787e-07,
+      "loss": 0.355,
+      "step": 82813
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1609871529983174e-07,
+      "loss": 0.3831,
+      "step": 82814
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.160762519663513e-07,
+      "loss": 0.4009,
+      "step": 82815
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1605378907995467e-07,
+      "loss": 0.479,
+      "step": 82816
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1603132664065855e-07,
+      "loss": 0.4819,
+      "step": 82817
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.160088646484804e-07,
+      "loss": 0.4243,
+      "step": 82818
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.159864031034377e-07,
+      "loss": 0.459,
+      "step": 82819
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.15963942005547e-07,
+      "loss": 0.3784,
+      "step": 82820
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.159414813548262e-07,
+      "loss": 0.4468,
+      "step": 82821
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1591902115129187e-07,
+      "loss": 0.3994,
+      "step": 82822
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1589656139496144e-07,
+      "loss": 0.4441,
+      "step": 82823
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1587410208585206e-07,
+      "loss": 0.4294,
+      "step": 82824
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.158516432239815e-07,
+      "loss": 0.3643,
+      "step": 82825
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.158291848093661e-07,
+      "loss": 0.4502,
+      "step": 82826
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1580672684202343e-07,
+      "loss": 0.5098,
+      "step": 82827
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.157842693219706e-07,
+      "loss": 0.4341,
+      "step": 82828
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.157618122492258e-07,
+      "loss": 0.3916,
+      "step": 82829
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.157393556238045e-07,
+      "loss": 0.479,
+      "step": 82830
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.157168994457246e-07,
+      "loss": 0.5293,
+      "step": 82831
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.156944437150035e-07,
+      "loss": 0.5757,
+      "step": 82832
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1567198843165876e-07,
+      "loss": 0.3542,
+      "step": 82833
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1564953359570667e-07,
+      "loss": 0.2935,
+      "step": 82834
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.156270792071648e-07,
+      "loss": 0.3816,
+      "step": 82835
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1560462526605044e-07,
+      "loss": 0.3057,
+      "step": 82836
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.155821717723811e-07,
+      "loss": 0.5869,
+      "step": 82837
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1555971872617324e-07,
+      "loss": 0.2991,
+      "step": 82838
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.155372661274449e-07,
+      "loss": 0.5552,
+      "step": 82839
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1551481397621234e-07,
+      "loss": 0.3813,
+      "step": 82840
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1549236227249316e-07,
+      "loss": 0.5,
+      "step": 82841
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.154699110163051e-07,
+      "loss": 0.4941,
+      "step": 82842
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.154474602076644e-07,
+      "loss": 0.4692,
+      "step": 82843
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.154250098465888e-07,
+      "loss": 0.4233,
+      "step": 82844
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1540255993309525e-07,
+      "loss": 0.4478,
+      "step": 82845
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1538011046720154e-07,
+      "loss": 0.4453,
+      "step": 82846
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1535766144892404e-07,
+      "loss": 0.3772,
+      "step": 82847
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.153352128782802e-07,
+      "loss": 0.4526,
+      "step": 82848
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1531276475528774e-07,
+      "loss": 0.458,
+      "step": 82849
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1529031707996345e-07,
+      "loss": 0.3566,
+      "step": 82850
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.152678698523241e-07,
+      "loss": 0.5703,
+      "step": 82851
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.152454230723872e-07,
+      "loss": 0.3706,
+      "step": 82852
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1522297674017003e-07,
+      "loss": 0.4272,
+      "step": 82853
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.152005308556901e-07,
+      "loss": 0.4238,
+      "step": 82854
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.151780854189638e-07,
+      "loss": 0.3038,
+      "step": 82855
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1515564043000886e-07,
+      "loss": 0.4888,
+      "step": 82856
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.151331958888423e-07,
+      "loss": 0.457,
+      "step": 82857
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.151107517954817e-07,
+      "loss": 0.3403,
+      "step": 82858
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.150883081499439e-07,
+      "loss": 0.343,
+      "step": 82859
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.150658649522458e-07,
+      "loss": 0.215,
+      "step": 82860
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1504342220240484e-07,
+      "loss": 0.4377,
+      "step": 82861
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.150209799004382e-07,
+      "loss": 0.3733,
+      "step": 82862
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1499853804636345e-07,
+      "loss": 0.3374,
+      "step": 82863
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.149760966401971e-07,
+      "loss": 0.3901,
+      "step": 82864
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.149536556819566e-07,
+      "loss": 0.3513,
+      "step": 82865
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1493121517165963e-07,
+      "loss": 0.4492,
+      "step": 82866
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.149087751093225e-07,
+      "loss": 0.3215,
+      "step": 82867
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.148863354949629e-07,
+      "loss": 0.4824,
+      "step": 82868
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.148638963285982e-07,
+      "loss": 0.2946,
+      "step": 82869
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.148414576102449e-07,
+      "loss": 0.2722,
+      "step": 82870
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1481901933992104e-07,
+      "loss": 0.2947,
+      "step": 82871
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.14796581517643e-07,
+      "loss": 0.3589,
+      "step": 82872
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1477414414342825e-07,
+      "loss": 0.4492,
+      "step": 82873
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1475170721729416e-07,
+      "loss": 0.4233,
+      "step": 82874
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.14729270739258e-07,
+      "loss": 0.2872,
+      "step": 82875
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1470683470933645e-07,
+      "loss": 0.5024,
+      "step": 82876
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1468439912754683e-07,
+      "loss": 0.5117,
+      "step": 82877
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.146619639939066e-07,
+      "loss": 0.3579,
+      "step": 82878
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1463952930843307e-07,
+      "loss": 0.4858,
+      "step": 82879
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.146170950711432e-07,
+      "loss": 0.405,
+      "step": 82880
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.145946612820537e-07,
+      "loss": 0.415,
+      "step": 82881
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1457222794118205e-07,
+      "loss": 0.3696,
+      "step": 82882
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.14549795048546e-07,
+      "loss": 0.4575,
+      "step": 82883
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.14527362604162e-07,
+      "loss": 0.2693,
+      "step": 82884
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1450493060804726e-07,
+      "loss": 0.2902,
+      "step": 82885
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.144824990602193e-07,
+      "loss": 0.4741,
+      "step": 82886
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1446006796069556e-07,
+      "loss": 0.222,
+      "step": 82887
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1443763730949234e-07,
+      "loss": 0.6152,
+      "step": 82888
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.144152071066277e-07,
+      "loss": 0.3163,
+      "step": 82889
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1439277735211806e-07,
+      "loss": 0.4033,
+      "step": 82890
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1437034804598104e-07,
+      "loss": 0.4307,
+      "step": 82891
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.14347919188234e-07,
+      "loss": 0.3992,
+      "step": 82892
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1432549077889336e-07,
+      "loss": 0.3835,
+      "step": 82893
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.143030628179769e-07,
+      "loss": 0.3706,
+      "step": 82894
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.142806353055016e-07,
+      "loss": 0.4746,
+      "step": 82895
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1425820824148503e-07,
+      "loss": 0.2032,
+      "step": 82896
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1423578162594377e-07,
+      "loss": 0.3481,
+      "step": 82897
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1421335545889513e-07,
+      "loss": 0.4908,
+      "step": 82898
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1419092974035673e-07,
+      "loss": 0.5044,
+      "step": 82899
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1416850447034547e-07,
+      "loss": 0.3345,
+      "step": 82900
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1414607964887806e-07,
+      "loss": 0.415,
+      "step": 82901
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1412365527597205e-07,
+      "loss": 0.4106,
+      "step": 82902
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1410123135164467e-07,
+      "loss": 0.502,
+      "step": 82903
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.140788078759133e-07,
+      "loss": 0.4751,
+      "step": 82904
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1405638484879457e-07,
+      "loss": 0.3896,
+      "step": 82905
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.140339622703058e-07,
+      "loss": 0.5156,
+      "step": 82906
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1401154014046444e-07,
+      "loss": 0.4858,
+      "step": 82907
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.1398911845928773e-07,
+      "loss": 0.3875,
+      "step": 82908
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 4.139666972267927e-07,
+      "loss": 0.2994,
+      "step": 82909
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.13944276442996e-07,
+      "loss": 0.4351,
+      "step": 82910
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.139218561079153e-07,
+      "loss": 0.3923,
+      "step": 82911
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1389943622156766e-07,
+      "loss": 0.4136,
+      "step": 82912
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.138770167839706e-07,
+      "loss": 0.3403,
+      "step": 82913
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1385459779514065e-07,
+      "loss": 0.3325,
+      "step": 82914
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1383217925509515e-07,
+      "loss": 0.4351,
+      "step": 82915
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.138097611638516e-07,
+      "loss": 0.498,
+      "step": 82916
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.137873435214272e-07,
+      "loss": 0.4775,
+      "step": 82917
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1376492632783843e-07,
+      "loss": 0.2937,
+      "step": 82918
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.137425095831034e-07,
+      "loss": 0.4292,
+      "step": 82919
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1372009328723834e-07,
+      "loss": 0.4453,
+      "step": 82920
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.136976774402613e-07,
+      "loss": 0.5107,
+      "step": 82921
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.136752620421885e-07,
+      "loss": 0.3464,
+      "step": 82922
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1365284709303773e-07,
+      "loss": 0.4917,
+      "step": 82923
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1363043259282593e-07,
+      "loss": 0.3379,
+      "step": 82924
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.136080185415707e-07,
+      "loss": 0.3452,
+      "step": 82925
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.135856049392885e-07,
+      "loss": 0.2823,
+      "step": 82926
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1356319178599685e-07,
+      "loss": 0.4922,
+      "step": 82927
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.135407790817129e-07,
+      "loss": 0.415,
+      "step": 82928
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1351836682645414e-07,
+      "loss": 0.3604,
+      "step": 82929
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.134959550202375e-07,
+      "loss": 0.4429,
+      "step": 82930
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.134735436630795e-07,
+      "loss": 0.3984,
+      "step": 82931
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1345113275499796e-07,
+      "loss": 0.4849,
+      "step": 82932
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.134287222960099e-07,
+      "loss": 0.436,
+      "step": 82933
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.134063122861329e-07,
+      "loss": 0.4771,
+      "step": 82934
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1338390272538327e-07,
+      "loss": 0.375,
+      "step": 82935
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.133614936137787e-07,
+      "loss": 0.3657,
+      "step": 82936
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.133390849513362e-07,
+      "loss": 0.3765,
+      "step": 82937
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1331667673807334e-07,
+      "loss": 0.2132,
+      "step": 82938
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.132942689740069e-07,
+      "loss": 0.3573,
+      "step": 82939
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1327186165915376e-07,
+      "loss": 0.4399,
+      "step": 82940
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1324945479353133e-07,
+      "loss": 0.4604,
+      "step": 82941
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.132270483771573e-07,
+      "loss": 0.4214,
+      "step": 82942
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1320464241004783e-07,
+      "loss": 0.439,
+      "step": 82943
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1318223689222063e-07,
+      "loss": 0.2936,
+      "step": 82944
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1315983182369285e-07,
+      "loss": 0.3651,
+      "step": 82945
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1313742720448184e-07,
+      "loss": 0.3807,
+      "step": 82946
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1311502303460435e-07,
+      "loss": 0.4961,
+      "step": 82947
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1309261931407757e-07,
+      "loss": 0.4149,
+      "step": 82948
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1307021604291913e-07,
+      "loss": 0.4814,
+      "step": 82949
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1304781322114544e-07,
+      "loss": 0.4128,
+      "step": 82950
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.130254108487745e-07,
+      "loss": 0.3241,
+      "step": 82951
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1300300892582275e-07,
+      "loss": 0.5166,
+      "step": 82952
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1298060745230746e-07,
+      "loss": 0.1882,
+      "step": 82953
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1295820642824595e-07,
+      "loss": 0.4702,
+      "step": 82954
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.129358058536557e-07,
+      "loss": 0.5869,
+      "step": 82955
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1291340572855313e-07,
+      "loss": 0.3978,
+      "step": 82956
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.128910060529559e-07,
+      "loss": 0.5542,
+      "step": 82957
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.128686068268813e-07,
+      "loss": 0.4468,
+      "step": 82958
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.128462080503462e-07,
+      "loss": 0.3953,
+      "step": 82959
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1282380972336737e-07,
+      "loss": 0.3376,
+      "step": 82960
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.128014118459624e-07,
+      "loss": 0.4419,
+      "step": 82961
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1277901441814845e-07,
+      "loss": 0.5098,
+      "step": 82962
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1275661743994283e-07,
+      "loss": 0.3828,
+      "step": 82963
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.127342209113622e-07,
+      "loss": 0.4604,
+      "step": 82964
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1271182483242395e-07,
+      "loss": 0.3823,
+      "step": 82965
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.126894292031452e-07,
+      "loss": 0.3094,
+      "step": 82966
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1266703402354364e-07,
+      "loss": 0.3708,
+      "step": 82967
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1264463929363547e-07,
+      "loss": 0.4546,
+      "step": 82968
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.126222450134387e-07,
+      "loss": 0.3735,
+      "step": 82969
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1259985118296956e-07,
+      "loss": 0.46,
+      "step": 82970
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1257745780224586e-07,
+      "loss": 0.4155,
+      "step": 82971
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.125550648712849e-07,
+      "loss": 0.5103,
+      "step": 82972
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1253267239010325e-07,
+      "loss": 0.4565,
+      "step": 82973
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.125102803587182e-07,
+      "loss": 0.436,
+      "step": 82974
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1248788877714735e-07,
+      "loss": 0.3591,
+      "step": 82975
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.124654976454073e-07,
+      "loss": 0.3071,
+      "step": 82976
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.124431069635153e-07,
+      "loss": 0.4487,
+      "step": 82977
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.124207167314886e-07,
+      "loss": 0.4727,
+      "step": 82978
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.123983269493447e-07,
+      "loss": 0.2476,
+      "step": 82979
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.123759376171004e-07,
+      "loss": 0.3599,
+      "step": 82980
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1235354873477246e-07,
+      "loss": 0.3208,
+      "step": 82981
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1233116030237833e-07,
+      "loss": 0.3882,
+      "step": 82982
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.123087723199353e-07,
+      "loss": 0.4487,
+      "step": 82983
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.122863847874607e-07,
+      "loss": 0.3518,
+      "step": 82984
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1226399770497113e-07,
+      "loss": 0.4121,
+      "step": 82985
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.122416110724839e-07,
+      "loss": 0.4482,
+      "step": 82986
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.122192248900163e-07,
+      "loss": 0.397,
+      "step": 82987
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1219683915758574e-07,
+      "loss": 0.3525,
+      "step": 82988
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.121744538752091e-07,
+      "loss": 0.2573,
+      "step": 82989
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.12152069042903e-07,
+      "loss": 0.406,
+      "step": 82990
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.121296846606851e-07,
+      "loss": 0.3861,
+      "step": 82991
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.121073007285728e-07,
+      "loss": 0.4541,
+      "step": 82992
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.120849172465825e-07,
+      "loss": 0.3799,
+      "step": 82993
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1206253421473183e-07,
+      "loss": 0.3867,
+      "step": 82994
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1204015163303774e-07,
+      "loss": 0.52,
+      "step": 82995
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.120177695015179e-07,
+      "loss": 0.4785,
+      "step": 82996
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.119953878201886e-07,
+      "loss": 0.3132,
+      "step": 82997
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1197300658906753e-07,
+      "loss": 0.4951,
+      "step": 82998
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1195062580817195e-07,
+      "loss": 0.5195,
+      "step": 82999
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1192824547751837e-07,
+      "loss": 0.417,
+      "step": 83000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.119058655971247e-07,
+      "loss": 0.4248,
+      "step": 83001
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.118834861670073e-07,
+      "loss": 0.356,
+      "step": 83002
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1186110718718356e-07,
+      "loss": 0.3394,
+      "step": 83003
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.118387286576709e-07,
+      "loss": 0.4407,
+      "step": 83004
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1181635057848654e-07,
+      "loss": 0.252,
+      "step": 83005
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1179397294964704e-07,
+      "loss": 0.4194,
+      "step": 83006
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.117715957711698e-07,
+      "loss": 0.4136,
+      "step": 83007
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1174921904307204e-07,
+      "loss": 0.4292,
+      "step": 83008
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1172684276537164e-07,
+      "loss": 0.4368,
+      "step": 83009
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1170446693808413e-07,
+      "loss": 0.4556,
+      "step": 83010
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1168209156122747e-07,
+      "loss": 0.4736,
+      "step": 83011
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.116597166348188e-07,
+      "loss": 0.3511,
+      "step": 83012
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.116373421588756e-07,
+      "loss": 0.4658,
+      "step": 83013
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1161496813341424e-07,
+      "loss": 0.4292,
+      "step": 83014
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.115925945584522e-07,
+      "loss": 0.3711,
+      "step": 83015
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1157022143400674e-07,
+      "loss": 0.3994,
+      "step": 83016
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1154784876009516e-07,
+      "loss": 0.4624,
+      "step": 83017
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.115254765367341e-07,
+      "loss": 0.4202,
+      "step": 83018
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.115031047639412e-07,
+      "loss": 0.4214,
+      "step": 83019
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1148073344173307e-07,
+      "loss": 0.3887,
+      "step": 83020
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.11458362570127e-07,
+      "loss": 0.4282,
+      "step": 83021
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.114359921491405e-07,
+      "loss": 0.4993,
+      "step": 83022
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1141362217879014e-07,
+      "loss": 0.4019,
+      "step": 83023
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1139125265909325e-07,
+      "loss": 0.347,
+      "step": 83024
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1136888359006706e-07,
+      "loss": 0.3547,
+      "step": 83025
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.11346514971729e-07,
+      "loss": 0.395,
+      "step": 83026
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1132414680409543e-07,
+      "loss": 0.3541,
+      "step": 83027
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1130177908718386e-07,
+      "loss": 0.4331,
+      "step": 83028
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.112794118210119e-07,
+      "loss": 0.3376,
+      "step": 83029
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.112570450055962e-07,
+      "loss": 0.4209,
+      "step": 83030
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1123467864095363e-07,
+      "loss": 0.519,
+      "step": 83031
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1121231272710146e-07,
+      "loss": 0.3594,
+      "step": 83032
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.11189947264057e-07,
+      "loss": 0.3245,
+      "step": 83033
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.111675822518377e-07,
+      "loss": 0.4629,
+      "step": 83034
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1114521769045995e-07,
+      "loss": 0.4038,
+      "step": 83035
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1112285357994124e-07,
+      "loss": 0.3984,
+      "step": 83036
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1110048992029865e-07,
+      "loss": 0.4146,
+      "step": 83037
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1107812671154963e-07,
+      "loss": 0.457,
+      "step": 83038
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.110557639537111e-07,
+      "loss": 0.4297,
+      "step": 83039
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1103340164679976e-07,
+      "loss": 0.4177,
+      "step": 83040
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.11011039790833e-07,
+      "loss": 0.3159,
+      "step": 83041
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.109886783858281e-07,
+      "loss": 0.4312,
+      "step": 83042
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.109663174318023e-07,
+      "loss": 0.3975,
+      "step": 83043
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1094395692877216e-07,
+      "loss": 0.4468,
+      "step": 83044
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.109215968767552e-07,
+      "loss": 0.447,
+      "step": 83045
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1089923727576893e-07,
+      "loss": 0.3708,
+      "step": 83046
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.108768781258295e-07,
+      "loss": 0.407,
+      "step": 83047
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1085451942695504e-07,
+      "loss": 0.3679,
+      "step": 83048
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1083216117916175e-07,
+      "loss": 0.3955,
+      "step": 83049
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.108098033824673e-07,
+      "loss": 0.4399,
+      "step": 83050
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1078744603688896e-07,
+      "loss": 0.3162,
+      "step": 83051
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1076508914244325e-07,
+      "loss": 0.4468,
+      "step": 83052
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.107427326991475e-07,
+      "loss": 0.4585,
+      "step": 83053
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.107203767070191e-07,
+      "loss": 0.4873,
+      "step": 83054
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1069802116607533e-07,
+      "loss": 0.2744,
+      "step": 83055
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.106756660763325e-07,
+      "loss": 0.3135,
+      "step": 83056
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1065331143780844e-07,
+      "loss": 0.4741,
+      "step": 83057
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.106309572505202e-07,
+      "loss": 0.4399,
+      "step": 83058
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1060860351448454e-07,
+      "loss": 0.519,
+      "step": 83059
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1058625022971907e-07,
+      "loss": 0.3208,
+      "step": 83060
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.105638973962402e-07,
+      "loss": 0.458,
+      "step": 83061
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1054154501406547e-07,
+      "loss": 0.4646,
+      "step": 83062
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1051919308321234e-07,
+      "loss": 0.4038,
+      "step": 83063
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.104968416036972e-07,
+      "loss": 0.3723,
+      "step": 83064
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1047449057553753e-07,
+      "loss": 0.4253,
+      "step": 83065
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1045213999875053e-07,
+      "loss": 0.2243,
+      "step": 83066
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.104297898733534e-07,
+      "loss": 0.4423,
+      "step": 83067
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1040744019936314e-07,
+      "loss": 0.5171,
+      "step": 83068
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.103850909767964e-07,
+      "loss": 0.2869,
+      "step": 83069
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.103627422056708e-07,
+      "loss": 0.4048,
+      "step": 83070
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.103403938860033e-07,
+      "loss": 0.3235,
+      "step": 83071
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.103180460178113e-07,
+      "loss": 0.4075,
+      "step": 83072
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1029569860111136e-07,
+      "loss": 0.4497,
+      "step": 83073
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.102733516359209e-07,
+      "loss": 0.5298,
+      "step": 83074
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.102510051222571e-07,
+      "loss": 0.563,
+      "step": 83075
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1022865906013727e-07,
+      "loss": 0.3354,
+      "step": 83076
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.102063134495778e-07,
+      "loss": 0.2767,
+      "step": 83077
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1018396829059665e-07,
+      "loss": 0.4043,
+      "step": 83078
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1016162358321027e-07,
+      "loss": 0.4723,
+      "step": 83079
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1013927932743586e-07,
+      "loss": 0.347,
+      "step": 83080
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.101169355232911e-07,
+      "loss": 0.4775,
+      "step": 83081
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1009459217079236e-07,
+      "loss": 0.4385,
+      "step": 83082
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.10072249269957e-07,
+      "loss": 0.4575,
+      "step": 83083
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1004990682080253e-07,
+      "loss": 0.501,
+      "step": 83084
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.1002756482334544e-07,
+      "loss": 0.3926,
+      "step": 83085
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.100052232776031e-07,
+      "loss": 0.4087,
+      "step": 83086
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.099828821835926e-07,
+      "loss": 0.3324,
+      "step": 83087
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.099605415413314e-07,
+      "loss": 0.4126,
+      "step": 83088
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0993820135083635e-07,
+      "loss": 0.498,
+      "step": 83089
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.099158616121241e-07,
+      "loss": 0.4448,
+      "step": 83090
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0989352232521226e-07,
+      "loss": 0.3461,
+      "step": 83091
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.098711834901176e-07,
+      "loss": 0.4683,
+      "step": 83092
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.098488451068579e-07,
+      "loss": 0.4312,
+      "step": 83093
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0982650717544953e-07,
+      "loss": 0.4143,
+      "step": 83094
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.098041696959098e-07,
+      "loss": 0.4707,
+      "step": 83095
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0978183266825583e-07,
+      "loss": 0.3738,
+      "step": 83096
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0975949609250517e-07,
+      "loss": 0.4666,
+      "step": 83097
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0973715996867453e-07,
+      "loss": 0.2965,
+      "step": 83098
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.097148242967806e-07,
+      "loss": 0.2426,
+      "step": 83099
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0969248907684096e-07,
+      "loss": 0.4761,
+      "step": 83100
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0967015430887296e-07,
+      "loss": 0.3804,
+      "step": 83101
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.09647819992893e-07,
+      "loss": 0.4668,
+      "step": 83102
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0962548612891866e-07,
+      "loss": 0.4102,
+      "step": 83103
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0960315271696687e-07,
+      "loss": 0.397,
+      "step": 83104
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.095808197570552e-07,
+      "loss": 0.2809,
+      "step": 83105
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0955848724919983e-07,
+      "loss": 0.5127,
+      "step": 83106
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0953615519341857e-07,
+      "loss": 0.4001,
+      "step": 83107
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0951382358972855e-07,
+      "loss": 0.3174,
+      "step": 83108
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.094914924381463e-07,
+      "loss": 0.46,
+      "step": 83109
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.094691617386897e-07,
+      "loss": 0.3065,
+      "step": 83110
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0944683149137503e-07,
+      "loss": 0.4214,
+      "step": 83111
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.094245016962197e-07,
+      "loss": 0.416,
+      "step": 83112
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0940217235324095e-07,
+      "loss": 0.4126,
+      "step": 83113
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0937984346245616e-07,
+      "loss": 0.4534,
+      "step": 83114
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.093575150238816e-07,
+      "loss": 0.3156,
+      "step": 83115
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.093351870375349e-07,
+      "loss": 0.4678,
+      "step": 83116
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.093128595034331e-07,
+      "loss": 0.4238,
+      "step": 83117
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.092905324215941e-07,
+      "loss": 0.4047,
+      "step": 83118
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0926820579203326e-07,
+      "loss": 0.4463,
+      "step": 83119
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0924587961476875e-07,
+      "loss": 0.4153,
+      "step": 83120
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0922355388981746e-07,
+      "loss": 0.3105,
+      "step": 83121
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0920122861719685e-07,
+      "loss": 0.4272,
+      "step": 83122
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0917890379692333e-07,
+      "loss": 0.4536,
+      "step": 83123
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.091565794290144e-07,
+      "loss": 0.2917,
+      "step": 83124
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0913425551348714e-07,
+      "loss": 0.373,
+      "step": 83125
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0911193205035887e-07,
+      "loss": 0.4619,
+      "step": 83126
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0908960903964604e-07,
+      "loss": 0.4324,
+      "step": 83127
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.090672864813667e-07,
+      "loss": 0.4084,
+      "step": 83128
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.090449643755368e-07,
+      "loss": 0.3264,
+      "step": 83129
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0902264272217414e-07,
+      "loss": 0.4341,
+      "step": 83130
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.090003215212959e-07,
+      "loss": 0.2644,
+      "step": 83131
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.089780007729187e-07,
+      "loss": 0.3489,
+      "step": 83132
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0895568047705973e-07,
+      "loss": 0.3804,
+      "step": 83133
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.089333606337363e-07,
+      "loss": 0.4524,
+      "step": 83134
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0891104124296583e-07,
+      "loss": 0.4888,
+      "step": 83135
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0888872230476455e-07,
+      "loss": 0.4517,
+      "step": 83136
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0886640381915006e-07,
+      "loss": 0.4458,
+      "step": 83137
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.088440857861397e-07,
+      "loss": 0.3013,
+      "step": 83138
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.088217682057502e-07,
+      "loss": 0.4229,
+      "step": 83139
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.087994510779984e-07,
+      "loss": 0.3479,
+      "step": 83140
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.087771344029016e-07,
+      "loss": 0.3872,
+      "step": 83141
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0875481818047706e-07,
+      "loss": 0.3605,
+      "step": 83142
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.087325024107421e-07,
+      "loss": 0.5718,
+      "step": 83143
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.087101870937131e-07,
+      "loss": 0.3655,
+      "step": 83144
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.086878722294075e-07,
+      "loss": 0.3687,
+      "step": 83145
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.086655578178425e-07,
+      "loss": 0.5435,
+      "step": 83146
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0864324385903535e-07,
+      "loss": 0.4272,
+      "step": 83147
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0862093035300294e-07,
+      "loss": 0.5083,
+      "step": 83148
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0859861729976184e-07,
+      "loss": 0.3706,
+      "step": 83149
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.085763046993297e-07,
+      "loss": 0.293,
+      "step": 83150
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.085539925517234e-07,
+      "loss": 0.4048,
+      "step": 83151
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0853168085696057e-07,
+      "loss": 0.3564,
+      "step": 83152
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0850936961505734e-07,
+      "loss": 0.3638,
+      "step": 83153
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.084870588260314e-07,
+      "loss": 0.4292,
+      "step": 83154
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.084647484899001e-07,
+      "loss": 0.3119,
+      "step": 83155
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.084424386066797e-07,
+      "loss": 0.3955,
+      "step": 83156
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0842012917638767e-07,
+      "loss": 0.4263,
+      "step": 83157
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.083978201990417e-07,
+      "loss": 0.5073,
+      "step": 83158
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.083755116746578e-07,
+      "loss": 0.3506,
+      "step": 83159
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.08353203603254e-07,
+      "loss": 0.4204,
+      "step": 83160
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.083308959848465e-07,
+      "loss": 0.3601,
+      "step": 83161
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0830858881945286e-07,
+      "loss": 0.439,
+      "step": 83162
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0828628210709014e-07,
+      "loss": 0.3564,
+      "step": 83163
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0826397584777585e-07,
+      "loss": 0.3708,
+      "step": 83164
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.082416700415262e-07,
+      "loss": 0.447,
+      "step": 83165
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.082193646883586e-07,
+      "loss": 0.3799,
+      "step": 83166
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0819705978829035e-07,
+      "loss": 0.4565,
+      "step": 83167
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0817475534133874e-07,
+      "loss": 0.4678,
+      "step": 83168
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0815245134752053e-07,
+      "loss": 0.4376,
+      "step": 83169
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.081301478068523e-07,
+      "loss": 0.3809,
+      "step": 83170
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0810784471935166e-07,
+      "loss": 0.3955,
+      "step": 83171
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.08085542085036e-07,
+      "loss": 0.5132,
+      "step": 83172
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0806323990392165e-07,
+      "loss": 0.4077,
+      "step": 83173
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0804093817602616e-07,
+      "loss": 0.314,
+      "step": 83174
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.080186369013664e-07,
+      "loss": 0.252,
+      "step": 83175
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0799633607995985e-07,
+      "loss": 0.5044,
+      "step": 83176
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.07974035711823e-07,
+      "loss": 0.5347,
+      "step": 83177
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.079517357969736e-07,
+      "loss": 0.3833,
+      "step": 83178
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.07929436335428e-07,
+      "loss": 0.3999,
+      "step": 83179
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.079071373272035e-07,
+      "loss": 0.3838,
+      "step": 83180
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.078848387723176e-07,
+      "loss": 0.354,
+      "step": 83181
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0786254067078683e-07,
+      "loss": 0.3136,
+      "step": 83182
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0784024302262844e-07,
+      "loss": 0.3691,
+      "step": 83183
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.078179458278596e-07,
+      "loss": 0.458,
+      "step": 83184
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0779564908649755e-07,
+      "loss": 0.386,
+      "step": 83185
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0777335279855883e-07,
+      "loss": 0.3672,
+      "step": 83186
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0775105696406087e-07,
+      "loss": 0.4619,
+      "step": 83187
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0772876158302105e-07,
+      "loss": 0.3092,
+      "step": 83188
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0770646665545615e-07,
+      "loss": 0.4294,
+      "step": 83189
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0768417218138274e-07,
+      "loss": 0.4463,
+      "step": 83190
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0766187816081833e-07,
+      "loss": 0.3696,
+      "step": 83191
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.076395845937799e-07,
+      "loss": 0.4263,
+      "step": 83192
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.07617291480285e-07,
+      "loss": 0.5293,
+      "step": 83193
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0759499882035007e-07,
+      "loss": 0.4844,
+      "step": 83194
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.075727066139922e-07,
+      "loss": 0.3809,
+      "step": 83195
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.075504148612289e-07,
+      "loss": 0.3264,
+      "step": 83196
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.075281235620772e-07,
+      "loss": 0.3569,
+      "step": 83197
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0750583271655403e-07,
+      "loss": 0.3701,
+      "step": 83198
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0748354232467607e-07,
+      "loss": 0.4082,
+      "step": 83199
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.074612523864607e-07,
+      "loss": 0.5229,
+      "step": 83200
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.074389629019249e-07,
+      "loss": 0.363,
+      "step": 83201
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0741667387108633e-07,
+      "loss": 0.3264,
+      "step": 83202
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.073943852939612e-07,
+      "loss": 0.4849,
+      "step": 83203
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.073720971705669e-07,
+      "loss": 0.3733,
+      "step": 83204
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0734980950092046e-07,
+      "loss": 0.374,
+      "step": 83205
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0732752228503955e-07,
+      "loss": 0.3726,
+      "step": 83206
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.073052355229402e-07,
+      "loss": 0.4221,
+      "step": 83207
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0728294921464035e-07,
+      "loss": 0.5073,
+      "step": 83208
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0726066336015643e-07,
+      "loss": 0.3848,
+      "step": 83209
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.07238377959506e-07,
+      "loss": 0.4268,
+      "step": 83210
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0721609301270563e-07,
+      "loss": 0.3779,
+      "step": 83211
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.071938085197726e-07,
+      "loss": 0.3662,
+      "step": 83212
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.07171524480724e-07,
+      "loss": 0.4443,
+      "step": 83213
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0714924089557734e-07,
+      "loss": 0.3882,
+      "step": 83214
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.071269577643488e-07,
+      "loss": 0.3184,
+      "step": 83215
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0710467508705593e-07,
+      "loss": 0.3633,
+      "step": 83216
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.070823928637157e-07,
+      "loss": 0.4026,
+      "step": 83217
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.070601110943457e-07,
+      "loss": 0.3416,
+      "step": 83218
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.070378297789624e-07,
+      "loss": 0.4893,
+      "step": 83219
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.070155489175826e-07,
+      "loss": 0.4351,
+      "step": 83220
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0699326851022376e-07,
+      "loss": 0.3874,
+      "step": 83221
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.069709885569029e-07,
+      "loss": 0.3058,
+      "step": 83222
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0694870905763747e-07,
+      "loss": 0.3589,
+      "step": 83223
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0692643001244373e-07,
+      "loss": 0.4199,
+      "step": 83224
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0690415142133917e-07,
+      "loss": 0.446,
+      "step": 83225
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.068818732843412e-07,
+      "loss": 0.321,
+      "step": 83226
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0685959560146623e-07,
+      "loss": 0.4458,
+      "step": 83227
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0683731837273184e-07,
+      "loss": 0.3337,
+      "step": 83228
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.068150415981545e-07,
+      "loss": 0.5283,
+      "step": 83229
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0679276527775166e-07,
+      "loss": 0.282,
+      "step": 83230
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.067704894115407e-07,
+      "loss": 0.375,
+      "step": 83231
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0674821399953783e-07,
+      "loss": 0.5376,
+      "step": 83232
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0672593904176066e-07,
+      "loss": 0.4426,
+      "step": 83233
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0670366453822613e-07,
+      "loss": 0.4236,
+      "step": 83234
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.066813904889517e-07,
+      "loss": 0.3948,
+      "step": 83235
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.066591168939536e-07,
+      "loss": 0.2944,
+      "step": 83236
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.066368437532495e-07,
+      "loss": 0.4194,
+      "step": 83237
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.066145710668565e-07,
+      "loss": 0.4263,
+      "step": 83238
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0659229883479117e-07,
+      "loss": 0.4531,
+      "step": 83239
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0657002705707113e-07,
+      "loss": 0.5151,
+      "step": 83240
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.065477557337128e-07,
+      "loss": 0.3707,
+      "step": 83241
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.065254848647336e-07,
+      "loss": 0.4321,
+      "step": 83242
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.065032144501505e-07,
+      "loss": 0.4033,
+      "step": 83243
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0648094448998093e-07,
+      "loss": 0.4023,
+      "step": 83244
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0645867498424123e-07,
+      "loss": 0.3216,
+      "step": 83245
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0643640593294894e-07,
+      "loss": 0.3733,
+      "step": 83246
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.064141373361213e-07,
+      "loss": 0.4014,
+      "step": 83247
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.063918691937751e-07,
+      "loss": 0.4419,
+      "step": 83248
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0636960150592694e-07,
+      "loss": 0.2581,
+      "step": 83249
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0634733427259426e-07,
+      "loss": 0.4761,
+      "step": 83250
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0632506749379423e-07,
+      "loss": 0.4575,
+      "step": 83251
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0630280116954405e-07,
+      "loss": 0.4097,
+      "step": 83252
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.062805352998603e-07,
+      "loss": 0.3047,
+      "step": 83253
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0625826988476007e-07,
+      "loss": 0.3608,
+      "step": 83254
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0623600492426067e-07,
+      "loss": 0.3574,
+      "step": 83255
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0621374041837943e-07,
+      "loss": 0.417,
+      "step": 83256
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.061914763671326e-07,
+      "loss": 0.3225,
+      "step": 83257
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.061692127705381e-07,
+      "loss": 0.3848,
+      "step": 83258
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0614694962861207e-07,
+      "loss": 0.3008,
+      "step": 83259
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0612468694137213e-07,
+      "loss": 0.4355,
+      "step": 83260
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0610242470883547e-07,
+      "loss": 0.4014,
+      "step": 83261
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0608016293101854e-07,
+      "loss": 0.4282,
+      "step": 83262
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.060579016079387e-07,
+      "loss": 0.4868,
+      "step": 83263
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.060356407396135e-07,
+      "loss": 0.2797,
+      "step": 83264
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0601338032605903e-07,
+      "loss": 0.3059,
+      "step": 83265
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.059911203672929e-07,
+      "loss": 0.4172,
+      "step": 83266
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0596886086333195e-07,
+      "loss": 0.4502,
+      "step": 83267
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.059466018141937e-07,
+      "loss": 0.4839,
+      "step": 83268
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.059243432198948e-07,
+      "loss": 0.5562,
+      "step": 83269
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.05902085080452e-07,
+      "loss": 0.5615,
+      "step": 83270
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0587982739588265e-07,
+      "loss": 0.3302,
+      "step": 83271
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0585757016620383e-07,
+      "loss": 0.4141,
+      "step": 83272
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0583531339143286e-07,
+      "loss": 0.3828,
+      "step": 83273
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0581305707158607e-07,
+      "loss": 0.2847,
+      "step": 83274
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.05790801206681e-07,
+      "loss": 0.4131,
+      "step": 83275
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0576854579673457e-07,
+      "loss": 0.5278,
+      "step": 83276
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0574629084176417e-07,
+      "loss": 0.4187,
+      "step": 83277
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.057240363417865e-07,
+      "loss": 0.4019,
+      "step": 83278
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.057017822968183e-07,
+      "loss": 0.3721,
+      "step": 83279
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0567952870687684e-07,
+      "loss": 0.4102,
+      "step": 83280
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.056572755719797e-07,
+      "loss": 0.2705,
+      "step": 83281
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0563502289214303e-07,
+      "loss": 0.4521,
+      "step": 83282
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0561277066738433e-07,
+      "loss": 0.4946,
+      "step": 83283
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0559051889772056e-07,
+      "loss": 0.4507,
+      "step": 83284
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.055682675831692e-07,
+      "loss": 0.4375,
+      "step": 83285
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0554601672374643e-07,
+      "loss": 0.502,
+      "step": 83286
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0552376631946986e-07,
+      "loss": 0.3187,
+      "step": 83287
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.055015163703567e-07,
+      "loss": 0.3848,
+      "step": 83288
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0547926687642334e-07,
+      "loss": 0.5356,
+      "step": 83289
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.054570178376876e-07,
+      "loss": 0.3408,
+      "step": 83290
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0543476925416567e-07,
+      "loss": 0.481,
+      "step": 83291
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.05412521125875e-07,
+      "loss": 0.4766,
+      "step": 83292
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.053902734528325e-07,
+      "loss": 0.343,
+      "step": 83293
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.0536802623505585e-07,
+      "loss": 0.4624,
+      "step": 83294
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.053457794725611e-07,
+      "loss": 0.3701,
+      "step": 83295
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.053235331653658e-07,
+      "loss": 0.4756,
+      "step": 83296
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.05301287313487e-07,
+      "loss": 0.501,
+      "step": 83297
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.052790419169423e-07,
+      "loss": 0.3893,
+      "step": 83298
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 4.052567969757473e-07,
+      "loss": 0.4119,
+      "step": 83299
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0523455248991986e-07,
+      "loss": 0.4839,
+      "step": 83300
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0521230845947707e-07,
+      "loss": 0.376,
+      "step": 83301
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.051900648844362e-07,
+      "loss": 0.3848,
+      "step": 83302
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.051678217648136e-07,
+      "loss": 0.3245,
+      "step": 83303
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.051455791006265e-07,
+      "loss": 0.3804,
+      "step": 83304
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.051233368918923e-07,
+      "loss": 0.308,
+      "step": 83305
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.05101095138628e-07,
+      "loss": 0.5176,
+      "step": 83306
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.050788538408501e-07,
+      "loss": 0.4614,
+      "step": 83307
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.050566129985763e-07,
+      "loss": 0.3486,
+      "step": 83308
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0503437261182307e-07,
+      "loss": 0.3419,
+      "step": 83309
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0501213268060755e-07,
+      "loss": 0.3076,
+      "step": 83310
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0498989320494726e-07,
+      "loss": 0.4609,
+      "step": 83311
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0496765418485855e-07,
+      "loss": 0.308,
+      "step": 83312
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0494541562035865e-07,
+      "loss": 0.4529,
+      "step": 83313
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.049231775114648e-07,
+      "loss": 0.4263,
+      "step": 83314
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.049009398581942e-07,
+      "loss": 0.3438,
+      "step": 83315
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.048787026605631e-07,
+      "loss": 0.459,
+      "step": 83316
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0485646591858926e-07,
+      "loss": 0.3022,
+      "step": 83317
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0483422963228965e-07,
+      "loss": 0.4194,
+      "step": 83318
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.048119938016812e-07,
+      "loss": 0.4596,
+      "step": 83319
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0478975842678044e-07,
+      "loss": 0.4785,
+      "step": 83320
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.047675235076048e-07,
+      "loss": 0.3916,
+      "step": 83321
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0474528904417126e-07,
+      "loss": 0.3528,
+      "step": 83322
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.047230550364973e-07,
+      "loss": 0.4727,
+      "step": 83323
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0470082148459907e-07,
+      "loss": 0.4976,
+      "step": 83324
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.046785883884941e-07,
+      "loss": 0.4448,
+      "step": 83325
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0465635574819933e-07,
+      "loss": 0.2217,
+      "step": 83326
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0463412356373226e-07,
+      "loss": 0.5034,
+      "step": 83327
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.046118918351095e-07,
+      "loss": 0.3962,
+      "step": 83328
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0458966056234754e-07,
+      "loss": 0.3594,
+      "step": 83329
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.04567429745464e-07,
+      "loss": 0.3682,
+      "step": 83330
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0454519938447573e-07,
+      "loss": 0.3511,
+      "step": 83331
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.045229694794002e-07,
+      "loss": 0.3728,
+      "step": 83332
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.045007400302537e-07,
+      "loss": 0.5015,
+      "step": 83333
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0447851103705354e-07,
+      "loss": 0.478,
+      "step": 83334
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0445628249981733e-07,
+      "loss": 0.3931,
+      "step": 83335
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0443405441856103e-07,
+      "loss": 0.5327,
+      "step": 83336
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.044118267933023e-07,
+      "loss": 0.3411,
+      "step": 83337
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0438959962405836e-07,
+      "loss": 0.3726,
+      "step": 83338
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0436737291084556e-07,
+      "loss": 0.4839,
+      "step": 83339
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.043451466536816e-07,
+      "loss": 0.4722,
+      "step": 83340
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.043229208525829e-07,
+      "loss": 0.4341,
+      "step": 83341
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.043006955075666e-07,
+      "loss": 0.3055,
+      "step": 83342
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.042784706186501e-07,
+      "loss": 0.4585,
+      "step": 83343
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.042562461858503e-07,
+      "loss": 0.3916,
+      "step": 83344
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.042340222091839e-07,
+      "loss": 0.344,
+      "step": 83345
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0421179868866807e-07,
+      "loss": 0.438,
+      "step": 83346
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.041895756243202e-07,
+      "loss": 0.418,
+      "step": 83347
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0416735301615666e-07,
+      "loss": 0.4033,
+      "step": 83348
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.041451308641951e-07,
+      "loss": 0.3691,
+      "step": 83349
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0412290916845184e-07,
+      "loss": 0.3205,
+      "step": 83350
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0410068792894434e-07,
+      "loss": 0.3793,
+      "step": 83351
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.040784671456898e-07,
+      "loss": 0.4122,
+      "step": 83352
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0405624681870475e-07,
+      "loss": 0.3501,
+      "step": 83353
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0403402694800636e-07,
+      "loss": 0.3669,
+      "step": 83354
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0401180753361177e-07,
+      "loss": 0.3511,
+      "step": 83355
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0398958857553813e-07,
+      "loss": 0.4326,
+      "step": 83356
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0396737007380243e-07,
+      "loss": 0.4067,
+      "step": 83357
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0394515202842103e-07,
+      "loss": 0.4521,
+      "step": 83358
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.039229344394115e-07,
+      "loss": 0.3966,
+      "step": 83359
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0390071730679074e-07,
+      "loss": 0.4058,
+      "step": 83360
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0387850063057616e-07,
+      "loss": 0.3055,
+      "step": 83361
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.03856284410784e-07,
+      "loss": 0.3254,
+      "step": 83362
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.038340686474317e-07,
+      "loss": 0.4185,
+      "step": 83363
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0381185334053627e-07,
+      "loss": 0.4448,
+      "step": 83364
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.03789638490115e-07,
+      "loss": 0.4878,
+      "step": 83365
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0376742409618434e-07,
+      "loss": 0.377,
+      "step": 83366
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0374521015876186e-07,
+      "loss": 0.4106,
+      "step": 83367
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.037229966778638e-07,
+      "loss": 0.3843,
+      "step": 83368
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.03700783653508e-07,
+      "loss": 0.3779,
+      "step": 83369
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.036785710857108e-07,
+      "loss": 0.4985,
+      "step": 83370
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0365635897448946e-07,
+      "loss": 0.4102,
+      "step": 83371
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.036341473198611e-07,
+      "loss": 0.4595,
+      "step": 83372
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0361193612184297e-07,
+      "loss": 0.5005,
+      "step": 83373
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0358972538045134e-07,
+      "loss": 0.4507,
+      "step": 83374
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0356751509570376e-07,
+      "loss": 0.3896,
+      "step": 83375
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.03545305267617e-07,
+      "loss": 0.4229,
+      "step": 83376
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0352309589620846e-07,
+      "loss": 0.4634,
+      "step": 83377
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.03500886981495e-07,
+      "loss": 0.4644,
+      "step": 83378
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0347867852349313e-07,
+      "loss": 0.2947,
+      "step": 83379
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.034564705222201e-07,
+      "loss": 0.4785,
+      "step": 83380
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0343426297769313e-07,
+      "loss": 0.3762,
+      "step": 83381
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.034120558899294e-07,
+      "loss": 0.4276,
+      "step": 83382
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0338984925894527e-07,
+      "loss": 0.5078,
+      "step": 83383
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.033676430847581e-07,
+      "loss": 0.4751,
+      "step": 83384
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.033454373673849e-07,
+      "loss": 0.3233,
+      "step": 83385
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0332323210684305e-07,
+      "loss": 0.3662,
+      "step": 83386
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.033010273031492e-07,
+      "loss": 0.3416,
+      "step": 83387
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.032788229563199e-07,
+      "loss": 0.4272,
+      "step": 83388
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0325661906637254e-07,
+      "loss": 0.3394,
+      "step": 83389
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.032344156333246e-07,
+      "loss": 0.3247,
+      "step": 83390
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0321221265719216e-07,
+      "loss": 0.3767,
+      "step": 83391
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.031900101379928e-07,
+      "loss": 0.3245,
+      "step": 83392
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0316780807574335e-07,
+      "loss": 0.4751,
+      "step": 83393
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0314560647046116e-07,
+      "loss": 0.4392,
+      "step": 83394
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.031234053221626e-07,
+      "loss": 0.4624,
+      "step": 83395
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0310120463086503e-07,
+      "loss": 0.3552,
+      "step": 83396
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.030790043965858e-07,
+      "loss": 0.2489,
+      "step": 83397
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0305680461934124e-07,
+      "loss": 0.4917,
+      "step": 83398
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.030346052991489e-07,
+      "loss": 0.385,
+      "step": 83399
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.030124064360252e-07,
+      "loss": 0.3462,
+      "step": 83400
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.029902080299874e-07,
+      "loss": 0.4663,
+      "step": 83401
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.029680100810525e-07,
+      "loss": 0.4033,
+      "step": 83402
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02945812589238e-07,
+      "loss": 0.3517,
+      "step": 83403
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.029236155545601e-07,
+      "loss": 0.4199,
+      "step": 83404
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02901418977036e-07,
+      "loss": 0.5518,
+      "step": 83405
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02879222856683e-07,
+      "loss": 0.3926,
+      "step": 83406
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0285702719351856e-07,
+      "loss": 0.3228,
+      "step": 83407
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.028348319875583e-07,
+      "loss": 0.54,
+      "step": 83408
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0281263723881986e-07,
+      "loss": 0.3784,
+      "step": 83409
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0279044294732047e-07,
+      "loss": 0.4438,
+      "step": 83410
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0276824911307726e-07,
+      "loss": 0.4282,
+      "step": 83411
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0274605573610664e-07,
+      "loss": 0.3021,
+      "step": 83412
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.027238628164259e-07,
+      "loss": 0.3901,
+      "step": 83413
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02701670354052e-07,
+      "loss": 0.4055,
+      "step": 83414
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.026794783490023e-07,
+      "loss": 0.4561,
+      "step": 83415
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.026572868012932e-07,
+      "loss": 0.4531,
+      "step": 83416
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0263509571094223e-07,
+      "loss": 0.5225,
+      "step": 83417
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0261290507796576e-07,
+      "loss": 0.5513,
+      "step": 83418
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0259071490238105e-07,
+      "loss": 0.2632,
+      "step": 83419
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.025685251842056e-07,
+      "loss": 0.4678,
+      "step": 83420
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0254633592345555e-07,
+      "loss": 0.4189,
+      "step": 83421
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.025241471201484e-07,
+      "loss": 0.5244,
+      "step": 83422
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0250195877430095e-07,
+      "loss": 0.4272,
+      "step": 83423
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0247977088593066e-07,
+      "loss": 0.4771,
+      "step": 83424
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0245758345505377e-07,
+      "loss": 0.3195,
+      "step": 83425
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0243539648168765e-07,
+      "loss": 0.3813,
+      "step": 83426
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.024132099658496e-07,
+      "loss": 0.439,
+      "step": 83427
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.023910239075563e-07,
+      "loss": 0.3892,
+      "step": 83428
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0236883830682433e-07,
+      "loss": 0.4111,
+      "step": 83429
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.023466531636711e-07,
+      "loss": 0.4009,
+      "step": 83430
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0232446847811354e-07,
+      "loss": 0.4824,
+      "step": 83431
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02302284250169e-07,
+      "loss": 0.4299,
+      "step": 83432
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.022801004798537e-07,
+      "loss": 0.4722,
+      "step": 83433
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0225791716718516e-07,
+      "loss": 0.4062,
+      "step": 83434
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0223573431218015e-07,
+      "loss": 0.3877,
+      "step": 83435
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0221355191485616e-07,
+      "loss": 0.3417,
+      "step": 83436
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.021913699752297e-07,
+      "loss": 0.4607,
+      "step": 83437
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.021691884933176e-07,
+      "loss": 0.4019,
+      "step": 83438
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.02147007469137e-07,
+      "loss": 0.498,
+      "step": 83439
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0212482690270486e-07,
+      "loss": 0.3831,
+      "step": 83440
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.021026467940387e-07,
+      "loss": 0.2504,
+      "step": 83441
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0208046714315467e-07,
+      "loss": 0.3143,
+      "step": 83442
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0205828795007013e-07,
+      "loss": 0.5215,
+      "step": 83443
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0203610921480246e-07,
+      "loss": 0.4473,
+      "step": 83444
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.020139309373679e-07,
+      "loss": 0.384,
+      "step": 83445
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0199175311778366e-07,
+      "loss": 0.3191,
+      "step": 83446
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0196957575606726e-07,
+      "loss": 0.4009,
+      "step": 83447
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0194739885223495e-07,
+      "loss": 0.4102,
+      "step": 83448
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.019252224063043e-07,
+      "loss": 0.3994,
+      "step": 83449
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0190304641829177e-07,
+      "loss": 0.3628,
+      "step": 83450
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0188087088821445e-07,
+      "loss": 0.4722,
+      "step": 83451
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0185869581608946e-07,
+      "loss": 0.3828,
+      "step": 83452
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.018365212019341e-07,
+      "loss": 0.5024,
+      "step": 83453
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.018143470457648e-07,
+      "loss": 0.4883,
+      "step": 83454
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0179217334759865e-07,
+      "loss": 0.4114,
+      "step": 83455
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0177000010745267e-07,
+      "loss": 0.3368,
+      "step": 83456
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.017478273253443e-07,
+      "loss": 0.2852,
+      "step": 83457
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.017256550012901e-07,
+      "loss": 0.4404,
+      "step": 83458
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.017034831353068e-07,
+      "loss": 0.3838,
+      "step": 83459
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.016813117274115e-07,
+      "loss": 0.3672,
+      "step": 83460
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0165914077762174e-07,
+      "loss": 0.2881,
+      "step": 83461
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0163697028595367e-07,
+      "loss": 0.4883,
+      "step": 83462
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0161480025242466e-07,
+      "loss": 0.3774,
+      "step": 83463
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0159263067705164e-07,
+      "loss": 0.4326,
+      "step": 83464
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.015704615598521e-07,
+      "loss": 0.4253,
+      "step": 83465
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0154829290084214e-07,
+      "loss": 0.3566,
+      "step": 83466
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.015261247000394e-07,
+      "loss": 0.3882,
+      "step": 83467
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.015039569574603e-07,
+      "loss": 0.5488,
+      "step": 83468
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.014817896731221e-07,
+      "loss": 0.3568,
+      "step": 83469
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0145962284704216e-07,
+      "loss": 0.3268,
+      "step": 83470
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0143745647923664e-07,
+      "loss": 0.4917,
+      "step": 83471
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.01415290569723e-07,
+      "loss": 0.4717,
+      "step": 83472
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.013931251185182e-07,
+      "loss": 0.4951,
+      "step": 83473
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0137096012563943e-07,
+      "loss": 0.3979,
+      "step": 83474
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.01348795591103e-07,
+      "loss": 0.4146,
+      "step": 83475
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0132663151492627e-07,
+      "loss": 0.4961,
+      "step": 83476
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0130446789712667e-07,
+      "loss": 0.5059,
+      "step": 83477
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0128230473772064e-07,
+      "loss": 0.3936,
+      "step": 83478
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.012601420367249e-07,
+      "loss": 0.3706,
+      "step": 83479
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0123797979415674e-07,
+      "loss": 0.4351,
+      "step": 83480
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.01215818010033e-07,
+      "loss": 0.4888,
+      "step": 83481
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.011936566843713e-07,
+      "loss": 0.1868,
+      "step": 83482
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.011714958171877e-07,
+      "loss": 0.3472,
+      "step": 83483
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0114933540849946e-07,
+      "loss": 0.385,
+      "step": 83484
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0112717545832367e-07,
+      "loss": 0.4819,
+      "step": 83485
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0110501596667767e-07,
+      "loss": 0.4011,
+      "step": 83486
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0108285693357804e-07,
+      "loss": 0.2612,
+      "step": 83487
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0106069835904124e-07,
+      "loss": 0.2828,
+      "step": 83488
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.010385402430848e-07,
+      "loss": 0.2788,
+      "step": 83489
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.010163825857257e-07,
+      "loss": 0.4912,
+      "step": 83490
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.00994225386981e-07,
+      "loss": 0.3597,
+      "step": 83491
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0097206864686726e-07,
+      "loss": 0.3494,
+      "step": 83492
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0094991236540154e-07,
+      "loss": 0.3535,
+      "step": 83493
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0092775654260104e-07,
+      "loss": 0.4146,
+      "step": 83494
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0090560117848296e-07,
+      "loss": 0.3125,
+      "step": 83495
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.008834462730635e-07,
+      "loss": 0.3835,
+      "step": 83496
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.008612918263604e-07,
+      "loss": 0.3828,
+      "step": 83497
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0083913783838996e-07,
+      "loss": 0.4194,
+      "step": 83498
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0081698430916976e-07,
+      "loss": 0.439,
+      "step": 83499
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0079483123871605e-07,
+      "loss": 0.5142,
+      "step": 83500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0077267862704624e-07,
+      "loss": 0.4722,
+      "step": 83501
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.007505264741772e-07,
+      "loss": 0.522,
+      "step": 83502
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0072837478012634e-07,
+      "loss": 0.3157,
+      "step": 83503
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.007062235449098e-07,
+      "loss": 0.4583,
+      "step": 83504
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.00684072768545e-07,
+      "loss": 0.3931,
+      "step": 83505
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.006619224510488e-07,
+      "loss": 0.4526,
+      "step": 83506
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0063977259243855e-07,
+      "loss": 0.333,
+      "step": 83507
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0061762319273086e-07,
+      "loss": 0.4746,
+      "step": 83508
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.005954742519424e-07,
+      "loss": 0.4639,
+      "step": 83509
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.005733257700904e-07,
+      "loss": 0.4453,
+      "step": 83510
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.005511777471918e-07,
+      "loss": 0.321,
+      "step": 83511
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.00529030183264e-07,
+      "loss": 0.4663,
+      "step": 83512
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.005068830783231e-07,
+      "loss": 0.4619,
+      "step": 83513
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.004847364323867e-07,
+      "loss": 0.4565,
+      "step": 83514
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.004625902454718e-07,
+      "loss": 0.5518,
+      "step": 83515
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0044044451759486e-07,
+      "loss": 0.4731,
+      "step": 83516
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0041829924877337e-07,
+      "loss": 0.3892,
+      "step": 83517
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.003961544390237e-07,
+      "loss": 0.5098,
+      "step": 83518
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0037401008836313e-07,
+      "loss": 0.4814,
+      "step": 83519
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0035186619680895e-07,
+      "loss": 0.394,
+      "step": 83520
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.003297227643775e-07,
+      "loss": 0.4731,
+      "step": 83521
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0030757979108586e-07,
+      "loss": 0.4976,
+      "step": 83522
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.002854372769512e-07,
+      "loss": 0.4492,
+      "step": 83523
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0026329522199077e-07,
+      "loss": 0.2637,
+      "step": 83524
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.002411536262208e-07,
+      "loss": 0.3438,
+      "step": 83525
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.002190124896586e-07,
+      "loss": 0.4321,
+      "step": 83526
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.001968718123215e-07,
+      "loss": 0.3823,
+      "step": 83527
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0017473159422567e-07,
+      "loss": 0.3208,
+      "step": 83528
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0015259183538886e-07,
+      "loss": 0.4309,
+      "step": 83529
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0013045253582724e-07,
+      "loss": 0.3113,
+      "step": 83530
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0010831369555817e-07,
+      "loss": 0.4702,
+      "step": 83531
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0008617531459897e-07,
+      "loss": 0.3762,
+      "step": 83532
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0006403739296577e-07,
+      "loss": 0.335,
+      "step": 83533
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0004189993067606e-07,
+      "loss": 0.4277,
+      "step": 83534
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 4.0001976292774665e-07,
+      "loss": 0.4219,
+      "step": 83535
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9999762638419477e-07,
+      "loss": 0.3396,
+      "step": 83536
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.999754903000372e-07,
+      "loss": 0.4341,
+      "step": 83537
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.999533546752905e-07,
+      "loss": 0.4255,
+      "step": 83538
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.999312195099718e-07,
+      "loss": 0.3752,
+      "step": 83539
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.999090848040982e-07,
+      "loss": 0.4448,
+      "step": 83540
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9988695055768704e-07,
+      "loss": 0.4185,
+      "step": 83541
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9986481677075444e-07,
+      "loss": 0.5811,
+      "step": 83542
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9984268344331783e-07,
+      "loss": 0.4717,
+      "step": 83543
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.99820550575394e-07,
+      "loss": 0.3445,
+      "step": 83544
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9979841816700043e-07,
+      "loss": 0.3735,
+      "step": 83545
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.997762862181532e-07,
+      "loss": 0.395,
+      "step": 83546
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9975415472887e-07,
+      "loss": 0.437,
+      "step": 83547
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.997320236991671e-07,
+      "loss": 0.5366,
+      "step": 83548
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.997098931290621e-07,
+      "loss": 0.355,
+      "step": 83549
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9968776301857134e-07,
+      "loss": 0.4021,
+      "step": 83550
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.996656333677121e-07,
+      "loss": 0.4336,
+      "step": 83551
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.996435041765012e-07,
+      "loss": 0.2737,
+      "step": 83552
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.99621375444956e-07,
+      "loss": 0.3611,
+      "step": 83553
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.995992471730929e-07,
+      "loss": 0.2714,
+      "step": 83554
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9957711936092885e-07,
+      "loss": 0.5063,
+      "step": 83555
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.995549920084811e-07,
+      "loss": 0.2288,
+      "step": 83556
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.995328651157669e-07,
+      "loss": 0.3466,
+      "step": 83557
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.995107386828026e-07,
+      "loss": 0.4375,
+      "step": 83558
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.994886127096051e-07,
+      "loss": 0.3669,
+      "step": 83559
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.994664871961916e-07,
+      "loss": 0.4907,
+      "step": 83560
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.994443621425788e-07,
+      "loss": 0.3064,
+      "step": 83561
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9942223754878436e-07,
+      "loss": 0.4438,
+      "step": 83562
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9940011341482425e-07,
+      "loss": 0.3491,
+      "step": 83563
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.993779897407159e-07,
+      "loss": 0.5103,
+      "step": 83564
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.993558665264762e-07,
+      "loss": 0.3425,
+      "step": 83565
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9933374377212245e-07,
+      "loss": 0.4541,
+      "step": 83566
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9931162147767126e-07,
+      "loss": 0.3333,
+      "step": 83567
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.992894996431392e-07,
+      "loss": 0.3062,
+      "step": 83568
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.992673782685435e-07,
+      "loss": 0.2939,
+      "step": 83569
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9924525735390146e-07,
+      "loss": 0.3508,
+      "step": 83570
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.992231368992295e-07,
+      "loss": 0.4771,
+      "step": 83571
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9920101690454464e-07,
+      "loss": 0.3489,
+      "step": 83572
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.99178897369864e-07,
+      "loss": 0.4248,
+      "step": 83573
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.991567782952047e-07,
+      "loss": 0.4561,
+      "step": 83574
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9913465968058315e-07,
+      "loss": 0.4487,
+      "step": 83575
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.991125415260165e-07,
+      "loss": 0.4062,
+      "step": 83576
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9909042383152214e-07,
+      "loss": 0.4082,
+      "step": 83577
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9906830659711633e-07,
+      "loss": 0.501,
+      "step": 83578
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9904618982281655e-07,
+      "loss": 0.4061,
+      "step": 83579
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.990240735086392e-07,
+      "loss": 0.5117,
+      "step": 83580
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.990019576546014e-07,
+      "loss": 0.4258,
+      "step": 83581
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9897984226072013e-07,
+      "loss": 0.3556,
+      "step": 83582
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.989577273270128e-07,
+      "loss": 0.4893,
+      "step": 83583
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.989356128534955e-07,
+      "loss": 0.5054,
+      "step": 83584
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9891349884018564e-07,
+      "loss": 0.3708,
+      "step": 83585
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9889138528709995e-07,
+      "loss": 0.4668,
+      "step": 83586
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9886927219425634e-07,
+      "loss": 0.4849,
+      "step": 83587
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9884715956167016e-07,
+      "loss": 0.4858,
+      "step": 83588
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.98825047389359e-07,
+      "loss": 0.5176,
+      "step": 83589
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.988029356773398e-07,
+      "loss": 0.2875,
+      "step": 83590
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9878082442563e-07,
+      "loss": 0.3313,
+      "step": 83591
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.987587136342457e-07,
+      "loss": 0.4521,
+      "step": 83592
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.987366033032042e-07,
+      "loss": 0.3898,
+      "step": 83593
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.987144934325225e-07,
+      "loss": 0.4771,
+      "step": 83594
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.986923840222177e-07,
+      "loss": 0.438,
+      "step": 83595
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9867027507230623e-07,
+      "loss": 0.6152,
+      "step": 83596
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9864816658280563e-07,
+      "loss": 0.3892,
+      "step": 83597
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.986260585537321e-07,
+      "loss": 0.2844,
+      "step": 83598
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9860395098510294e-07,
+      "loss": 0.4307,
+      "step": 83599
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.985818438769354e-07,
+      "loss": 0.4678,
+      "step": 83600
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9855973722924575e-07,
+      "loss": 0.4287,
+      "step": 83601
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9853763104205127e-07,
+      "loss": 0.3955,
+      "step": 83602
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.985155253153689e-07,
+      "loss": 0.2395,
+      "step": 83603
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9849342004921593e-07,
+      "loss": 0.3442,
+      "step": 83604
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.984713152436084e-07,
+      "loss": 0.5049,
+      "step": 83605
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9844921089856386e-07,
+      "loss": 0.4971,
+      "step": 83606
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9842710701409933e-07,
+      "loss": 0.3708,
+      "step": 83607
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.984050035902316e-07,
+      "loss": 0.4253,
+      "step": 83608
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.983829006269771e-07,
+      "loss": 0.4707,
+      "step": 83609
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.983607981243532e-07,
+      "loss": 0.3463,
+      "step": 83610
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.983386960823767e-07,
+      "loss": 0.3098,
+      "step": 83611
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.983165945010649e-07,
+      "loss": 0.4775,
+      "step": 83612
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.982944933804342e-07,
+      "loss": 0.3489,
+      "step": 83613
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.982723927205017e-07,
+      "loss": 0.3051,
+      "step": 83614
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9825029252128426e-07,
+      "loss": 0.436,
+      "step": 83615
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.982281927827994e-07,
+      "loss": 0.3413,
+      "step": 83616
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.982060935050635e-07,
+      "loss": 0.4771,
+      "step": 83617
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9818399468809316e-07,
+      "loss": 0.4292,
+      "step": 83618
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.981618963319057e-07,
+      "loss": 0.2808,
+      "step": 83619
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.981397984365179e-07,
+      "loss": 0.3892,
+      "step": 83620
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.981177010019472e-07,
+      "loss": 0.4609,
+      "step": 83621
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9809560402820973e-07,
+      "loss": 0.4067,
+      "step": 83622
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9807350751532275e-07,
+      "loss": 0.3792,
+      "step": 83623
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9805141146330365e-07,
+      "loss": 0.3853,
+      "step": 83624
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9802931587216847e-07,
+      "loss": 0.4512,
+      "step": 83625
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.980072207419346e-07,
+      "loss": 0.3281,
+      "step": 83626
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9798512607261924e-07,
+      "loss": 0.5308,
+      "step": 83627
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9796303186423876e-07,
+      "loss": 0.3643,
+      "step": 83628
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.979409381168105e-07,
+      "loss": 0.542,
+      "step": 83629
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.979188448303509e-07,
+      "loss": 0.3341,
+      "step": 83630
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9789675200487717e-07,
+      "loss": 0.501,
+      "step": 83631
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.978746596404061e-07,
+      "loss": 0.384,
+      "step": 83632
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9785256773695517e-07,
+      "loss": 0.562,
+      "step": 83633
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.978304762945406e-07,
+      "loss": 0.4404,
+      "step": 83634
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9780838531317937e-07,
+      "loss": 0.4053,
+      "step": 83635
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.977862947928886e-07,
+      "loss": 0.4302,
+      "step": 83636
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.977642047336855e-07,
+      "loss": 0.2266,
+      "step": 83637
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9774211513558674e-07,
+      "loss": 0.4238,
+      "step": 83638
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.977200259986088e-07,
+      "loss": 0.424,
+      "step": 83639
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.976979373227689e-07,
+      "loss": 0.4409,
+      "step": 83640
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.976758491080844e-07,
+      "loss": 0.3806,
+      "step": 83641
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.976537613545713e-07,
+      "loss": 0.4077,
+      "step": 83642
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9763167406224716e-07,
+      "loss": 0.4712,
+      "step": 83643
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9760958723112867e-07,
+      "loss": 0.3643,
+      "step": 83644
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.975875008612333e-07,
+      "loss": 0.5508,
+      "step": 83645
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9756541495257735e-07,
+      "loss": 0.3643,
+      "step": 83646
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9754332950517755e-07,
+      "loss": 0.4478,
+      "step": 83647
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9752124451905114e-07,
+      "loss": 0.3287,
+      "step": 83648
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.97499159994215e-07,
+      "loss": 0.4797,
+      "step": 83649
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9747707593068636e-07,
+      "loss": 0.5312,
+      "step": 83650
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9745499232848155e-07,
+      "loss": 0.5029,
+      "step": 83651
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9743290918761774e-07,
+      "loss": 0.3984,
+      "step": 83652
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9741082650811176e-07,
+      "loss": 0.272,
+      "step": 83653
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.97388744289981e-07,
+      "loss": 0.5767,
+      "step": 83654
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.973666625332416e-07,
+      "loss": 0.3306,
+      "step": 83655
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9734458123791126e-07,
+      "loss": 0.5435,
+      "step": 83656
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9732250040400605e-07,
+      "loss": 0.4766,
+      "step": 83657
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9730042003154364e-07,
+      "loss": 0.3237,
+      "step": 83658
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.972783401205402e-07,
+      "loss": 0.4585,
+      "step": 83659
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9725626067101313e-07,
+      "loss": 0.5054,
+      "step": 83660
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9723418168297917e-07,
+      "loss": 0.3567,
+      "step": 83661
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.972121031564556e-07,
+      "loss": 0.3018,
+      "step": 83662
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.971900250914586e-07,
+      "loss": 0.3328,
+      "step": 83663
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9716794748800566e-07,
+      "loss": 0.4219,
+      "step": 83664
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9714587034611337e-07,
+      "loss": 0.3958,
+      "step": 83665
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.971237936657992e-07,
+      "loss": 0.2384,
+      "step": 83666
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9710171744707956e-07,
+      "loss": 0.3124,
+      "step": 83667
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.97079641689971e-07,
+      "loss": 0.4634,
+      "step": 83668
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.970575663944909e-07,
+      "loss": 0.5449,
+      "step": 83669
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.970354915606561e-07,
+      "loss": 0.3823,
+      "step": 83670
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9701341718848393e-07,
+      "loss": 0.3208,
+      "step": 83671
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9699134327799045e-07,
+      "loss": 0.4263,
+      "step": 83672
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.969692698291929e-07,
+      "loss": 0.4355,
+      "step": 83673
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.969471968421083e-07,
+      "loss": 0.342,
+      "step": 83674
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9692512431675384e-07,
+      "loss": 0.3845,
+      "step": 83675
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9690305225314614e-07,
+      "loss": 0.3276,
+      "step": 83676
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.968809806513016e-07,
+      "loss": 0.353,
+      "step": 83677
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.968589095112377e-07,
+      "loss": 0.3438,
+      "step": 83678
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9683683883297137e-07,
+      "loss": 0.5444,
+      "step": 83679
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.968147686165191e-07,
+      "loss": 0.3008,
+      "step": 83680
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9679269886189805e-07,
+      "loss": 0.2686,
+      "step": 83681
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.96770629569125e-07,
+      "loss": 0.3679,
+      "step": 83682
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.967485607382174e-07,
+      "loss": 0.4551,
+      "step": 83683
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.967264923691913e-07,
+      "loss": 0.3433,
+      "step": 83684
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.96704424462064e-07,
+      "loss": 0.3589,
+      "step": 83685
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.966823570168526e-07,
+      "loss": 0.4849,
+      "step": 83686
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.966602900335736e-07,
+      "loss": 0.3306,
+      "step": 83687
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.966382235122443e-07,
+      "loss": 0.3745,
+      "step": 83688
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.9661615745288114e-07,
+      "loss": 0.4473,
+      "step": 83689
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.965940918555012e-07,
+      "loss": 0.582,
+      "step": 83690
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9657202672012137e-07,
+      "loss": 0.3447,
+      "step": 83691
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9654996204675893e-07,
+      "loss": 0.4478,
+      "step": 83692
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9652789783543014e-07,
+      "loss": 0.5552,
+      "step": 83693
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.965058340861521e-07,
+      "loss": 0.418,
+      "step": 83694
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9648377079894226e-07,
+      "loss": 0.4565,
+      "step": 83695
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9646170797381704e-07,
+      "loss": 0.5366,
+      "step": 83696
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.964396456107929e-07,
+      "loss": 0.4302,
+      "step": 83697
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.964175837098873e-07,
+      "loss": 0.4731,
+      "step": 83698
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.963955222711168e-07,
+      "loss": 0.3505,
+      "step": 83699
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.963734612944991e-07,
+      "loss": 0.2896,
+      "step": 83700
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.963514007800499e-07,
+      "loss": 0.4893,
+      "step": 83701
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.963293407277869e-07,
+      "loss": 0.5327,
+      "step": 83702
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.963072811377266e-07,
+      "loss": 0.3616,
+      "step": 83703
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9628522200988645e-07,
+      "loss": 0.3638,
+      "step": 83704
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9626316334428254e-07,
+      "loss": 0.4229,
+      "step": 83705
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9624110514093267e-07,
+      "loss": 0.4438,
+      "step": 83706
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.962190473998528e-07,
+      "loss": 0.438,
+      "step": 83707
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.961969901210602e-07,
+      "loss": 0.3999,
+      "step": 83708
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9617493330457217e-07,
+      "loss": 0.3589,
+      "step": 83709
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.961528769504049e-07,
+      "loss": 0.2773,
+      "step": 83710
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9613082105857556e-07,
+      "loss": 0.4302,
+      "step": 83711
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9610876562910147e-07,
+      "loss": 0.4307,
+      "step": 83712
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9608671066199873e-07,
+      "loss": 0.297,
+      "step": 83713
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9606465615728466e-07,
+      "loss": 0.5435,
+      "step": 83714
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.960426021149762e-07,
+      "loss": 0.4824,
+      "step": 83715
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9602054853509035e-07,
+      "loss": 0.3425,
+      "step": 83716
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9599849541764383e-07,
+      "loss": 0.4624,
+      "step": 83717
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.959764427626533e-07,
+      "loss": 0.4341,
+      "step": 83718
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.959543905701357e-07,
+      "loss": 0.2916,
+      "step": 83719
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.95932338840108e-07,
+      "loss": 0.2756,
+      "step": 83720
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.959102875725876e-07,
+      "loss": 0.2732,
+      "step": 83721
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.958882367675904e-07,
+      "loss": 0.5947,
+      "step": 83722
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9586618642513393e-07,
+      "loss": 0.4097,
+      "step": 83723
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9584413654523495e-07,
+      "loss": 0.354,
+      "step": 83724
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9582208712791066e-07,
+      "loss": 0.3955,
+      "step": 83725
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9580003817317765e-07,
+      "loss": 0.3005,
+      "step": 83726
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9577798968105237e-07,
+      "loss": 0.4001,
+      "step": 83727
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9575594165155213e-07,
+      "loss": 0.4673,
+      "step": 83728
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9573389408469403e-07,
+      "loss": 0.3354,
+      "step": 83729
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9571184698049444e-07,
+      "loss": 0.5034,
+      "step": 83730
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.956898003389705e-07,
+      "loss": 0.3961,
+      "step": 83731
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.956677541601391e-07,
+      "loss": 0.3647,
+      "step": 83732
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9564570844401744e-07,
+      "loss": 0.3574,
+      "step": 83733
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.956236631906218e-07,
+      "loss": 0.4287,
+      "step": 83734
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.956016183999692e-07,
+      "loss": 0.3948,
+      "step": 83735
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.955795740720771e-07,
+      "loss": 0.5586,
+      "step": 83736
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.955575302069616e-07,
+      "loss": 0.5054,
+      "step": 83737
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.955354868046401e-07,
+      "loss": 0.4355,
+      "step": 83738
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.955134438651291e-07,
+      "loss": 0.584,
+      "step": 83739
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.954914013884456e-07,
+      "loss": 0.3718,
+      "step": 83740
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9546935937460645e-07,
+      "loss": 0.4746,
+      "step": 83741
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.954473178236291e-07,
+      "loss": 0.522,
+      "step": 83742
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.954252767355295e-07,
+      "loss": 0.3721,
+      "step": 83743
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.95403236110325e-07,
+      "loss": 0.4893,
+      "step": 83744
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9538119594803245e-07,
+      "loss": 0.2104,
+      "step": 83745
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.953591562486691e-07,
+      "loss": 0.2144,
+      "step": 83746
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9533711701225134e-07,
+      "loss": 0.4248,
+      "step": 83747
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9531507823879596e-07,
+      "loss": 0.3535,
+      "step": 83748
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9529303992831984e-07,
+      "loss": 0.4321,
+      "step": 83749
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.952710020808405e-07,
+      "loss": 0.3394,
+      "step": 83750
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9524896469637405e-07,
+      "loss": 0.2925,
+      "step": 83751
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9522692777493747e-07,
+      "loss": 0.342,
+      "step": 83752
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.95204891316548e-07,
+      "loss": 0.3311,
+      "step": 83753
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9518285532122253e-07,
+      "loss": 0.3721,
+      "step": 83754
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9516081978897753e-07,
+      "loss": 0.4229,
+      "step": 83755
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.951387847198303e-07,
+      "loss": 0.3687,
+      "step": 83756
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.951167501137973e-07,
+      "loss": 0.46,
+      "step": 83757
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.950947159708955e-07,
+      "loss": 0.3511,
+      "step": 83758
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.950726822911422e-07,
+      "loss": 0.5098,
+      "step": 83759
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9505064907455357e-07,
+      "loss": 0.4453,
+      "step": 83760
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.950286163211468e-07,
+      "loss": 0.3264,
+      "step": 83761
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9500658403093887e-07,
+      "loss": 0.417,
+      "step": 83762
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94984552203947e-07,
+      "loss": 0.3921,
+      "step": 83763
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.949625208401872e-07,
+      "loss": 0.4819,
+      "step": 83764
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.949404899396768e-07,
+      "loss": 0.4614,
+      "step": 83765
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9491845950243296e-07,
+      "loss": 0.4004,
+      "step": 83766
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9489642952847236e-07,
+      "loss": 0.2859,
+      "step": 83767
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9487440001781126e-07,
+      "loss": 0.4087,
+      "step": 83768
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9485237097046697e-07,
+      "loss": 0.3672,
+      "step": 83769
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.948303423864565e-07,
+      "loss": 0.46,
+      "step": 83770
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94808314265797e-07,
+      "loss": 0.4102,
+      "step": 83771
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.947862866085044e-07,
+      "loss": 0.5527,
+      "step": 83772
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.947642594145962e-07,
+      "loss": 0.4043,
+      "step": 83773
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.947422326840894e-07,
+      "loss": 0.3432,
+      "step": 83774
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9472020641700066e-07,
+      "loss": 0.4551,
+      "step": 83775
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94698180613347e-07,
+      "loss": 0.3513,
+      "step": 83776
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.946761552731448e-07,
+      "loss": 0.3279,
+      "step": 83777
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9465413039641116e-07,
+      "loss": 0.5303,
+      "step": 83778
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94632105983163e-07,
+      "loss": 0.478,
+      "step": 83779
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.946100820334175e-07,
+      "loss": 0.394,
+      "step": 83780
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94588058547191e-07,
+      "loss": 0.3579,
+      "step": 83781
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.945660355245005e-07,
+      "loss": 0.3547,
+      "step": 83782
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.94544012965363e-07,
+      "loss": 0.397,
+      "step": 83783
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9452199086979567e-07,
+      "loss": 0.3828,
+      "step": 83784
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.944999692378146e-07,
+      "loss": 0.4636,
+      "step": 83785
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.944779480694375e-07,
+      "loss": 0.4927,
+      "step": 83786
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9445592736468034e-07,
+      "loss": 0.4438,
+      "step": 83787
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9443390712356083e-07,
+      "loss": 0.3662,
+      "step": 83788
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.944118873460951e-07,
+      "loss": 0.522,
+      "step": 83789
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9438986803230035e-07,
+      "loss": 0.3921,
+      "step": 83790
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9436784918219333e-07,
+      "loss": 0.4653,
+      "step": 83791
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.943458307957914e-07,
+      "loss": 0.4189,
+      "step": 83792
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9432381287311077e-07,
+      "loss": 0.3555,
+      "step": 83793
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9430179541416844e-07,
+      "loss": 0.3887,
+      "step": 83794
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.942797784189814e-07,
+      "loss": 0.3787,
+      "step": 83795
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.942577618875669e-07,
+      "loss": 0.3672,
+      "step": 83796
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.942357458199412e-07,
+      "loss": 0.4458,
+      "step": 83797
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9421373021612116e-07,
+      "loss": 0.3945,
+      "step": 83798
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.941917150761237e-07,
+      "loss": 0.425,
+      "step": 83799
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.941697003999659e-07,
+      "loss": 0.3618,
+      "step": 83800
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.941476861876648e-07,
+      "loss": 0.3433,
+      "step": 83801
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.941256724392366e-07,
+      "loss": 0.4436,
+      "step": 83802
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9410365915469844e-07,
+      "loss": 0.416,
+      "step": 83803
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.940816463340677e-07,
+      "loss": 0.4443,
+      "step": 83804
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.940596339773604e-07,
+      "loss": 0.3076,
+      "step": 83805
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9403762208459424e-07,
+      "loss": 0.3711,
+      "step": 83806
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.940156106557851e-07,
+      "loss": 0.4456,
+      "step": 83807
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9399359969095037e-07,
+      "loss": 0.4092,
+      "step": 83808
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.939715891901073e-07,
+      "loss": 0.2858,
+      "step": 83809
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.939495791532719e-07,
+      "loss": 0.3289,
+      "step": 83810
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9392756958046147e-07,
+      "loss": 0.5068,
+      "step": 83811
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9390556047169274e-07,
+      "loss": 0.3279,
+      "step": 83812
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.938835518269832e-07,
+      "loss": 0.4531,
+      "step": 83813
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.938615436463486e-07,
+      "loss": 0.3438,
+      "step": 83814
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9383953592980643e-07,
+      "loss": 0.4087,
+      "step": 83815
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9381752867737383e-07,
+      "loss": 0.2812,
+      "step": 83816
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9379552188906695e-07,
+      "loss": 0.314,
+      "step": 83817
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.937735155649032e-07,
+      "loss": 0.3799,
+      "step": 83818
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9375150970489903e-07,
+      "loss": 0.2712,
+      "step": 83819
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.937295043090713e-07,
+      "loss": 0.3096,
+      "step": 83820
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9370749937743743e-07,
+      "loss": 0.4404,
+      "step": 83821
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.936854949100135e-07,
+      "loss": 0.3998,
+      "step": 83822
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.936634909068166e-07,
+      "loss": 0.4355,
+      "step": 83823
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.936414873678637e-07,
+      "loss": 0.4128,
+      "step": 83824
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9361948429317217e-07,
+      "loss": 0.3735,
+      "step": 83825
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.935974816827582e-07,
+      "loss": 0.5117,
+      "step": 83826
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.935754795366384e-07,
+      "loss": 0.4165,
+      "step": 83827
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.935534778548299e-07,
+      "loss": 0.5024,
+      "step": 83828
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9353147663734974e-07,
+      "loss": 0.5259,
+      "step": 83829
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.935094758842149e-07,
+      "loss": 0.2935,
+      "step": 83830
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.934874755954417e-07,
+      "loss": 0.4688,
+      "step": 83831
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9346547577104714e-07,
+      "loss": 0.3982,
+      "step": 83832
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9344347641104827e-07,
+      "loss": 0.4219,
+      "step": 83833
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9342147751546216e-07,
+      "loss": 0.5024,
+      "step": 83834
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9339947908430506e-07,
+      "loss": 0.3916,
+      "step": 83835
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.933774811175943e-07,
+      "loss": 0.4199,
+      "step": 83836
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.933554836153462e-07,
+      "loss": 0.4512,
+      "step": 83837
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9333348657757825e-07,
+      "loss": 0.3325,
+      "step": 83838
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9331149000430676e-07,
+      "loss": 0.3337,
+      "step": 83839
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9328949389554857e-07,
+      "loss": 0.4585,
+      "step": 83840
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9326749825132077e-07,
+      "loss": 0.3546,
+      "step": 83841
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.932455030716405e-07,
+      "loss": 0.3877,
+      "step": 83842
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9322350835652394e-07,
+      "loss": 0.3477,
+      "step": 83843
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.932015141059882e-07,
+      "loss": 0.4121,
+      "step": 83844
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.931795203200502e-07,
+      "loss": 0.4458,
+      "step": 83845
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9315752699872705e-07,
+      "loss": 0.3945,
+      "step": 83846
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.931355341420354e-07,
+      "loss": 0.3135,
+      "step": 83847
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9311354174999144e-07,
+      "loss": 0.3926,
+      "step": 83848
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.930915498226126e-07,
+      "loss": 0.3252,
+      "step": 83849
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9306955835991573e-07,
+      "loss": 0.4233,
+      "step": 83850
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.930475673619178e-07,
+      "loss": 0.3442,
+      "step": 83851
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9302557682863524e-07,
+      "loss": 0.4109,
+      "step": 83852
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9300358676008493e-07,
+      "loss": 0.4971,
+      "step": 83853
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9298159715628397e-07,
+      "loss": 0.3725,
+      "step": 83854
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9295960801724935e-07,
+      "loss": 0.4492,
+      "step": 83855
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9293761934299775e-07,
+      "loss": 0.4517,
+      "step": 83856
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.929156311335455e-07,
+      "loss": 0.3391,
+      "step": 83857
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9289364338890983e-07,
+      "loss": 0.4775,
+      "step": 83858
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.92871656109108e-07,
+      "loss": 0.3984,
+      "step": 83859
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9284966929415606e-07,
+      "loss": 0.244,
+      "step": 83860
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9282768294407107e-07,
+      "loss": 0.4197,
+      "step": 83861
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.928056970588702e-07,
+      "loss": 0.4146,
+      "step": 83862
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9278371163857037e-07,
+      "loss": 0.3281,
+      "step": 83863
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9276172668318776e-07,
+      "loss": 0.4175,
+      "step": 83864
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.927397421927395e-07,
+      "loss": 0.3757,
+      "step": 83865
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9271775816724296e-07,
+      "loss": 0.3594,
+      "step": 83866
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.926957746067141e-07,
+      "loss": 0.4746,
+      "step": 83867
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9267379151117054e-07,
+      "loss": 0.3667,
+      "step": 83868
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9265180888062845e-07,
+      "loss": 0.3892,
+      "step": 83869
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.926298267151048e-07,
+      "loss": 0.4203,
+      "step": 83870
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9260784501461665e-07,
+      "loss": 0.4785,
+      "step": 83871
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9258586377918103e-07,
+      "loss": 0.396,
+      "step": 83872
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9256388300881414e-07,
+      "loss": 0.4263,
+      "step": 83873
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9254190270353324e-07,
+      "loss": 0.3364,
+      "step": 83874
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9251992286335544e-07,
+      "loss": 0.3142,
+      "step": 83875
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9249794348829713e-07,
+      "loss": 0.4487,
+      "step": 83876
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9247596457837485e-07,
+      "loss": 0.3912,
+      "step": 83877
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.924539861336058e-07,
+      "loss": 0.4912,
+      "step": 83878
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.924320081540068e-07,
+      "loss": 0.5146,
+      "step": 83879
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.92410030639595e-07,
+      "loss": 0.5073,
+      "step": 83880
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9238805359038664e-07,
+      "loss": 0.3354,
+      "step": 83881
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9236607700639867e-07,
+      "loss": 0.3691,
+      "step": 83882
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9234410088764826e-07,
+      "loss": 0.3936,
+      "step": 83883
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.923221252341522e-07,
+      "loss": 0.4111,
+      "step": 83884
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9230015004592687e-07,
+      "loss": 0.2786,
+      "step": 83885
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9227817532298966e-07,
+      "loss": 0.3496,
+      "step": 83886
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9225620106535683e-07,
+      "loss": 0.4771,
+      "step": 83887
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9223422727304546e-07,
+      "loss": 0.3921,
+      "step": 83888
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.922122539460727e-07,
+      "loss": 0.4746,
+      "step": 83889
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9219028108445483e-07,
+      "loss": 0.3735,
+      "step": 83890
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.921683086882088e-07,
+      "loss": 0.3959,
+      "step": 83891
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.92146336757352e-07,
+      "loss": 0.4417,
+      "step": 83892
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.921243652919004e-07,
+      "loss": 0.4277,
+      "step": 83893
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.921023942918712e-07,
+      "loss": 0.4297,
+      "step": 83894
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9208042375728123e-07,
+      "loss": 0.2795,
+      "step": 83895
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.920584536881477e-07,
+      "loss": 0.2692,
+      "step": 83896
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.920364840844871e-07,
+      "loss": 0.4551,
+      "step": 83897
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.920145149463158e-07,
+      "loss": 0.4697,
+      "step": 83898
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9199254627365107e-07,
+      "loss": 0.2656,
+      "step": 83899
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.919705780665097e-07,
+      "loss": 0.4385,
+      "step": 83900
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9194861032490876e-07,
+      "loss": 0.3801,
+      "step": 83901
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.919266430488646e-07,
+      "loss": 0.3735,
+      "step": 83902
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9190467623839417e-07,
+      "loss": 0.4556,
+      "step": 83903
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.918827098935144e-07,
+      "loss": 0.3208,
+      "step": 83904
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.918607440142423e-07,
+      "loss": 0.4163,
+      "step": 83905
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9183877860059465e-07,
+      "loss": 0.4382,
+      "step": 83906
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9181681365258754e-07,
+      "loss": 0.2817,
+      "step": 83907
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9179484917023854e-07,
+      "loss": 0.3208,
+      "step": 83908
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.917728851535641e-07,
+      "loss": 0.3503,
+      "step": 83909
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.917509216025816e-07,
+      "loss": 0.4316,
+      "step": 83910
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.917289585173071e-07,
+      "loss": 0.4312,
+      "step": 83911
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.917069958977578e-07,
+      "loss": 0.3469,
+      "step": 83912
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.916850337439508e-07,
+      "loss": 0.3906,
+      "step": 83913
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.916630720559023e-07,
+      "loss": 0.3577,
+      "step": 83914
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.916411108336294e-07,
+      "loss": 0.4312,
+      "step": 83915
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.916191500771492e-07,
+      "loss": 0.5381,
+      "step": 83916
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9159718978647793e-07,
+      "loss": 0.3645,
+      "step": 83917
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9157522996163303e-07,
+      "loss": 0.4424,
+      "step": 83918
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.915532706026308e-07,
+      "loss": 0.4324,
+      "step": 83919
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9153131170948804e-07,
+      "loss": 0.5146,
+      "step": 83920
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.915093532822219e-07,
+      "loss": 0.2793,
+      "step": 83921
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.914873953208494e-07,
+      "loss": 0.4614,
+      "step": 83922
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9146543782538677e-07,
+      "loss": 0.4907,
+      "step": 83923
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9144348079585086e-07,
+      "loss": 0.3901,
+      "step": 83924
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9142152423225894e-07,
+      "loss": 0.5046,
+      "step": 83925
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.913995681346277e-07,
+      "loss": 0.3724,
+      "step": 83926
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9137761250297396e-07,
+      "loss": 0.437,
+      "step": 83927
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9135565733731404e-07,
+      "loss": 0.5156,
+      "step": 83928
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9133370263766507e-07,
+      "loss": 0.4131,
+      "step": 83929
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9131174840404426e-07,
+      "loss": 0.4277,
+      "step": 83930
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.912897946364677e-07,
+      "loss": 0.4917,
+      "step": 83931
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.912678413349526e-07,
+      "loss": 0.3677,
+      "step": 83932
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.912458884995157e-07,
+      "loss": 0.373,
+      "step": 83933
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.912239361301741e-07,
+      "loss": 0.4355,
+      "step": 83934
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.912019842269444e-07,
+      "loss": 0.3523,
+      "step": 83935
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.91180032789843e-07,
+      "loss": 0.4012,
+      "step": 83936
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.911580818188871e-07,
+      "loss": 0.377,
+      "step": 83937
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.911361313140934e-07,
+      "loss": 0.5093,
+      "step": 83938
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9111418127547913e-07,
+      "loss": 0.2825,
+      "step": 83939
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9109223170306036e-07,
+      "loss": 0.3287,
+      "step": 83940
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.910702825968543e-07,
+      "loss": 0.3926,
+      "step": 83941
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9104833395687775e-07,
+      "loss": 0.4082,
+      "step": 83942
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9102638578314786e-07,
+      "loss": 0.3625,
+      "step": 83943
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9100443807568065e-07,
+      "loss": 0.3101,
+      "step": 83944
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9098249083449363e-07,
+      "loss": 0.325,
+      "step": 83945
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.909605440596031e-07,
+      "loss": 0.4031,
+      "step": 83946
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9093859775102634e-07,
+      "loss": 0.356,
+      "step": 83947
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.909166519087795e-07,
+      "loss": 0.3555,
+      "step": 83948
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.908947065328799e-07,
+      "loss": 0.3073,
+      "step": 83949
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9087276162334416e-07,
+      "loss": 0.3292,
+      "step": 83950
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.908508171801895e-07,
+      "loss": 0.3663,
+      "step": 83951
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9082887320343206e-07,
+      "loss": 0.4053,
+      "step": 83952
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.908069296930888e-07,
+      "loss": 0.2779,
+      "step": 83953
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.907849866491768e-07,
+      "loss": 0.4048,
+      "step": 83954
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.90763044071713e-07,
+      "loss": 0.5762,
+      "step": 83955
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9074110196071397e-07,
+      "loss": 0.395,
+      "step": 83956
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.90719160316196e-07,
+      "loss": 0.4688,
+      "step": 83957
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9069721913817643e-07,
+      "loss": 0.2743,
+      "step": 83958
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.90675278426672e-07,
+      "loss": 0.363,
+      "step": 83959
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.906533381816999e-07,
+      "loss": 0.3369,
+      "step": 83960
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.906313984032761e-07,
+      "loss": 0.3452,
+      "step": 83961
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9060945909141775e-07,
+      "loss": 0.4438,
+      "step": 83962
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9058752024614185e-07,
+      "loss": 0.2382,
+      "step": 83963
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9056558186746533e-07,
+      "loss": 0.3105,
+      "step": 83964
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9054364395540476e-07,
+      "loss": 0.356,
+      "step": 83965
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.905217065099765e-07,
+      "loss": 0.4624,
+      "step": 83966
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.904997695311978e-07,
+      "loss": 0.5181,
+      "step": 83967
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.904778330190857e-07,
+      "loss": 0.3618,
+      "step": 83968
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9045589697365633e-07,
+      "loss": 0.4385,
+      "step": 83969
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.904339613949269e-07,
+      "loss": 0.2594,
+      "step": 83970
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9041202628291414e-07,
+      "loss": 0.437,
+      "step": 83971
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9039009163763524e-07,
+      "loss": 0.4873,
+      "step": 83972
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.903681574591062e-07,
+      "loss": 0.3872,
+      "step": 83973
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.903462237473444e-07,
+      "loss": 0.4746,
+      "step": 83974
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.903242905023666e-07,
+      "loss": 0.3799,
+      "step": 83975
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.903023577241893e-07,
+      "loss": 0.4897,
+      "step": 83976
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.902804254128297e-07,
+      "loss": 0.3011,
+      "step": 83977
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.90258493568304e-07,
+      "loss": 0.4438,
+      "step": 83978
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9023656219062927e-07,
+      "loss": 0.4829,
+      "step": 83979
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.902146312798225e-07,
+      "loss": 0.5605,
+      "step": 83980
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.901927008359005e-07,
+      "loss": 0.4561,
+      "step": 83981
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.901707708588797e-07,
+      "loss": 0.3809,
+      "step": 83982
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.901488413487771e-07,
+      "loss": 0.4092,
+      "step": 83983
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.901269123056098e-07,
+      "loss": 0.4297,
+      "step": 83984
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9010498372939437e-07,
+      "loss": 0.3945,
+      "step": 83985
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.900830556201471e-07,
+      "loss": 0.4429,
+      "step": 83986
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.9006112797788514e-07,
+      "loss": 0.4121,
+      "step": 83987
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.900392008026254e-07,
+      "loss": 0.5039,
+      "step": 83988
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.900172740943849e-07,
+      "loss": 0.4673,
+      "step": 83989
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.899953478531798e-07,
+      "loss": 0.3221,
+      "step": 83990
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.899734220790272e-07,
+      "loss": 0.3726,
+      "step": 83991
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8995149677194395e-07,
+      "loss": 0.459,
+      "step": 83992
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.899295719319471e-07,
+      "loss": 0.4585,
+      "step": 83993
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.899076475590527e-07,
+      "loss": 0.4526,
+      "step": 83994
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8988572365327845e-07,
+      "loss": 0.4351,
+      "step": 83995
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8986380021464015e-07,
+      "loss": 0.3823,
+      "step": 83996
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.898418772431552e-07,
+      "loss": 0.3931,
+      "step": 83997
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8981995473884055e-07,
+      "loss": 0.4546,
+      "step": 83998
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.897980327017124e-07,
+      "loss": 0.4695,
+      "step": 83999
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8977611113178786e-07,
+      "loss": 0.3994,
+      "step": 84000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8975419002908395e-07,
+      "loss": 0.4224,
+      "step": 84001
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.897322693936169e-07,
+      "loss": 0.4482,
+      "step": 84002
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.897103492254038e-07,
+      "loss": 0.4507,
+      "step": 84003
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.896884295244615e-07,
+      "loss": 0.4141,
+      "step": 84004
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8966651029080686e-07,
+      "loss": 0.3857,
+      "step": 84005
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.896445915244567e-07,
+      "loss": 0.4248,
+      "step": 84006
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.896226732254272e-07,
+      "loss": 0.5239,
+      "step": 84007
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.896007553937355e-07,
+      "loss": 0.5049,
+      "step": 84008
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.895788380293985e-07,
+      "loss": 0.436,
+      "step": 84009
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8955692113243323e-07,
+      "loss": 0.4658,
+      "step": 84010
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8953500470285583e-07,
+      "loss": 0.3501,
+      "step": 84011
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8951308874068333e-07,
+      "loss": 0.4116,
+      "step": 84012
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8949117324593273e-07,
+      "loss": 0.2777,
+      "step": 84013
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8946925821862086e-07,
+      "loss": 0.4331,
+      "step": 84014
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.894473436587644e-07,
+      "loss": 0.4092,
+      "step": 84015
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.894254295663797e-07,
+      "loss": 0.5278,
+      "step": 84016
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.894035159414839e-07,
+      "loss": 0.4663,
+      "step": 84017
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8938160278409406e-07,
+      "loss": 0.4478,
+      "step": 84018
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.893596900942262e-07,
+      "loss": 0.4509,
+      "step": 84019
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.893377778718977e-07,
+      "loss": 0.3828,
+      "step": 84020
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.893158661171251e-07,
+      "loss": 0.4185,
+      "step": 84021
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8929395482992554e-07,
+      "loss": 0.313,
+      "step": 84022
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.892720440103153e-07,
+      "loss": 0.364,
+      "step": 84023
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.892501336583114e-07,
+      "loss": 0.479,
+      "step": 84024
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8922822377393084e-07,
+      "loss": 0.3535,
+      "step": 84025
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8920631435718977e-07,
+      "loss": 0.3579,
+      "step": 84026
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.891844054081057e-07,
+      "loss": 0.4819,
+      "step": 84027
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.891624969266948e-07,
+      "loss": 0.3838,
+      "step": 84028
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8914058891297406e-07,
+      "loss": 0.3799,
+      "step": 84029
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8911868136696026e-07,
+      "loss": 0.4702,
+      "step": 84030
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.890967742886705e-07,
+      "loss": 0.3787,
+      "step": 84031
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.890748676781209e-07,
+      "loss": 0.282,
+      "step": 84032
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8905296153532863e-07,
+      "loss": 0.4263,
+      "step": 84033
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.890310558603104e-07,
+      "loss": 0.2653,
+      "step": 84034
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.890091506530837e-07,
+      "loss": 0.418,
+      "step": 84035
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8898724591366395e-07,
+      "loss": 0.4053,
+      "step": 84036
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.889653416420685e-07,
+      "loss": 0.4141,
+      "step": 84037
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.889434378383142e-07,
+      "loss": 0.4106,
+      "step": 84038
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8892153450241815e-07,
+      "loss": 0.4482,
+      "step": 84039
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8889963163439653e-07,
+      "loss": 0.4482,
+      "step": 84040
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8887772923426633e-07,
+      "loss": 0.4082,
+      "step": 84041
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.888558273020442e-07,
+      "loss": 0.458,
+      "step": 84042
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.888339258377475e-07,
+      "loss": 0.4585,
+      "step": 84043
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.888120248413923e-07,
+      "loss": 0.4922,
+      "step": 84044
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8879012431299593e-07,
+      "loss": 0.4902,
+      "step": 84045
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.887682242525745e-07,
+      "loss": 0.3845,
+      "step": 84046
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.887463246601451e-07,
+      "loss": 0.4404,
+      "step": 84047
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8872442553572495e-07,
+      "loss": 0.2937,
+      "step": 84048
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8870252687933003e-07,
+      "loss": 0.3457,
+      "step": 84049
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.886806286909774e-07,
+      "loss": 0.3435,
+      "step": 84050
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.88658730970684e-07,
+      "loss": 0.4287,
+      "step": 84051
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8863683371846676e-07,
+      "loss": 0.4429,
+      "step": 84052
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8861493693434187e-07,
+      "loss": 0.394,
+      "step": 84053
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8859304061832656e-07,
+      "loss": 0.4851,
+      "step": 84054
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.885711447704376e-07,
+      "loss": 0.4966,
+      "step": 84055
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8854924939069167e-07,
+      "loss": 0.4915,
+      "step": 84056
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.885273544791051e-07,
+      "loss": 0.4561,
+      "step": 84057
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8850546003569505e-07,
+      "loss": 0.4229,
+      "step": 84058
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8848356606047817e-07,
+      "loss": 0.4829,
+      "step": 84059
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.884616725534717e-07,
+      "loss": 0.4736,
+      "step": 84060
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8843977951469173e-07,
+      "loss": 0.394,
+      "step": 84061
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.884178869441552e-07,
+      "loss": 0.4839,
+      "step": 84062
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.88395994841879e-07,
+      "loss": 0.4941,
+      "step": 84063
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8837410320788035e-07,
+      "loss": 0.3828,
+      "step": 84064
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.883522120421754e-07,
+      "loss": 0.4448,
+      "step": 84065
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.883303213447806e-07,
+      "loss": 0.4038,
+      "step": 84066
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.883084311157133e-07,
+      "loss": 0.4429,
+      "step": 84067
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8828654135499007e-07,
+      "loss": 0.415,
+      "step": 84068
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.88264652062628e-07,
+      "loss": 0.3235,
+      "step": 84069
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.882427632386432e-07,
+      "loss": 0.4468,
+      "step": 84070
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.882208748830529e-07,
+      "loss": 0.3787,
+      "step": 84071
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.881989869958736e-07,
+      "loss": 0.457,
+      "step": 84072
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8817709957712265e-07,
+      "loss": 0.3518,
+      "step": 84073
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.881552126268161e-07,
+      "loss": 0.3569,
+      "step": 84074
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.881333261449712e-07,
+      "loss": 0.3582,
+      "step": 84075
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.881114401316041e-07,
+      "loss": 0.2848,
+      "step": 84076
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.880895545867323e-07,
+      "loss": 0.4912,
+      "step": 84077
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.8806766951037184e-07,
+      "loss": 0.3691,
+      "step": 84078
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.880457849025399e-07,
+      "loss": 0.3809,
+      "step": 84079
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.880239007632531e-07,
+      "loss": 0.416,
+      "step": 84080
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8800201709252865e-07,
+      "loss": 0.3635,
+      "step": 84081
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.879801338903825e-07,
+      "loss": 0.2853,
+      "step": 84082
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.879582511568319e-07,
+      "loss": 0.3396,
+      "step": 84083
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8793636889189353e-07,
+      "loss": 0.3779,
+      "step": 84084
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.879144870955844e-07,
+      "loss": 0.5312,
+      "step": 84085
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8789260576792106e-07,
+      "loss": 0.3428,
+      "step": 84086
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.878707249089198e-07,
+      "loss": 0.3549,
+      "step": 84087
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.878488445185979e-07,
+      "loss": 0.4238,
+      "step": 84088
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8782696459697193e-07,
+      "loss": 0.4053,
+      "step": 84089
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8780508514405896e-07,
+      "loss": 0.311,
+      "step": 84090
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.877832061598753e-07,
+      "loss": 0.3167,
+      "step": 84091
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8776132764443783e-07,
+      "loss": 0.5474,
+      "step": 84092
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.877394495977636e-07,
+      "loss": 0.5249,
+      "step": 84093
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8771757201986887e-07,
+      "loss": 0.4272,
+      "step": 84094
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.87695694910771e-07,
+      "loss": 0.4629,
+      "step": 84095
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8767381827048606e-07,
+      "loss": 0.3474,
+      "step": 84096
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.876519420990311e-07,
+      "loss": 0.392,
+      "step": 84097
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8763006639642313e-07,
+      "loss": 0.3286,
+      "step": 84098
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.876081911626784e-07,
+      "loss": 0.3979,
+      "step": 84099
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.875863163978139e-07,
+      "loss": 0.3367,
+      "step": 84100
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.875644421018465e-07,
+      "loss": 0.519,
+      "step": 84101
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8754256827479304e-07,
+      "loss": 0.4224,
+      "step": 84102
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8752069491666985e-07,
+      "loss": 0.4829,
+      "step": 84103
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8749882202749385e-07,
+      "loss": 0.436,
+      "step": 84104
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8747694960728215e-07,
+      "loss": 0.5405,
+      "step": 84105
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.874550776560509e-07,
+      "loss": 0.3066,
+      "step": 84106
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8743320617381747e-07,
+      "loss": 0.4155,
+      "step": 84107
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.874113351605979e-07,
+      "loss": 0.4868,
+      "step": 84108
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.873894646164093e-07,
+      "loss": 0.3877,
+      "step": 84109
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.873675945412688e-07,
+      "loss": 0.4849,
+      "step": 84110
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8734572493519233e-07,
+      "loss": 0.5166,
+      "step": 84111
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8732385579819717e-07,
+      "loss": 0.377,
+      "step": 84112
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.873019871302998e-07,
+      "loss": 0.3536,
+      "step": 84113
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8728011893151767e-07,
+      "loss": 0.2193,
+      "step": 84114
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8725825120186695e-07,
+      "loss": 0.4253,
+      "step": 84115
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.872363839413639e-07,
+      "loss": 0.4199,
+      "step": 84116
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8721451715002596e-07,
+      "loss": 0.437,
+      "step": 84117
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8719265082786956e-07,
+      "loss": 0.5093,
+      "step": 84118
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.871707849749121e-07,
+      "loss": 0.478,
+      "step": 84119
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8714891959116925e-07,
+      "loss": 0.4072,
+      "step": 84120
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.871270546766583e-07,
+      "loss": 0.312,
+      "step": 84121
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.871051902313961e-07,
+      "loss": 0.4512,
+      "step": 84122
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8708332625539954e-07,
+      "loss": 0.4111,
+      "step": 84123
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.870614627486848e-07,
+      "loss": 0.4702,
+      "step": 84124
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.870395997112692e-07,
+      "loss": 0.3296,
+      "step": 84125
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.870177371431689e-07,
+      "loss": 0.3599,
+      "step": 84126
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8699587504440123e-07,
+      "loss": 0.3357,
+      "step": 84127
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.869740134149824e-07,
+      "loss": 0.3319,
+      "step": 84128
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8695215225492926e-07,
+      "loss": 0.3258,
+      "step": 84129
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.869302915642587e-07,
+      "loss": 0.3745,
+      "step": 84130
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8690843134298777e-07,
+      "loss": 0.4321,
+      "step": 84131
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.868865715911325e-07,
+      "loss": 0.4521,
+      "step": 84132
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8686471230871e-07,
+      "loss": 0.3521,
+      "step": 84133
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.868428534957371e-07,
+      "loss": 0.2231,
+      "step": 84134
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8682099515223064e-07,
+      "loss": 0.3696,
+      "step": 84135
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.867991372782072e-07,
+      "loss": 0.4092,
+      "step": 84136
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8677727987368315e-07,
+      "loss": 0.4321,
+      "step": 84137
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8675542293867547e-07,
+      "loss": 0.322,
+      "step": 84138
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.86733566473201e-07,
+      "loss": 0.3213,
+      "step": 84139
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.867117104772768e-07,
+      "loss": 0.4453,
+      "step": 84140
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8668985495091887e-07,
+      "loss": 0.3911,
+      "step": 84141
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8666799989414434e-07,
+      "loss": 0.2468,
+      "step": 84142
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.866461453069699e-07,
+      "loss": 0.4648,
+      "step": 84143
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8662429118941267e-07,
+      "loss": 0.4775,
+      "step": 84144
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8660243754148903e-07,
+      "loss": 0.3755,
+      "step": 84145
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8658058436321526e-07,
+      "loss": 0.3895,
+      "step": 84146
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.865587316546086e-07,
+      "loss": 0.4087,
+      "step": 84147
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8653687941568616e-07,
+      "loss": 0.4097,
+      "step": 84148
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.865150276464637e-07,
+      "loss": 0.3828,
+      "step": 84149
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8649317634695856e-07,
+      "loss": 0.4023,
+      "step": 84150
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.864713255171874e-07,
+      "loss": 0.4072,
+      "step": 84151
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.864494751571673e-07,
+      "loss": 0.3999,
+      "step": 84152
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.864276252669142e-07,
+      "loss": 0.2937,
+      "step": 84153
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8640577584644527e-07,
+      "loss": 0.2554,
+      "step": 84154
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8638392689577757e-07,
+      "loss": 0.3821,
+      "step": 84155
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.863620784149272e-07,
+      "loss": 0.3945,
+      "step": 84156
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.863402304039115e-07,
+      "loss": 0.356,
+      "step": 84157
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8631838286274656e-07,
+      "loss": 0.2657,
+      "step": 84158
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8629653579144926e-07,
+      "loss": 0.3125,
+      "step": 84159
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.862746891900366e-07,
+      "loss": 0.4238,
+      "step": 84160
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.862528430585256e-07,
+      "loss": 0.3542,
+      "step": 84161
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.862309973969321e-07,
+      "loss": 0.4688,
+      "step": 84162
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.862091522052734e-07,
+      "loss": 0.4048,
+      "step": 84163
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.861873074835664e-07,
+      "loss": 0.3904,
+      "step": 84164
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8616546323182755e-07,
+      "loss": 0.2668,
+      "step": 84165
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.861436194500732e-07,
+      "loss": 0.3872,
+      "step": 84166
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8612177613832064e-07,
+      "loss": 0.4146,
+      "step": 84167
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8609993329658617e-07,
+      "loss": 0.3656,
+      "step": 84168
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.860780909248873e-07,
+      "loss": 0.1659,
+      "step": 84169
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8605624902323973e-07,
+      "loss": 0.4902,
+      "step": 84170
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8603440759166074e-07,
+      "loss": 0.3481,
+      "step": 84171
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.860125666301669e-07,
+      "loss": 0.4047,
+      "step": 84172
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8599072613877546e-07,
+      "loss": 0.3826,
+      "step": 84173
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.859688861175022e-07,
+      "loss": 0.4326,
+      "step": 84174
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.859470465663648e-07,
+      "loss": 0.2064,
+      "step": 84175
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8592520748537915e-07,
+      "loss": 0.3818,
+      "step": 84176
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8590336887456233e-07,
+      "loss": 0.3674,
+      "step": 84177
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.858815307339314e-07,
+      "loss": 0.4756,
+      "step": 84178
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.858596930635023e-07,
+      "loss": 0.3796,
+      "step": 84179
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.858378558632923e-07,
+      "loss": 0.3787,
+      "step": 84180
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.858160191333183e-07,
+      "loss": 0.2665,
+      "step": 84181
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.857941828735964e-07,
+      "loss": 0.4634,
+      "step": 84182
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8577234708414376e-07,
+      "loss": 0.4521,
+      "step": 84183
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8575051176497687e-07,
+      "loss": 0.3511,
+      "step": 84184
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.85728676916113e-07,
+      "loss": 0.3774,
+      "step": 84185
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8570684253756836e-07,
+      "loss": 0.4287,
+      "step": 84186
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8568500862935935e-07,
+      "loss": 0.4263,
+      "step": 84187
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8566317519150314e-07,
+      "loss": 0.3887,
+      "step": 84188
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8564134222401645e-07,
+      "loss": 0.4502,
+      "step": 84189
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8561950972691626e-07,
+      "loss": 0.276,
+      "step": 84190
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.855976777002186e-07,
+      "loss": 0.4351,
+      "step": 84191
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8557584614394055e-07,
+      "loss": 0.3682,
+      "step": 84192
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.855540150580988e-07,
+      "loss": 0.3306,
+      "step": 84193
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.855321844427104e-07,
+      "loss": 0.2659,
+      "step": 84194
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.855103542977919e-07,
+      "loss": 0.3265,
+      "step": 84195
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.854885246233593e-07,
+      "loss": 0.4409,
+      "step": 84196
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8546669541943e-07,
+      "loss": 0.3264,
+      "step": 84197
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8544486668602094e-07,
+      "loss": 0.3774,
+      "step": 84198
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8542303842314815e-07,
+      "loss": 0.4482,
+      "step": 84199
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8540121063082873e-07,
+      "loss": 0.4492,
+      "step": 84200
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8537938330907926e-07,
+      "loss": 0.3538,
+      "step": 84201
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.853575564579169e-07,
+      "loss": 0.4058,
+      "step": 84202
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8533573007735767e-07,
+      "loss": 0.562,
+      "step": 84203
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.853139041674186e-07,
+      "loss": 0.3722,
+      "step": 84204
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8529207872811677e-07,
+      "loss": 0.3877,
+      "step": 84205
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.852702537594681e-07,
+      "loss": 0.4854,
+      "step": 84206
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8524842926149014e-07,
+      "loss": 0.436,
+      "step": 84207
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8522660523419893e-07,
+      "loss": 0.4888,
+      "step": 84208
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.852047816776113e-07,
+      "loss": 0.3679,
+      "step": 84209
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.851829585917442e-07,
+      "loss": 0.4048,
+      "step": 84210
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8516113597661446e-07,
+      "loss": 0.4272,
+      "step": 84211
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.851393138322383e-07,
+      "loss": 0.4248,
+      "step": 84212
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8511749215863276e-07,
+      "loss": 0.3979,
+      "step": 84213
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.850956709558144e-07,
+      "loss": 0.3203,
+      "step": 84214
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.850738502238008e-07,
+      "loss": 0.3835,
+      "step": 84215
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.85052029962607e-07,
+      "loss": 0.4102,
+      "step": 84216
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.850302101722508e-07,
+      "loss": 0.4878,
+      "step": 84217
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.850083908527485e-07,
+      "loss": 0.3857,
+      "step": 84218
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8498657200411744e-07,
+      "loss": 0.4873,
+      "step": 84219
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8496475362637357e-07,
+      "loss": 0.354,
+      "step": 84220
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8494293571953385e-07,
+      "loss": 0.3892,
+      "step": 84221
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8492111828361507e-07,
+      "loss": 0.3813,
+      "step": 84222
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.848993013186342e-07,
+      "loss": 0.4187,
+      "step": 84223
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8487748482460735e-07,
+      "loss": 0.3774,
+      "step": 84224
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8485566880155186e-07,
+      "loss": 0.4526,
+      "step": 84225
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.848338532494838e-07,
+      "loss": 0.4541,
+      "step": 84226
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.848120381684201e-07,
+      "loss": 0.4839,
+      "step": 84227
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.847902235583779e-07,
+      "loss": 0.4021,
+      "step": 84228
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.847684094193733e-07,
+      "loss": 0.4072,
+      "step": 84229
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.847465957514231e-07,
+      "loss": 0.2708,
+      "step": 84230
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.847247825545442e-07,
+      "loss": 0.2715,
+      "step": 84231
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.847029698287535e-07,
+      "loss": 0.4561,
+      "step": 84232
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8468115757406715e-07,
+      "loss": 0.2569,
+      "step": 84233
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.846593457905026e-07,
+      "loss": 0.5361,
+      "step": 84234
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.846375344780757e-07,
+      "loss": 0.4097,
+      "step": 84235
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.846157236368038e-07,
+      "loss": 0.4106,
+      "step": 84236
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.845939132667031e-07,
+      "loss": 0.3967,
+      "step": 84237
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8457210336779045e-07,
+      "loss": 0.478,
+      "step": 84238
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8455029394008263e-07,
+      "loss": 0.3999,
+      "step": 84239
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.845284849835968e-07,
+      "loss": 0.3646,
+      "step": 84240
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8450667649834877e-07,
+      "loss": 0.3723,
+      "step": 84241
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.844848684843558e-07,
+      "loss": 0.4331,
+      "step": 84242
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.844630609416344e-07,
+      "loss": 0.3784,
+      "step": 84243
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8444125387020164e-07,
+      "loss": 0.438,
+      "step": 84244
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.844194472700739e-07,
+      "loss": 0.437,
+      "step": 84245
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.843976411412676e-07,
+      "loss": 0.3406,
+      "step": 84246
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.843758354837997e-07,
+      "loss": 0.4204,
+      "step": 84247
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8435403029768685e-07,
+      "loss": 0.3081,
+      "step": 84248
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.843322255829463e-07,
+      "loss": 0.4331,
+      "step": 84249
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.843104213395938e-07,
+      "loss": 0.386,
+      "step": 84250
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.842886175676465e-07,
+      "loss": 0.4028,
+      "step": 84251
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8426681426712113e-07,
+      "loss": 0.4663,
+      "step": 84252
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.842450114380347e-07,
+      "loss": 0.4873,
+      "step": 84253
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.842232090804035e-07,
+      "loss": 0.3774,
+      "step": 84254
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.84201407194244e-07,
+      "loss": 0.3594,
+      "step": 84255
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.841796057795731e-07,
+      "loss": 0.4937,
+      "step": 84256
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.841578048364079e-07,
+      "loss": 0.3438,
+      "step": 84257
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.841360043647645e-07,
+      "loss": 0.3179,
+      "step": 84258
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.841142043646598e-07,
+      "loss": 0.3424,
+      "step": 84259
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.840924048361105e-07,
+      "loss": 0.4575,
+      "step": 84260
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.840706057791338e-07,
+      "loss": 0.3901,
+      "step": 84261
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.840488071937453e-07,
+      "loss": 0.479,
+      "step": 84262
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8402700907996254e-07,
+      "loss": 0.479,
+      "step": 84263
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8400521143780216e-07,
+      "loss": 0.5093,
+      "step": 84264
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.839834142672804e-07,
+      "loss": 0.3833,
+      "step": 84265
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.839616175684146e-07,
+      "loss": 0.4897,
+      "step": 84266
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8393982134122073e-07,
+      "loss": 0.4707,
+      "step": 84267
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.839180255857157e-07,
+      "loss": 0.3882,
+      "step": 84268
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8389623030191645e-07,
+      "loss": 0.4805,
+      "step": 84269
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.838744354898398e-07,
+      "loss": 0.429,
+      "step": 84270
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.838526411495018e-07,
+      "loss": 0.3398,
+      "step": 84271
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.838308472809195e-07,
+      "loss": 0.5142,
+      "step": 84272
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8380905388411e-07,
+      "loss": 0.3932,
+      "step": 84273
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.837872609590895e-07,
+      "loss": 0.4172,
+      "step": 84274
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.837654685058744e-07,
+      "loss": 0.3262,
+      "step": 84275
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.837436765244818e-07,
+      "loss": 0.418,
+      "step": 84276
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8372188501492843e-07,
+      "loss": 0.2736,
+      "step": 84277
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.83700093977231e-07,
+      "loss": 0.3618,
+      "step": 84278
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8367830341140596e-07,
+      "loss": 0.323,
+      "step": 84279
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8365651331746995e-07,
+      "loss": 0.3608,
+      "step": 84280
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.836347236954398e-07,
+      "loss": 0.4932,
+      "step": 84281
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.836129345453326e-07,
+      "loss": 0.4487,
+      "step": 84282
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8359114586716425e-07,
+      "loss": 0.3433,
+      "step": 84283
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8356935766095223e-07,
+      "loss": 0.4009,
+      "step": 84284
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.835475699267123e-07,
+      "loss": 0.328,
+      "step": 84285
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8352578266446187e-07,
+      "loss": 0.2615,
+      "step": 84286
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.835039958742177e-07,
+      "loss": 0.4229,
+      "step": 84287
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8348220955599567e-07,
+      "loss": 0.3984,
+      "step": 84288
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8346042370981315e-07,
+      "loss": 0.4204,
+      "step": 84289
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.834386383356868e-07,
+      "loss": 0.3882,
+      "step": 84290
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8341685343363295e-07,
+      "loss": 0.3834,
+      "step": 84291
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8339506900366835e-07,
+      "loss": 0.5098,
+      "step": 84292
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.833732850458099e-07,
+      "loss": 0.3262,
+      "step": 84293
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8335150156007443e-07,
+      "loss": 0.3896,
+      "step": 84294
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8332971854647836e-07,
+      "loss": 0.3452,
+      "step": 84295
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.833079360050381e-07,
+      "loss": 0.4727,
+      "step": 84296
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8328615393577056e-07,
+      "loss": 0.4517,
+      "step": 84297
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8326437233869235e-07,
+      "loss": 0.3682,
+      "step": 84298
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.832425912138207e-07,
+      "loss": 0.2932,
+      "step": 84299
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.832208105611715e-07,
+      "loss": 0.353,
+      "step": 84300
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.831990303807617e-07,
+      "loss": 0.4316,
+      "step": 84301
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.831772506726082e-07,
+      "loss": 0.3386,
+      "step": 84302
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.831554714367278e-07,
+      "loss": 0.3462,
+      "step": 84303
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.831336926731368e-07,
+      "loss": 0.48,
+      "step": 84304
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8311191438185167e-07,
+      "loss": 0.4585,
+      "step": 84305
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.830901365628894e-07,
+      "loss": 0.344,
+      "step": 84306
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.830683592162669e-07,
+      "loss": 0.3119,
+      "step": 84307
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8304658234200037e-07,
+      "loss": 0.4624,
+      "step": 84308
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.830248059401067e-07,
+      "loss": 0.4058,
+      "step": 84309
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.830030300106024e-07,
+      "loss": 0.3367,
+      "step": 84310
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8298125455350485e-07,
+      "loss": 0.2988,
+      "step": 84311
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.829594795688298e-07,
+      "loss": 0.3341,
+      "step": 84312
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.829377050565942e-07,
+      "loss": 0.3652,
+      "step": 84313
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.829159310168152e-07,
+      "loss": 0.4761,
+      "step": 84314
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.828941574495088e-07,
+      "loss": 0.4883,
+      "step": 84315
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.828723843546924e-07,
+      "loss": 0.2949,
+      "step": 84316
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8285061173238175e-07,
+      "loss": 0.439,
+      "step": 84317
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8282883958259404e-07,
+      "loss": 0.3008,
+      "step": 84318
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.828070679053459e-07,
+      "loss": 0.4746,
+      "step": 84319
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.827852967006544e-07,
+      "loss": 0.502,
+      "step": 84320
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.827635259685354e-07,
+      "loss": 0.4512,
+      "step": 84321
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8274175570900603e-07,
+      "loss": 0.2839,
+      "step": 84322
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.82719985922083e-07,
+      "loss": 0.3569,
+      "step": 84323
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8269821660778345e-07,
+      "loss": 0.2767,
+      "step": 84324
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8267644776612284e-07,
+      "loss": 0.3823,
+      "step": 84325
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.826546793971186e-07,
+      "loss": 0.48,
+      "step": 84326
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.826329115007871e-07,
+      "loss": 0.4512,
+      "step": 84327
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8261114407714567e-07,
+      "loss": 0.3555,
+      "step": 84328
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.825893771262101e-07,
+      "loss": 0.4399,
+      "step": 84329
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.825676106479975e-07,
+      "loss": 0.3472,
+      "step": 84330
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8254584464252447e-07,
+      "loss": 0.4023,
+      "step": 84331
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.82524079109808e-07,
+      "loss": 0.4951,
+      "step": 84332
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.825023140498641e-07,
+      "loss": 0.4331,
+      "step": 84333
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.824805494627102e-07,
+      "loss": 0.416,
+      "step": 84334
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.824587853483622e-07,
+      "loss": 0.3064,
+      "step": 84335
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.82437021706837e-07,
+      "loss": 0.4326,
+      "step": 84336
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.824152585381518e-07,
+      "loss": 0.4487,
+      "step": 84337
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.823934958423225e-07,
+      "loss": 0.4648,
+      "step": 84338
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8237173361936615e-07,
+      "loss": 0.4604,
+      "step": 84339
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.823499718692993e-07,
+      "loss": 0.3467,
+      "step": 84340
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8232821059213906e-07,
+      "loss": 0.4546,
+      "step": 84341
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.823064497879013e-07,
+      "loss": 0.4556,
+      "step": 84342
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.822846894566031e-07,
+      "loss": 0.4673,
+      "step": 84343
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8226292959826156e-07,
+      "loss": 0.3625,
+      "step": 84344
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.822411702128928e-07,
+      "loss": 0.46,
+      "step": 84345
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.822194113005132e-07,
+      "loss": 0.3496,
+      "step": 84346
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8219765286113983e-07,
+      "loss": 0.4102,
+      "step": 84347
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.821758948947893e-07,
+      "loss": 0.449,
+      "step": 84348
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.821541374014786e-07,
+      "loss": 0.3338,
+      "step": 84349
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8213238038122375e-07,
+      "loss": 0.3125,
+      "step": 84350
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.821106238340418e-07,
+      "loss": 0.3848,
+      "step": 84351
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.820888677599492e-07,
+      "loss": 0.4443,
+      "step": 84352
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8206711215896314e-07,
+      "loss": 0.2826,
+      "step": 84353
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.820453570311e-07,
+      "loss": 0.281,
+      "step": 84354
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.820236023763758e-07,
+      "loss": 0.3745,
+      "step": 84355
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.820018481948077e-07,
+      "loss": 0.3584,
+      "step": 84356
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8198009448641244e-07,
+      "loss": 0.3933,
+      "step": 84357
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.81958341251207e-07,
+      "loss": 0.5293,
+      "step": 84358
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8193658848920727e-07,
+      "loss": 0.3752,
+      "step": 84359
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.819148362004302e-07,
+      "loss": 0.2527,
+      "step": 84360
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.81893084384893e-07,
+      "loss": 0.457,
+      "step": 84361
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.818713330426113e-07,
+      "loss": 0.4912,
+      "step": 84362
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8184958217360243e-07,
+      "loss": 0.3926,
+      "step": 84363
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8182783177788323e-07,
+      "loss": 0.3691,
+      "step": 84364
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.818060818554697e-07,
+      "loss": 0.478,
+      "step": 84365
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.817843324063792e-07,
+      "loss": 0.4385,
+      "step": 84366
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8176258343062763e-07,
+      "loss": 0.4072,
+      "step": 84367
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8174083492823205e-07,
+      "loss": 0.3174,
+      "step": 84368
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8171908689920904e-07,
+      "loss": 0.4243,
+      "step": 84369
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8169733934357575e-07,
+      "loss": 0.4414,
+      "step": 84370
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8167559226134795e-07,
+      "loss": 0.396,
+      "step": 84371
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8165384565254265e-07,
+      "loss": 0.4272,
+      "step": 84372
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8163209951717667e-07,
+      "loss": 0.3478,
+      "step": 84373
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8161035385526685e-07,
+      "loss": 0.3782,
+      "step": 84374
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.815886086668295e-07,
+      "loss": 0.5181,
+      "step": 84375
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8156686395188107e-07,
+      "loss": 0.5161,
+      "step": 84376
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8154511971043846e-07,
+      "loss": 0.3273,
+      "step": 84377
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.815233759425186e-07,
+      "loss": 0.4141,
+      "step": 84378
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.815016326481375e-07,
+      "loss": 0.3896,
+      "step": 84379
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.814798898273122e-07,
+      "loss": 0.4136,
+      "step": 84380
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8145814748005934e-07,
+      "loss": 0.2812,
+      "step": 84381
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.814364056063958e-07,
+      "loss": 0.3561,
+      "step": 84382
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.814146642063376e-07,
+      "loss": 0.3828,
+      "step": 84383
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.813929232799021e-07,
+      "loss": 0.2969,
+      "step": 84384
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8137118282710536e-07,
+      "loss": 0.4287,
+      "step": 84385
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.813494428479642e-07,
+      "loss": 0.3525,
+      "step": 84386
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.813277033424956e-07,
+      "loss": 0.3008,
+      "step": 84387
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.813059643107157e-07,
+      "loss": 0.3318,
+      "step": 84388
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.812842257526414e-07,
+      "loss": 0.4492,
+      "step": 84389
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.812624876682892e-07,
+      "loss": 0.2203,
+      "step": 84390
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.812407500576763e-07,
+      "loss": 0.418,
+      "step": 84391
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.812190129208185e-07,
+      "loss": 0.2435,
+      "step": 84392
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8119727625773277e-07,
+      "loss": 0.3525,
+      "step": 84393
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8117554006843634e-07,
+      "loss": 0.4321,
+      "step": 84394
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.811538043529453e-07,
+      "loss": 0.2795,
+      "step": 84395
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8113206911127595e-07,
+      "loss": 0.2905,
+      "step": 84396
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.811103343434453e-07,
+      "loss": 0.2708,
+      "step": 84397
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.810886000494701e-07,
+      "loss": 0.3076,
+      "step": 84398
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.810668662293671e-07,
+      "loss": 0.4604,
+      "step": 84399
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8104513288315256e-07,
+      "loss": 0.3696,
+      "step": 84400
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.810234000108432e-07,
+      "loss": 0.4302,
+      "step": 84401
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.810016676124558e-07,
+      "loss": 0.4097,
+      "step": 84402
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8097993568800737e-07,
+      "loss": 0.3563,
+      "step": 84403
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.80958204237514e-07,
+      "loss": 0.5146,
+      "step": 84404
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.809364732609922e-07,
+      "loss": 0.3999,
+      "step": 84405
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.809147427584589e-07,
+      "loss": 0.4065,
+      "step": 84406
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.808930127299307e-07,
+      "loss": 0.4985,
+      "step": 84407
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.808712831754245e-07,
+      "loss": 0.4106,
+      "step": 84408
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8084955409495647e-07,
+      "loss": 0.4321,
+      "step": 84409
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.808278254885434e-07,
+      "loss": 0.353,
+      "step": 84410
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.808060973562021e-07,
+      "loss": 0.4829,
+      "step": 84411
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8078436969794924e-07,
+      "loss": 0.5312,
+      "step": 84412
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8076264251380106e-07,
+      "loss": 0.4604,
+      "step": 84413
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8074091580377486e-07,
+      "loss": 0.395,
+      "step": 84414
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.807191895678865e-07,
+      "loss": 0.3743,
+      "step": 84415
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.806974638061533e-07,
+      "loss": 0.4053,
+      "step": 84416
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8067573851859125e-07,
+      "loss": 0.5068,
+      "step": 84417
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.806540137052173e-07,
+      "loss": 0.323,
+      "step": 84418
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.806322893660482e-07,
+      "loss": 0.4058,
+      "step": 84419
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.806105655011007e-07,
+      "loss": 0.4346,
+      "step": 84420
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8058884211039087e-07,
+      "loss": 0.4287,
+      "step": 84421
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.805671191939357e-07,
+      "loss": 0.215,
+      "step": 84422
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.805453967517518e-07,
+      "loss": 0.3416,
+      "step": 84423
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8052367478385616e-07,
+      "loss": 0.4121,
+      "step": 84424
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.80501953290265e-07,
+      "loss": 0.438,
+      "step": 84425
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8048023227099477e-07,
+      "loss": 0.4038,
+      "step": 84426
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8045851172606225e-07,
+      "loss": 0.2635,
+      "step": 84427
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8043679165548427e-07,
+      "loss": 0.522,
+      "step": 84428
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8041507205927757e-07,
+      "loss": 0.478,
+      "step": 84429
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8039335293745823e-07,
+      "loss": 0.5332,
+      "step": 84430
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8037163429004325e-07,
+      "loss": 0.3965,
+      "step": 84431
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8034991611704926e-07,
+      "loss": 0.4783,
+      "step": 84432
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.803281984184932e-07,
+      "loss": 0.481,
+      "step": 84433
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8030648119439124e-07,
+      "loss": 0.3708,
+      "step": 84434
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.802847644447598e-07,
+      "loss": 0.3208,
+      "step": 84435
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8026304816961584e-07,
+      "loss": 0.4463,
+      "step": 84436
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.802413323689764e-07,
+      "loss": 0.3569,
+      "step": 84437
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.802196170428572e-07,
+      "loss": 0.3711,
+      "step": 84438
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8019790219127533e-07,
+      "loss": 0.4185,
+      "step": 84439
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.801761878142475e-07,
+      "loss": 0.4546,
+      "step": 84440
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.801544739117906e-07,
+      "loss": 0.3613,
+      "step": 84441
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.801327604839205e-07,
+      "loss": 0.4104,
+      "step": 84442
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8011104753065436e-07,
+      "loss": 0.4502,
+      "step": 84443
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.80089335052009e-07,
+      "loss": 0.3201,
+      "step": 84444
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8006762304800035e-07,
+      "loss": 0.4172,
+      "step": 84445
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.8004591151864575e-07,
+      "loss": 0.4683,
+      "step": 84446
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.800242004639612e-07,
+      "loss": 0.4238,
+      "step": 84447
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.800024898839636e-07,
+      "loss": 0.4121,
+      "step": 84448
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.799807797786695e-07,
+      "loss": 0.4219,
+      "step": 84449
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.79959070148096e-07,
+      "loss": 0.3511,
+      "step": 84450
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7993736099225903e-07,
+      "loss": 0.4312,
+      "step": 84451
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.799156523111755e-07,
+      "loss": 0.4839,
+      "step": 84452
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.798939441048624e-07,
+      "loss": 0.3479,
+      "step": 84453
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.79872236373336e-07,
+      "loss": 0.3074,
+      "step": 84454
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.798505291166125e-07,
+      "loss": 0.3755,
+      "step": 84455
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7982882233470893e-07,
+      "loss": 0.5366,
+      "step": 84456
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7980711602764205e-07,
+      "loss": 0.3506,
+      "step": 84457
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.797854101954285e-07,
+      "loss": 0.3647,
+      "step": 84458
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.797637048380845e-07,
+      "loss": 0.3625,
+      "step": 84459
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7974199995562695e-07,
+      "loss": 0.3369,
+      "step": 84460
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.797202955480724e-07,
+      "loss": 0.3911,
+      "step": 84461
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7969859161543785e-07,
+      "loss": 0.3202,
+      "step": 84462
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7967688815773914e-07,
+      "loss": 0.4009,
+      "step": 84463
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7965518517499373e-07,
+      "loss": 0.498,
+      "step": 84464
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.796334826672175e-07,
+      "loss": 0.3225,
+      "step": 84465
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.7961178063442735e-07,
+      "loss": 0.1563,
+      "step": 84466
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.795900790766402e-07,
+      "loss": 0.521,
+      "step": 84467
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.795683779938722e-07,
+      "loss": 0.4233,
+      "step": 84468
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.795466773861401e-07,
+      "loss": 0.4492,
+      "step": 84469
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.795249772534609e-07,
+      "loss": 0.5039,
+      "step": 84470
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.795032775958505e-07,
+      "loss": 0.5156,
+      "step": 84471
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7948157841332605e-07,
+      "loss": 0.4766,
+      "step": 84472
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.79459879705904e-07,
+      "loss": 0.3071,
+      "step": 84473
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.794381814736013e-07,
+      "loss": 0.4795,
+      "step": 84474
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.794164837164342e-07,
+      "loss": 0.4761,
+      "step": 84475
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.79394786434419e-07,
+      "loss": 0.2471,
+      "step": 84476
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7937308962757277e-07,
+      "loss": 0.355,
+      "step": 84477
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7935139329591206e-07,
+      "loss": 0.4022,
+      "step": 84478
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.793296974394536e-07,
+      "loss": 0.4243,
+      "step": 84479
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.793080020582137e-07,
+      "loss": 0.301,
+      "step": 84480
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.792863071522091e-07,
+      "loss": 0.3804,
+      "step": 84481
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7926461272145637e-07,
+      "loss": 0.3289,
+      "step": 84482
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.792429187659726e-07,
+      "loss": 0.417,
+      "step": 84483
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.792212252857738e-07,
+      "loss": 0.3364,
+      "step": 84484
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7919953228087666e-07,
+      "loss": 0.3848,
+      "step": 84485
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7917783975129766e-07,
+      "loss": 0.4243,
+      "step": 84486
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.791561476970542e-07,
+      "loss": 0.5098,
+      "step": 84487
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7913445611816184e-07,
+      "loss": 0.5591,
+      "step": 84488
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.791127650146377e-07,
+      "loss": 0.4402,
+      "step": 84489
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7909107438649845e-07,
+      "loss": 0.448,
+      "step": 84490
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.790693842337609e-07,
+      "loss": 0.3923,
+      "step": 84491
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7904769455644103e-07,
+      "loss": 0.3823,
+      "step": 84492
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7902600535455586e-07,
+      "loss": 0.4002,
+      "step": 84493
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7900431662812213e-07,
+      "loss": 0.3857,
+      "step": 84494
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.78982628377156e-07,
+      "loss": 0.4094,
+      "step": 84495
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7896094060167463e-07,
+      "loss": 0.5796,
+      "step": 84496
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7893925330169394e-07,
+      "loss": 0.5063,
+      "step": 84497
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7891756647723093e-07,
+      "loss": 0.606,
+      "step": 84498
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.788958801283022e-07,
+      "loss": 0.3701,
+      "step": 84499
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7887419425492475e-07,
+      "loss": 0.46,
+      "step": 84500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.788525088571143e-07,
+      "loss": 0.3066,
+      "step": 84501
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.78830823934888e-07,
+      "loss": 0.4714,
+      "step": 84502
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7880913948826234e-07,
+      "loss": 0.4026,
+      "step": 84503
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7878745551725464e-07,
+      "loss": 0.3147,
+      "step": 84504
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.787657720218801e-07,
+      "loss": 0.4409,
+      "step": 84505
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7874408900215604e-07,
+      "loss": 0.4893,
+      "step": 84506
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7872240645809917e-07,
+      "loss": 0.4814,
+      "step": 84507
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.787007243897262e-07,
+      "loss": 0.3101,
+      "step": 84508
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.786790427970533e-07,
+      "loss": 0.3357,
+      "step": 84509
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.786573616800972e-07,
+      "loss": 0.4561,
+      "step": 84510
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7863568103887464e-07,
+      "loss": 0.374,
+      "step": 84511
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7861400087340244e-07,
+      "loss": 0.3665,
+      "step": 84512
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.785923211836967e-07,
+      "loss": 0.3745,
+      "step": 84513
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.785706419697745e-07,
+      "loss": 0.5669,
+      "step": 84514
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7854896323165187e-07,
+      "loss": 0.4404,
+      "step": 84515
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.785272849693457e-07,
+      "loss": 0.46,
+      "step": 84516
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7850560718287295e-07,
+      "loss": 0.3757,
+      "step": 84517
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7848392987224964e-07,
+      "loss": 0.4243,
+      "step": 84518
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7846225303749247e-07,
+      "loss": 0.395,
+      "step": 84519
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.784405766786183e-07,
+      "loss": 0.3293,
+      "step": 84520
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.784189007956439e-07,
+      "loss": 0.4316,
+      "step": 84521
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.783972253885851e-07,
+      "loss": 0.3052,
+      "step": 84522
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.783755504574591e-07,
+      "loss": 0.5083,
+      "step": 84523
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.783538760022826e-07,
+      "loss": 0.5322,
+      "step": 84524
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7833220202307203e-07,
+      "loss": 0.4165,
+      "step": 84525
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7831052851984356e-07,
+      "loss": 0.3866,
+      "step": 84526
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7828885549261415e-07,
+      "loss": 0.2363,
+      "step": 84527
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7826718294140037e-07,
+      "loss": 0.3796,
+      "step": 84528
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7824551086621915e-07,
+      "loss": 0.4355,
+      "step": 84529
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7822383926708633e-07,
+      "loss": 0.4736,
+      "step": 84530
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7820216814401907e-07,
+      "loss": 0.3982,
+      "step": 84531
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.781804974970337e-07,
+      "loss": 0.3511,
+      "step": 84532
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7815882732614736e-07,
+      "loss": 0.4146,
+      "step": 84533
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.781371576313761e-07,
+      "loss": 0.2974,
+      "step": 84534
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7811548841273644e-07,
+      "loss": 0.3965,
+      "step": 84535
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.78093819670245e-07,
+      "loss": 0.541,
+      "step": 84536
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.780721514039186e-07,
+      "loss": 0.3308,
+      "step": 84537
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7805048361377413e-07,
+      "loss": 0.5005,
+      "step": 84538
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.780288162998274e-07,
+      "loss": 0.3833,
+      "step": 84539
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7800714946209545e-07,
+      "loss": 0.4849,
+      "step": 84540
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.779854831005952e-07,
+      "loss": 0.3791,
+      "step": 84541
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.779638172153424e-07,
+      "loss": 0.4009,
+      "step": 84542
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.779421518063546e-07,
+      "loss": 0.3094,
+      "step": 84543
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.779204868736474e-07,
+      "loss": 0.4482,
+      "step": 84544
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.778988224172379e-07,
+      "loss": 0.3998,
+      "step": 84545
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.778771584371432e-07,
+      "loss": 0.3789,
+      "step": 84546
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.778554949333788e-07,
+      "loss": 0.2621,
+      "step": 84547
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7783383190596195e-07,
+      "loss": 0.3048,
+      "step": 84548
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7781216935490903e-07,
+      "loss": 0.3618,
+      "step": 84549
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.777905072802371e-07,
+      "loss": 0.4453,
+      "step": 84550
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.777688456819621e-07,
+      "loss": 0.3627,
+      "step": 84551
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.777471845601008e-07,
+      "loss": 0.4082,
+      "step": 84552
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7772552391467027e-07,
+      "loss": 0.541,
+      "step": 84553
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.777038637456863e-07,
+      "loss": 0.3506,
+      "step": 84554
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7768220405316633e-07,
+      "loss": 0.3296,
+      "step": 84555
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7766054483712597e-07,
+      "loss": 0.3494,
+      "step": 84556
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7763888609758245e-07,
+      "loss": 0.3412,
+      "step": 84557
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.776172278345524e-07,
+      "loss": 0.3398,
+      "step": 84558
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.77595570048052e-07,
+      "loss": 0.4255,
+      "step": 84559
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.77573912738098e-07,
+      "loss": 0.2875,
+      "step": 84560
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7755225590470707e-07,
+      "loss": 0.4414,
+      "step": 84561
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7753059954789614e-07,
+      "loss": 0.3633,
+      "step": 84562
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7750894366768137e-07,
+      "loss": 0.3293,
+      "step": 84563
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.774872882640791e-07,
+      "loss": 0.261,
+      "step": 84564
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7746563333710603e-07,
+      "loss": 0.5552,
+      "step": 84565
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.77443978886779e-07,
+      "loss": 0.4829,
+      "step": 84566
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7742232491311487e-07,
+      "loss": 0.4292,
+      "step": 84567
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.774006714161294e-07,
+      "loss": 0.3037,
+      "step": 84568
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.773790183958396e-07,
+      "loss": 0.3801,
+      "step": 84569
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7735736585226217e-07,
+      "loss": 0.3481,
+      "step": 84570
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7733571378541387e-07,
+      "loss": 0.3711,
+      "step": 84571
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.773140621953106e-07,
+      "loss": 0.3865,
+      "step": 84572
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7729241108196973e-07,
+      "loss": 0.4299,
+      "step": 84573
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.772707604454071e-07,
+      "loss": 0.5508,
+      "step": 84574
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7724911028563986e-07,
+      "loss": 0.436,
+      "step": 84575
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7722746060268405e-07,
+      "loss": 0.4028,
+      "step": 84576
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.772058113965565e-07,
+      "loss": 0.3977,
+      "step": 84577
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.771841626672739e-07,
+      "loss": 0.4341,
+      "step": 84578
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.77162514414853e-07,
+      "loss": 0.6299,
+      "step": 84579
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.771408666393099e-07,
+      "loss": 0.5884,
+      "step": 84580
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7711921934066125e-07,
+      "loss": 0.4056,
+      "step": 84581
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.770975725189239e-07,
+      "loss": 0.4712,
+      "step": 84582
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7707592617411453e-07,
+      "loss": 0.4609,
+      "step": 84583
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7705428030624965e-07,
+      "loss": 0.4155,
+      "step": 84584
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.770326349153452e-07,
+      "loss": 0.4116,
+      "step": 84585
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.770109900014182e-07,
+      "loss": 0.3977,
+      "step": 84586
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7698934556448525e-07,
+      "loss": 0.4039,
+      "step": 84587
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7696770160456325e-07,
+      "loss": 0.4114,
+      "step": 84588
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.769460581216681e-07,
+      "loss": 0.3553,
+      "step": 84589
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7692441511581676e-07,
+      "loss": 0.3511,
+      "step": 84590
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7690277258702575e-07,
+      "loss": 0.2913,
+      "step": 84591
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.768811305353119e-07,
+      "loss": 0.4023,
+      "step": 84592
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7685948896069153e-07,
+      "loss": 0.4185,
+      "step": 84593
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7683784786318084e-07,
+      "loss": 0.3792,
+      "step": 84594
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.768162072427968e-07,
+      "loss": 0.3728,
+      "step": 84595
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7679456709955625e-07,
+      "loss": 0.3987,
+      "step": 84596
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7677292743347513e-07,
+      "loss": 0.3271,
+      "step": 84597
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.767512882445704e-07,
+      "loss": 0.2659,
+      "step": 84598
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.767296495328583e-07,
+      "loss": 0.325,
+      "step": 84599
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7670801129835627e-07,
+      "loss": 0.397,
+      "step": 84600
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7668637354107977e-07,
+      "loss": 0.4624,
+      "step": 84601
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.766647362610459e-07,
+      "loss": 0.5234,
+      "step": 84602
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7664309945827143e-07,
+      "loss": 0.5015,
+      "step": 84603
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.766214631327724e-07,
+      "loss": 0.4067,
+      "step": 84604
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.765998272845661e-07,
+      "loss": 0.3562,
+      "step": 84605
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7657819191366813e-07,
+      "loss": 0.4109,
+      "step": 84606
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.765565570200957e-07,
+      "loss": 0.2994,
+      "step": 84607
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.765349226038652e-07,
+      "loss": 0.4009,
+      "step": 84608
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7651328866499354e-07,
+      "loss": 0.3679,
+      "step": 84609
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.764916552034967e-07,
+      "loss": 0.4312,
+      "step": 84610
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.764700222193915e-07,
+      "loss": 0.3489,
+      "step": 84611
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7644838971269457e-07,
+      "loss": 0.3726,
+      "step": 84612
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.764267576834231e-07,
+      "loss": 0.4048,
+      "step": 84613
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7640512613159225e-07,
+      "loss": 0.4375,
+      "step": 84614
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7638349505721947e-07,
+      "loss": 0.5381,
+      "step": 84615
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.76361864460321e-07,
+      "loss": 0.4058,
+      "step": 84616
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.76340234340914e-07,
+      "loss": 0.3276,
+      "step": 84617
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.763186046990142e-07,
+      "loss": 0.3284,
+      "step": 84618
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7629697553463866e-07,
+      "loss": 0.355,
+      "step": 84619
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7627534684780384e-07,
+      "loss": 0.4067,
+      "step": 84620
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.762537186385266e-07,
+      "loss": 0.3448,
+      "step": 84621
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.76232090906823e-07,
+      "loss": 0.4688,
+      "step": 84622
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.762104636527099e-07,
+      "loss": 0.4268,
+      "step": 84623
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7618883687620365e-07,
+      "loss": 0.3193,
+      "step": 84624
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.761672105773208e-07,
+      "loss": 0.5254,
+      "step": 84625
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.761455847560784e-07,
+      "loss": 0.5107,
+      "step": 84626
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.761239594124923e-07,
+      "loss": 0.2996,
+      "step": 84627
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.761023345465794e-07,
+      "loss": 0.3818,
+      "step": 84628
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.760807101583563e-07,
+      "loss": 0.3403,
+      "step": 84629
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.760590862478398e-07,
+      "loss": 0.3234,
+      "step": 84630
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.760374628150458e-07,
+      "loss": 0.4614,
+      "step": 84631
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.760158398599913e-07,
+      "loss": 0.4917,
+      "step": 84632
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.75994217382693e-07,
+      "loss": 0.2502,
+      "step": 84633
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.759725953831674e-07,
+      "loss": 0.4272,
+      "step": 84634
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.759509738614304e-07,
+      "loss": 0.2732,
+      "step": 84635
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7592935281749917e-07,
+      "loss": 0.4751,
+      "step": 84636
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.759077322513901e-07,
+      "loss": 0.2689,
+      "step": 84637
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7588611216312006e-07,
+      "loss": 0.4087,
+      "step": 84638
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.75864492552705e-07,
+      "loss": 0.4272,
+      "step": 84639
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7584287342016185e-07,
+      "loss": 0.3263,
+      "step": 84640
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7582125476550707e-07,
+      "loss": 0.478,
+      "step": 84641
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.757996365887576e-07,
+      "loss": 0.564,
+      "step": 84642
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7577801888992966e-07,
+      "loss": 0.3547,
+      "step": 84643
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7575640166903945e-07,
+      "loss": 0.5049,
+      "step": 84644
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.757347849261039e-07,
+      "loss": 0.3542,
+      "step": 84645
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7571316866113947e-07,
+      "loss": 0.3582,
+      "step": 84646
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.756915528741631e-07,
+      "loss": 0.2747,
+      "step": 84647
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7566993756519074e-07,
+      "loss": 0.4648,
+      "step": 84648
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.756483227342391e-07,
+      "loss": 0.3264,
+      "step": 84649
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7562670838132513e-07,
+      "loss": 0.4233,
+      "step": 84650
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.756050945064648e-07,
+      "loss": 0.2933,
+      "step": 84651
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.755834811096751e-07,
+      "loss": 0.259,
+      "step": 84652
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.755618681909726e-07,
+      "loss": 0.4214,
+      "step": 84653
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.755402557503733e-07,
+      "loss": 0.4346,
+      "step": 84654
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7551864378789445e-07,
+      "loss": 0.5557,
+      "step": 84655
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.75497032303552e-07,
+      "loss": 0.3275,
+      "step": 84656
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7547542129736265e-07,
+      "loss": 0.3848,
+      "step": 84657
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7545381076934313e-07,
+      "loss": 0.4062,
+      "step": 84658
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.754322007195103e-07,
+      "loss": 0.4712,
+      "step": 84659
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7541059114787985e-07,
+      "loss": 0.3774,
+      "step": 84660
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.753889820544689e-07,
+      "loss": 0.2917,
+      "step": 84661
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.753673734392938e-07,
+      "loss": 0.4016,
+      "step": 84662
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.753457653023716e-07,
+      "loss": 0.3945,
+      "step": 84663
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.753241576437184e-07,
+      "loss": 0.3799,
+      "step": 84664
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7530255046335044e-07,
+      "loss": 0.4167,
+      "step": 84665
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.752809437612845e-07,
+      "loss": 0.3349,
+      "step": 84666
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.752593375375377e-07,
+      "loss": 0.4351,
+      "step": 84667
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.752377317921257e-07,
+      "loss": 0.3789,
+      "step": 84668
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.752161265250654e-07,
+      "loss": 0.3486,
+      "step": 84669
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7519452173637334e-07,
+      "loss": 0.479,
+      "step": 84670
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.751729174260666e-07,
+      "loss": 0.2671,
+      "step": 84671
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7515131359416076e-07,
+      "loss": 0.3613,
+      "step": 84672
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.751297102406732e-07,
+      "loss": 0.5107,
+      "step": 84673
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.751081073656198e-07,
+      "loss": 0.5649,
+      "step": 84674
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7508650496901726e-07,
+      "loss": 0.3904,
+      "step": 84675
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.750649030508828e-07,
+      "loss": 0.36,
+      "step": 84676
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7504330161123185e-07,
+      "loss": 0.3782,
+      "step": 84677
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7502170065008165e-07,
+      "loss": 0.2546,
+      "step": 84678
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.750001001674485e-07,
+      "loss": 0.3726,
+      "step": 84679
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.749785001633494e-07,
+      "loss": 0.4072,
+      "step": 84680
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.749569006378003e-07,
+      "loss": 0.5273,
+      "step": 84681
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7493530159081786e-07,
+      "loss": 0.458,
+      "step": 84682
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.749137030224191e-07,
+      "loss": 0.4536,
+      "step": 84683
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.748921049326201e-07,
+      "loss": 0.353,
+      "step": 84684
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.748705073214372e-07,
+      "loss": 0.3923,
+      "step": 84685
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.748489101888872e-07,
+      "loss": 0.4146,
+      "step": 84686
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.748273135349865e-07,
+      "loss": 0.4106,
+      "step": 84687
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.748057173597523e-07,
+      "loss": 0.3882,
+      "step": 84688
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7478412166320027e-07,
+      "loss": 0.4336,
+      "step": 84689
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.747625264453472e-07,
+      "loss": 0.3655,
+      "step": 84690
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7474093170620977e-07,
+      "loss": 0.3457,
+      "step": 84691
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7471933744580476e-07,
+      "loss": 0.5137,
+      "step": 84692
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7469774366414853e-07,
+      "loss": 0.4209,
+      "step": 84693
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.74676150361257e-07,
+      "loss": 0.4033,
+      "step": 84694
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7465455753714726e-07,
+      "loss": 0.4878,
+      "step": 84695
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7463296519183573e-07,
+      "loss": 0.3315,
+      "step": 84696
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.746113733253393e-07,
+      "loss": 0.4995,
+      "step": 84697
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7458978193767386e-07,
+      "loss": 0.3379,
+      "step": 84698
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.745681910288563e-07,
+      "loss": 0.479,
+      "step": 84699
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.745466005989031e-07,
+      "loss": 0.323,
+      "step": 84700
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7452501064783114e-07,
+      "loss": 0.437,
+      "step": 84701
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7450342117565636e-07,
+      "loss": 0.4194,
+      "step": 84702
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.744818321823958e-07,
+      "loss": 0.3352,
+      "step": 84703
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.744602436680654e-07,
+      "loss": 0.2561,
+      "step": 84704
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.744386556326824e-07,
+      "loss": 0.4741,
+      "step": 84705
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.744170680762626e-07,
+      "loss": 0.416,
+      "step": 84706
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.743954809988229e-07,
+      "loss": 0.4546,
+      "step": 84707
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.743738944003798e-07,
+      "loss": 0.4565,
+      "step": 84708
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.743523082809502e-07,
+      "loss": 0.3687,
+      "step": 84709
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.743307226405499e-07,
+      "loss": 0.5786,
+      "step": 84710
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.743091374791958e-07,
+      "loss": 0.4507,
+      "step": 84711
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.742875527969045e-07,
+      "loss": 0.3911,
+      "step": 84712
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7426596859369277e-07,
+      "loss": 0.5259,
+      "step": 84713
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7424438486957664e-07,
+      "loss": 0.4092,
+      "step": 84714
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7422280162457263e-07,
+      "loss": 0.4072,
+      "step": 84715
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7420121885869746e-07,
+      "loss": 0.3657,
+      "step": 84716
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.741796365719676e-07,
+      "loss": 0.4309,
+      "step": 84717
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.741580547644e-07,
+      "loss": 0.3153,
+      "step": 84718
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.741364734360104e-07,
+      "loss": 0.4834,
+      "step": 84719
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.741148925868157e-07,
+      "loss": 0.5493,
+      "step": 84720
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7409331221683284e-07,
+      "loss": 0.3979,
+      "step": 84721
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7407173232607756e-07,
+      "loss": 0.459,
+      "step": 84722
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.740501529145672e-07,
+      "loss": 0.439,
+      "step": 84723
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.740285739823175e-07,
+      "loss": 0.4326,
+      "step": 84724
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7400699552934536e-07,
+      "loss": 0.3799,
+      "step": 84725
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.739854175556676e-07,
+      "loss": 0.4227,
+      "step": 84726
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.739638400613e-07,
+      "loss": 0.4502,
+      "step": 84727
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.739422630462595e-07,
+      "loss": 0.3982,
+      "step": 84728
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7392068651056277e-07,
+      "loss": 0.4067,
+      "step": 84729
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.738991104542264e-07,
+      "loss": 0.3618,
+      "step": 84730
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7387753487726637e-07,
+      "loss": 0.4321,
+      "step": 84731
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7385595977969954e-07,
+      "loss": 0.4082,
+      "step": 84732
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7383438516154277e-07,
+      "loss": 0.3065,
+      "step": 84733
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7381281102281183e-07,
+      "loss": 0.2838,
+      "step": 84734
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7379123736352404e-07,
+      "loss": 0.2582,
+      "step": 84735
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.73769664183695e-07,
+      "loss": 0.2854,
+      "step": 84736
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7374809148334186e-07,
+      "loss": 0.2902,
+      "step": 84737
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7372651926248135e-07,
+      "loss": 0.4097,
+      "step": 84738
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7370494752112936e-07,
+      "loss": 0.4473,
+      "step": 84739
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.736833762593026e-07,
+      "loss": 0.3267,
+      "step": 84740
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7366180547701773e-07,
+      "loss": 0.4414,
+      "step": 84741
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7364023517429154e-07,
+      "loss": 0.2566,
+      "step": 84742
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7361866535114016e-07,
+      "loss": 0.3716,
+      "step": 84743
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7359709600757995e-07,
+      "loss": 0.5322,
+      "step": 84744
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.735755271436275e-07,
+      "loss": 0.5723,
+      "step": 84745
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.735539587592996e-07,
+      "loss": 0.4658,
+      "step": 84746
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7353239085461286e-07,
+      "loss": 0.3896,
+      "step": 84747
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.735108234295833e-07,
+      "loss": 0.3877,
+      "step": 84748
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.734892564842277e-07,
+      "loss": 0.4604,
+      "step": 84749
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7346769001856247e-07,
+      "loss": 0.4194,
+      "step": 84750
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7344612403260465e-07,
+      "loss": 0.2966,
+      "step": 84751
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7342455852637e-07,
+      "loss": 0.377,
+      "step": 84752
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.734029934998757e-07,
+      "loss": 0.4219,
+      "step": 84753
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.733814289531375e-07,
+      "loss": 0.4924,
+      "step": 84754
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.733598648861723e-07,
+      "loss": 0.4258,
+      "step": 84755
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7333830129899714e-07,
+      "loss": 0.3794,
+      "step": 84756
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.733167381916276e-07,
+      "loss": 0.3882,
+      "step": 84757
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7329517556408054e-07,
+      "loss": 0.3195,
+      "step": 84758
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.73273613416373e-07,
+      "loss": 0.4434,
+      "step": 84759
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.732520517485206e-07,
+      "loss": 0.4277,
+      "step": 84760
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7323049056054035e-07,
+      "loss": 0.3767,
+      "step": 84761
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.732089298524487e-07,
+      "loss": 0.5054,
+      "step": 84762
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7318736962426244e-07,
+      "loss": 0.4541,
+      "step": 84763
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7316580987599776e-07,
+      "loss": 0.2854,
+      "step": 84764
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7314425060767094e-07,
+      "loss": 0.5415,
+      "step": 84765
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7312269181929867e-07,
+      "loss": 0.3949,
+      "step": 84766
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7310113351089767e-07,
+      "loss": 0.3833,
+      "step": 84767
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.730795756824845e-07,
+      "loss": 0.5723,
+      "step": 84768
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7305801833407524e-07,
+      "loss": 0.3901,
+      "step": 84769
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.730364614656866e-07,
+      "loss": 0.4214,
+      "step": 84770
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7301490507733504e-07,
+      "loss": 0.3138,
+      "step": 84771
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.729933491690375e-07,
+      "loss": 0.3546,
+      "step": 84772
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.729717937408101e-07,
+      "loss": 0.3655,
+      "step": 84773
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.729502387926692e-07,
+      "loss": 0.4648,
+      "step": 84774
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7292868432463134e-07,
+      "loss": 0.4897,
+      "step": 84775
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7290713033671337e-07,
+      "loss": 0.421,
+      "step": 84776
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.728855768289314e-07,
+      "loss": 0.4048,
+      "step": 84777
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.728640238013021e-07,
+      "loss": 0.3921,
+      "step": 84778
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.728424712538419e-07,
+      "loss": 0.3896,
+      "step": 84779
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7282091918656777e-07,
+      "loss": 0.3877,
+      "step": 84780
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.727993675994955e-07,
+      "loss": 0.2597,
+      "step": 84781
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7277781649264193e-07,
+      "loss": 0.3379,
+      "step": 84782
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.727562658660238e-07,
+      "loss": 0.2605,
+      "step": 84783
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7273471571965707e-07,
+      "loss": 0.5,
+      "step": 84784
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7271316605355896e-07,
+      "loss": 0.4424,
+      "step": 84785
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.726916168677451e-07,
+      "loss": 0.5049,
+      "step": 84786
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7267006816223243e-07,
+      "loss": 0.4224,
+      "step": 84787
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7264851993703736e-07,
+      "loss": 0.4663,
+      "step": 84788
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.72626972192177e-07,
+      "loss": 0.4644,
+      "step": 84789
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.726054249276668e-07,
+      "loss": 0.4321,
+      "step": 84790
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7258387814352386e-07,
+      "loss": 0.4438,
+      "step": 84791
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.725623318397646e-07,
+      "loss": 0.4172,
+      "step": 84792
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7254078601640615e-07,
+      "loss": 0.4238,
+      "step": 84793
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.725192406734637e-07,
+      "loss": 0.3722,
+      "step": 84794
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7249769581095446e-07,
+      "loss": 0.3252,
+      "step": 84795
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.724761514288949e-07,
+      "loss": 0.4893,
+      "step": 84796
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7245460752730174e-07,
+      "loss": 0.4209,
+      "step": 84797
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7243306410619104e-07,
+      "loss": 0.5132,
+      "step": 84798
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7241152116557927e-07,
+      "loss": 0.3945,
+      "step": 84799
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.723899787054833e-07,
+      "loss": 0.4844,
+      "step": 84800
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7236843672591977e-07,
+      "loss": 0.3821,
+      "step": 84801
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7234689522690454e-07,
+      "loss": 0.4197,
+      "step": 84802
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.723253542084548e-07,
+      "loss": 0.3938,
+      "step": 84803
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.723038136705863e-07,
+      "loss": 0.3706,
+      "step": 84804
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.722822736133159e-07,
+      "loss": 0.4282,
+      "step": 84805
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7226073403666037e-07,
+      "loss": 0.3386,
+      "step": 84806
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.722391949406356e-07,
+      "loss": 0.3064,
+      "step": 84807
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.722176563252585e-07,
+      "loss": 0.3682,
+      "step": 84808
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.721961181905455e-07,
+      "loss": 0.4443,
+      "step": 84809
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.721745805365133e-07,
+      "loss": 0.3572,
+      "step": 84810
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7215304336317785e-07,
+      "loss": 0.2302,
+      "step": 84811
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7213150667055606e-07,
+      "loss": 0.3623,
+      "step": 84812
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7210997045866454e-07,
+      "loss": 0.4243,
+      "step": 84813
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.720884347275195e-07,
+      "loss": 0.4727,
+      "step": 84814
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.720668994771372e-07,
+      "loss": 0.3975,
+      "step": 84815
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.720453647075343e-07,
+      "loss": 0.479,
+      "step": 84816
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7202383041872755e-07,
+      "loss": 0.311,
+      "step": 84817
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.720022966107336e-07,
+      "loss": 0.3118,
+      "step": 84818
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.719807632835682e-07,
+      "loss": 0.3784,
+      "step": 84819
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.719592304372483e-07,
+      "loss": 0.3906,
+      "step": 84820
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.719376980717903e-07,
+      "loss": 0.4502,
+      "step": 84821
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.71916166187211e-07,
+      "loss": 0.4127,
+      "step": 84822
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7189463478352676e-07,
+      "loss": 0.3881,
+      "step": 84823
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7187310386075354e-07,
+      "loss": 0.4021,
+      "step": 84824
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7185157341890816e-07,
+      "loss": 0.3862,
+      "step": 84825
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.718300434580072e-07,
+      "loss": 0.5664,
+      "step": 84826
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7180851397806735e-07,
+      "loss": 0.4604,
+      "step": 84827
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7178698497910454e-07,
+      "loss": 0.3279,
+      "step": 84828
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7176545646113546e-07,
+      "loss": 0.374,
+      "step": 84829
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7174392842417713e-07,
+      "loss": 0.3374,
+      "step": 84830
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7172240086824527e-07,
+      "loss": 0.416,
+      "step": 84831
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.717008737933569e-07,
+      "loss": 0.5447,
+      "step": 84832
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.716793471995281e-07,
+      "loss": 0.396,
+      "step": 84833
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7165782108677537e-07,
+      "loss": 0.3301,
+      "step": 84834
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7163629545511574e-07,
+      "loss": 0.3264,
+      "step": 84835
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.71614770304565e-07,
+      "loss": 0.4302,
+      "step": 84836
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7159324563513995e-07,
+      "loss": 0.2568,
+      "step": 84837
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7157172144685714e-07,
+      "loss": 0.458,
+      "step": 84838
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.715501977397332e-07,
+      "loss": 0.4526,
+      "step": 84839
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.71528674513784e-07,
+      "loss": 0.4326,
+      "step": 84840
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7150715176902646e-07,
+      "loss": 0.4121,
+      "step": 84841
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7148562950547726e-07,
+      "loss": 0.3381,
+      "step": 84842
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.714641077231524e-07,
+      "loss": 0.3418,
+      "step": 84843
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7144258642206885e-07,
+      "loss": 0.4111,
+      "step": 84844
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.714210656022424e-07,
+      "loss": 0.3413,
+      "step": 84845
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7139954526369e-07,
+      "loss": 0.5479,
+      "step": 84846
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.713780254064284e-07,
+      "loss": 0.3625,
+      "step": 84847
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7135650603047343e-07,
+      "loss": 0.4761,
+      "step": 84848
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7133498713584176e-07,
+      "loss": 0.4614,
+      "step": 84849
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.713134687225501e-07,
+      "loss": 0.417,
+      "step": 84850
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7129195079061517e-07,
+      "loss": 0.4624,
+      "step": 84851
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.71270433340053e-07,
+      "loss": 0.415,
+      "step": 84852
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7124891637087987e-07,
+      "loss": 0.4106,
+      "step": 84853
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.712273998831125e-07,
+      "loss": 0.3723,
+      "step": 84854
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7120588387676745e-07,
+      "loss": 0.4312,
+      "step": 84855
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.711843683518614e-07,
+      "loss": 0.5679,
+      "step": 84856
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.711628533084102e-07,
+      "loss": 0.3704,
+      "step": 84857
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.7114133874643085e-07,
+      "loss": 0.3945,
+      "step": 84858
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.711198246659395e-07,
+      "loss": 0.3896,
+      "step": 84859
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7109831106695324e-07,
+      "loss": 0.3921,
+      "step": 84860
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7107679794948766e-07,
+      "loss": 0.3752,
+      "step": 84861
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.710552853135601e-07,
+      "loss": 0.3501,
+      "step": 84862
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7103377315918617e-07,
+      "loss": 0.4297,
+      "step": 84863
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.710122614863831e-07,
+      "loss": 0.31,
+      "step": 84864
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7099075029516667e-07,
+      "loss": 0.3501,
+      "step": 84865
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.709692395855537e-07,
+      "loss": 0.4053,
+      "step": 84866
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.709477293575607e-07,
+      "loss": 0.3428,
+      "step": 84867
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.709262196112044e-07,
+      "loss": 0.3894,
+      "step": 84868
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.709047103465006e-07,
+      "loss": 0.377,
+      "step": 84869
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.708832015634662e-07,
+      "loss": 0.3558,
+      "step": 84870
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.708616932621176e-07,
+      "loss": 0.3123,
+      "step": 84871
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7084018544247163e-07,
+      "loss": 0.3513,
+      "step": 84872
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7081867810454436e-07,
+      "loss": 0.365,
+      "step": 84873
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.70797171248352e-07,
+      "loss": 0.3391,
+      "step": 84874
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.707756648739113e-07,
+      "loss": 0.4174,
+      "step": 84875
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7075415898123876e-07,
+      "loss": 0.3152,
+      "step": 84876
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.707326535703511e-07,
+      "loss": 0.4482,
+      "step": 84877
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.707111486412642e-07,
+      "loss": 0.3678,
+      "step": 84878
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7068964419399487e-07,
+      "loss": 0.4893,
+      "step": 84879
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.706681402285594e-07,
+      "loss": 0.4707,
+      "step": 84880
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.706466367449749e-07,
+      "loss": 0.4961,
+      "step": 84881
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7062513374325733e-07,
+      "loss": 0.3752,
+      "step": 84882
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7060363122342277e-07,
+      "loss": 0.3481,
+      "step": 84883
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7058212918548806e-07,
+      "loss": 0.4912,
+      "step": 84884
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7056062762947005e-07,
+      "loss": 0.4458,
+      "step": 84885
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7053912655538446e-07,
+      "loss": 0.3611,
+      "step": 84886
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.705176259632481e-07,
+      "loss": 0.5264,
+      "step": 84887
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7049612585307744e-07,
+      "loss": 0.3589,
+      "step": 84888
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.704746262248893e-07,
+      "loss": 0.4805,
+      "step": 84889
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7045312707869947e-07,
+      "loss": 0.3997,
+      "step": 84890
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.704316284145247e-07,
+      "loss": 0.3589,
+      "step": 84891
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.704101302323819e-07,
+      "loss": 0.342,
+      "step": 84892
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.703886325322867e-07,
+      "loss": 0.4712,
+      "step": 84893
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7036713531425635e-07,
+      "loss": 0.417,
+      "step": 84894
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.703456385783066e-07,
+      "loss": 0.3735,
+      "step": 84895
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.703241423244542e-07,
+      "loss": 0.3618,
+      "step": 84896
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.703026465527157e-07,
+      "loss": 0.3984,
+      "step": 84897
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7028115126310775e-07,
+      "loss": 0.4702,
+      "step": 84898
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.702596564556464e-07,
+      "loss": 0.479,
+      "step": 84899
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.702381621303482e-07,
+      "loss": 0.3645,
+      "step": 84900
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7021666828722996e-07,
+      "loss": 0.5024,
+      "step": 84901
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.701951749263079e-07,
+      "loss": 0.5166,
+      "step": 84902
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.701736820475981e-07,
+      "loss": 0.4012,
+      "step": 84903
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7015218965111737e-07,
+      "loss": 0.4121,
+      "step": 84904
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.701306977368821e-07,
+      "loss": 0.5027,
+      "step": 84905
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7010920630490915e-07,
+      "loss": 0.3196,
+      "step": 84906
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.700877153552143e-07,
+      "loss": 0.4624,
+      "step": 84907
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7006622488781424e-07,
+      "loss": 0.3766,
+      "step": 84908
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.700447349027256e-07,
+      "loss": 0.5195,
+      "step": 84909
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.700232453999651e-07,
+      "loss": 0.418,
+      "step": 84910
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.7000175637954844e-07,
+      "loss": 0.4624,
+      "step": 84911
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.699802678414928e-07,
+      "loss": 0.4136,
+      "step": 84912
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6995877978581404e-07,
+      "loss": 0.3635,
+      "step": 84913
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.699372922125288e-07,
+      "loss": 0.5186,
+      "step": 84914
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.699158051216539e-07,
+      "loss": 0.4438,
+      "step": 84915
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6989431851320517e-07,
+      "loss": 0.4521,
+      "step": 84916
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.698728323871995e-07,
+      "loss": 0.4766,
+      "step": 84917
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6985134674365316e-07,
+      "loss": 0.2822,
+      "step": 84918
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.69829861582583e-07,
+      "loss": 0.5386,
+      "step": 84919
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.698083769040048e-07,
+      "loss": 0.3784,
+      "step": 84920
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.697868927079353e-07,
+      "loss": 0.4409,
+      "step": 84921
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6976540899439135e-07,
+      "loss": 0.344,
+      "step": 84922
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.697439257633891e-07,
+      "loss": 0.4087,
+      "step": 84923
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.697224430149446e-07,
+      "loss": 0.3525,
+      "step": 84924
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.697009607490746e-07,
+      "loss": 0.3542,
+      "step": 84925
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6967947896579567e-07,
+      "loss": 0.3098,
+      "step": 84926
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.696579976651245e-07,
+      "loss": 0.3613,
+      "step": 84927
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6963651684707687e-07,
+      "loss": 0.4019,
+      "step": 84928
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6961503651166956e-07,
+      "loss": 0.3691,
+      "step": 84929
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.69593556658919e-07,
+      "loss": 0.502,
+      "step": 84930
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.695720772888421e-07,
+      "loss": 0.4468,
+      "step": 84931
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6955059840145473e-07,
+      "loss": 0.2987,
+      "step": 84932
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6952911999677327e-07,
+      "loss": 0.4402,
+      "step": 84933
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6950764207481434e-07,
+      "loss": 0.4531,
+      "step": 84934
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.694861646355943e-07,
+      "loss": 0.3845,
+      "step": 84935
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.694646876791302e-07,
+      "loss": 0.4321,
+      "step": 84936
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6944321120543763e-07,
+      "loss": 0.4331,
+      "step": 84937
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6942173521453336e-07,
+      "loss": 0.2831,
+      "step": 84938
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6940025970643416e-07,
+      "loss": 0.3511,
+      "step": 84939
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.69378784681156e-07,
+      "loss": 0.4456,
+      "step": 84940
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.693573101387154e-07,
+      "loss": 0.4243,
+      "step": 84941
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.693358360791292e-07,
+      "loss": 0.4131,
+      "step": 84942
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.693143625024133e-07,
+      "loss": 0.3798,
+      "step": 84943
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.692928894085847e-07,
+      "loss": 0.3855,
+      "step": 84944
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6927141679765926e-07,
+      "loss": 0.3765,
+      "step": 84945
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6924994466965367e-07,
+      "loss": 0.5146,
+      "step": 84946
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.692284730245844e-07,
+      "loss": 0.395,
+      "step": 84947
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.692070018624681e-07,
+      "loss": 0.6177,
+      "step": 84948
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6918553118332083e-07,
+      "loss": 0.3694,
+      "step": 84949
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.691640609871591e-07,
+      "loss": 0.4844,
+      "step": 84950
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.691425912739994e-07,
+      "loss": 0.3057,
+      "step": 84951
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.691211220438586e-07,
+      "loss": 0.3988,
+      "step": 84952
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6909965329675275e-07,
+      "loss": 0.4539,
+      "step": 84953
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6907818503269796e-07,
+      "loss": 0.3652,
+      "step": 84954
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6905671725171103e-07,
+      "loss": 0.3398,
+      "step": 84955
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6903524995380865e-07,
+      "loss": 0.4668,
+      "step": 84956
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6901378313900653e-07,
+      "loss": 0.4346,
+      "step": 84957
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6899231680732167e-07,
+      "loss": 0.4066,
+      "step": 84958
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.689708509587703e-07,
+      "loss": 0.4521,
+      "step": 84959
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.689493855933694e-07,
+      "loss": 0.4141,
+      "step": 84960
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6892792071113454e-07,
+      "loss": 0.4863,
+      "step": 84961
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.689064563120828e-07,
+      "loss": 0.5464,
+      "step": 84962
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6888499239623016e-07,
+      "loss": 0.4795,
+      "step": 84963
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6886352896359317e-07,
+      "loss": 0.4937,
+      "step": 84964
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6884206601418874e-07,
+      "loss": 0.4229,
+      "step": 84965
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6882060354803255e-07,
+      "loss": 0.3218,
+      "step": 84966
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6879914156514147e-07,
+      "loss": 0.292,
+      "step": 84967
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6877768006553176e-07,
+      "loss": 0.4419,
+      "step": 84968
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6875621904922037e-07,
+      "loss": 0.4185,
+      "step": 84969
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6873475851622306e-07,
+      "loss": 0.4927,
+      "step": 84970
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.687132984665564e-07,
+      "loss": 0.395,
+      "step": 84971
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6869183890023726e-07,
+      "loss": 0.5894,
+      "step": 84972
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.686703798172818e-07,
+      "loss": 0.3887,
+      "step": 84973
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6864892121770606e-07,
+      "loss": 0.46,
+      "step": 84974
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.686274631015268e-07,
+      "loss": 0.3683,
+      "step": 84975
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.686060054687604e-07,
+      "loss": 0.4087,
+      "step": 84976
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6858454831942376e-07,
+      "loss": 0.3707,
+      "step": 84977
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.685630916535326e-07,
+      "loss": 0.4319,
+      "step": 84978
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.685416354711035e-07,
+      "loss": 0.4243,
+      "step": 84979
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.685201797721532e-07,
+      "loss": 0.303,
+      "step": 84980
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.684987245566982e-07,
+      "loss": 0.333,
+      "step": 84981
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.684772698247547e-07,
+      "loss": 0.4526,
+      "step": 84982
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6845581557633886e-07,
+      "loss": 0.3184,
+      "step": 84983
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6843436181146727e-07,
+      "loss": 0.3727,
+      "step": 84984
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.684129085301565e-07,
+      "loss": 0.3406,
+      "step": 84985
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6839145573242326e-07,
+      "loss": 0.2668,
+      "step": 84986
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6837000341828326e-07,
+      "loss": 0.4746,
+      "step": 84987
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6834855158775333e-07,
+      "loss": 0.4458,
+      "step": 84988
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.683271002408499e-07,
+      "loss": 0.4419,
+      "step": 84989
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.683056493775896e-07,
+      "loss": 0.4014,
+      "step": 84990
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6828419899798835e-07,
+      "loss": 0.5601,
+      "step": 84991
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6826274910206324e-07,
+      "loss": 0.4678,
+      "step": 84992
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6824129968982985e-07,
+      "loss": 0.5112,
+      "step": 84993
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6821985076130547e-07,
+      "loss": 0.5312,
+      "step": 84994
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6819840231650567e-07,
+      "loss": 0.408,
+      "step": 84995
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.681769543554474e-07,
+      "loss": 0.222,
+      "step": 84996
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.681555068781469e-07,
+      "loss": 0.3833,
+      "step": 84997
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.681340598846211e-07,
+      "loss": 0.5103,
+      "step": 84998
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.6811261337488565e-07,
+      "loss": 0.4241,
+      "step": 84999
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.680911673489572e-07,
+      "loss": 0.4688,
+      "step": 85000
     }
   ],
   "logging_steps": 1.0,
   "max_steps": 117048,
   "num_train_epochs": 3,
   "save_steps": 5000,
-  "total_flos": 1687867611955200.0,
+  "total_flos": 2207165233029120.0,
   "trial_name": null,
   "trial_params": null
 }