diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,6016 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 5.452394476972822,
+  "eval_steps": 1000000,
+  "global_step": 499134,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 10.2449,
+      "step": 500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 9.9641,
+      "step": 1000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5e-06,
+      "loss": 9.8029,
+      "step": 1500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 9.599,
+      "step": 2000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.5e-06,
+      "loss": 9.3409,
+      "step": 2500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3e-06,
+      "loss": 9.0312,
+      "step": 3000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.5e-06,
+      "loss": 8.6908,
+      "step": 3500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 8.3383,
+      "step": 4000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.5e-06,
+      "loss": 8.0041,
+      "step": 4500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5e-06,
+      "loss": 7.7006,
+      "step": 5000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.500000000000001e-06,
+      "loss": 7.4574,
+      "step": 5500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 6e-06,
+      "loss": 7.2801,
+      "step": 6000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 7.1686,
+      "step": 6500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 7e-06,
+      "loss": 7.0915,
+      "step": 7000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 7.500000000000001e-06,
+      "loss": 7.0348,
+      "step": 7500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 6.9833,
+      "step": 8000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 8.5e-06,
+      "loss": 6.9317,
+      "step": 8500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9e-06,
+      "loss": 6.9026,
+      "step": 9000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 9.5e-06,
+      "loss": 6.8661,
+      "step": 9500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1e-05,
+      "loss": 6.8333,
+      "step": 10000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.999453217269587e-06,
+      "loss": 6.8129,
+      "step": 10500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.998906434539172e-06,
+      "loss": 6.7814,
+      "step": 11000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.998359651808758e-06,
+      "loss": 6.7639,
+      "step": 11500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.997812869078344e-06,
+      "loss": 6.7428,
+      "step": 12000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.99726608634793e-06,
+      "loss": 6.7265,
+      "step": 12500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.996719303617516e-06,
+      "loss": 6.7047,
+      "step": 13000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.996172520887101e-06,
+      "loss": 6.6934,
+      "step": 13500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.995625738156687e-06,
+      "loss": 6.6698,
+      "step": 14000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.995078955426273e-06,
+      "loss": 6.6658,
+      "step": 14500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.994532172695859e-06,
+      "loss": 6.6426,
+      "step": 15000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.993985389965445e-06,
+      "loss": 6.6317,
+      "step": 15500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.99343860723503e-06,
+      "loss": 6.6221,
+      "step": 16000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.992891824504616e-06,
+      "loss": 6.6111,
+      "step": 16500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.992345041774202e-06,
+      "loss": 6.6005,
+      "step": 17000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.991798259043786e-06,
+      "loss": 6.5898,
+      "step": 17500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.991251476313374e-06,
+      "loss": 6.5796,
+      "step": 18000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.99070469358296e-06,
+      "loss": 6.5629,
+      "step": 18500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.990157910852545e-06,
+      "loss": 6.5594,
+      "step": 19000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.98961112812213e-06,
+      "loss": 6.5604,
+      "step": 19500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.989064345391715e-06,
+      "loss": 6.5431,
+      "step": 20000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.988517562661302e-06,
+      "loss": 6.53,
+      "step": 20500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.987970779930888e-06,
+      "loss": 6.5237,
+      "step": 21000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.987423997200472e-06,
+      "loss": 6.5207,
+      "step": 21500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.986877214470058e-06,
+      "loss": 6.513,
+      "step": 22000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.986330431739644e-06,
+      "loss": 6.5085,
+      "step": 22500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.985783649009231e-06,
+      "loss": 6.4952,
+      "step": 23000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.985236866278816e-06,
+      "loss": 6.4856,
+      "step": 23500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.984690083548401e-06,
+      "loss": 6.4856,
+      "step": 24000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.984143300817987e-06,
+      "loss": 6.4789,
+      "step": 24500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.983596518087575e-06,
+      "loss": 6.464,
+      "step": 25000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.98304973535716e-06,
+      "loss": 6.4641,
+      "step": 25500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.982502952626745e-06,
+      "loss": 6.4555,
+      "step": 26000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.98195616989633e-06,
+      "loss": 6.4573,
+      "step": 26500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.981409387165916e-06,
+      "loss": 6.4442,
+      "step": 27000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.980862604435504e-06,
+      "loss": 6.4336,
+      "step": 27500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.980315821705088e-06,
+      "loss": 6.4362,
+      "step": 28000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.979769038974673e-06,
+      "loss": 6.4239,
+      "step": 28500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.97922225624426e-06,
+      "loss": 6.4254,
+      "step": 29000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.978675473513845e-06,
+      "loss": 6.4172,
+      "step": 29500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.97812869078343e-06,
+      "loss": 6.4088,
+      "step": 30000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.977581908053017e-06,
+      "loss": 6.409,
+      "step": 30500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.977035125322602e-06,
+      "loss": 6.4082,
+      "step": 31000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.976488342592188e-06,
+      "loss": 6.3967,
+      "step": 31500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.975941559861774e-06,
+      "loss": 6.3971,
+      "step": 32000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.97539477713136e-06,
+      "loss": 6.3906,
+      "step": 32500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.974847994400946e-06,
+      "loss": 6.3855,
+      "step": 33000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.974301211670531e-06,
+      "loss": 6.3789,
+      "step": 33500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.973754428940117e-06,
+      "loss": 6.3731,
+      "step": 34000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.973207646209703e-06,
+      "loss": 6.3685,
+      "step": 34500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.972660863479289e-06,
+      "loss": 6.3639,
+      "step": 35000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.972114080748875e-06,
+      "loss": 6.3644,
+      "step": 35500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.97156729801846e-06,
+      "loss": 6.3619,
+      "step": 36000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.971020515288046e-06,
+      "loss": 6.3512,
+      "step": 36500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.970473732557632e-06,
+      "loss": 6.3534,
+      "step": 37000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.969926949827218e-06,
+      "loss": 6.3491,
+      "step": 37500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.969380167096804e-06,
+      "loss": 6.3459,
+      "step": 38000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.96883338436639e-06,
+      "loss": 6.3426,
+      "step": 38500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.968286601635975e-06,
+      "loss": 6.3301,
+      "step": 39000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.967739818905561e-06,
+      "loss": 6.3277,
+      "step": 39500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.967193036175147e-06,
+      "loss": 6.3262,
+      "step": 40000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.966646253444732e-06,
+      "loss": 6.327,
+      "step": 40500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.966099470714318e-06,
+      "loss": 6.3185,
+      "step": 41000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.965552687983904e-06,
+      "loss": 6.3233,
+      "step": 41500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.96500590525349e-06,
+      "loss": 6.3169,
+      "step": 42000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.964459122523074e-06,
+      "loss": 6.3163,
+      "step": 42500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.963912339792661e-06,
+      "loss": 6.3121,
+      "step": 43000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.963365557062247e-06,
+      "loss": 6.2977,
+      "step": 43500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.962818774331833e-06,
+      "loss": 6.3044,
+      "step": 44000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.962271991601417e-06,
+      "loss": 6.3029,
+      "step": 44500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.961725208871003e-06,
+      "loss": 6.2973,
+      "step": 45000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.96117842614059e-06,
+      "loss": 6.2979,
+      "step": 45500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.960631643410176e-06,
+      "loss": 6.2926,
+      "step": 46000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.96008486067976e-06,
+      "loss": 6.2899,
+      "step": 46500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.959538077949346e-06,
+      "loss": 6.2846,
+      "step": 47000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.958991295218932e-06,
+      "loss": 6.2828,
+      "step": 47500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.95844451248852e-06,
+      "loss": 6.2788,
+      "step": 48000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.957897729758103e-06,
+      "loss": 6.2782,
+      "step": 48500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.95735094702769e-06,
+      "loss": 6.2797,
+      "step": 49000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.956804164297275e-06,
+      "loss": 6.2758,
+      "step": 49500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.956257381566863e-06,
+      "loss": 6.274,
+      "step": 50000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.955710598836447e-06,
+      "loss": 6.2679,
+      "step": 50500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.955163816106032e-06,
+      "loss": 6.2632,
+      "step": 51000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.954617033375618e-06,
+      "loss": 6.2601,
+      "step": 51500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.954070250645204e-06,
+      "loss": 6.2611,
+      "step": 52000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.953523467914792e-06,
+      "loss": 6.2554,
+      "step": 52500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.952976685184376e-06,
+      "loss": 6.2618,
+      "step": 53000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.952429902453961e-06,
+      "loss": 6.2543,
+      "step": 53500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.951883119723547e-06,
+      "loss": 6.2554,
+      "step": 54000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.951336336993133e-06,
+      "loss": 6.2506,
+      "step": 54500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.950789554262719e-06,
+      "loss": 6.2466,
+      "step": 55000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.950242771532305e-06,
+      "loss": 6.2484,
+      "step": 55500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.94969598880189e-06,
+      "loss": 6.246,
+      "step": 56000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.949149206071476e-06,
+      "loss": 6.2427,
+      "step": 56500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.948602423341062e-06,
+      "loss": 6.2403,
+      "step": 57000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.948055640610648e-06,
+      "loss": 6.2398,
+      "step": 57500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.947508857880234e-06,
+      "loss": 6.237,
+      "step": 58000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.94696207514982e-06,
+      "loss": 6.2337,
+      "step": 58500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.946415292419405e-06,
+      "loss": 6.2312,
+      "step": 59000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 9.945868509688991e-06,
+      "loss": 6.225,
+      "step": 59500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.945321726958577e-06,
+      "loss": 6.2348,
+      "step": 60000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.944774944228163e-06,
+      "loss": 6.2262,
+      "step": 60500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.944228161497748e-06,
+      "loss": 6.224,
+      "step": 61000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.943681378767334e-06,
+      "loss": 6.2259,
+      "step": 61500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.94313459603692e-06,
+      "loss": 6.2191,
+      "step": 62000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.942587813306506e-06,
+      "loss": 6.213,
+      "step": 62500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.94204103057609e-06,
+      "loss": 6.2173,
+      "step": 63000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.941494247845677e-06,
+      "loss": 6.2081,
+      "step": 63500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.940947465115263e-06,
+      "loss": 6.2135,
+      "step": 64000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.940400682384849e-06,
+      "loss": 6.2155,
+      "step": 64500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 9.939853899654433e-06,
+      "loss": 6.2033,
+      "step": 65000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 9.93930711692402e-06,
+      "loss": 6.2047,
+      "step": 65500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 9.938760334193606e-06,
+      "loss": 6.2025,
+      "step": 66000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 9.938213551463192e-06,
+      "loss": 6.2045,
+      "step": 66500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 9.937666768732778e-06,
+      "loss": 6.2012,
+      "step": 67000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 9.937119986002362e-06,
+      "loss": 6.2052,
+      "step": 67500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 9.93657320327195e-06,
+      "loss": 6.2001,
+      "step": 68000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.936026420541535e-06,
+      "loss": 6.2009,
+      "step": 68500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 9.935479637811121e-06,
+      "loss": 6.1945,
+      "step": 69000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.934932855080705e-06,
+      "loss": 6.1954,
+      "step": 69500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 9.934386072350291e-06,
+      "loss": 6.1971,
+      "step": 70000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 9.933839289619878e-06,
+      "loss": 6.1938,
+      "step": 70500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 9.933292506889464e-06,
+      "loss": 6.1866,
+      "step": 71000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 9.932745724159048e-06,
+      "loss": 6.1868,
+      "step": 71500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 9.932198941428634e-06,
+      "loss": 6.1902,
+      "step": 72000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 9.93165215869822e-06,
+      "loss": 6.1847,
+      "step": 72500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 9.931105375967807e-06,
+      "loss": 6.1802,
+      "step": 73000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 9.930558593237391e-06,
+      "loss": 6.1833,
+      "step": 73500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 9.930011810506977e-06,
+      "loss": 6.1882,
+      "step": 74000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 9.929465027776563e-06,
+      "loss": 6.1862,
+      "step": 74500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 9.928918245046149e-06,
+      "loss": 6.1832,
+      "step": 75000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 9.928371462315735e-06,
+      "loss": 6.1728,
+      "step": 75500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 9.92782467958532e-06,
+      "loss": 6.177,
+      "step": 76000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 9.927277896854906e-06,
+      "loss": 6.1793,
+      "step": 76500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 9.926731114124492e-06,
+      "loss": 6.1756,
+      "step": 77000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 9.926184331394078e-06,
+      "loss": 6.1778,
+      "step": 77500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 9.925637548663664e-06,
+      "loss": 6.1703,
+      "step": 78000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.92509076593325e-06,
+      "loss": 6.1681,
+      "step": 78500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.924543983202835e-06,
+      "loss": 6.1696,
+      "step": 79000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.923997200472421e-06,
+      "loss": 6.1652,
+      "step": 79500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.923450417742007e-06,
+      "loss": 6.1739,
+      "step": 80000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 9.922903635011593e-06,
+      "loss": 6.1762,
+      "step": 80500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 9.922356852281178e-06,
+      "loss": 6.1652,
+      "step": 81000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 9.921810069550764e-06,
+      "loss": 6.1645,
+      "step": 81500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 9.92126328682035e-06,
+      "loss": 6.1654,
+      "step": 82000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 9.920716504089936e-06,
+      "loss": 6.166,
+      "step": 82500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 9.920169721359522e-06,
+      "loss": 6.1605,
+      "step": 83000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 9.919622938629107e-06,
+      "loss": 6.1605,
+      "step": 83500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 9.919076155898693e-06,
+      "loss": 6.1662,
+      "step": 84000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 9.918529373168279e-06,
+      "loss": 6.1545,
+      "step": 84500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.917982590437865e-06,
+      "loss": 6.1607,
+      "step": 85000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 9.91743580770745e-06,
+      "loss": 6.1628,
+      "step": 85500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 9.916889024977036e-06,
+      "loss": 6.1552,
+      "step": 86000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 9.916342242246622e-06,
+      "loss": 6.1548,
+      "step": 86500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 9.915795459516208e-06,
+      "loss": 6.1551,
+      "step": 87000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.915248676785794e-06,
+      "loss": 6.1551,
+      "step": 87500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.914701894055378e-06,
+      "loss": 6.1477,
+      "step": 88000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.914155111324965e-06,
+      "loss": 6.1528,
+      "step": 88500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 9.913608328594551e-06,
+      "loss": 6.1504,
+      "step": 89000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.913061545864137e-06,
+      "loss": 6.1519,
+      "step": 89500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 9.912514763133721e-06,
+      "loss": 6.1468,
+      "step": 90000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.911967980403307e-06,
+      "loss": 6.1448,
+      "step": 90500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.911421197672894e-06,
+      "loss": 6.1437,
+      "step": 91000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.91087441494248e-06,
+      "loss": 6.1491,
+      "step": 91500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.910327632212064e-06,
+      "loss": 6.1429,
+      "step": 92000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 9.90978084948165e-06,
+      "loss": 6.1391,
+      "step": 92500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 9.909234066751237e-06,
+      "loss": 6.1401,
+      "step": 93000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 9.908687284020823e-06,
+      "loss": 6.146,
+      "step": 93500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.908140501290409e-06,
+      "loss": 6.1388,
+      "step": 94000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.907593718559993e-06,
+      "loss": 6.1438,
+      "step": 94500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.907046935829579e-06,
+      "loss": 6.1377,
+      "step": 95000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.906500153099166e-06,
+      "loss": 6.1376,
+      "step": 95500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.905953370368752e-06,
+      "loss": 6.1339,
+      "step": 96000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.905406587638336e-06,
+      "loss": 6.1361,
+      "step": 96500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.904859804907922e-06,
+      "loss": 6.1351,
+      "step": 97000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.904313022177508e-06,
+      "loss": 6.1361,
+      "step": 97500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.903766239447095e-06,
+      "loss": 6.1307,
+      "step": 98000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.90321945671668e-06,
+      "loss": 6.1374,
+      "step": 98500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.902672673986265e-06,
+      "loss": 6.1289,
+      "step": 99000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.902125891255851e-06,
+      "loss": 6.1304,
+      "step": 99500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.901579108525437e-06,
+      "loss": 6.1266,
+      "step": 100000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 9.901032325795023e-06,
+      "loss": 6.1303,
+      "step": 100500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 9.900485543064608e-06,
+      "loss": 6.1279,
+      "step": 101000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 9.899938760334194e-06,
+      "loss": 6.1267,
+      "step": 101500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 9.89939197760378e-06,
+      "loss": 6.1268,
+      "step": 102000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 9.898845194873366e-06,
+      "loss": 6.1295,
+      "step": 102500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 9.898298412142952e-06,
+      "loss": 6.1208,
+      "step": 103000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 9.897751629412537e-06,
+      "loss": 6.1241,
+      "step": 103500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 9.897204846682123e-06,
+      "loss": 6.1203,
+      "step": 104000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 9.896658063951709e-06,
+      "loss": 6.1226,
+      "step": 104500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 9.896111281221295e-06,
+      "loss": 6.1188,
+      "step": 105000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 9.89556449849088e-06,
+      "loss": 6.1232,
+      "step": 105500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 9.895017715760466e-06,
+      "loss": 6.1173,
+      "step": 106000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 9.894470933030052e-06,
+      "loss": 6.1177,
+      "step": 106500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 9.893924150299638e-06,
+      "loss": 6.1184,
+      "step": 107000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 9.893377367569224e-06,
+      "loss": 6.1181,
+      "step": 107500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 9.89283058483881e-06,
+      "loss": 6.1131,
+      "step": 108000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 9.892283802108395e-06,
+      "loss": 6.1101,
+      "step": 108500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 9.891737019377981e-06,
+      "loss": 6.1131,
+      "step": 109000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 9.891190236647567e-06,
+      "loss": 6.1156,
+      "step": 109500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 9.890643453917153e-06,
+      "loss": 6.1099,
+      "step": 110000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 9.890096671186738e-06,
+      "loss": 6.1086,
+      "step": 110500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 9.889549888456324e-06,
+      "loss": 6.1074,
+      "step": 111000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 9.88900310572591e-06,
+      "loss": 6.1081,
+      "step": 111500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 9.888456322995496e-06,
+      "loss": 6.1094,
+      "step": 112000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 9.887909540265082e-06,
+      "loss": 6.1017,
+      "step": 112500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 9.887362757534666e-06,
+      "loss": 6.1055,
+      "step": 113000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 9.886815974804253e-06,
+      "loss": 6.1034,
+      "step": 113500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 9.886269192073839e-06,
+      "loss": 6.1052,
+      "step": 114000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 9.885722409343425e-06,
+      "loss": 6.1086,
+      "step": 114500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 9.885175626613009e-06,
+      "loss": 6.1014,
+      "step": 115000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 9.884628843882595e-06,
+      "loss": 6.1119,
+      "step": 115500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.884082061152182e-06,
+      "loss": 6.1035,
+      "step": 116000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 9.883535278421768e-06,
+      "loss": 6.1041,
+      "step": 116500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.882988495691352e-06,
+      "loss": 6.1056,
+      "step": 117000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 9.882441712960938e-06,
+      "loss": 6.1026,
+      "step": 117500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.881894930230524e-06,
+      "loss": 6.1016,
+      "step": 118000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 9.881348147500111e-06,
+      "loss": 6.1038,
+      "step": 118500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 9.880801364769695e-06,
+      "loss": 6.0953,
+      "step": 119000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 9.880254582039281e-06,
+      "loss": 6.1028,
+      "step": 119500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 9.879707799308867e-06,
+      "loss": 6.1048,
+      "step": 120000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 9.879161016578454e-06,
+      "loss": 6.1006,
+      "step": 120500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 9.878614233848038e-06,
+      "loss": 6.0935,
+      "step": 121000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 9.878067451117624e-06,
+      "loss": 6.1022,
+      "step": 121500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 9.87752066838721e-06,
+      "loss": 6.0975,
+      "step": 122000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 9.876973885656796e-06,
+      "loss": 6.0926,
+      "step": 122500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 9.876427102926383e-06,
+      "loss": 6.0949,
+      "step": 123000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 9.875880320195967e-06,
+      "loss": 6.0921,
+      "step": 123500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 9.875333537465553e-06,
+      "loss": 6.0974,
+      "step": 124000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 9.874786754735139e-06,
+      "loss": 6.0933,
+      "step": 124500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 9.874239972004725e-06,
+      "loss": 6.0887,
+      "step": 125000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 9.87369318927431e-06,
+      "loss": 6.0903,
+      "step": 125500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 9.873146406543896e-06,
+      "loss": 6.0933,
+      "step": 126000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 9.872599623813482e-06,
+      "loss": 6.0877,
+      "step": 126500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 9.872052841083068e-06,
+      "loss": 6.0895,
+      "step": 127000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 9.871506058352654e-06,
+      "loss": 6.0885,
+      "step": 127500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 9.87095927562224e-06,
+      "loss": 6.0868,
+      "step": 128000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 9.870412492891825e-06,
+      "loss": 6.0899,
+      "step": 128500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 9.869865710161411e-06,
+      "loss": 6.0871,
+      "step": 129000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 9.869318927430997e-06,
+      "loss": 6.0824,
+      "step": 129500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 9.868772144700583e-06,
+      "loss": 6.0877,
+      "step": 130000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 9.868225361970168e-06,
+      "loss": 6.0817,
+      "step": 130500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 9.867678579239754e-06,
+      "loss": 6.0842,
+      "step": 131000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 9.86713179650934e-06,
+      "loss": 6.0824,
+      "step": 131500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 9.866585013778926e-06,
+      "loss": 6.0864,
+      "step": 132000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 9.866038231048512e-06,
+      "loss": 6.0826,
+      "step": 132500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 9.865491448318097e-06,
+      "loss": 6.0841,
+      "step": 133000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 9.864944665587683e-06,
+      "loss": 6.082,
+      "step": 133500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 9.864397882857269e-06,
+      "loss": 6.0828,
+      "step": 134000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 9.863851100126855e-06,
+      "loss": 6.0759,
+      "step": 134500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 9.86330431739644e-06,
+      "loss": 6.0817,
+      "step": 135000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 9.862757534666025e-06,
+      "loss": 6.0814,
+      "step": 135500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 9.862210751935612e-06,
+      "loss": 6.0795,
+      "step": 136000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 9.861663969205198e-06,
+      "loss": 6.0682,
+      "step": 136500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.861117186474784e-06,
+      "loss": 6.073,
+      "step": 137000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 9.86057040374437e-06,
+      "loss": 6.076,
+      "step": 137500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.860023621013954e-06,
+      "loss": 6.0731,
+      "step": 138000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 9.859476838283541e-06,
+      "loss": 6.0701,
+      "step": 138500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.858930055553127e-06,
+      "loss": 6.0733,
+      "step": 139000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 9.858383272822713e-06,
+      "loss": 6.0692,
+      "step": 139500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.857836490092297e-06,
+      "loss": 6.0732,
+      "step": 140000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 9.857289707361883e-06,
+      "loss": 6.0694,
+      "step": 140500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 9.85674292463147e-06,
+      "loss": 6.0765,
+      "step": 141000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.856196141901056e-06,
+      "loss": 6.0727,
+      "step": 141500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 9.85564935917064e-06,
+      "loss": 6.07,
+      "step": 142000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.855102576440226e-06,
+      "loss": 6.0789,
+      "step": 142500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 9.854555793709812e-06,
+      "loss": 6.0758,
+      "step": 143000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.854009010979399e-06,
+      "loss": 6.0656,
+      "step": 143500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 9.853462228248983e-06,
+      "loss": 6.0701,
+      "step": 144000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.852915445518569e-06,
+      "loss": 6.0727,
+      "step": 144500
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 9.852368662788155e-06,
+      "loss": 6.065,
+      "step": 145000
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.851821880057742e-06,
+      "loss": 6.0763,
+      "step": 145500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 9.851275097327326e-06,
+      "loss": 6.0738,
+      "step": 146000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 9.850728314596912e-06,
+      "loss": 6.0683,
+      "step": 146500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.850181531866498e-06,
+      "loss": 6.0692,
+      "step": 147000
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 9.849634749136084e-06,
+      "loss": 6.0706,
+      "step": 147500
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.84908796640567e-06,
+      "loss": 6.0698,
+      "step": 148000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 9.848541183675255e-06,
+      "loss": 6.0608,
+      "step": 148500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.847994400944841e-06,
+      "loss": 6.0646,
+      "step": 149000
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 9.847447618214427e-06,
+      "loss": 6.0669,
+      "step": 149500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.846900835484013e-06,
+      "loss": 6.0667,
+      "step": 150000
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 9.846354052753598e-06,
+      "loss": 6.0602,
+      "step": 150500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 9.845807270023184e-06,
+      "loss": 6.071,
+      "step": 151000
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 9.84526048729277e-06,
+      "loss": 6.068,
+      "step": 151500
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 9.844713704562356e-06,
+      "loss": 6.0632,
+      "step": 152000
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 9.844166921831942e-06,
+      "loss": 6.0588,
+      "step": 152500
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 9.843620139101527e-06,
+      "loss": 6.0601,
+      "step": 153000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 9.843073356371113e-06,
+      "loss": 6.061,
+      "step": 153500
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 9.842526573640699e-06,
+      "loss": 6.0643,
+      "step": 154000
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 9.841979790910285e-06,
+      "loss": 6.059,
+      "step": 154500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 9.84143300817987e-06,
+      "loss": 6.0543,
+      "step": 155000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 9.840886225449456e-06,
+      "loss": 6.0581,
+      "step": 155500
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 9.840339442719042e-06,
+      "loss": 6.0551,
+      "step": 156000
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 9.839792659988628e-06,
+      "loss": 6.0629,
+      "step": 156500
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.839245877258214e-06,
+      "loss": 6.058,
+      "step": 157000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.8386990945278e-06,
+      "loss": 6.0519,
+      "step": 157500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.838152311797385e-06,
+      "loss": 6.0579,
+      "step": 158000
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.83760552906697e-06,
+      "loss": 6.0554,
+      "step": 158500
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 9.837058746336557e-06,
+      "loss": 6.0574,
+      "step": 159000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 9.836511963606143e-06,
+      "loss": 6.0523,
+      "step": 159500
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 9.835965180875729e-06,
+      "loss": 6.0542,
+      "step": 160000
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 9.835418398145313e-06,
+      "loss": 6.0508,
+      "step": 160500
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 9.8348716154149e-06,
+      "loss": 6.051,
+      "step": 161000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 9.834324832684486e-06,
+      "loss": 6.0519,
+      "step": 161500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 9.833778049954072e-06,
+      "loss": 6.0508,
+      "step": 162000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 9.833231267223656e-06,
+      "loss": 6.0545,
+      "step": 162500
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 9.832684484493242e-06,
+      "loss": 6.0438,
+      "step": 163000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 9.832137701762829e-06,
+      "loss": 6.0509,
+      "step": 163500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 9.831590919032415e-06,
+      "loss": 6.0441,
+      "step": 164000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 9.831044136302e-06,
+      "loss": 6.0523,
+      "step": 164500
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 9.830497353571585e-06,
+      "loss": 6.043,
+      "step": 165000
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.82995057084117e-06,
+      "loss": 6.052,
+      "step": 165500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 9.829403788110758e-06,
+      "loss": 6.0464,
+      "step": 166000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 9.828857005380344e-06,
+      "loss": 6.0454,
+      "step": 166500
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 9.828310222649928e-06,
+      "loss": 6.0511,
+      "step": 167000
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 9.827763439919514e-06,
+      "loss": 6.0474,
+      "step": 167500
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 9.8272166571891e-06,
+      "loss": 6.0478,
+      "step": 168000
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 9.826669874458687e-06,
+      "loss": 6.0464,
+      "step": 168500
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 9.826123091728271e-06,
+      "loss": 6.0558,
+      "step": 169000
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 9.825576308997857e-06,
+      "loss": 6.0403,
+      "step": 169500
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 9.825029526267443e-06,
+      "loss": 6.0497,
+      "step": 170000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 9.824482743537029e-06,
+      "loss": 6.0485,
+      "step": 170500
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 9.823935960806614e-06,
+      "loss": 6.0438,
+      "step": 171000
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 9.8233891780762e-06,
+      "loss": 6.0416,
+      "step": 171500
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 9.822842395345786e-06,
+      "loss": 6.0455,
+      "step": 172000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 9.822295612615372e-06,
+      "loss": 6.0431,
+      "step": 172500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 9.821748829884957e-06,
+      "loss": 6.0393,
+      "step": 173000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 9.821202047154543e-06,
+      "loss": 6.0398,
+      "step": 173500
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 9.820655264424129e-06,
+      "loss": 6.045,
+      "step": 174000
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.820108481693715e-06,
+      "loss": 6.0376,
+      "step": 174500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.8195616989633e-06,
+      "loss": 6.0369,
+      "step": 175000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 9.819014916232886e-06,
+      "loss": 6.0405,
+      "step": 175500
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 9.818468133502472e-06,
+      "loss": 6.04,
+      "step": 176000
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 9.817921350772058e-06,
+      "loss": 6.0418,
+      "step": 176500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 9.817374568041644e-06,
+      "loss": 6.0396,
+      "step": 177000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 9.81682778531123e-06,
+      "loss": 6.0363,
+      "step": 177500
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 9.816281002580815e-06,
+      "loss": 6.041,
+      "step": 178000
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 9.815734219850401e-06,
+      "loss": 6.0362,
+      "step": 178500
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 9.815187437119987e-06,
+      "loss": 6.0332,
+      "step": 179000
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 9.814640654389573e-06,
+      "loss": 6.0404,
+      "step": 179500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 9.814093871659159e-06,
+      "loss": 6.0369,
+      "step": 180000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 9.813547088928744e-06,
+      "loss": 6.0322,
+      "step": 180500
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 9.81300030619833e-06,
+      "loss": 6.0371,
+      "step": 181000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 9.812453523467916e-06,
+      "loss": 6.0334,
+      "step": 181500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 9.811906740737502e-06,
+      "loss": 6.0323,
+      "step": 182000
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 9.811359958007088e-06,
+      "loss": 6.0371,
+      "step": 182500
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 9.810813175276673e-06,
+      "loss": 6.0354,
+      "step": 183000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 9.810266392546257e-06,
+      "loss": 6.0329,
+      "step": 183500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 9.809719609815845e-06,
+      "loss": 6.0319,
+      "step": 184000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 9.80917282708543e-06,
+      "loss": 6.0346,
+      "step": 184500
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 9.808626044355017e-06,
+      "loss": 6.0337,
+      "step": 185000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 9.8080792616246e-06,
+      "loss": 6.0276,
+      "step": 185500
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 9.807532478894186e-06,
+      "loss": 6.0353,
+      "step": 186000
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 9.806985696163774e-06,
+      "loss": 6.0336,
+      "step": 186500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 9.80643891343336e-06,
+      "loss": 6.0328,
+      "step": 187000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 9.805892130702944e-06,
+      "loss": 6.0295,
+      "step": 187500
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 9.80534534797253e-06,
+      "loss": 6.0233,
+      "step": 188000
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 9.804798565242117e-06,
+      "loss": 6.0265,
+      "step": 188500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 9.804251782511703e-06,
+      "loss": 6.0288,
+      "step": 189000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 9.803704999781287e-06,
+      "loss": 6.0271,
+      "step": 189500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 9.803158217050873e-06,
+      "loss": 6.0309,
+      "step": 190000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 9.802611434320459e-06,
+      "loss": 6.0193,
+      "step": 190500
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 9.802064651590046e-06,
+      "loss": 6.0251,
+      "step": 191000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 9.80151786885963e-06,
+      "loss": 6.0209,
+      "step": 191500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 9.800971086129216e-06,
+      "loss": 6.019,
+      "step": 192000
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 9.800424303398802e-06,
+      "loss": 6.0223,
+      "step": 192500
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 9.799877520668388e-06,
+      "loss": 6.0231,
+      "step": 193000
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 9.799330737937975e-06,
+      "loss": 6.0117,
+      "step": 193500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 9.798783955207559e-06,
+      "loss": 6.0128,
+      "step": 194000
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 9.798237172477145e-06,
+      "loss": 6.0028,
+      "step": 194500
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 9.79769038974673e-06,
+      "loss": 6.0038,
+      "step": 195000
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.797143607016316e-06,
+      "loss": 5.9958,
+      "step": 195500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.796596824285902e-06,
+      "loss": 5.9925,
+      "step": 196000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.796050041555488e-06,
+      "loss": 5.9917,
+      "step": 196500
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.795503258825074e-06,
+      "loss": 5.9822,
+      "step": 197000
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.79495647609466e-06,
+      "loss": 5.9781,
+      "step": 197500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.794409693364245e-06,
+      "loss": 5.9781,
+      "step": 198000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.793862910633831e-06,
+      "loss": 5.9723,
+      "step": 198500
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.793316127903417e-06,
+      "loss": 5.9693,
+      "step": 199000
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.792769345173003e-06,
+      "loss": 5.9595,
+      "step": 199500
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.792222562442589e-06,
+      "loss": 5.9611,
+      "step": 200000
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 9.791675779712174e-06,
+      "loss": 5.9564,
+      "step": 200500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 9.79112899698176e-06,
+      "loss": 5.9508,
+      "step": 201000
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 9.790582214251346e-06,
+      "loss": 5.9467,
+      "step": 201500
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 9.790035431520932e-06,
+      "loss": 5.9424,
+      "step": 202000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 9.789488648790518e-06,
+      "loss": 5.9431,
+      "step": 202500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 9.788941866060103e-06,
+      "loss": 5.9397,
+      "step": 203000
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 9.78839508332969e-06,
+      "loss": 5.9345,
+      "step": 203500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 9.787848300599275e-06,
+      "loss": 5.9282,
+      "step": 204000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 9.78730151786886e-06,
+      "loss": 5.9267,
+      "step": 204500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.786754735138447e-06,
+      "loss": 5.9228,
+      "step": 205000
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 9.786207952408032e-06,
+      "loss": 5.9158,
+      "step": 205500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 9.785661169677618e-06,
+      "loss": 5.9132,
+      "step": 206000
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 9.785114386947204e-06,
+      "loss": 5.9134,
+      "step": 206500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 9.78456760421679e-06,
+      "loss": 5.911,
+      "step": 207000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 9.784020821486376e-06,
+      "loss": 5.9075,
+      "step": 207500
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 9.783474038755961e-06,
+      "loss": 5.907,
+      "step": 208000
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 9.782927256025545e-06,
+      "loss": 5.9109,
+      "step": 208500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 9.782380473295133e-06,
+      "loss": 5.9009,
+      "step": 209000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 9.781833690564719e-06,
+      "loss": 5.8982,
+      "step": 209500
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 9.781286907834304e-06,
+      "loss": 5.8992,
+      "step": 210000
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 9.780740125103889e-06,
+      "loss": 5.8952,
+      "step": 210500
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 9.780193342373474e-06,
+      "loss": 5.8974,
+      "step": 211000
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 9.779646559643062e-06,
+      "loss": 5.8929,
+      "step": 211500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 9.779099776912648e-06,
+      "loss": 5.89,
+      "step": 212000
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 9.778552994182232e-06,
+      "loss": 5.8803,
+      "step": 212500
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 9.778006211451818e-06,
+      "loss": 5.8817,
+      "step": 213000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 9.777459428721405e-06,
+      "loss": 5.8825,
+      "step": 213500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 9.77691264599099e-06,
+      "loss": 5.8891,
+      "step": 214000
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 9.776365863260575e-06,
+      "loss": 5.8805,
+      "step": 214500
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 9.77581908053016e-06,
+      "loss": 5.8725,
+      "step": 215000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 9.775272297799747e-06,
+      "loss": 5.8775,
+      "step": 215500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 9.774725515069334e-06,
+      "loss": 5.8702,
+      "step": 216000
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 9.774178732338918e-06,
+      "loss": 5.8678,
+      "step": 216500
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 9.773631949608504e-06,
+      "loss": 5.8773,
+      "step": 217000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 9.77308516687809e-06,
+      "loss": 5.8643,
+      "step": 217500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 9.772538384147675e-06,
+      "loss": 5.8694,
+      "step": 218000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 9.771991601417261e-06,
+      "loss": 5.8675,
+      "step": 218500
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 9.771444818686847e-06,
+      "loss": 5.8613,
+      "step": 219000
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 9.770898035956433e-06,
+      "loss": 5.8626,
+      "step": 219500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 9.770351253226019e-06,
+      "loss": 5.8643,
+      "step": 220000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 9.769804470495604e-06,
+      "loss": 5.8604,
+      "step": 220500
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 9.76925768776519e-06,
+      "loss": 5.8536,
+      "step": 221000
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 9.768710905034776e-06,
+      "loss": 5.8558,
+      "step": 221500
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.768164122304362e-06,
+      "loss": 5.8479,
+      "step": 222000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.767617339573948e-06,
+      "loss": 5.8541,
+      "step": 222500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 9.767070556843533e-06,
+      "loss": 5.846,
+      "step": 223000
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 9.76652377411312e-06,
+      "loss": 5.8451,
+      "step": 223500
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 9.765976991382705e-06,
+      "loss": 5.8491,
+      "step": 224000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 9.76543020865229e-06,
+      "loss": 5.8432,
+      "step": 224500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 9.764883425921877e-06,
+      "loss": 5.8439,
+      "step": 225000
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 9.764336643191462e-06,
+      "loss": 5.8381,
+      "step": 225500
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 9.763789860461048e-06,
+      "loss": 5.8377,
+      "step": 226000
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 9.763243077730634e-06,
+      "loss": 5.84,
+      "step": 226500
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 9.76269629500022e-06,
+      "loss": 5.8378,
+      "step": 227000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 9.762149512269806e-06,
+      "loss": 5.8322,
+      "step": 227500
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 9.761602729539391e-06,
+      "loss": 5.8287,
+      "step": 228000
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 9.761055946808977e-06,
+      "loss": 5.8315,
+      "step": 228500
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 9.760509164078563e-06,
+      "loss": 5.8268,
+      "step": 229000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 9.759962381348149e-06,
+      "loss": 5.8293,
+      "step": 229500
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 9.759415598617735e-06,
+      "loss": 5.8319,
+      "step": 230000
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 9.75886881588732e-06,
+      "loss": 5.818,
+      "step": 230500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 9.758322033156904e-06,
+      "loss": 5.8173,
+      "step": 231000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 9.757775250426492e-06,
+      "loss": 5.817,
+      "step": 231500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 9.757228467696078e-06,
+      "loss": 5.8195,
+      "step": 232000
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 9.756681684965663e-06,
+      "loss": 5.8123,
+      "step": 232500
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 9.756134902235248e-06,
+      "loss": 5.8148,
+      "step": 233000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 9.755588119504833e-06,
+      "loss": 5.8127,
+      "step": 233500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 9.75504133677442e-06,
+      "loss": 5.8122,
+      "step": 234000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 9.754494554044007e-06,
+      "loss": 5.8093,
+      "step": 234500
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 9.753947771313592e-06,
+      "loss": 5.8096,
+      "step": 235000
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 9.753400988583177e-06,
+      "loss": 5.8091,
+      "step": 235500
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 9.752854205852762e-06,
+      "loss": 5.8033,
+      "step": 236000
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 9.75230742312235e-06,
+      "loss": 5.8024,
+      "step": 236500
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 9.751760640391936e-06,
+      "loss": 5.8088,
+      "step": 237000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 9.75121385766152e-06,
+      "loss": 5.7974,
+      "step": 237500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 9.750667074931105e-06,
+      "loss": 5.7945,
+      "step": 238000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.750120292200691e-06,
+      "loss": 5.7975,
+      "step": 238500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.749573509470279e-06,
+      "loss": 5.7948,
+      "step": 239000
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.749026726739863e-06,
+      "loss": 5.7952,
+      "step": 239500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 9.748479944009449e-06,
+      "loss": 5.7934,
+      "step": 240000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.747933161279034e-06,
+      "loss": 5.7922,
+      "step": 240500
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 9.747386378548622e-06,
+      "loss": 5.7882,
+      "step": 241000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 9.746839595818206e-06,
+      "loss": 5.7848,
+      "step": 241500
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 9.746292813087792e-06,
+      "loss": 5.7818,
+      "step": 242000
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.745746030357378e-06,
+      "loss": 5.784,
+      "step": 242500
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 9.745199247626963e-06,
+      "loss": 5.7778,
+      "step": 243000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 9.74465246489655e-06,
+      "loss": 5.7741,
+      "step": 243500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.744105682166135e-06,
+      "loss": 5.7778,
+      "step": 244000
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 9.74355889943572e-06,
+      "loss": 5.7808,
+      "step": 244500
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.743012116705307e-06,
+      "loss": 5.7732,
+      "step": 245000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 9.742465333974892e-06,
+      "loss": 5.7754,
+      "step": 245500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.741918551244478e-06,
+      "loss": 5.776,
+      "step": 246000
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 9.741371768514064e-06,
+      "loss": 5.7701,
+      "step": 246500
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.74082498578365e-06,
+      "loss": 5.7699,
+      "step": 247000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 9.740278203053236e-06,
+      "loss": 5.7676,
+      "step": 247500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.739731420322821e-06,
+      "loss": 5.7654,
+      "step": 248000
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 9.739184637592407e-06,
+      "loss": 5.7587,
+      "step": 248500
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 9.738637854861993e-06,
+      "loss": 5.7649,
+      "step": 249000
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.738091072131579e-06,
+      "loss": 5.7664,
+      "step": 249500
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 9.737544289401165e-06,
+      "loss": 5.7565,
+      "step": 250000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.73699750667075e-06,
+      "loss": 5.7551,
+      "step": 250500
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.736450723940336e-06,
+      "loss": 5.755,
+      "step": 251000
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.735903941209922e-06,
+      "loss": 5.7547,
+      "step": 251500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.735357158479508e-06,
+      "loss": 5.7536,
+      "step": 252000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.734810375749093e-06,
+      "loss": 5.7485,
+      "step": 252500
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 9.73426359301868e-06,
+      "loss": 5.7443,
+      "step": 253000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.733716810288265e-06,
+      "loss": 5.7412,
+      "step": 253500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 9.73317002755785e-06,
+      "loss": 5.7483,
+      "step": 254000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 9.732623244827437e-06,
+      "loss": 5.7448,
+      "step": 254500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.732076462097022e-06,
+      "loss": 5.7401,
+      "step": 255000
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 9.731529679366608e-06,
+      "loss": 5.7405,
+      "step": 255500
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.730982896636192e-06,
+      "loss": 5.7389,
+      "step": 256000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 9.73043611390578e-06,
+      "loss": 5.7385,
+      "step": 256500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.729889331175366e-06,
+      "loss": 5.7367,
+      "step": 257000
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 9.729342548444951e-06,
+      "loss": 5.7324,
+      "step": 257500
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.728795765714536e-06,
+      "loss": 5.7367,
+      "step": 258000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.728248982984121e-06,
+      "loss": 5.7376,
+      "step": 258500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.727702200253709e-06,
+      "loss": 5.7256,
+      "step": 259000
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 9.727155417523295e-06,
+      "loss": 5.7302,
+      "step": 259500
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 9.726608634792879e-06,
+      "loss": 5.7245,
+      "step": 260000
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 9.726061852062464e-06,
+      "loss": 5.7238,
+      "step": 260500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 9.72551506933205e-06,
+      "loss": 5.72,
+      "step": 261000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 9.724968286601638e-06,
+      "loss": 5.7209,
+      "step": 261500
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 9.724421503871224e-06,
+      "loss": 5.7219,
+      "step": 262000
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 9.723874721140808e-06,
+      "loss": 5.7158,
+      "step": 262500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 9.723327938410393e-06,
+      "loss": 5.7134,
+      "step": 263000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 9.72278115567998e-06,
+      "loss": 5.708,
+      "step": 263500
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 9.722234372949567e-06,
+      "loss": 5.7111,
+      "step": 264000
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 9.72168759021915e-06,
+      "loss": 5.7126,
+      "step": 264500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 9.721140807488737e-06,
+      "loss": 5.7079,
+      "step": 265000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 9.720594024758322e-06,
+      "loss": 5.7069,
+      "step": 265500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 9.72004724202791e-06,
+      "loss": 5.7005,
+      "step": 266000
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 9.719500459297494e-06,
+      "loss": 5.6974,
+      "step": 266500
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.71895367656708e-06,
+      "loss": 5.6986,
+      "step": 267000
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.718406893836666e-06,
+      "loss": 5.7036,
+      "step": 267500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 9.717860111106251e-06,
+      "loss": 5.6978,
+      "step": 268000
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 9.717313328375837e-06,
+      "loss": 5.6924,
+      "step": 268500
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 9.716766545645423e-06,
+      "loss": 5.6914,
+      "step": 269000
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 9.716219762915009e-06,
+      "loss": 5.6911,
+      "step": 269500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 9.715672980184595e-06,
+      "loss": 5.6978,
+      "step": 270000
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 9.71512619745418e-06,
+      "loss": 5.6868,
+      "step": 270500
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 9.714579414723766e-06,
+      "loss": 5.681,
+      "step": 271000
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 9.714032631993352e-06,
+      "loss": 5.6875,
+      "step": 271500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 9.713485849262938e-06,
+      "loss": 5.6777,
+      "step": 272000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 9.712939066532524e-06,
+      "loss": 5.681,
+      "step": 272500
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 9.71239228380211e-06,
+      "loss": 5.6747,
+      "step": 273000
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 9.711845501071695e-06,
+      "loss": 5.6748,
+      "step": 273500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 9.711298718341281e-06,
+      "loss": 5.6775,
+      "step": 274000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 9.710751935610867e-06,
+      "loss": 5.6695,
+      "step": 274500
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 9.710205152880452e-06,
+      "loss": 5.673,
+      "step": 275000
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 9.709658370150038e-06,
+      "loss": 5.6728,
+      "step": 275500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 9.709111587419624e-06,
+      "loss": 5.6676,
+      "step": 276000
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 9.70856480468921e-06,
+      "loss": 5.6651,
+      "step": 276500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 9.708018021958796e-06,
+      "loss": 5.6635,
+      "step": 277000
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 9.707471239228381e-06,
+      "loss": 5.6688,
+      "step": 277500
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 9.706924456497967e-06,
+      "loss": 5.6595,
+      "step": 278000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 9.706377673767553e-06,
+      "loss": 5.6608,
+      "step": 278500
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 9.705830891037137e-06,
+      "loss": 5.6523,
+      "step": 279000
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 9.705284108306725e-06,
+      "loss": 5.6539,
+      "step": 279500
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 9.70473732557631e-06,
+      "loss": 5.6539,
+      "step": 280000
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 9.704190542845896e-06,
+      "loss": 5.6493,
+      "step": 280500
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 9.70364376011548e-06,
+      "loss": 5.6498,
+      "step": 281000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 9.703096977385068e-06,
+      "loss": 5.6476,
+      "step": 281500
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 9.702550194654654e-06,
+      "loss": 5.6485,
+      "step": 282000
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 9.70200341192424e-06,
+      "loss": 5.643,
+      "step": 282500
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 9.701456629193823e-06,
+      "loss": 5.6398,
+      "step": 283000
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 9.70090984646341e-06,
+      "loss": 5.6433,
+      "step": 283500
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 9.700363063732997e-06,
+      "loss": 5.6331,
+      "step": 284000
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 9.699816281002583e-06,
+      "loss": 5.6353,
+      "step": 284500
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 9.699269498272167e-06,
+      "loss": 5.634,
+      "step": 285000
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 9.698722715541752e-06,
+      "loss": 5.631,
+      "step": 285500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 9.698175932811338e-06,
+      "loss": 5.6272,
+      "step": 286000
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 9.697629150080926e-06,
+      "loss": 5.6288,
+      "step": 286500
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 9.69708236735051e-06,
+      "loss": 5.622,
+      "step": 287000
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 9.696535584620096e-06,
+      "loss": 5.6179,
+      "step": 287500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 9.695988801889681e-06,
+      "loss": 5.6209,
+      "step": 288000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 9.695442019159267e-06,
+      "loss": 5.6191,
+      "step": 288500
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 9.694895236428853e-06,
+      "loss": 5.6182,
+      "step": 289000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 9.694348453698439e-06,
+      "loss": 5.6188,
+      "step": 289500
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 9.693801670968025e-06,
+      "loss": 5.6138,
+      "step": 290000
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 9.69325488823761e-06,
+      "loss": 5.6069,
+      "step": 290500
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 9.692708105507196e-06,
+      "loss": 5.6132,
+      "step": 291000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 9.692161322776782e-06,
+      "loss": 5.6059,
+      "step": 291500
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 9.691614540046368e-06,
+      "loss": 5.604,
+      "step": 292000
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 9.691067757315954e-06,
+      "loss": 5.5996,
+      "step": 292500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 9.69052097458554e-06,
+      "loss": 5.6005,
+      "step": 293000
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 9.689974191855125e-06,
+      "loss": 5.5953,
+      "step": 293500
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 9.689427409124711e-06,
+      "loss": 5.5977,
+      "step": 294000
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 9.688880626394297e-06,
+      "loss": 5.5874,
+      "step": 294500
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 9.688333843663883e-06,
+      "loss": 5.5894,
+      "step": 295000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 9.687787060933468e-06,
+      "loss": 5.5816,
+      "step": 295500
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 9.687240278203054e-06,
+      "loss": 5.584,
+      "step": 296000
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 9.68669349547264e-06,
+      "loss": 5.5811,
+      "step": 296500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 9.686146712742226e-06,
+      "loss": 5.5806,
+      "step": 297000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 9.685599930011811e-06,
+      "loss": 5.5779,
+      "step": 297500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 9.685053147281397e-06,
+      "loss": 5.5738,
+      "step": 298000
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 9.684506364550983e-06,
+      "loss": 5.576,
+      "step": 298500
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 9.683959581820569e-06,
+      "loss": 5.5674,
+      "step": 299000
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 9.683412799090155e-06,
+      "loss": 5.5655,
+      "step": 299500
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 9.68286601635974e-06,
+      "loss": 5.5591,
+      "step": 300000
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 9.682319233629326e-06,
+      "loss": 5.5554,
+      "step": 300500
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.681772450898912e-06,
+      "loss": 5.5602,
+      "step": 301000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.681225668168496e-06,
+      "loss": 5.555,
+      "step": 301500
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.680678885438084e-06,
+      "loss": 5.5522,
+      "step": 302000
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.68013210270767e-06,
+      "loss": 5.554,
+      "step": 302500
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 9.679585319977255e-06,
+      "loss": 5.5482,
+      "step": 303000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.67903853724684e-06,
+      "loss": 5.5451,
+      "step": 303500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.678491754516425e-06,
+      "loss": 5.5419,
+      "step": 304000
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.677944971786013e-06,
+      "loss": 5.5377,
+      "step": 304500
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.677398189055598e-06,
+      "loss": 5.5387,
+      "step": 305000
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.676851406325184e-06,
+      "loss": 5.533,
+      "step": 305500
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.676304623594768e-06,
+      "loss": 5.5316,
+      "step": 306000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 9.675757840864354e-06,
+      "loss": 5.5313,
+      "step": 306500
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 9.675211058133942e-06,
+      "loss": 5.5239,
+      "step": 307000
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 9.674664275403527e-06,
+      "loss": 5.5251,
+      "step": 307500
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 9.674117492673111e-06,
+      "loss": 5.5175,
+      "step": 308000
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 9.673570709942697e-06,
+      "loss": 5.5172,
+      "step": 308500
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 9.673023927212285e-06,
+      "loss": 5.5158,
+      "step": 309000
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 9.67247714448187e-06,
+      "loss": 5.511,
+      "step": 309500
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 9.671930361751455e-06,
+      "loss": 5.5059,
+      "step": 310000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 9.67138357902104e-06,
+      "loss": 5.506,
+      "step": 310500
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 9.670836796290626e-06,
+      "loss": 5.5045,
+      "step": 311000
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 9.670290013560214e-06,
+      "loss": 5.5064,
+      "step": 311500
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 9.669743230829798e-06,
+      "loss": 5.499,
+      "step": 312000
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 9.669196448099384e-06,
+      "loss": 5.4937,
+      "step": 312500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 9.66864966536897e-06,
+      "loss": 5.4939,
+      "step": 313000
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 9.668102882638555e-06,
+      "loss": 5.4958,
+      "step": 313500
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 9.667556099908141e-06,
+      "loss": 5.4849,
+      "step": 314000
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 9.667009317177727e-06,
+      "loss": 5.4902,
+      "step": 314500
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 9.666462534447313e-06,
+      "loss": 5.4883,
+      "step": 315000
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 9.665915751716898e-06,
+      "loss": 5.4783,
+      "step": 315500
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 9.665368968986484e-06,
+      "loss": 5.4801,
+      "step": 316000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 9.66482218625607e-06,
+      "loss": 5.48,
+      "step": 316500
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 9.664275403525656e-06,
+      "loss": 5.4714,
+      "step": 317000
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 9.663728620795242e-06,
+      "loss": 5.471,
+      "step": 317500
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 9.663181838064827e-06,
+      "loss": 5.4671,
+      "step": 318000
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 9.662635055334413e-06,
+      "loss": 5.4648,
+      "step": 318500
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 9.662088272603999e-06,
+      "loss": 5.4635,
+      "step": 319000
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 9.661541489873585e-06,
+      "loss": 5.461,
+      "step": 319500
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 9.66099470714317e-06,
+      "loss": 5.4565,
+      "step": 320000
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 9.660447924412756e-06,
+      "loss": 5.4529,
+      "step": 320500
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 9.659901141682342e-06,
+      "loss": 5.4506,
+      "step": 321000
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 9.659354358951928e-06,
+      "loss": 5.4503,
+      "step": 321500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 9.658807576221514e-06,
+      "loss": 5.4435,
+      "step": 322000
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 9.6582607934911e-06,
+      "loss": 5.4479,
+      "step": 322500
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 9.657714010760685e-06,
+      "loss": 5.4402,
+      "step": 323000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 9.657167228030271e-06,
+      "loss": 5.4387,
+      "step": 323500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 9.656620445299857e-06,
+      "loss": 5.4311,
+      "step": 324000
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 9.656073662569443e-06,
+      "loss": 5.4352,
+      "step": 324500
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 9.655526879839028e-06,
+      "loss": 5.4251,
+      "step": 325000
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 9.654980097108614e-06,
+      "loss": 5.4239,
+      "step": 325500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 9.6544333143782e-06,
+      "loss": 5.4237,
+      "step": 326000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 9.653886531647784e-06,
+      "loss": 5.4187,
+      "step": 326500
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 9.653339748917372e-06,
+      "loss": 5.4143,
+      "step": 327000
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 9.652792966186957e-06,
+      "loss": 5.4097,
+      "step": 327500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 9.652246183456543e-06,
+      "loss": 5.4088,
+      "step": 328000
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 9.651699400726127e-06,
+      "loss": 5.4111,
+      "step": 328500
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 9.651152617995713e-06,
+      "loss": 5.4016,
+      "step": 329000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 9.6506058352653e-06,
+      "loss": 5.4059,
+      "step": 329500
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 9.650059052534886e-06,
+      "loss": 5.399,
+      "step": 330000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 9.64951226980447e-06,
+      "loss": 5.3932,
+      "step": 330500
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 9.648965487074056e-06,
+      "loss": 5.3881,
+      "step": 331000
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 9.648418704343642e-06,
+      "loss": 5.3839,
+      "step": 331500
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 9.64787192161323e-06,
+      "loss": 5.3917,
+      "step": 332000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 9.647325138882815e-06,
+      "loss": 5.3823,
+      "step": 332500
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 9.6467783561524e-06,
+      "loss": 5.3785,
+      "step": 333000
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 9.646231573421985e-06,
+      "loss": 5.3776,
+      "step": 333500
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 9.645684790691571e-06,
+      "loss": 5.379,
+      "step": 334000
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 9.645138007961158e-06,
+      "loss": 5.3698,
+      "step": 334500
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 9.644591225230743e-06,
+      "loss": 5.3655,
+      "step": 335000
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 9.644044442500328e-06,
+      "loss": 5.3631,
+      "step": 335500
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 9.643497659769914e-06,
+      "loss": 5.3626,
+      "step": 336000
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 9.642950877039502e-06,
+      "loss": 5.3566,
+      "step": 336500
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 9.642404094309086e-06,
+      "loss": 5.3527,
+      "step": 337000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 9.641857311578672e-06,
+      "loss": 5.3513,
+      "step": 337500
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 9.641310528848257e-06,
+      "loss": 5.3456,
+      "step": 338000
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 9.640763746117843e-06,
+      "loss": 5.3431,
+      "step": 338500
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 9.640216963387429e-06,
+      "loss": 5.3489,
+      "step": 339000
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 9.639670180657015e-06,
+      "loss": 5.3341,
+      "step": 339500
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 9.6391233979266e-06,
+      "loss": 5.3336,
+      "step": 340000
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 9.638576615196186e-06,
+      "loss": 5.3278,
+      "step": 340500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 9.638029832465772e-06,
+      "loss": 5.332,
+      "step": 341000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 9.637483049735358e-06,
+      "loss": 5.325,
+      "step": 341500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 9.636936267004944e-06,
+      "loss": 5.3214,
+      "step": 342000
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 9.63638948427453e-06,
+      "loss": 5.3183,
+      "step": 342500
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 9.635842701544115e-06,
+      "loss": 5.3144,
+      "step": 343000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 9.635295918813701e-06,
+      "loss": 5.3065,
+      "step": 343500
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 9.634749136083287e-06,
+      "loss": 5.3047,
+      "step": 344000
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 9.634202353352873e-06,
+      "loss": 5.3022,
+      "step": 344500
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 9.633655570622458e-06,
+      "loss": 5.2952,
+      "step": 345000
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 9.633108787892044e-06,
+      "loss": 5.2922,
+      "step": 345500
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 9.63256200516163e-06,
+      "loss": 5.293,
+      "step": 346000
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 9.632015222431216e-06,
+      "loss": 5.2943,
+      "step": 346500
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 9.631468439700802e-06,
+      "loss": 5.2873,
+      "step": 347000
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 9.630921656970387e-06,
+      "loss": 5.272,
+      "step": 347500
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 9.630374874239973e-06,
+      "loss": 5.2803,
+      "step": 348000
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 9.629828091509559e-06,
+      "loss": 5.2719,
+      "step": 348500
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 9.629281308779145e-06,
+      "loss": 5.2677,
+      "step": 349000
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 9.628734526048729e-06,
+      "loss": 5.2678,
+      "step": 349500
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 9.628187743318316e-06,
+      "loss": 5.262,
+      "step": 350000
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 9.627640960587902e-06,
+      "loss": 5.2579,
+      "step": 350500
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 9.627094177857488e-06,
+      "loss": 5.2476,
+      "step": 351000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 9.626547395127072e-06,
+      "loss": 5.2479,
+      "step": 351500
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 9.62600061239666e-06,
+      "loss": 5.2507,
+      "step": 352000
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 9.625453829666245e-06,
+      "loss": 5.2419,
+      "step": 352500
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 9.624907046935831e-06,
+      "loss": 5.2424,
+      "step": 353000
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 9.624360264205415e-06,
+      "loss": 5.2317,
+      "step": 353500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 9.623813481475001e-06,
+      "loss": 5.2343,
+      "step": 354000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 9.623266698744588e-06,
+      "loss": 5.2251,
+      "step": 354500
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 9.622719916014174e-06,
+      "loss": 5.2228,
+      "step": 355000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 9.622173133283758e-06,
+      "loss": 5.2196,
+      "step": 355500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 9.621626350553344e-06,
+      "loss": 5.211,
+      "step": 356000
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 9.62107956782293e-06,
+      "loss": 5.2129,
+      "step": 356500
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 9.620532785092517e-06,
+      "loss": 5.2012,
+      "step": 357000
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 9.619986002362102e-06,
+      "loss": 5.1988,
+      "step": 357500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 9.619439219631687e-06,
+      "loss": 5.1954,
+      "step": 358000
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 9.618892436901273e-06,
+      "loss": 5.1955,
+      "step": 358500
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 9.618345654170859e-06,
+      "loss": 5.1848,
+      "step": 359000
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 9.617798871440445e-06,
+      "loss": 5.188,
+      "step": 359500
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 9.61725208871003e-06,
+      "loss": 5.1867,
+      "step": 360000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 9.616705305979616e-06,
+      "loss": 5.1689,
+      "step": 360500
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 9.616158523249202e-06,
+      "loss": 5.1738,
+      "step": 361000
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 9.61561174051879e-06,
+      "loss": 5.1671,
+      "step": 361500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 9.615064957788374e-06,
+      "loss": 5.1646,
+      "step": 362000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 9.61451817505796e-06,
+      "loss": 5.1562,
+      "step": 362500
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 9.613971392327545e-06,
+      "loss": 5.1568,
+      "step": 363000
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 9.613424609597131e-06,
+      "loss": 5.1498,
+      "step": 363500
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 9.612877826866717e-06,
+      "loss": 5.1435,
+      "step": 364000
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 9.612331044136303e-06,
+      "loss": 5.1395,
+      "step": 364500
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 9.611784261405888e-06,
+      "loss": 5.1349,
+      "step": 365000
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 9.611237478675474e-06,
+      "loss": 5.1349,
+      "step": 365500
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 9.61069069594506e-06,
+      "loss": 5.1275,
+      "step": 366000
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 9.610143913214646e-06,
+      "loss": 5.1228,
+      "step": 366500
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 9.609597130484232e-06,
+      "loss": 5.1191,
+      "step": 367000
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 9.609050347753817e-06,
+      "loss": 5.1151,
+      "step": 367500
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 9.608503565023403e-06,
+      "loss": 5.1089,
+      "step": 368000
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 9.607956782292989e-06,
+      "loss": 5.1041,
+      "step": 368500
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 9.607409999562575e-06,
+      "loss": 5.0985,
+      "step": 369000
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 9.60686321683216e-06,
+      "loss": 5.102,
+      "step": 369500
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 9.606316434101746e-06,
+      "loss": 5.0946,
+      "step": 370000
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 9.605769651371332e-06,
+      "loss": 5.0812,
+      "step": 370500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 9.605222868640918e-06,
+      "loss": 5.0835,
+      "step": 371000
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 9.604676085910504e-06,
+      "loss": 5.0752,
+      "step": 371500
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 9.604129303180088e-06,
+      "loss": 5.0734,
+      "step": 372000
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 9.603582520449675e-06,
+      "loss": 5.0628,
+      "step": 372500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 9.603035737719261e-06,
+      "loss": 5.0641,
+      "step": 373000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 9.602488954988847e-06,
+      "loss": 5.0574,
+      "step": 373500
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 9.601942172258433e-06,
+      "loss": 5.05,
+      "step": 374000
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 9.601395389528017e-06,
+      "loss": 5.0549,
+      "step": 374500
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 9.600848606797604e-06,
+      "loss": 5.0453,
+      "step": 375000
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 9.60030182406719e-06,
+      "loss": 5.04,
+      "step": 375500
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 9.599755041336776e-06,
+      "loss": 5.0348,
+      "step": 376000
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 9.59920825860636e-06,
+      "loss": 5.0326,
+      "step": 376500
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 9.598661475875947e-06,
+      "loss": 5.0251,
+      "step": 377000
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 9.598114693145533e-06,
+      "loss": 5.0185,
+      "step": 377500
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 9.597567910415119e-06,
+      "loss": 5.0179,
+      "step": 378000
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 9.597021127684703e-06,
+      "loss": 5.0095,
+      "step": 378500
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 9.596474344954289e-06,
+      "loss": 5.0071,
+      "step": 379000
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 9.595927562223876e-06,
+      "loss": 5.0017,
+      "step": 379500
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 9.595380779493462e-06,
+      "loss": 4.9998,
+      "step": 380000
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 9.594833996763046e-06,
+      "loss": 4.9936,
+      "step": 380500
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 9.594287214032632e-06,
+      "loss": 4.9838,
+      "step": 381000
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 9.593740431302218e-06,
+      "loss": 4.9839,
+      "step": 381500
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 9.593193648571805e-06,
+      "loss": 4.9739,
+      "step": 382000
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 9.59264686584139e-06,
+      "loss": 4.9714,
+      "step": 382500
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 9.592100083110975e-06,
+      "loss": 4.963,
+      "step": 383000
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 9.591553300380561e-06,
+      "loss": 4.9637,
+      "step": 383500
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 9.591006517650147e-06,
+      "loss": 4.9556,
+      "step": 384000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 9.590459734919733e-06,
+      "loss": 4.9552,
+      "step": 384500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 9.589912952189318e-06,
+      "loss": 4.9493,
+      "step": 385000
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 9.589366169458904e-06,
+      "loss": 4.944,
+      "step": 385500
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 9.58881938672849e-06,
+      "loss": 4.9345,
+      "step": 386000
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 9.588272603998076e-06,
+      "loss": 4.9343,
+      "step": 386500
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 9.587725821267662e-06,
+      "loss": 4.9298,
+      "step": 387000
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 9.587179038537247e-06,
+      "loss": 4.9261,
+      "step": 387500
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 9.586632255806833e-06,
+      "loss": 4.9172,
+      "step": 388000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 9.586085473076419e-06,
+      "loss": 4.9179,
+      "step": 388500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 9.585538690346005e-06,
+      "loss": 4.9155,
+      "step": 389000
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 9.58499190761559e-06,
+      "loss": 4.9035,
+      "step": 389500
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 9.584445124885176e-06,
+      "loss": 4.9015,
+      "step": 390000
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 9.583898342154762e-06,
+      "loss": 4.8933,
+      "step": 390500
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 9.583351559424348e-06,
+      "loss": 4.8894,
+      "step": 391000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 9.582804776693934e-06,
+      "loss": 4.8848,
+      "step": 391500
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 9.58225799396352e-06,
+      "loss": 4.8809,
+      "step": 392000
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 9.581711211233105e-06,
+      "loss": 4.8738,
+      "step": 392500
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 9.581164428502691e-06,
+      "loss": 4.8704,
+      "step": 393000
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 9.580617645772277e-06,
+      "loss": 4.8696,
+      "step": 393500
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 9.580070863041863e-06,
+      "loss": 4.8608,
+      "step": 394000
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 9.579524080311449e-06,
+      "loss": 4.8579,
+      "step": 394500
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 9.578977297581034e-06,
+      "loss": 4.8539,
+      "step": 395000
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 9.57843051485062e-06,
+      "loss": 4.8503,
+      "step": 395500
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.577883732120206e-06,
+      "loss": 4.8417,
+      "step": 396000
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 9.577336949389792e-06,
+      "loss": 4.8364,
+      "step": 396500
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.576790166659376e-06,
+      "loss": 4.83,
+      "step": 397000
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 9.576243383928963e-06,
+      "loss": 4.8264,
+      "step": 397500
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 9.575696601198549e-06,
+      "loss": 4.8215,
+      "step": 398000
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 9.575149818468135e-06,
+      "loss": 4.8176,
+      "step": 398500
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 9.574603035737719e-06,
+      "loss": 4.8142,
+      "step": 399000
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 9.574056253007305e-06,
+      "loss": 4.8088,
+      "step": 399500
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 9.573509470276892e-06,
+      "loss": 4.8048,
+      "step": 400000
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 9.572962687546478e-06,
+      "loss": 4.7965,
+      "step": 400500
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 9.572415904816062e-06,
+      "loss": 4.7899,
+      "step": 401000
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 9.571869122085648e-06,
+      "loss": 4.7891,
+      "step": 401500
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 9.571322339355234e-06,
+      "loss": 4.7805,
+      "step": 402000
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 9.570775556624821e-06,
+      "loss": 4.7789,
+      "step": 402500
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 9.570228773894407e-06,
+      "loss": 4.7797,
+      "step": 403000
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 9.569681991163991e-06,
+      "loss": 4.7659,
+      "step": 403500
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 9.569135208433577e-06,
+      "loss": 4.7702,
+      "step": 404000
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 9.568588425703164e-06,
+      "loss": 4.7696,
+      "step": 404500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 9.56804164297275e-06,
+      "loss": 4.7572,
+      "step": 405000
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 9.567494860242334e-06,
+      "loss": 4.7501,
+      "step": 405500
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 9.56694807751192e-06,
+      "loss": 4.7428,
+      "step": 406000
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 9.566401294781506e-06,
+      "loss": 4.7433,
+      "step": 406500
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 9.565854512051093e-06,
+      "loss": 4.7368,
+      "step": 407000
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 9.565307729320677e-06,
+      "loss": 4.7326,
+      "step": 407500
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 9.564760946590263e-06,
+      "loss": 4.7201,
+      "step": 408000
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 9.564214163859849e-06,
+      "loss": 4.7242,
+      "step": 408500
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 9.563667381129435e-06,
+      "loss": 4.7201,
+      "step": 409000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 9.56312059839902e-06,
+      "loss": 4.7188,
+      "step": 409500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 9.562573815668606e-06,
+      "loss": 4.7074,
+      "step": 410000
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 9.562027032938192e-06,
+      "loss": 4.7049,
+      "step": 410500
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 9.561480250207778e-06,
+      "loss": 4.6953,
+      "step": 411000
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 9.560933467477364e-06,
+      "loss": 4.6977,
+      "step": 411500
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 9.56038668474695e-06,
+      "loss": 4.6932,
+      "step": 412000
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 9.559839902016535e-06,
+      "loss": 4.6914,
+      "step": 412500
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 9.559293119286121e-06,
+      "loss": 4.6764,
+      "step": 413000
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 9.558746336555707e-06,
+      "loss": 4.6721,
+      "step": 413500
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 9.558199553825293e-06,
+      "loss": 4.6633,
+      "step": 414000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 9.557652771094879e-06,
+      "loss": 4.6613,
+      "step": 414500
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 9.557105988364464e-06,
+      "loss": 4.6675,
+      "step": 415000
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 9.55655920563405e-06,
+      "loss": 4.6553,
+      "step": 415500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 9.556012422903636e-06,
+      "loss": 4.6479,
+      "step": 416000
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 9.555465640173222e-06,
+      "loss": 4.6436,
+      "step": 416500
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 9.554918857442808e-06,
+      "loss": 4.641,
+      "step": 417000
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 9.554372074712393e-06,
+      "loss": 4.6309,
+      "step": 417500
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 9.553825291981979e-06,
+      "loss": 4.6319,
+      "step": 418000
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 9.553278509251565e-06,
+      "loss": 4.6301,
+      "step": 418500
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 9.55273172652115e-06,
+      "loss": 4.619,
+      "step": 419000
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 9.552184943790737e-06,
+      "loss": 4.6225,
+      "step": 419500
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 9.551638161060322e-06,
+      "loss": 4.6094,
+      "step": 420000
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 9.551091378329908e-06,
+      "loss": 4.6124,
+      "step": 420500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 9.550544595599494e-06,
+      "loss": 4.6105,
+      "step": 421000
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 9.54999781286908e-06,
+      "loss": 4.6087,
+      "step": 421500
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 9.549451030138664e-06,
+      "loss": 4.5916,
+      "step": 422000
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 9.548904247408251e-06,
+      "loss": 4.5895,
+      "step": 422500
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 9.548357464677837e-06,
+      "loss": 4.5792,
+      "step": 423000
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 9.547810681947423e-06,
+      "loss": 4.5795,
+      "step": 423500
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 9.547263899217007e-06,
+      "loss": 4.5767,
+      "step": 424000
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 9.546717116486593e-06,
+      "loss": 4.5714,
+      "step": 424500
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 9.54617033375618e-06,
+      "loss": 4.5716,
+      "step": 425000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 9.545623551025766e-06,
+      "loss": 4.5677,
+      "step": 425500
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 9.54507676829535e-06,
+      "loss": 4.5558,
+      "step": 426000
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 9.544529985564936e-06,
+      "loss": 4.5525,
+      "step": 426500
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 9.543983202834522e-06,
+      "loss": 4.5524,
+      "step": 427000
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 9.54343642010411e-06,
+      "loss": 4.5458,
+      "step": 427500
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 9.542889637373693e-06,
+      "loss": 4.5347,
+      "step": 428000
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 9.542342854643279e-06,
+      "loss": 4.5299,
+      "step": 428500
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 9.541796071912865e-06,
+      "loss": 4.5315,
+      "step": 429000
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 9.541249289182452e-06,
+      "loss": 4.5166,
+      "step": 429500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 9.540702506452038e-06,
+      "loss": 4.5221,
+      "step": 430000
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 9.540155723721622e-06,
+      "loss": 4.5206,
+      "step": 430500
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 9.539608940991208e-06,
+      "loss": 4.5156,
+      "step": 431000
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 9.539062158260794e-06,
+      "loss": 4.5059,
+      "step": 431500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 9.538515375530381e-06,
+      "loss": 4.5065,
+      "step": 432000
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 9.537968592799965e-06,
+      "loss": 4.4984,
+      "step": 432500
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 9.537421810069551e-06,
+      "loss": 4.4922,
+      "step": 433000
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 9.536875027339137e-06,
+      "loss": 4.4847,
+      "step": 433500
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 9.536328244608723e-06,
+      "loss": 4.486,
+      "step": 434000
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 9.535781461878309e-06,
+      "loss": 4.4765,
+      "step": 434500
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 9.535234679147894e-06,
+      "loss": 4.4729,
+      "step": 435000
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 9.53468789641748e-06,
+      "loss": 4.4674,
+      "step": 435500
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 9.534141113687066e-06,
+      "loss": 4.4694,
+      "step": 436000
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 9.533594330956652e-06,
+      "loss": 4.462,
+      "step": 436500
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 9.533047548226238e-06,
+      "loss": 4.4627,
+      "step": 437000
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 9.532500765495823e-06,
+      "loss": 4.4552,
+      "step": 437500
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 9.531953982765409e-06,
+      "loss": 4.4394,
+      "step": 438000
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 9.531407200034995e-06,
+      "loss": 4.4471,
+      "step": 438500
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.53086041730458e-06,
+      "loss": 4.438,
+      "step": 439000
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 9.530313634574167e-06,
+      "loss": 4.4304,
+      "step": 439500
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 9.529766851843752e-06,
+      "loss": 4.4337,
+      "step": 440000
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 9.529220069113338e-06,
+      "loss": 4.434,
+      "step": 440500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 9.528673286382924e-06,
+      "loss": 4.4226,
+      "step": 441000
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 9.52812650365251e-06,
+      "loss": 4.4071,
+      "step": 441500
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 9.527579720922095e-06,
+      "loss": 4.4089,
+      "step": 442000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 9.52703293819168e-06,
+      "loss": 4.4118,
+      "step": 442500
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 9.526486155461267e-06,
+      "loss": 4.4082,
+      "step": 443000
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 9.525939372730853e-06,
+      "loss": 4.3927,
+      "step": 443500
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 9.525392590000439e-06,
+      "loss": 4.4015,
+      "step": 444000
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 9.524845807270024e-06,
+      "loss": 4.3857,
+      "step": 444500
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 9.52429902453961e-06,
+      "loss": 4.383,
+      "step": 445000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 9.523752241809196e-06,
+      "loss": 4.3945,
+      "step": 445500
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 9.523205459078782e-06,
+      "loss": 4.3776,
+      "step": 446000
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 9.522658676348368e-06,
+      "loss": 4.3746,
+      "step": 446500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 9.522111893617952e-06,
+      "loss": 4.3645,
+      "step": 447000
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 9.52156511088754e-06,
+      "loss": 4.3641,
+      "step": 447500
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 9.521018328157125e-06,
+      "loss": 4.3621,
+      "step": 448000
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 9.52047154542671e-06,
+      "loss": 4.3562,
+      "step": 448500
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 9.519924762696295e-06,
+      "loss": 4.3482,
+      "step": 449000
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 9.51937797996588e-06,
+      "loss": 4.3439,
+      "step": 449500
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 9.518831197235468e-06,
+      "loss": 4.342,
+      "step": 450000
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 9.518284414505054e-06,
+      "loss": 4.3389,
+      "step": 450500
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 9.517737631774638e-06,
+      "loss": 4.3359,
+      "step": 451000
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 9.517190849044224e-06,
+      "loss": 4.3261,
+      "step": 451500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 9.51664406631381e-06,
+      "loss": 4.3196,
+      "step": 452000
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 9.516097283583397e-06,
+      "loss": 4.3216,
+      "step": 452500
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 9.515550500852981e-06,
+      "loss": 4.3231,
+      "step": 453000
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 9.515003718122567e-06,
+      "loss": 4.3164,
+      "step": 453500
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 9.514456935392153e-06,
+      "loss": 4.3188,
+      "step": 454000
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 9.513910152661739e-06,
+      "loss": 4.3195,
+      "step": 454500
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 9.513363369931324e-06,
+      "loss": 4.3036,
+      "step": 455000
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 9.51281658720091e-06,
+      "loss": 4.2979,
+      "step": 455500
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 9.512269804470496e-06,
+      "loss": 4.302,
+      "step": 456000
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 9.511723021740082e-06,
+      "loss": 4.2918,
+      "step": 456500
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 9.511176239009668e-06,
+      "loss": 4.2948,
+      "step": 457000
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 9.510629456279253e-06,
+      "loss": 4.2925,
+      "step": 457500
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 9.51008267354884e-06,
+      "loss": 4.2762,
+      "step": 458000
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 9.509535890818425e-06,
+      "loss": 4.2798,
+      "step": 458500
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 9.50898910808801e-06,
+      "loss": 4.2699,
+      "step": 459000
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 9.508442325357597e-06,
+      "loss": 4.2632,
+      "step": 459500
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 9.507895542627182e-06,
+      "loss": 4.2615,
+      "step": 460000
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 9.507348759896768e-06,
+      "loss": 4.2606,
+      "step": 460500
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 9.506801977166354e-06,
+      "loss": 4.258,
+      "step": 461000
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 9.50625519443594e-06,
+      "loss": 4.2495,
+      "step": 461500
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 9.505708411705526e-06,
+      "loss": 4.2521,
+      "step": 462000
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 9.505161628975111e-06,
+      "loss": 4.2388,
+      "step": 462500
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 9.504614846244697e-06,
+      "loss": 4.2454,
+      "step": 463000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 9.504068063514283e-06,
+      "loss": 4.234,
+      "step": 463500
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 9.503521280783869e-06,
+      "loss": 4.2306,
+      "step": 464000
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 9.502974498053454e-06,
+      "loss": 4.2225,
+      "step": 464500
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 9.50242771532304e-06,
+      "loss": 4.2125,
+      "step": 465000
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 9.501880932592626e-06,
+      "loss": 4.2172,
+      "step": 465500
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 9.501334149862212e-06,
+      "loss": 4.2187,
+      "step": 466000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 9.500787367131798e-06,
+      "loss": 4.2111,
+      "step": 466500
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 9.500240584401383e-06,
+      "loss": 4.2081,
+      "step": 467000
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 9.499693801670968e-06,
+      "loss": 4.202,
+      "step": 467500
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 9.499147018940555e-06,
+      "loss": 4.2048,
+      "step": 468000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 9.49860023621014e-06,
+      "loss": 4.2082,
+      "step": 468500
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 9.498053453479727e-06,
+      "loss": 4.1995,
+      "step": 469000
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 9.49750667074931e-06,
+      "loss": 4.1859,
+      "step": 469500
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 9.496959888018897e-06,
+      "loss": 4.1834,
+      "step": 470000
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 9.496413105288484e-06,
+      "loss": 4.1902,
+      "step": 470500
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 9.49586632255807e-06,
+      "loss": 4.1773,
+      "step": 471000
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 9.495319539827654e-06,
+      "loss": 4.1776,
+      "step": 471500
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.49477275709724e-06,
+      "loss": 4.1778,
+      "step": 472000
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.494225974366827e-06,
+      "loss": 4.1757,
+      "step": 472500
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.493679191636413e-06,
+      "loss": 4.1702,
+      "step": 473000
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.493132408905999e-06,
+      "loss": 4.1599,
+      "step": 473500
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.492585626175583e-06,
+      "loss": 4.1559,
+      "step": 474000
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.492038843445169e-06,
+      "loss": 4.1565,
+      "step": 474500
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.491492060714756e-06,
+      "loss": 4.1562,
+      "step": 475000
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.490945277984342e-06,
+      "loss": 4.1492,
+      "step": 475500
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 9.490398495253926e-06,
+      "loss": 4.1425,
+      "step": 476000
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.489851712523512e-06,
+      "loss": 4.1448,
+      "step": 476500
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.489304929793098e-06,
+      "loss": 4.1395,
+      "step": 477000
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 9.488758147062685e-06,
+      "loss": 4.1331,
+      "step": 477500
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 9.48821136433227e-06,
+      "loss": 4.1383,
+      "step": 478000
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.487664581601855e-06,
+      "loss": 4.1344,
+      "step": 478500
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.48711779887144e-06,
+      "loss": 4.1179,
+      "step": 479000
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.486571016141027e-06,
+      "loss": 4.1171,
+      "step": 479500
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.486024233410612e-06,
+      "loss": 4.1132,
+      "step": 480000
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.485477450680198e-06,
+      "loss": 4.1087,
+      "step": 480500
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.484930667949784e-06,
+      "loss": 4.1064,
+      "step": 481000
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 9.48438388521937e-06,
+      "loss": 4.1014,
+      "step": 481500
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.483837102488956e-06,
+      "loss": 4.1004,
+      "step": 482000
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.483290319758541e-06,
+      "loss": 4.0989,
+      "step": 482500
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.482743537028127e-06,
+      "loss": 4.0915,
+      "step": 483000
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.482196754297713e-06,
+      "loss": 4.098,
+      "step": 483500
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.481649971567299e-06,
+      "loss": 4.0871,
+      "step": 484000
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.481103188836885e-06,
+      "loss": 4.0871,
+      "step": 484500
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.48055640610647e-06,
+      "loss": 4.0845,
+      "step": 485000
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.480009623376056e-06,
+      "loss": 4.0715,
+      "step": 485500
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.479462840645642e-06,
+      "loss": 4.0853,
+      "step": 486000
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.478916057915228e-06,
+      "loss": 4.0699,
+      "step": 486500
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 9.478369275184813e-06,
+      "loss": 4.0738,
+      "step": 487000
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 9.4778224924544e-06,
+      "loss": 4.067,
+      "step": 487500
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 9.477275709723985e-06,
+      "loss": 4.0589,
+      "step": 488000
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.476728926993571e-06,
+      "loss": 4.0534,
+      "step": 488500
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.476182144263157e-06,
+      "loss": 4.0577,
+      "step": 489000
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.475635361532742e-06,
+      "loss": 4.0538,
+      "step": 489500
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.475088578802328e-06,
+      "loss": 4.048,
+      "step": 490000
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.474541796071914e-06,
+      "loss": 4.0448,
+      "step": 490500
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.4739950133415e-06,
+      "loss": 4.0399,
+      "step": 491000
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.473448230611086e-06,
+      "loss": 4.0329,
+      "step": 491500
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.472901447880671e-06,
+      "loss": 4.0383,
+      "step": 492000
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 9.472354665150256e-06,
+      "loss": 4.028,
+      "step": 492500
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.471807882419843e-06,
+      "loss": 4.0282,
+      "step": 493000
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.471261099689429e-06,
+      "loss": 4.0316,
+      "step": 493500
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.470714316959015e-06,
+      "loss": 4.0219,
+      "step": 494000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.470167534228599e-06,
+      "loss": 4.022,
+      "step": 494500
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.469620751498184e-06,
+      "loss": 4.016,
+      "step": 495000
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.469073968767772e-06,
+      "loss": 4.0138,
+      "step": 495500
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.468527186037358e-06,
+      "loss": 4.0019,
+      "step": 496000
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.467980403306942e-06,
+      "loss": 4.002,
+      "step": 496500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.467433620576528e-06,
+      "loss": 3.994,
+      "step": 497000
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.466886837846115e-06,
+      "loss": 3.997,
+      "step": 497500
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 9.466340055115701e-06,
+      "loss": 3.9945,
+      "step": 498000
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.465793272385285e-06,
+      "loss": 3.9883,
+      "step": 498500
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.46524648965487e-06,
+      "loss": 3.9824,
+      "step": 499000
+    },
+    {
+      "epoch": 5.45,
+      "step": 499134,
+      "total_flos": 1.0578981712518185e+18,
+      "train_loss": 5.580446514515485,
+      "train_runtime": 197998.7388,
+      "train_samples_per_second": 2958.988,
+      "train_steps_per_second": 46.235
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 9154400,
+  "num_train_epochs": 100,
+  "save_steps": 1000000,
+  "total_flos": 1.0578981712518185e+18,
+  "trial_name": null,
+  "trial_params": null
+}