diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,8146 @@
+{
+  "best_metric": 0.8621688895493554,
+  "best_model_checkpoint": "convnext-base-224-22k-finetuned-combinedSpiders/checkpoint-13456",
+  "epoch": 3.9991084032989077,
+  "eval_steps": 500,
+  "global_step": 13456,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.7147102526002973e-07,
+      "loss": 5.4769,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.429420505200595e-07,
+      "loss": 5.4671,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1144130757800891e-06,
+      "loss": 5.4571,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.485884101040119e-06,
+      "loss": 5.4551,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.8573551263001487e-06,
+      "loss": 5.4249,
+      "step": 50
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.2288261515601783e-06,
+      "loss": 5.4008,
+      "step": 60
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.600297176820208e-06,
+      "loss": 5.3555,
+      "step": 70
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 2.971768202080238e-06,
+      "loss": 5.3237,
+      "step": 80
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.3432392273402677e-06,
+      "loss": 5.2974,
+      "step": 90
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 3.7147102526002974e-06,
+      "loss": 5.2394,
+      "step": 100
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.086181277860327e-06,
+      "loss": 5.1621,
+      "step": 110
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.457652303120357e-06,
+      "loss": 5.1085,
+      "step": 120
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.829123328380386e-06,
+      "loss": 5.031,
+      "step": 130
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.200594353640416e-06,
+      "loss": 4.9637,
+      "step": 140
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 5.572065378900446e-06,
+      "loss": 4.9248,
+      "step": 150
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 5.943536404160476e-06,
+      "loss": 4.7622,
+      "step": 160
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6.3150074294205055e-06,
+      "loss": 4.6425,
+      "step": 170
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 6.686478454680535e-06,
+      "loss": 4.461,
+      "step": 180
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.057949479940565e-06,
+      "loss": 4.2733,
+      "step": 190
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.429420505200595e-06,
+      "loss": 4.317,
+      "step": 200
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 7.800891530460625e-06,
+      "loss": 4.174,
+      "step": 210
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.172362555720654e-06,
+      "loss": 4.0508,
+      "step": 220
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.543833580980684e-06,
+      "loss": 3.9818,
+      "step": 230
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 8.915304606240713e-06,
+      "loss": 3.9572,
+      "step": 240
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 9.286775631500744e-06,
+      "loss": 3.8253,
+      "step": 250
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 9.658246656760773e-06,
+      "loss": 3.7814,
+      "step": 260
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.0029717682020803e-05,
+      "loss": 3.6116,
+      "step": 270
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.0401188707280832e-05,
+      "loss": 3.5507,
+      "step": 280
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.0772659732540863e-05,
+      "loss": 3.6112,
+      "step": 290
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.1144130757800892e-05,
+      "loss": 3.5481,
+      "step": 300
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.1515601783060921e-05,
+      "loss": 3.3788,
+      "step": 310
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.1887072808320951e-05,
+      "loss": 3.3373,
+      "step": 320
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.225854383358098e-05,
+      "loss": 3.2484,
+      "step": 330
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.2630014858841011e-05,
+      "loss": 3.1009,
+      "step": 340
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.3001485884101042e-05,
+      "loss": 3.2515,
+      "step": 350
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.337295690936107e-05,
+      "loss": 3.0072,
+      "step": 360
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.3744427934621101e-05,
+      "loss": 3.2075,
+      "step": 370
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.411589895988113e-05,
+      "loss": 3.0212,
+      "step": 380
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.4487369985141161e-05,
+      "loss": 3.0027,
+      "step": 390
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.485884101040119e-05,
+      "loss": 3.127,
+      "step": 400
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.5230312035661217e-05,
+      "loss": 2.8748,
+      "step": 410
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.560178306092125e-05,
+      "loss": 2.8177,
+      "step": 420
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.5973254086181278e-05,
+      "loss": 2.9048,
+      "step": 430
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.6344725111441307e-05,
+      "loss": 2.8571,
+      "step": 440
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.6716196136701336e-05,
+      "loss": 2.8192,
+      "step": 450
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.708766716196137e-05,
+      "loss": 2.7917,
+      "step": 460
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7459138187221397e-05,
+      "loss": 2.7966,
+      "step": 470
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.7830609212481426e-05,
+      "loss": 2.7529,
+      "step": 480
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.8202080237741455e-05,
+      "loss": 2.7381,
+      "step": 490
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.8573551263001488e-05,
+      "loss": 2.6934,
+      "step": 500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.8945022288261517e-05,
+      "loss": 2.6434,
+      "step": 510
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9316493313521546e-05,
+      "loss": 2.6754,
+      "step": 520
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9687964338781574e-05,
+      "loss": 2.5138,
+      "step": 530
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.0059435364041607e-05,
+      "loss": 2.4553,
+      "step": 540
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 2.0430906389301636e-05,
+      "loss": 2.5362,
+      "step": 550
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.0802377414561665e-05,
+      "loss": 2.5306,
+      "step": 560
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.1173848439821694e-05,
+      "loss": 2.5154,
+      "step": 570
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 2.1545319465081726e-05,
+      "loss": 2.452,
+      "step": 580
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.1916790490341755e-05,
+      "loss": 2.4911,
+      "step": 590
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.2288261515601784e-05,
+      "loss": 2.4497,
+      "step": 600
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.2659732540861813e-05,
+      "loss": 2.5367,
+      "step": 610
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 2.3031203566121842e-05,
+      "loss": 2.4149,
+      "step": 620
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.3402674591381874e-05,
+      "loss": 2.3358,
+      "step": 630
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.3774145616641903e-05,
+      "loss": 2.3967,
+      "step": 640
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 2.4145616641901932e-05,
+      "loss": 2.4463,
+      "step": 650
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.451708766716196e-05,
+      "loss": 2.3694,
+      "step": 660
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.4888558692421993e-05,
+      "loss": 2.444,
+      "step": 670
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 2.5260029717682022e-05,
+      "loss": 2.3637,
+      "step": 680
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.563150074294205e-05,
+      "loss": 2.4219,
+      "step": 690
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.6002971768202083e-05,
+      "loss": 2.3928,
+      "step": 700
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.6374442793462112e-05,
+      "loss": 2.3344,
+      "step": 710
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 2.674591381872214e-05,
+      "loss": 2.3485,
+      "step": 720
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.711738484398217e-05,
+      "loss": 2.3023,
+      "step": 730
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.7488855869242203e-05,
+      "loss": 2.3978,
+      "step": 740
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 2.786032689450223e-05,
+      "loss": 2.2595,
+      "step": 750
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.823179791976226e-05,
+      "loss": 2.1633,
+      "step": 760
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.860326894502229e-05,
+      "loss": 2.2674,
+      "step": 770
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.8974739970282322e-05,
+      "loss": 2.2895,
+      "step": 780
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 2.934621099554235e-05,
+      "loss": 2.4147,
+      "step": 790
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 2.971768202080238e-05,
+      "loss": 2.2049,
+      "step": 800
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.008915304606241e-05,
+      "loss": 2.1035,
+      "step": 810
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.0460624071322434e-05,
+      "loss": 2.068,
+      "step": 820
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.083209509658247e-05,
+      "loss": 2.0606,
+      "step": 830
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.12035661218425e-05,
+      "loss": 2.1117,
+      "step": 840
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.157503714710253e-05,
+      "loss": 2.2307,
+      "step": 850
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.1946508172362557e-05,
+      "loss": 2.0279,
+      "step": 860
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.2317979197622586e-05,
+      "loss": 2.2121,
+      "step": 870
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.2689450222882614e-05,
+      "loss": 2.2117,
+      "step": 880
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.3060921248142643e-05,
+      "loss": 2.0072,
+      "step": 890
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.343239227340267e-05,
+      "loss": 2.0412,
+      "step": 900
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.380386329866271e-05,
+      "loss": 1.9867,
+      "step": 910
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.417533432392274e-05,
+      "loss": 2.1701,
+      "step": 920
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.4546805349182766e-05,
+      "loss": 2.0865,
+      "step": 930
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.4918276374442795e-05,
+      "loss": 2.0244,
+      "step": 940
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5289747399702824e-05,
+      "loss": 2.0317,
+      "step": 950
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.566121842496285e-05,
+      "loss": 2.1421,
+      "step": 960
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.603268945022288e-05,
+      "loss": 1.985,
+      "step": 970
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.640416047548291e-05,
+      "loss": 1.9338,
+      "step": 980
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.6775631500742946e-05,
+      "loss": 2.0607,
+      "step": 990
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.7147102526002975e-05,
+      "loss": 2.0481,
+      "step": 1000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.7518573551263004e-05,
+      "loss": 1.9969,
+      "step": 1010
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.789004457652303e-05,
+      "loss": 1.9012,
+      "step": 1020
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.826151560178306e-05,
+      "loss": 2.0606,
+      "step": 1030
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.863298662704309e-05,
+      "loss": 1.9283,
+      "step": 1040
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.900445765230312e-05,
+      "loss": 1.9245,
+      "step": 1050
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.937592867756315e-05,
+      "loss": 1.9724,
+      "step": 1060
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.974739970282318e-05,
+      "loss": 1.8486,
+      "step": 1070
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.0118870728083214e-05,
+      "loss": 1.9798,
+      "step": 1080
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.049034175334324e-05,
+      "loss": 2.019,
+      "step": 1090
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.086181277860327e-05,
+      "loss": 1.931,
+      "step": 1100
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.12332838038633e-05,
+      "loss": 1.7966,
+      "step": 1110
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.160475482912333e-05,
+      "loss": 1.9558,
+      "step": 1120
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.197622585438336e-05,
+      "loss": 1.8243,
+      "step": 1130
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.234769687964339e-05,
+      "loss": 1.8345,
+      "step": 1140
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.2719167904903416e-05,
+      "loss": 1.8198,
+      "step": 1150
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.309063893016345e-05,
+      "loss": 1.7667,
+      "step": 1160
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.346210995542348e-05,
+      "loss": 1.7788,
+      "step": 1170
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.383358098068351e-05,
+      "loss": 1.8643,
+      "step": 1180
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.420505200594354e-05,
+      "loss": 1.8101,
+      "step": 1190
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.457652303120357e-05,
+      "loss": 1.7821,
+      "step": 1200
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.49479940564636e-05,
+      "loss": 1.8136,
+      "step": 1210
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.5319465081723626e-05,
+      "loss": 1.7873,
+      "step": 1220
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.5690936106983654e-05,
+      "loss": 1.8414,
+      "step": 1230
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6062407132243683e-05,
+      "loss": 1.8443,
+      "step": 1240
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.643387815750372e-05,
+      "loss": 1.5641,
+      "step": 1250
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.680534918276375e-05,
+      "loss": 1.6055,
+      "step": 1260
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.717682020802378e-05,
+      "loss": 1.8514,
+      "step": 1270
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.7548291233283806e-05,
+      "loss": 1.7509,
+      "step": 1280
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.7919762258543835e-05,
+      "loss": 1.681,
+      "step": 1290
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.8291233283803864e-05,
+      "loss": 1.9008,
+      "step": 1300
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.866270430906389e-05,
+      "loss": 1.6932,
+      "step": 1310
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.903417533432392e-05,
+      "loss": 1.6835,
+      "step": 1320
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.940564635958396e-05,
+      "loss": 1.7406,
+      "step": 1330
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9777117384843986e-05,
+      "loss": 1.7638,
+      "step": 1340
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.9983484723369115e-05,
+      "loss": 1.6864,
+      "step": 1350
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.994219653179191e-05,
+      "loss": 1.6977,
+      "step": 1360
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.9900908340214704e-05,
+      "loss": 1.7116,
+      "step": 1370
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.985962014863749e-05,
+      "loss": 1.5635,
+      "step": 1380
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.981833195706028e-05,
+      "loss": 1.5219,
+      "step": 1390
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.9777043765483074e-05,
+      "loss": 1.652,
+      "step": 1400
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.973575557390587e-05,
+      "loss": 1.6614,
+      "step": 1410
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.9694467382328656e-05,
+      "loss": 1.6868,
+      "step": 1420
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.965317919075145e-05,
+      "loss": 1.7126,
+      "step": 1430
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.961189099917424e-05,
+      "loss": 1.6174,
+      "step": 1440
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.9570602807597026e-05,
+      "loss": 1.6435,
+      "step": 1450
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.952931461601982e-05,
+      "loss": 1.7329,
+      "step": 1460
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.9488026424442615e-05,
+      "loss": 1.656,
+      "step": 1470
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.94467382328654e-05,
+      "loss": 1.6349,
+      "step": 1480
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.940545004128819e-05,
+      "loss": 1.6277,
+      "step": 1490
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.9364161849710985e-05,
+      "loss": 1.6437,
+      "step": 1500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.932287365813378e-05,
+      "loss": 1.6557,
+      "step": 1510
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.928158546655657e-05,
+      "loss": 1.607,
+      "step": 1520
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.9240297274979355e-05,
+      "loss": 1.6604,
+      "step": 1530
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.919900908340215e-05,
+      "loss": 1.7543,
+      "step": 1540
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.915772089182494e-05,
+      "loss": 1.4938,
+      "step": 1550
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.911643270024773e-05,
+      "loss": 1.5508,
+      "step": 1560
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.9075144508670526e-05,
+      "loss": 1.5558,
+      "step": 1570
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.9033856317093314e-05,
+      "loss": 1.5854,
+      "step": 1580
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.89925681255161e-05,
+      "loss": 1.5201,
+      "step": 1590
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8951279933938896e-05,
+      "loss": 1.646,
+      "step": 1600
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.890999174236169e-05,
+      "loss": 1.5494,
+      "step": 1610
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.886870355078448e-05,
+      "loss": 1.6561,
+      "step": 1620
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.8827415359207266e-05,
+      "loss": 1.5297,
+      "step": 1630
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.878612716763006e-05,
+      "loss": 1.589,
+      "step": 1640
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.8744838976052855e-05,
+      "loss": 1.4322,
+      "step": 1650
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.870355078447564e-05,
+      "loss": 1.4366,
+      "step": 1660
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.866226259289843e-05,
+      "loss": 1.5548,
+      "step": 1670
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.8620974401321225e-05,
+      "loss": 1.6049,
+      "step": 1680
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.857968620974401e-05,
+      "loss": 1.4819,
+      "step": 1690
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.853839801816681e-05,
+      "loss": 1.676,
+      "step": 1700
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.8497109826589595e-05,
+      "loss": 1.6156,
+      "step": 1710
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.845582163501239e-05,
+      "loss": 1.486,
+      "step": 1720
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.841453344343518e-05,
+      "loss": 1.4179,
+      "step": 1730
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8373245251857965e-05,
+      "loss": 1.4969,
+      "step": 1740
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8331957060280766e-05,
+      "loss": 1.512,
+      "step": 1750
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.8290668868703554e-05,
+      "loss": 1.5026,
+      "step": 1760
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.824938067712634e-05,
+      "loss": 1.4611,
+      "step": 1770
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.8208092485549136e-05,
+      "loss": 1.5097,
+      "step": 1780
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.816680429397193e-05,
+      "loss": 1.4507,
+      "step": 1790
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.812551610239472e-05,
+      "loss": 1.5534,
+      "step": 1800
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.8084227910817506e-05,
+      "loss": 1.4548,
+      "step": 1810
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.80429397192403e-05,
+      "loss": 1.4948,
+      "step": 1820
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.800165152766309e-05,
+      "loss": 1.4844,
+      "step": 1830
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.796036333608588e-05,
+      "loss": 1.4406,
+      "step": 1840
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.791907514450867e-05,
+      "loss": 1.4048,
+      "step": 1850
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.7877786952931465e-05,
+      "loss": 1.4718,
+      "step": 1860
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.783649876135425e-05,
+      "loss": 1.4017,
+      "step": 1870
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.779521056977704e-05,
+      "loss": 1.5017,
+      "step": 1880
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.775392237819984e-05,
+      "loss": 1.3266,
+      "step": 1890
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.771263418662263e-05,
+      "loss": 1.5309,
+      "step": 1900
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.767134599504542e-05,
+      "loss": 1.3523,
+      "step": 1910
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.763005780346821e-05,
+      "loss": 1.3471,
+      "step": 1920
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.7588769611891e-05,
+      "loss": 1.3926,
+      "step": 1930
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.7547481420313794e-05,
+      "loss": 1.3736,
+      "step": 1940
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.750619322873658e-05,
+      "loss": 1.5417,
+      "step": 1950
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.7464905037159377e-05,
+      "loss": 1.37,
+      "step": 1960
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.7423616845582164e-05,
+      "loss": 1.2312,
+      "step": 1970
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.738232865400496e-05,
+      "loss": 1.415,
+      "step": 1980
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.7341040462427747e-05,
+      "loss": 1.3694,
+      "step": 1990
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.729975227085054e-05,
+      "loss": 1.2538,
+      "step": 2000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.725846407927333e-05,
+      "loss": 1.4002,
+      "step": 2010
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.7217175887696117e-05,
+      "loss": 1.4004,
+      "step": 2020
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.717588769611892e-05,
+      "loss": 1.2891,
+      "step": 2030
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.7134599504541706e-05,
+      "loss": 1.1726,
+      "step": 2040
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.709331131296449e-05,
+      "loss": 1.3932,
+      "step": 2050
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.705202312138728e-05,
+      "loss": 1.432,
+      "step": 2060
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.7010734929810076e-05,
+      "loss": 1.4299,
+      "step": 2070
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.696944673823287e-05,
+      "loss": 1.397,
+      "step": 2080
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.692815854665566e-05,
+      "loss": 1.3712,
+      "step": 2090
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.688687035507845e-05,
+      "loss": 1.3233,
+      "step": 2100
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.684558216350124e-05,
+      "loss": 1.4045,
+      "step": 2110
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.680429397192403e-05,
+      "loss": 1.419,
+      "step": 2120
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.676300578034682e-05,
+      "loss": 1.2554,
+      "step": 2130
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.672171758876962e-05,
+      "loss": 1.4685,
+      "step": 2140
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.6680429397192405e-05,
+      "loss": 1.2991,
+      "step": 2150
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.663914120561519e-05,
+      "loss": 1.3892,
+      "step": 2160
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.659785301403799e-05,
+      "loss": 1.3421,
+      "step": 2170
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.655656482246078e-05,
+      "loss": 1.3872,
+      "step": 2180
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.651527663088357e-05,
+      "loss": 1.2716,
+      "step": 2190
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.647398843930636e-05,
+      "loss": 1.2481,
+      "step": 2200
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.643270024772915e-05,
+      "loss": 1.3937,
+      "step": 2210
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.6391412056151946e-05,
+      "loss": 1.2905,
+      "step": 2220
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.6350123864574733e-05,
+      "loss": 1.2108,
+      "step": 2230
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.630883567299753e-05,
+      "loss": 1.2396,
+      "step": 2240
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.6267547481420316e-05,
+      "loss": 1.3102,
+      "step": 2250
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.6226259289843103e-05,
+      "loss": 1.263,
+      "step": 2260
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.61849710982659e-05,
+      "loss": 1.4224,
+      "step": 2270
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.614368290668869e-05,
+      "loss": 1.3365,
+      "step": 2280
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.610239471511148e-05,
+      "loss": 1.2659,
+      "step": 2290
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.606110652353427e-05,
+      "loss": 1.3508,
+      "step": 2300
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.601981833195706e-05,
+      "loss": 1.4464,
+      "step": 2310
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.597853014037986e-05,
+      "loss": 1.2285,
+      "step": 2320
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.5937241948802645e-05,
+      "loss": 1.324,
+      "step": 2330
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.589595375722543e-05,
+      "loss": 1.29,
+      "step": 2340
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.585466556564823e-05,
+      "loss": 1.2455,
+      "step": 2350
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.581337737407102e-05,
+      "loss": 1.2505,
+      "step": 2360
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.577208918249381e-05,
+      "loss": 1.2991,
+      "step": 2370
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.57308009909166e-05,
+      "loss": 1.3561,
+      "step": 2380
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.568951279933939e-05,
+      "loss": 1.2466,
+      "step": 2390
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.564822460776218e-05,
+      "loss": 1.3424,
+      "step": 2400
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.5606936416184974e-05,
+      "loss": 1.2386,
+      "step": 2410
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.556564822460777e-05,
+      "loss": 1.2878,
+      "step": 2420
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.5524360033030556e-05,
+      "loss": 1.2825,
+      "step": 2430
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.5483071841453344e-05,
+      "loss": 1.3185,
+      "step": 2440
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.544178364987614e-05,
+      "loss": 1.3218,
+      "step": 2450
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.540049545829893e-05,
+      "loss": 1.2743,
+      "step": 2460
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.535920726672172e-05,
+      "loss": 1.2057,
+      "step": 2470
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.531791907514451e-05,
+      "loss": 1.35,
+      "step": 2480
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.52766308835673e-05,
+      "loss": 1.349,
+      "step": 2490
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.523534269199009e-05,
+      "loss": 1.3563,
+      "step": 2500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.5194054500412885e-05,
+      "loss": 1.2365,
+      "step": 2510
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.515276630883567e-05,
+      "loss": 1.1873,
+      "step": 2520
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.511147811725847e-05,
+      "loss": 1.3485,
+      "step": 2530
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.5070189925681255e-05,
+      "loss": 1.2511,
+      "step": 2540
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.502890173410405e-05,
+      "loss": 1.2134,
+      "step": 2550
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.4987613542526844e-05,
+      "loss": 1.2765,
+      "step": 2560
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.494632535094963e-05,
+      "loss": 1.1946,
+      "step": 2570
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.490503715937242e-05,
+      "loss": 1.2485,
+      "step": 2580
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.4863748967795214e-05,
+      "loss": 1.2155,
+      "step": 2590
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.482246077621801e-05,
+      "loss": 1.2722,
+      "step": 2600
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.4781172584640796e-05,
+      "loss": 1.1289,
+      "step": 2610
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.4739884393063584e-05,
+      "loss": 1.2266,
+      "step": 2620
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.469859620148638e-05,
+      "loss": 1.1599,
+      "step": 2630
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.4657308009909166e-05,
+      "loss": 1.1312,
+      "step": 2640
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.461601981833196e-05,
+      "loss": 1.3054,
+      "step": 2650
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.457473162675475e-05,
+      "loss": 1.3679,
+      "step": 2660
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.453344343517754e-05,
+      "loss": 1.118,
+      "step": 2670
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.449215524360033e-05,
+      "loss": 1.2677,
+      "step": 2680
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.4450867052023125e-05,
+      "loss": 1.2401,
+      "step": 2690
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.440957886044591e-05,
+      "loss": 1.1636,
+      "step": 2700
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.436829066886871e-05,
+      "loss": 1.1956,
+      "step": 2710
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.4327002477291495e-05,
+      "loss": 1.2319,
+      "step": 2720
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.428571428571428e-05,
+      "loss": 1.3061,
+      "step": 2730
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.4244426094137084e-05,
+      "loss": 1.1913,
+      "step": 2740
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.420313790255987e-05,
+      "loss": 1.3291,
+      "step": 2750
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.416184971098266e-05,
+      "loss": 1.0688,
+      "step": 2760
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.4120561519405454e-05,
+      "loss": 1.2088,
+      "step": 2770
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.407927332782824e-05,
+      "loss": 1.0837,
+      "step": 2780
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.4037985136251036e-05,
+      "loss": 1.2793,
+      "step": 2790
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.3996696944673824e-05,
+      "loss": 1.2985,
+      "step": 2800
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.395540875309662e-05,
+      "loss": 1.2168,
+      "step": 2810
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.3914120561519406e-05,
+      "loss": 1.2744,
+      "step": 2820
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.3872832369942194e-05,
+      "loss": 1.0637,
+      "step": 2830
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.383154417836499e-05,
+      "loss": 1.1661,
+      "step": 2840
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.379025598678778e-05,
+      "loss": 1.2258,
+      "step": 2850
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.374896779521057e-05,
+      "loss": 1.2925,
+      "step": 2860
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.370767960363336e-05,
+      "loss": 1.1818,
+      "step": 2870
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.366639141205616e-05,
+      "loss": 1.2175,
+      "step": 2880
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.362510322047895e-05,
+      "loss": 1.2059,
+      "step": 2890
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.3583815028901735e-05,
+      "loss": 1.185,
+      "step": 2900
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.354252683732453e-05,
+      "loss": 1.0776,
+      "step": 2910
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.350123864574732e-05,
+      "loss": 1.1255,
+      "step": 2920
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.345995045417011e-05,
+      "loss": 1.1448,
+      "step": 2930
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.34186622625929e-05,
+      "loss": 1.1163,
+      "step": 2940
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.3377374071015694e-05,
+      "loss": 1.1724,
+      "step": 2950
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.333608587943848e-05,
+      "loss": 1.1334,
+      "step": 2960
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.329479768786127e-05,
+      "loss": 1.2145,
+      "step": 2970
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.3253509496284064e-05,
+      "loss": 1.0618,
+      "step": 2980
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.321222130470686e-05,
+      "loss": 1.0882,
+      "step": 2990
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.3170933113129646e-05,
+      "loss": 1.1081,
+      "step": 3000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.3129644921552434e-05,
+      "loss": 1.2089,
+      "step": 3010
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.308835672997523e-05,
+      "loss": 1.1133,
+      "step": 3020
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.304706853839802e-05,
+      "loss": 1.0688,
+      "step": 3030
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.300578034682081e-05,
+      "loss": 1.2481,
+      "step": 3040
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.29644921552436e-05,
+      "loss": 1.1571,
+      "step": 3050
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.292320396366639e-05,
+      "loss": 1.2975,
+      "step": 3060
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.288191577208919e-05,
+      "loss": 1.2199,
+      "step": 3070
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.2840627580511975e-05,
+      "loss": 1.1153,
+      "step": 3080
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.279933938893477e-05,
+      "loss": 1.0653,
+      "step": 3090
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.275805119735756e-05,
+      "loss": 1.0695,
+      "step": 3100
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.2716763005780345e-05,
+      "loss": 1.1104,
+      "step": 3110
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.267547481420314e-05,
+      "loss": 1.1751,
+      "step": 3120
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.2634186622625934e-05,
+      "loss": 1.0879,
+      "step": 3130
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.259289843104872e-05,
+      "loss": 1.1682,
+      "step": 3140
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.255161023947151e-05,
+      "loss": 1.0527,
+      "step": 3150
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.2510322047894304e-05,
+      "loss": 1.2085,
+      "step": 3160
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.24690338563171e-05,
+      "loss": 1.0835,
+      "step": 3170
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.242774566473989e-05,
+      "loss": 1.2115,
+      "step": 3180
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.2386457473162674e-05,
+      "loss": 1.2643,
+      "step": 3190
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.234516928158547e-05,
+      "loss": 1.0601,
+      "step": 3200
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.230388109000826e-05,
+      "loss": 1.2138,
+      "step": 3210
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.226259289843105e-05,
+      "loss": 1.1062,
+      "step": 3220
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.2221304706853846e-05,
+      "loss": 1.1669,
+      "step": 3230
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.2180016515276633e-05,
+      "loss": 1.2187,
+      "step": 3240
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.213872832369942e-05,
+      "loss": 1.1373,
+      "step": 3250
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.2097440132122216e-05,
+      "loss": 1.0939,
+      "step": 3260
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.205615194054501e-05,
+      "loss": 1.0179,
+      "step": 3270
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.20148637489678e-05,
+      "loss": 1.1837,
+      "step": 3280
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.1973575557390586e-05,
+      "loss": 1.0698,
+      "step": 3290
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.193228736581338e-05,
+      "loss": 1.0702,
+      "step": 3300
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.1890999174236175e-05,
+      "loss": 1.0625,
+      "step": 3310
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.184971098265896e-05,
+      "loss": 1.1888,
+      "step": 3320
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.180842279108175e-05,
+      "loss": 1.263,
+      "step": 3330
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.1767134599504545e-05,
+      "loss": 1.061,
+      "step": 3340
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.172584640792733e-05,
+      "loss": 1.0628,
+      "step": 3350
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.168455821635013e-05,
+      "loss": 1.1686,
+      "step": 3360
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.7767953338039157,
+      "eval_f1": 0.43128943900592315,
+      "eval_loss": 0.8614734411239624,
+      "eval_precision": 0.5558129023969989,
+      "eval_recall": 0.40616326210422404,
+      "eval_runtime": 166.8352,
+      "eval_samples_per_second": 161.339,
+      "eval_steps_per_second": 10.088,
+      "step": 3364
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.1643270024772915e-05,
+      "loss": 1.1091,
+      "step": 3370
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.160198183319571e-05,
+      "loss": 1.0051,
+      "step": 3380
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.15606936416185e-05,
+      "loss": 1.0469,
+      "step": 3390
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.1519405450041285e-05,
+      "loss": 1.1269,
+      "step": 3400
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.1478117258464086e-05,
+      "loss": 1.1028,
+      "step": 3410
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.1436829066886874e-05,
+      "loss": 0.958,
+      "step": 3420
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.139554087530966e-05,
+      "loss": 1.0645,
+      "step": 3430
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.1354252683732456e-05,
+      "loss": 1.1045,
+      "step": 3440
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.131296449215525e-05,
+      "loss": 1.0672,
+      "step": 3450
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.127167630057804e-05,
+      "loss": 1.1159,
+      "step": 3460
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.1230388109000826e-05,
+      "loss": 1.0095,
+      "step": 3470
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.118909991742362e-05,
+      "loss": 1.0087,
+      "step": 3480
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.114781172584641e-05,
+      "loss": 1.0704,
+      "step": 3490
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.11065235342692e-05,
+      "loss": 1.0943,
+      "step": 3500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.106523534269199e-05,
+      "loss": 1.0202,
+      "step": 3510
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.1023947151114785e-05,
+      "loss": 0.9964,
+      "step": 3520
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.098265895953757e-05,
+      "loss": 1.2246,
+      "step": 3530
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.094137076796036e-05,
+      "loss": 1.0729,
+      "step": 3540
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.090008257638316e-05,
+      "loss": 0.9676,
+      "step": 3550
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.085879438480595e-05,
+      "loss": 1.0512,
+      "step": 3560
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.081750619322874e-05,
+      "loss": 0.9898,
+      "step": 3570
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.077621800165153e-05,
+      "loss": 0.9965,
+      "step": 3580
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.073492981007432e-05,
+      "loss": 1.0716,
+      "step": 3590
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.0693641618497114e-05,
+      "loss": 0.9443,
+      "step": 3600
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.06523534269199e-05,
+      "loss": 1.0573,
+      "step": 3610
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.0611065235342696e-05,
+      "loss": 1.0709,
+      "step": 3620
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.0569777043765484e-05,
+      "loss": 0.9179,
+      "step": 3630
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.052848885218828e-05,
+      "loss": 1.0501,
+      "step": 3640
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.0487200660611066e-05,
+      "loss": 0.9926,
+      "step": 3650
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.044591246903386e-05,
+      "loss": 1.0323,
+      "step": 3660
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.040462427745665e-05,
+      "loss": 1.1282,
+      "step": 3670
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.0363336085879436e-05,
+      "loss": 0.9758,
+      "step": 3680
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.032204789430224e-05,
+      "loss": 0.9273,
+      "step": 3690
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.0280759702725025e-05,
+      "loss": 0.9133,
+      "step": 3700
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.023947151114781e-05,
+      "loss": 0.8501,
+      "step": 3710
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.01981833195706e-05,
+      "loss": 0.9788,
+      "step": 3720
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.0156895127993395e-05,
+      "loss": 1.1003,
+      "step": 3730
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.011560693641619e-05,
+      "loss": 1.0044,
+      "step": 3740
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.007431874483898e-05,
+      "loss": 0.9837,
+      "step": 3750
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.003303055326177e-05,
+      "loss": 0.9825,
+      "step": 3760
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.999174236168456e-05,
+      "loss": 0.9885,
+      "step": 3770
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.995045417010735e-05,
+      "loss": 0.9296,
+      "step": 3780
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.990916597853014e-05,
+      "loss": 1.0403,
+      "step": 3790
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.9867877786952936e-05,
+      "loss": 1.0269,
+      "step": 3800
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.9826589595375724e-05,
+      "loss": 1.0112,
+      "step": 3810
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.978530140379851e-05,
+      "loss": 1.0928,
+      "step": 3820
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.9744013212221306e-05,
+      "loss": 1.0016,
+      "step": 3830
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.97027250206441e-05,
+      "loss": 1.0688,
+      "step": 3840
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.966143682906689e-05,
+      "loss": 1.0394,
+      "step": 3850
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.9620148637489676e-05,
+      "loss": 1.0164,
+      "step": 3860
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.957886044591247e-05,
+      "loss": 1.0149,
+      "step": 3870
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.9537572254335265e-05,
+      "loss": 1.1226,
+      "step": 3880
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.949628406275805e-05,
+      "loss": 1.0587,
+      "step": 3890
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.945499587118085e-05,
+      "loss": 1.0624,
+      "step": 3900
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.9413707679603635e-05,
+      "loss": 0.9625,
+      "step": 3910
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.937241948802642e-05,
+      "loss": 0.9834,
+      "step": 3920
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.933113129644922e-05,
+      "loss": 1.0298,
+      "step": 3930
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.928984310487201e-05,
+      "loss": 0.9873,
+      "step": 3940
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.92485549132948e-05,
+      "loss": 0.8883,
+      "step": 3950
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.920726672171759e-05,
+      "loss": 0.8128,
+      "step": 3960
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.916597853014038e-05,
+      "loss": 0.9592,
+      "step": 3970
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.9124690338563176e-05,
+      "loss": 1.0227,
+      "step": 3980
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.9083402146985964e-05,
+      "loss": 1.0285,
+      "step": 3990
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.904211395540875e-05,
+      "loss": 1.0696,
+      "step": 4000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.9000825763831546e-05,
+      "loss": 1.0063,
+      "step": 4010
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.895953757225434e-05,
+      "loss": 0.8948,
+      "step": 4020
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.891824938067713e-05,
+      "loss": 1.0921,
+      "step": 4030
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.8876961189099916e-05,
+      "loss": 0.9952,
+      "step": 4040
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.883567299752271e-05,
+      "loss": 0.9707,
+      "step": 4050
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.87943848059455e-05,
+      "loss": 1.0392,
+      "step": 4060
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.875309661436829e-05,
+      "loss": 0.9669,
+      "step": 4070
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.871180842279109e-05,
+      "loss": 1.0256,
+      "step": 4080
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.8670520231213875e-05,
+      "loss": 0.8598,
+      "step": 4090
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.862923203963666e-05,
+      "loss": 1.0604,
+      "step": 4100
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.858794384805946e-05,
+      "loss": 1.1348,
+      "step": 4110
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.854665565648225e-05,
+      "loss": 0.9599,
+      "step": 4120
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.850536746490504e-05,
+      "loss": 1.0197,
+      "step": 4130
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.846407927332783e-05,
+      "loss": 0.9226,
+      "step": 4140
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.842279108175062e-05,
+      "loss": 0.8991,
+      "step": 4150
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.838150289017341e-05,
+      "loss": 0.9736,
+      "step": 4160
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.8340214698596204e-05,
+      "loss": 0.9621,
+      "step": 4170
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.829892650701899e-05,
+      "loss": 1.0498,
+      "step": 4180
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.8257638315441787e-05,
+      "loss": 1.0539,
+      "step": 4190
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.8216350123864574e-05,
+      "loss": 0.9569,
+      "step": 4200
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.817506193228737e-05,
+      "loss": 0.9329,
+      "step": 4210
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.813377374071016e-05,
+      "loss": 1.0702,
+      "step": 4220
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.809248554913295e-05,
+      "loss": 0.9069,
+      "step": 4230
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.805119735755574e-05,
+      "loss": 0.9541,
+      "step": 4240
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.800990916597853e-05,
+      "loss": 1.0066,
+      "step": 4250
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.796862097440133e-05,
+      "loss": 0.9346,
+      "step": 4260
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.7927332782824116e-05,
+      "loss": 1.0449,
+      "step": 4270
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.78860445912469e-05,
+      "loss": 0.934,
+      "step": 4280
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.78447563996697e-05,
+      "loss": 0.8796,
+      "step": 4290
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.7803468208092486e-05,
+      "loss": 1.0545,
+      "step": 4300
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.776218001651528e-05,
+      "loss": 0.9587,
+      "step": 4310
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.772089182493807e-05,
+      "loss": 0.9959,
+      "step": 4320
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.767960363336086e-05,
+      "loss": 1.0199,
+      "step": 4330
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.763831544178365e-05,
+      "loss": 0.9441,
+      "step": 4340
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.759702725020644e-05,
+      "loss": 0.8813,
+      "step": 4350
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.755573905862924e-05,
+      "loss": 1.0313,
+      "step": 4360
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.751445086705203e-05,
+      "loss": 0.9232,
+      "step": 4370
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.7473162675474815e-05,
+      "loss": 0.9363,
+      "step": 4380
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.74318744838976e-05,
+      "loss": 1.0835,
+      "step": 4390
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.7390586292320404e-05,
+      "loss": 0.9058,
+      "step": 4400
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.734929810074319e-05,
+      "loss": 0.9507,
+      "step": 4410
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.730800990916598e-05,
+      "loss": 0.9422,
+      "step": 4420
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.7266721717588773e-05,
+      "loss": 0.9739,
+      "step": 4430
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.722543352601156e-05,
+      "loss": 0.9393,
+      "step": 4440
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.7184145334434356e-05,
+      "loss": 1.0038,
+      "step": 4450
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.7142857142857143e-05,
+      "loss": 1.0644,
+      "step": 4460
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.710156895127994e-05,
+      "loss": 0.9235,
+      "step": 4470
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.7060280759702726e-05,
+      "loss": 0.9771,
+      "step": 4480
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.7018992568125513e-05,
+      "loss": 0.9349,
+      "step": 4490
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.697770437654831e-05,
+      "loss": 0.8724,
+      "step": 4500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.69364161849711e-05,
+      "loss": 0.8822,
+      "step": 4510
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.689512799339389e-05,
+      "loss": 0.817,
+      "step": 4520
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.685383980181668e-05,
+      "loss": 0.9419,
+      "step": 4530
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.681255161023947e-05,
+      "loss": 1.0354,
+      "step": 4540
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.677126341866227e-05,
+      "loss": 0.9066,
+      "step": 4550
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.6729975227085055e-05,
+      "loss": 0.8172,
+      "step": 4560
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.668868703550785e-05,
+      "loss": 1.0179,
+      "step": 4570
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.664739884393064e-05,
+      "loss": 0.9419,
+      "step": 4580
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.660611065235343e-05,
+      "loss": 0.9163,
+      "step": 4590
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.656482246077622e-05,
+      "loss": 0.929,
+      "step": 4600
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.6523534269199014e-05,
+      "loss": 0.942,
+      "step": 4610
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 3.64822460776218e-05,
+      "loss": 0.9966,
+      "step": 4620
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.644095788604459e-05,
+      "loss": 0.9118,
+      "step": 4630
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.6399669694467384e-05,
+      "loss": 0.906,
+      "step": 4640
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.635838150289018e-05,
+      "loss": 0.8901,
+      "step": 4650
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 3.6317093311312966e-05,
+      "loss": 0.8751,
+      "step": 4660
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.6275805119735754e-05,
+      "loss": 1.025,
+      "step": 4670
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.623451692815855e-05,
+      "loss": 0.846,
+      "step": 4680
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 3.619322873658134e-05,
+      "loss": 0.9562,
+      "step": 4690
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.615194054500413e-05,
+      "loss": 0.9231,
+      "step": 4700
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.611065235342692e-05,
+      "loss": 0.9017,
+      "step": 4710
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 3.606936416184971e-05,
+      "loss": 1.0509,
+      "step": 4720
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.60280759702725e-05,
+      "loss": 0.9465,
+      "step": 4730
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.5986787778695295e-05,
+      "loss": 1.0177,
+      "step": 4740
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.594549958711809e-05,
+      "loss": 0.8984,
+      "step": 4750
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 3.590421139554088e-05,
+      "loss": 0.8296,
+      "step": 4760
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.5862923203963665e-05,
+      "loss": 0.8144,
+      "step": 4770
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.582163501238646e-05,
+      "loss": 0.8979,
+      "step": 4780
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 3.5780346820809254e-05,
+      "loss": 0.8947,
+      "step": 4790
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.573905862923204e-05,
+      "loss": 0.9915,
+      "step": 4800
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.569777043765483e-05,
+      "loss": 0.8855,
+      "step": 4810
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.5656482246077624e-05,
+      "loss": 1.0126,
+      "step": 4820
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.561519405450042e-05,
+      "loss": 0.9607,
+      "step": 4830
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.5573905862923206e-05,
+      "loss": 0.8952,
+      "step": 4840
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.5532617671345994e-05,
+      "loss": 0.9283,
+      "step": 4850
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.549132947976879e-05,
+      "loss": 0.9253,
+      "step": 4860
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.5450041288191576e-05,
+      "loss": 0.9621,
+      "step": 4870
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.540875309661437e-05,
+      "loss": 0.9656,
+      "step": 4880
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.5367464905037165e-05,
+      "loss": 0.9215,
+      "step": 4890
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.532617671345995e-05,
+      "loss": 0.8536,
+      "step": 4900
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.528488852188274e-05,
+      "loss": 0.8631,
+      "step": 4910
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.5243600330305535e-05,
+      "loss": 0.8857,
+      "step": 4920
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.520231213872833e-05,
+      "loss": 0.8414,
+      "step": 4930
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.516102394715112e-05,
+      "loss": 0.9057,
+      "step": 4940
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.5119735755573905e-05,
+      "loss": 0.9631,
+      "step": 4950
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.50784475639967e-05,
+      "loss": 0.9654,
+      "step": 4960
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.5037159372419494e-05,
+      "loss": 0.9361,
+      "step": 4970
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.499587118084228e-05,
+      "loss": 1.0339,
+      "step": 4980
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.495458298926507e-05,
+      "loss": 0.929,
+      "step": 4990
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.4913294797687864e-05,
+      "loss": 0.962,
+      "step": 5000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.487200660611065e-05,
+      "loss": 0.8975,
+      "step": 5010
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.4830718414533446e-05,
+      "loss": 1.0198,
+      "step": 5020
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.478943022295624e-05,
+      "loss": 0.9201,
+      "step": 5030
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.474814203137903e-05,
+      "loss": 0.9794,
+      "step": 5040
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.4706853839801816e-05,
+      "loss": 0.8758,
+      "step": 5050
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.4665565648224604e-05,
+      "loss": 0.9724,
+      "step": 5060
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.4624277456647405e-05,
+      "loss": 0.9798,
+      "step": 5070
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.458298926507019e-05,
+      "loss": 0.8502,
+      "step": 5080
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.454170107349298e-05,
+      "loss": 1.0077,
+      "step": 5090
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.4500412881915775e-05,
+      "loss": 0.9431,
+      "step": 5100
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.445912469033856e-05,
+      "loss": 0.9932,
+      "step": 5110
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.441783649876136e-05,
+      "loss": 0.9065,
+      "step": 5120
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 3.4376548307184145e-05,
+      "loss": 0.8573,
+      "step": 5130
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.433526011560694e-05,
+      "loss": 0.8999,
+      "step": 5140
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.429397192402973e-05,
+      "loss": 0.9368,
+      "step": 5150
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 3.425268373245252e-05,
+      "loss": 0.9819,
+      "step": 5160
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.421139554087531e-05,
+      "loss": 0.8536,
+      "step": 5170
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.4170107349298104e-05,
+      "loss": 1.0131,
+      "step": 5180
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 3.412881915772089e-05,
+      "loss": 0.8958,
+      "step": 5190
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.408753096614368e-05,
+      "loss": 0.9812,
+      "step": 5200
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.404624277456648e-05,
+      "loss": 0.966,
+      "step": 5210
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.400495458298927e-05,
+      "loss": 0.8875,
+      "step": 5220
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 3.3963666391412056e-05,
+      "loss": 0.8652,
+      "step": 5230
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.392237819983485e-05,
+      "loss": 0.896,
+      "step": 5240
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.388109000825764e-05,
+      "loss": 0.8278,
+      "step": 5250
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 3.383980181668043e-05,
+      "loss": 0.8427,
+      "step": 5260
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.379851362510322e-05,
+      "loss": 0.9571,
+      "step": 5270
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.3757225433526015e-05,
+      "loss": 0.9,
+      "step": 5280
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 3.37159372419488e-05,
+      "loss": 0.9324,
+      "step": 5290
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.367464905037159e-05,
+      "loss": 0.9181,
+      "step": 5300
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.3633360858794385e-05,
+      "loss": 0.9352,
+      "step": 5310
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.359207266721718e-05,
+      "loss": 0.8436,
+      "step": 5320
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 3.355078447563997e-05,
+      "loss": 0.9244,
+      "step": 5330
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.3509496284062755e-05,
+      "loss": 0.9332,
+      "step": 5340
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.346820809248556e-05,
+      "loss": 0.8271,
+      "step": 5350
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 3.3426919900908344e-05,
+      "loss": 0.9262,
+      "step": 5360
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.338563170933113e-05,
+      "loss": 0.9419,
+      "step": 5370
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.334434351775392e-05,
+      "loss": 1.0383,
+      "step": 5380
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.3303055326176714e-05,
+      "loss": 0.9572,
+      "step": 5390
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 3.326176713459951e-05,
+      "loss": 1.03,
+      "step": 5400
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.32204789430223e-05,
+      "loss": 0.774,
+      "step": 5410
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.317919075144509e-05,
+      "loss": 0.8089,
+      "step": 5420
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 3.313790255986788e-05,
+      "loss": 0.9178,
+      "step": 5430
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.309661436829067e-05,
+      "loss": 0.9183,
+      "step": 5440
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.305532617671346e-05,
+      "loss": 0.8959,
+      "step": 5450
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 3.3014037985136256e-05,
+      "loss": 0.954,
+      "step": 5460
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.297274979355904e-05,
+      "loss": 0.9075,
+      "step": 5470
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.293146160198183e-05,
+      "loss": 0.8012,
+      "step": 5480
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.2890173410404626e-05,
+      "loss": 0.9808,
+      "step": 5490
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 3.284888521882742e-05,
+      "loss": 0.8637,
+      "step": 5500
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.280759702725021e-05,
+      "loss": 0.9188,
+      "step": 5510
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.2766308835672996e-05,
+      "loss": 0.8846,
+      "step": 5520
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 3.272502064409579e-05,
+      "loss": 0.8964,
+      "step": 5530
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.2683732452518585e-05,
+      "loss": 0.8624,
+      "step": 5540
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.264244426094137e-05,
+      "loss": 0.8673,
+      "step": 5550
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 3.260115606936417e-05,
+      "loss": 0.9758,
+      "step": 5560
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.2559867877786955e-05,
+      "loss": 0.9361,
+      "step": 5570
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.251857968620974e-05,
+      "loss": 0.9068,
+      "step": 5580
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.247729149463254e-05,
+      "loss": 0.9003,
+      "step": 5590
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 3.243600330305533e-05,
+      "loss": 0.9155,
+      "step": 5600
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.239471511147812e-05,
+      "loss": 0.8421,
+      "step": 5610
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.235342691990091e-05,
+      "loss": 0.8183,
+      "step": 5620
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 3.23121387283237e-05,
+      "loss": 0.8649,
+      "step": 5630
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.2270850536746496e-05,
+      "loss": 0.7638,
+      "step": 5640
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.2229562345169284e-05,
+      "loss": 0.887,
+      "step": 5650
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 3.218827415359207e-05,
+      "loss": 0.9771,
+      "step": 5660
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.2146985962014866e-05,
+      "loss": 0.9711,
+      "step": 5670
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.2105697770437654e-05,
+      "loss": 0.9672,
+      "step": 5680
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.206440957886045e-05,
+      "loss": 0.8884,
+      "step": 5690
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 3.202312138728324e-05,
+      "loss": 0.8634,
+      "step": 5700
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.198183319570603e-05,
+      "loss": 0.8293,
+      "step": 5710
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.194054500412882e-05,
+      "loss": 0.8303,
+      "step": 5720
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 3.189925681255161e-05,
+      "loss": 0.8107,
+      "step": 5730
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.185796862097441e-05,
+      "loss": 0.7962,
+      "step": 5740
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.1816680429397195e-05,
+      "loss": 0.8317,
+      "step": 5750
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.177539223781998e-05,
+      "loss": 0.8986,
+      "step": 5760
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 3.173410404624278e-05,
+      "loss": 0.8208,
+      "step": 5770
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.169281585466557e-05,
+      "loss": 0.8905,
+      "step": 5780
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.165152766308836e-05,
+      "loss": 0.8486,
+      "step": 5790
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 3.161023947151115e-05,
+      "loss": 0.8749,
+      "step": 5800
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.156895127993394e-05,
+      "loss": 0.9029,
+      "step": 5810
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.152766308835673e-05,
+      "loss": 0.8022,
+      "step": 5820
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 3.1486374896779524e-05,
+      "loss": 0.9101,
+      "step": 5830
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.144508670520231e-05,
+      "loss": 0.7723,
+      "step": 5840
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.1403798513625106e-05,
+      "loss": 0.9002,
+      "step": 5850
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.1362510322047894e-05,
+      "loss": 0.8328,
+      "step": 5860
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 3.132122213047068e-05,
+      "loss": 0.8291,
+      "step": 5870
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.127993393889348e-05,
+      "loss": 0.9002,
+      "step": 5880
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.123864574731627e-05,
+      "loss": 0.8614,
+      "step": 5890
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.119735755573906e-05,
+      "loss": 0.9562,
+      "step": 5900
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.115606936416185e-05,
+      "loss": 0.8864,
+      "step": 5910
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.111478117258465e-05,
+      "loss": 0.9403,
+      "step": 5920
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.1073492981007435e-05,
+      "loss": 0.7972,
+      "step": 5930
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.103220478943022e-05,
+      "loss": 0.9177,
+      "step": 5940
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.099091659785302e-05,
+      "loss": 0.8377,
+      "step": 5950
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.0949628406275805e-05,
+      "loss": 0.8981,
+      "step": 5960
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.09083402146986e-05,
+      "loss": 0.8754,
+      "step": 5970
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.086705202312139e-05,
+      "loss": 0.8468,
+      "step": 5980
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.082576383154418e-05,
+      "loss": 0.8447,
+      "step": 5990
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.078447563996697e-05,
+      "loss": 0.8517,
+      "step": 6000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.074318744838976e-05,
+      "loss": 0.9445,
+      "step": 6010
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.070189925681256e-05,
+      "loss": 0.9384,
+      "step": 6020
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.0660611065235346e-05,
+      "loss": 0.8515,
+      "step": 6030
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0619322873658134e-05,
+      "loss": 0.9311,
+      "step": 6040
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.057803468208092e-05,
+      "loss": 0.9923,
+      "step": 6050
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.0536746490503716e-05,
+      "loss": 0.8712,
+      "step": 6060
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.049545829892651e-05,
+      "loss": 0.9297,
+      "step": 6070
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.0454170107349302e-05,
+      "loss": 0.8686,
+      "step": 6080
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.041288191577209e-05,
+      "loss": 0.9146,
+      "step": 6090
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.037159372419488e-05,
+      "loss": 0.8934,
+      "step": 6100
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.0330305532617675e-05,
+      "loss": 0.8451,
+      "step": 6110
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.0289017341040466e-05,
+      "loss": 0.8825,
+      "step": 6120
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.0247729149463254e-05,
+      "loss": 0.8838,
+      "step": 6130
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.0206440957886045e-05,
+      "loss": 0.7717,
+      "step": 6140
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.0165152766308836e-05,
+      "loss": 0.9197,
+      "step": 6150
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.012386457473163e-05,
+      "loss": 0.934,
+      "step": 6160
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.0082576383154422e-05,
+      "loss": 0.901,
+      "step": 6170
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.004128819157721e-05,
+      "loss": 0.8956,
+      "step": 6180
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3e-05,
+      "loss": 0.9108,
+      "step": 6190
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 2.9958711808422792e-05,
+      "loss": 0.953,
+      "step": 6200
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.9917423616845586e-05,
+      "loss": 0.829,
+      "step": 6210
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.9876135425268374e-05,
+      "loss": 1.0245,
+      "step": 6220
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.9834847233691165e-05,
+      "loss": 0.8651,
+      "step": 6230
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.9793559042113956e-05,
+      "loss": 0.8764,
+      "step": 6240
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.9752270850536744e-05,
+      "loss": 0.8499,
+      "step": 6250
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.9710982658959542e-05,
+      "loss": 0.8117,
+      "step": 6260
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.966969446738233e-05,
+      "loss": 0.8739,
+      "step": 6270
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.962840627580512e-05,
+      "loss": 0.8066,
+      "step": 6280
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.9587118084227912e-05,
+      "loss": 0.8525,
+      "step": 6290
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.9545829892650706e-05,
+      "loss": 0.812,
+      "step": 6300
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.9504541701073494e-05,
+      "loss": 0.8929,
+      "step": 6310
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.9463253509496285e-05,
+      "loss": 0.7825,
+      "step": 6320
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.9421965317919076e-05,
+      "loss": 0.9447,
+      "step": 6330
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 2.9380677126341864e-05,
+      "loss": 0.7871,
+      "step": 6340
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.9339388934764662e-05,
+      "loss": 0.839,
+      "step": 6350
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.929810074318745e-05,
+      "loss": 0.8951,
+      "step": 6360
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 2.925681255161024e-05,
+      "loss": 0.768,
+      "step": 6370
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.9215524360033032e-05,
+      "loss": 0.9465,
+      "step": 6380
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.917423616845582e-05,
+      "loss": 0.9417,
+      "step": 6390
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 2.9132947976878618e-05,
+      "loss": 0.8009,
+      "step": 6400
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.9091659785301405e-05,
+      "loss": 0.8246,
+      "step": 6410
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.9050371593724197e-05,
+      "loss": 0.9381,
+      "step": 6420
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.9009083402146984e-05,
+      "loss": 0.9222,
+      "step": 6430
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 2.8967795210569775e-05,
+      "loss": 0.8203,
+      "step": 6440
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 2.892650701899257e-05,
+      "loss": 0.9185,
+      "step": 6450
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 2.888521882741536e-05,
+      "loss": 0.7349,
+      "step": 6460
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 2.8843930635838152e-05,
+      "loss": 0.9247,
+      "step": 6470
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.880264244426094e-05,
+      "loss": 0.8209,
+      "step": 6480
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.8761354252683738e-05,
+      "loss": 0.8536,
+      "step": 6490
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.8720066061106526e-05,
+      "loss": 0.9474,
+      "step": 6500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 2.8678777869529317e-05,
+      "loss": 0.9213,
+      "step": 6510
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.8637489677952108e-05,
+      "loss": 0.8548,
+      "step": 6520
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.8596201486374896e-05,
+      "loss": 0.8839,
+      "step": 6530
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.855491329479769e-05,
+      "loss": 0.7263,
+      "step": 6540
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.851362510322048e-05,
+      "loss": 0.8929,
+      "step": 6550
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.8472336911643272e-05,
+      "loss": 0.8107,
+      "step": 6560
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.843104872006606e-05,
+      "loss": 0.8435,
+      "step": 6570
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.838976052848885e-05,
+      "loss": 0.8725,
+      "step": 6580
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.8348472336911646e-05,
+      "loss": 0.8464,
+      "step": 6590
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.8307184145334437e-05,
+      "loss": 0.7975,
+      "step": 6600
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.8265895953757228e-05,
+      "loss": 0.8484,
+      "step": 6610
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.8224607762180016e-05,
+      "loss": 0.8886,
+      "step": 6620
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.8183319570602807e-05,
+      "loss": 0.902,
+      "step": 6630
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.81420313790256e-05,
+      "loss": 0.9232,
+      "step": 6640
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.8100743187448392e-05,
+      "loss": 0.8423,
+      "step": 6650
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.805945499587118e-05,
+      "loss": 0.8743,
+      "step": 6660
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.801816680429397e-05,
+      "loss": 0.8104,
+      "step": 6670
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7976878612716766e-05,
+      "loss": 0.8576,
+      "step": 6680
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7935590421139557e-05,
+      "loss": 0.891,
+      "step": 6690
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7894302229562348e-05,
+      "loss": 0.8237,
+      "step": 6700
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7853014037985136e-05,
+      "loss": 0.7509,
+      "step": 6710
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.7811725846407927e-05,
+      "loss": 0.8543,
+      "step": 6720
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.8304788795185198,
+      "eval_f1": 0.6079311715534371,
+      "eval_loss": 0.6249329447746277,
+      "eval_precision": 0.7376051528239177,
+      "eval_recall": 0.5654073825304893,
+      "eval_runtime": 166.509,
+      "eval_samples_per_second": 161.655,
+      "eval_steps_per_second": 10.108,
+      "step": 6729
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.777043765483072e-05,
+      "loss": 0.8504,
+      "step": 6730
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.7729149463253512e-05,
+      "loss": 0.7652,
+      "step": 6740
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.7687861271676304e-05,
+      "loss": 0.6734,
+      "step": 6750
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.764657308009909e-05,
+      "loss": 0.8217,
+      "step": 6760
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.7605284888521882e-05,
+      "loss": 0.7845,
+      "step": 6770
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.7563996696944677e-05,
+      "loss": 0.7187,
+      "step": 6780
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.7522708505367468e-05,
+      "loss": 0.7549,
+      "step": 6790
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.7481420313790256e-05,
+      "loss": 0.7929,
+      "step": 6800
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.7440132122213047e-05,
+      "loss": 0.7692,
+      "step": 6810
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.7398843930635838e-05,
+      "loss": 0.7389,
+      "step": 6820
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.7357555739058633e-05,
+      "loss": 0.8474,
+      "step": 6830
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.7316267547481424e-05,
+      "loss": 0.7811,
+      "step": 6840
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.727497935590421e-05,
+      "loss": 0.7893,
+      "step": 6850
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.7233691164327003e-05,
+      "loss": 0.7149,
+      "step": 6860
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.7192402972749797e-05,
+      "loss": 0.7456,
+      "step": 6870
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.7151114781172588e-05,
+      "loss": 0.7362,
+      "step": 6880
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.7109826589595376e-05,
+      "loss": 0.7188,
+      "step": 6890
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.7068538398018167e-05,
+      "loss": 0.789,
+      "step": 6900
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.7027250206440958e-05,
+      "loss": 0.774,
+      "step": 6910
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.6985962014863753e-05,
+      "loss": 0.9016,
+      "step": 6920
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.6944673823286544e-05,
+      "loss": 0.8115,
+      "step": 6930
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.690338563170933e-05,
+      "loss": 0.6993,
+      "step": 6940
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.6862097440132123e-05,
+      "loss": 0.8238,
+      "step": 6950
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.6820809248554914e-05,
+      "loss": 0.7522,
+      "step": 6960
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.6779521056977708e-05,
+      "loss": 0.7432,
+      "step": 6970
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.6738232865400496e-05,
+      "loss": 0.804,
+      "step": 6980
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.6696944673823287e-05,
+      "loss": 0.7927,
+      "step": 6990
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.6655656482246078e-05,
+      "loss": 0.7865,
+      "step": 7000
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.6614368290668866e-05,
+      "loss": 0.7768,
+      "step": 7010
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.6573080099091664e-05,
+      "loss": 0.7427,
+      "step": 7020
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.653179190751445e-05,
+      "loss": 0.7125,
+      "step": 7030
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.6490503715937243e-05,
+      "loss": 0.7971,
+      "step": 7040
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.6449215524360034e-05,
+      "loss": 0.6742,
+      "step": 7050
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.640792733278283e-05,
+      "loss": 0.6585,
+      "step": 7060
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.636663914120562e-05,
+      "loss": 0.8082,
+      "step": 7070
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.6325350949628407e-05,
+      "loss": 0.7317,
+      "step": 7080
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.62840627580512e-05,
+      "loss": 0.7291,
+      "step": 7090
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.6242774566473986e-05,
+      "loss": 0.6987,
+      "step": 7100
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.6201486374896784e-05,
+      "loss": 0.7836,
+      "step": 7110
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.616019818331957e-05,
+      "loss": 0.7545,
+      "step": 7120
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.6118909991742363e-05,
+      "loss": 0.7255,
+      "step": 7130
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.6077621800165154e-05,
+      "loss": 0.6977,
+      "step": 7140
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.603633360858794e-05,
+      "loss": 0.8611,
+      "step": 7150
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.599504541701074e-05,
+      "loss": 0.7036,
+      "step": 7160
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.5953757225433527e-05,
+      "loss": 0.7756,
+      "step": 7170
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.591246903385632e-05,
+      "loss": 0.783,
+      "step": 7180
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.587118084227911e-05,
+      "loss": 0.8338,
+      "step": 7190
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.5829892650701897e-05,
+      "loss": 0.7271,
+      "step": 7200
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.5788604459124692e-05,
+      "loss": 0.7701,
+      "step": 7210
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.5747316267547483e-05,
+      "loss": 0.7739,
+      "step": 7220
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.5706028075970274e-05,
+      "loss": 0.7502,
+      "step": 7230
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.5664739884393062e-05,
+      "loss": 0.8473,
+      "step": 7240
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.562345169281586e-05,
+      "loss": 0.6975,
+      "step": 7250
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.5582163501238647e-05,
+      "loss": 0.7591,
+      "step": 7260
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.554087530966144e-05,
+      "loss": 0.8999,
+      "step": 7270
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 2.549958711808423e-05,
+      "loss": 0.7156,
+      "step": 7280
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.5458298926507017e-05,
+      "loss": 0.7926,
+      "step": 7290
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.5417010734929815e-05,
+      "loss": 0.7987,
+      "step": 7300
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 2.5375722543352603e-05,
+      "loss": 0.8874,
+      "step": 7310
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.5334434351775394e-05,
+      "loss": 0.7652,
+      "step": 7320
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.5293146160198182e-05,
+      "loss": 0.8559,
+      "step": 7330
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.5251857968620973e-05,
+      "loss": 0.751,
+      "step": 7340
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 2.5210569777043767e-05,
+      "loss": 0.7002,
+      "step": 7350
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.516928158546656e-05,
+      "loss": 0.8741,
+      "step": 7360
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.512799339388935e-05,
+      "loss": 0.7771,
+      "step": 7370
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 2.5086705202312137e-05,
+      "loss": 0.743,
+      "step": 7380
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 2.504541701073493e-05,
+      "loss": 0.7294,
+      "step": 7390
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 2.5004128819157723e-05,
+      "loss": 0.7112,
+      "step": 7400
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 2.4962840627580514e-05,
+      "loss": 0.7345,
+      "step": 7410
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.4921552436003305e-05,
+      "loss": 0.6502,
+      "step": 7420
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.4880264244426096e-05,
+      "loss": 0.7167,
+      "step": 7430
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.4838976052848888e-05,
+      "loss": 0.7226,
+      "step": 7440
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 2.4797687861271675e-05,
+      "loss": 0.7613,
+      "step": 7450
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.475639966969447e-05,
+      "loss": 0.7681,
+      "step": 7460
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.4715111478117258e-05,
+      "loss": 0.7345,
+      "step": 7470
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 2.4673823286540052e-05,
+      "loss": 0.7511,
+      "step": 7480
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.463253509496284e-05,
+      "loss": 0.8039,
+      "step": 7490
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.4591246903385634e-05,
+      "loss": 0.7068,
+      "step": 7500
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.4549958711808425e-05,
+      "loss": 0.7475,
+      "step": 7510
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 2.4508670520231213e-05,
+      "loss": 0.6903,
+      "step": 7520
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.4467382328654008e-05,
+      "loss": 0.8104,
+      "step": 7530
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.4426094137076795e-05,
+      "loss": 0.8372,
+      "step": 7540
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 2.438480594549959e-05,
+      "loss": 0.827,
+      "step": 7550
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.4343517753922378e-05,
+      "loss": 0.7466,
+      "step": 7560
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.4302229562345172e-05,
+      "loss": 0.7456,
+      "step": 7570
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 2.4260941370767963e-05,
+      "loss": 0.818,
+      "step": 7580
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.421965317919075e-05,
+      "loss": 0.7931,
+      "step": 7590
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.4178364987613546e-05,
+      "loss": 0.7411,
+      "step": 7600
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.4137076796036333e-05,
+      "loss": 0.7329,
+      "step": 7610
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 2.4095788604459128e-05,
+      "loss": 0.7954,
+      "step": 7620
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.4054500412881916e-05,
+      "loss": 0.7708,
+      "step": 7630
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.4013212221304707e-05,
+      "loss": 0.881,
+      "step": 7640
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 2.3971924029727498e-05,
+      "loss": 0.6884,
+      "step": 7650
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.393063583815029e-05,
+      "loss": 0.7274,
+      "step": 7660
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.3889347646573083e-05,
+      "loss": 0.7424,
+      "step": 7670
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 2.384805945499587e-05,
+      "loss": 0.7605,
+      "step": 7680
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.3806771263418666e-05,
+      "loss": 0.7498,
+      "step": 7690
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.3765483071841453e-05,
+      "loss": 0.8044,
+      "step": 7700
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.3724194880264244e-05,
+      "loss": 0.6768,
+      "step": 7710
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 2.3682906688687036e-05,
+      "loss": 0.724,
+      "step": 7720
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.3641618497109827e-05,
+      "loss": 0.7697,
+      "step": 7730
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.360033030553262e-05,
+      "loss": 0.7927,
+      "step": 7740
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 2.355904211395541e-05,
+      "loss": 0.744,
+      "step": 7750
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.3517753922378203e-05,
+      "loss": 0.6831,
+      "step": 7760
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.347646573080099e-05,
+      "loss": 0.7111,
+      "step": 7770
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 2.3435177539223782e-05,
+      "loss": 0.8084,
+      "step": 7780
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.3393889347646573e-05,
+      "loss": 0.7344,
+      "step": 7790
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.3352601156069365e-05,
+      "loss": 0.6832,
+      "step": 7800
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.331131296449216e-05,
+      "loss": 0.8081,
+      "step": 7810
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 2.3270024772914947e-05,
+      "loss": 0.6632,
+      "step": 7820
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.3228736581337738e-05,
+      "loss": 0.7909,
+      "step": 7830
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.318744838976053e-05,
+      "loss": 0.7552,
+      "step": 7840
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 2.314616019818332e-05,
+      "loss": 0.7455,
+      "step": 7850
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.310487200660611e-05,
+      "loss": 0.7275,
+      "step": 7860
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.3063583815028902e-05,
+      "loss": 0.7466,
+      "step": 7870
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.3022295623451694e-05,
+      "loss": 0.8004,
+      "step": 7880
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 2.2981007431874485e-05,
+      "loss": 0.7485,
+      "step": 7890
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.2939719240297276e-05,
+      "loss": 0.7212,
+      "step": 7900
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.2898431048720067e-05,
+      "loss": 0.7117,
+      "step": 7910
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 2.2857142857142858e-05,
+      "loss": 0.6614,
+      "step": 7920
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.281585466556565e-05,
+      "loss": 0.7919,
+      "step": 7930
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.277456647398844e-05,
+      "loss": 0.6506,
+      "step": 7940
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 2.273327828241123e-05,
+      "loss": 0.7325,
+      "step": 7950
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.2691990090834023e-05,
+      "loss": 0.7863,
+      "step": 7960
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.2650701899256814e-05,
+      "loss": 0.7035,
+      "step": 7970
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.2609413707679605e-05,
+      "loss": 0.7584,
+      "step": 7980
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 2.2568125516102396e-05,
+      "loss": 0.7518,
+      "step": 7990
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.2526837324525187e-05,
+      "loss": 0.6969,
+      "step": 8000
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.2485549132947978e-05,
+      "loss": 0.7242,
+      "step": 8010
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 2.244426094137077e-05,
+      "loss": 0.7298,
+      "step": 8020
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.240297274979356e-05,
+      "loss": 0.8287,
+      "step": 8030
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.236168455821635e-05,
+      "loss": 0.7656,
+      "step": 8040
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 2.2320396366639143e-05,
+      "loss": 0.7894,
+      "step": 8050
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.2279108175061934e-05,
+      "loss": 0.7504,
+      "step": 8060
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.2237819983484725e-05,
+      "loss": 0.7108,
+      "step": 8070
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.2196531791907516e-05,
+      "loss": 0.6963,
+      "step": 8080
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 2.2155243600330307e-05,
+      "loss": 0.822,
+      "step": 8090
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 2.2113955408753098e-05,
+      "loss": 0.7002,
+      "step": 8100
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 2.207266721717589e-05,
+      "loss": 0.749,
+      "step": 8110
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 2.203137902559868e-05,
+      "loss": 0.801,
+      "step": 8120
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.199009083402147e-05,
+      "loss": 0.782,
+      "step": 8130
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.1948802642444263e-05,
+      "loss": 0.76,
+      "step": 8140
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 2.1907514450867054e-05,
+      "loss": 0.7202,
+      "step": 8150
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.186622625928984e-05,
+      "loss": 0.7081,
+      "step": 8160
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.1824938067712636e-05,
+      "loss": 0.7653,
+      "step": 8170
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.1783649876135427e-05,
+      "loss": 0.8273,
+      "step": 8180
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 2.174236168455822e-05,
+      "loss": 0.7862,
+      "step": 8190
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.170107349298101e-05,
+      "loss": 0.7815,
+      "step": 8200
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.1659785301403797e-05,
+      "loss": 0.7878,
+      "step": 8210
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 2.1618497109826592e-05,
+      "loss": 0.8142,
+      "step": 8220
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.157720891824938e-05,
+      "loss": 0.8047,
+      "step": 8230
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.1535920726672174e-05,
+      "loss": 0.704,
+      "step": 8240
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.1494632535094965e-05,
+      "loss": 0.6791,
+      "step": 8250
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 2.1453344343517756e-05,
+      "loss": 0.8343,
+      "step": 8260
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.1412056151940547e-05,
+      "loss": 0.8026,
+      "step": 8270
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.1370767960363335e-05,
+      "loss": 0.7884,
+      "step": 8280
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 2.132947976878613e-05,
+      "loss": 0.7175,
+      "step": 8290
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 2.1288191577208917e-05,
+      "loss": 0.7147,
+      "step": 8300
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 2.1246903385631712e-05,
+      "loss": 0.6906,
+      "step": 8310
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 2.12056151940545e-05,
+      "loss": 0.8273,
+      "step": 8320
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.1164327002477294e-05,
+      "loss": 0.7687,
+      "step": 8330
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.1123038810900085e-05,
+      "loss": 0.7898,
+      "step": 8340
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.1081750619322873e-05,
+      "loss": 0.7992,
+      "step": 8350
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 2.1040462427745667e-05,
+      "loss": 0.7308,
+      "step": 8360
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.0999174236168455e-05,
+      "loss": 0.7096,
+      "step": 8370
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.095788604459125e-05,
+      "loss": 0.7837,
+      "step": 8380
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 2.0916597853014037e-05,
+      "loss": 0.7196,
+      "step": 8390
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 2.087530966143683e-05,
+      "loss": 0.8125,
+      "step": 8400
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 2.0834021469859623e-05,
+      "loss": 0.7669,
+      "step": 8410
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 2.079273327828241e-05,
+      "loss": 0.6934,
+      "step": 8420
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.0751445086705205e-05,
+      "loss": 0.7889,
+      "step": 8430
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.0710156895127993e-05,
+      "loss": 0.7139,
+      "step": 8440
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.0668868703550787e-05,
+      "loss": 0.6912,
+      "step": 8450
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 2.0627580511973575e-05,
+      "loss": 0.6702,
+      "step": 8460
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.0586292320396366e-05,
+      "loss": 0.7794,
+      "step": 8470
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.0545004128819157e-05,
+      "loss": 0.6949,
+      "step": 8480
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 2.050371593724195e-05,
+      "loss": 0.7626,
+      "step": 8490
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.0462427745664743e-05,
+      "loss": 0.7661,
+      "step": 8500
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.042113955408753e-05,
+      "loss": 0.7504,
+      "step": 8510
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 2.0379851362510325e-05,
+      "loss": 0.6816,
+      "step": 8520
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.0338563170933113e-05,
+      "loss": 0.759,
+      "step": 8530
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.0297274979355904e-05,
+      "loss": 0.7479,
+      "step": 8540
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.0255986787778695e-05,
+      "loss": 0.7234,
+      "step": 8550
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.0214698596201486e-05,
+      "loss": 0.6785,
+      "step": 8560
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.017341040462428e-05,
+      "loss": 0.748,
+      "step": 8570
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.013212221304707e-05,
+      "loss": 0.7633,
+      "step": 8580
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.009083402146986e-05,
+      "loss": 0.7017,
+      "step": 8590
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.004954582989265e-05,
+      "loss": 0.8035,
+      "step": 8600
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.0008257638315442e-05,
+      "loss": 0.7041,
+      "step": 8610
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.9966969446738233e-05,
+      "loss": 0.6099,
+      "step": 8620
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.9925681255161024e-05,
+      "loss": 0.7198,
+      "step": 8630
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.988439306358382e-05,
+      "loss": 0.7127,
+      "step": 8640
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.9843104872006607e-05,
+      "loss": 0.6626,
+      "step": 8650
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.9801816680429398e-05,
+      "loss": 0.7983,
+      "step": 8660
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.976052848885219e-05,
+      "loss": 0.7042,
+      "step": 8670
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.971924029727498e-05,
+      "loss": 0.727,
+      "step": 8680
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.967795210569777e-05,
+      "loss": 0.8152,
+      "step": 8690
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.9636663914120562e-05,
+      "loss": 0.7634,
+      "step": 8700
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.9595375722543353e-05,
+      "loss": 0.7096,
+      "step": 8710
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.9554087530966144e-05,
+      "loss": 0.7092,
+      "step": 8720
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.9512799339388936e-05,
+      "loss": 0.7893,
+      "step": 8730
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.9471511147811727e-05,
+      "loss": 0.7785,
+      "step": 8740
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.9430222956234518e-05,
+      "loss": 0.689,
+      "step": 8750
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.938893476465731e-05,
+      "loss": 0.7484,
+      "step": 8760
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.93476465730801e-05,
+      "loss": 0.6882,
+      "step": 8770
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.930635838150289e-05,
+      "loss": 0.7023,
+      "step": 8780
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.9265070189925682e-05,
+      "loss": 0.7803,
+      "step": 8790
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.9223781998348473e-05,
+      "loss": 0.8714,
+      "step": 8800
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.9182493806771264e-05,
+      "loss": 0.7524,
+      "step": 8810
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.9141205615194056e-05,
+      "loss": 0.7236,
+      "step": 8820
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.9099917423616847e-05,
+      "loss": 0.7112,
+      "step": 8830
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9058629232039638e-05,
+      "loss": 0.6092,
+      "step": 8840
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.901734104046243e-05,
+      "loss": 0.7033,
+      "step": 8850
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.897605284888522e-05,
+      "loss": 0.7103,
+      "step": 8860
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.893476465730801e-05,
+      "loss": 0.682,
+      "step": 8870
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8893476465730802e-05,
+      "loss": 0.882,
+      "step": 8880
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8852188274153593e-05,
+      "loss": 0.6191,
+      "step": 8890
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8810900082576385e-05,
+      "loss": 0.7848,
+      "step": 8900
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8769611890999176e-05,
+      "loss": 0.7408,
+      "step": 8910
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8728323699421967e-05,
+      "loss": 0.6172,
+      "step": 8920
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.8687035507844758e-05,
+      "loss": 0.6989,
+      "step": 8930
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.864574731626755e-05,
+      "loss": 0.7122,
+      "step": 8940
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.860445912469034e-05,
+      "loss": 0.6999,
+      "step": 8950
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.856317093311313e-05,
+      "loss": 0.7758,
+      "step": 8960
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.852188274153592e-05,
+      "loss": 0.8661,
+      "step": 8970
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.8480594549958714e-05,
+      "loss": 0.7911,
+      "step": 8980
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.84393063583815e-05,
+      "loss": 0.6834,
+      "step": 8990
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.8398018166804296e-05,
+      "loss": 0.825,
+      "step": 9000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.8356729975227087e-05,
+      "loss": 0.7495,
+      "step": 9010
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.8315441783649878e-05,
+      "loss": 0.7184,
+      "step": 9020
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.827415359207267e-05,
+      "loss": 0.7649,
+      "step": 9030
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.8232865400495457e-05,
+      "loss": 0.7606,
+      "step": 9040
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.819157720891825e-05,
+      "loss": 0.7655,
+      "step": 9050
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.815028901734104e-05,
+      "loss": 0.6365,
+      "step": 9060
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.8109000825763834e-05,
+      "loss": 0.797,
+      "step": 9070
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.8067712634186625e-05,
+      "loss": 0.6802,
+      "step": 9080
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.8026424442609416e-05,
+      "loss": 0.6901,
+      "step": 9090
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.7985136251032207e-05,
+      "loss": 0.7206,
+      "step": 9100
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7943848059454995e-05,
+      "loss": 0.7308,
+      "step": 9110
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.790255986787779e-05,
+      "loss": 0.6648,
+      "step": 9120
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7861271676300577e-05,
+      "loss": 0.6895,
+      "step": 9130
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.781998348472337e-05,
+      "loss": 0.6852,
+      "step": 9140
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.777869529314616e-05,
+      "loss": 0.7206,
+      "step": 9150
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.773740710156895e-05,
+      "loss": 0.6799,
+      "step": 9160
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7696118909991745e-05,
+      "loss": 0.7507,
+      "step": 9170
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7654830718414533e-05,
+      "loss": 0.7931,
+      "step": 9180
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7613542526837327e-05,
+      "loss": 0.7268,
+      "step": 9190
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7572254335260115e-05,
+      "loss": 0.7366,
+      "step": 9200
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.753096614368291e-05,
+      "loss": 0.6607,
+      "step": 9210
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7489677952105697e-05,
+      "loss": 0.7259,
+      "step": 9220
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7448389760528488e-05,
+      "loss": 0.7273,
+      "step": 9230
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.7407101568951283e-05,
+      "loss": 0.8033,
+      "step": 9240
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.736581337737407e-05,
+      "loss": 0.7193,
+      "step": 9250
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.7324525185796865e-05,
+      "loss": 0.7263,
+      "step": 9260
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7283236994219653e-05,
+      "loss": 0.7458,
+      "step": 9270
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7241948802642447e-05,
+      "loss": 0.7089,
+      "step": 9280
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7200660611065235e-05,
+      "loss": 0.7897,
+      "step": 9290
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7159372419488026e-05,
+      "loss": 0.6724,
+      "step": 9300
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.711808422791082e-05,
+      "loss": 0.6919,
+      "step": 9310
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.707679603633361e-05,
+      "loss": 0.7537,
+      "step": 9320
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.7035507844756403e-05,
+      "loss": 0.827,
+      "step": 9330
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.699421965317919e-05,
+      "loss": 0.7782,
+      "step": 9340
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.695293146160198e-05,
+      "loss": 0.7163,
+      "step": 9350
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.6911643270024773e-05,
+      "loss": 0.7226,
+      "step": 9360
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.6870355078447564e-05,
+      "loss": 0.8097,
+      "step": 9370
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.6829066886870355e-05,
+      "loss": 0.8645,
+      "step": 9380
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.6787778695293146e-05,
+      "loss": 0.7522,
+      "step": 9390
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.674649050371594e-05,
+      "loss": 0.6823,
+      "step": 9400
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.670520231213873e-05,
+      "loss": 0.8494,
+      "step": 9410
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.666391412056152e-05,
+      "loss": 0.6533,
+      "step": 9420
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.662262592898431e-05,
+      "loss": 0.7531,
+      "step": 9430
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6581337737407102e-05,
+      "loss": 0.6884,
+      "step": 9440
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6540049545829893e-05,
+      "loss": 0.7403,
+      "step": 9450
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6498761354252684e-05,
+      "loss": 0.7691,
+      "step": 9460
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.645747316267548e-05,
+      "loss": 0.6903,
+      "step": 9470
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.6416184971098266e-05,
+      "loss": 0.8134,
+      "step": 9480
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.6374896779521057e-05,
+      "loss": 0.7332,
+      "step": 9490
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.633360858794385e-05,
+      "loss": 0.6905,
+      "step": 9500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.629232039636664e-05,
+      "loss": 0.7426,
+      "step": 9510
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.625103220478943e-05,
+      "loss": 0.7637,
+      "step": 9520
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.6209744013212222e-05,
+      "loss": 0.6984,
+      "step": 9530
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6168455821635013e-05,
+      "loss": 0.7183,
+      "step": 9540
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6127167630057804e-05,
+      "loss": 0.6763,
+      "step": 9550
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6085879438480595e-05,
+      "loss": 0.7123,
+      "step": 9560
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6044591246903386e-05,
+      "loss": 0.6449,
+      "step": 9570
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.6003303055326177e-05,
+      "loss": 0.6195,
+      "step": 9580
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.596201486374897e-05,
+      "loss": 0.7698,
+      "step": 9590
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.592072667217176e-05,
+      "loss": 0.7371,
+      "step": 9600
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.587943848059455e-05,
+      "loss": 0.6615,
+      "step": 9610
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.5838150289017342e-05,
+      "loss": 0.7133,
+      "step": 9620
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.5796862097440133e-05,
+      "loss": 0.7568,
+      "step": 9630
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.5755573905862924e-05,
+      "loss": 0.6927,
+      "step": 9640
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.5714285714285715e-05,
+      "loss": 0.7383,
+      "step": 9650
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.5672997522708506e-05,
+      "loss": 0.731,
+      "step": 9660
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.5631709331131298e-05,
+      "loss": 0.7375,
+      "step": 9670
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.559042113955409e-05,
+      "loss": 0.721,
+      "step": 9680
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.554913294797688e-05,
+      "loss": 0.7335,
+      "step": 9690
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.550784475639967e-05,
+      "loss": 0.7472,
+      "step": 9700
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.5466556564822462e-05,
+      "loss": 0.6682,
+      "step": 9710
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.5425268373245253e-05,
+      "loss": 0.7125,
+      "step": 9720
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.538398018166804e-05,
+      "loss": 0.7556,
+      "step": 9730
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.5342691990090835e-05,
+      "loss": 0.7346,
+      "step": 9740
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.5301403798513627e-05,
+      "loss": 0.6173,
+      "step": 9750
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.5260115606936418e-05,
+      "loss": 0.8171,
+      "step": 9760
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.5218827415359207e-05,
+      "loss": 0.752,
+      "step": 9770
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.5177539223782e-05,
+      "loss": 0.6987,
+      "step": 9780
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.513625103220479e-05,
+      "loss": 0.671,
+      "step": 9790
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.509496284062758e-05,
+      "loss": 0.7031,
+      "step": 9800
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.5053674649050373e-05,
+      "loss": 0.7316,
+      "step": 9810
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.5012386457473163e-05,
+      "loss": 0.8888,
+      "step": 9820
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.4971098265895956e-05,
+      "loss": 0.7623,
+      "step": 9830
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.4929810074318745e-05,
+      "loss": 0.6482,
+      "step": 9840
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.4888521882741538e-05,
+      "loss": 0.7015,
+      "step": 9850
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.4847233691164327e-05,
+      "loss": 0.6863,
+      "step": 9860
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.4805945499587118e-05,
+      "loss": 0.6682,
+      "step": 9870
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.4764657308009911e-05,
+      "loss": 0.6453,
+      "step": 9880
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.47233691164327e-05,
+      "loss": 0.6705,
+      "step": 9890
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.4682080924855493e-05,
+      "loss": 0.738,
+      "step": 9900
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.4640792733278283e-05,
+      "loss": 0.6809,
+      "step": 9910
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.4599504541701072e-05,
+      "loss": 0.6705,
+      "step": 9920
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.4558216350123865e-05,
+      "loss": 0.7164,
+      "step": 9930
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.4516928158546656e-05,
+      "loss": 0.7956,
+      "step": 9940
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.4475639966969447e-05,
+      "loss": 0.6928,
+      "step": 9950
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.4434351775392238e-05,
+      "loss": 0.7991,
+      "step": 9960
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.4393063583815031e-05,
+      "loss": 0.6658,
+      "step": 9970
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.435177539223782e-05,
+      "loss": 0.6758,
+      "step": 9980
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.431048720066061e-05,
+      "loss": 0.7344,
+      "step": 9990
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.4269199009083403e-05,
+      "loss": 0.6554,
+      "step": 10000
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.4227910817506192e-05,
+      "loss": 0.6538,
+      "step": 10010
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4186622625928985e-05,
+      "loss": 0.7178,
+      "step": 10020
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4145334434351776e-05,
+      "loss": 0.7525,
+      "step": 10030
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4104046242774569e-05,
+      "loss": 0.7357,
+      "step": 10040
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.4062758051197359e-05,
+      "loss": 0.6798,
+      "step": 10050
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.4021469859620148e-05,
+      "loss": 0.7581,
+      "step": 10060
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.398018166804294e-05,
+      "loss": 0.6251,
+      "step": 10070
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.393889347646573e-05,
+      "loss": 0.7929,
+      "step": 10080
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.3897605284888523e-05,
+      "loss": 0.6975,
+      "step": 10090
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.8520637515324888,
+      "eval_f1": 0.6676348279817839,
+      "eval_loss": 0.5358899235725403,
+      "eval_precision": 0.7613587226106678,
+      "eval_recall": 0.6312791692165124,
+      "eval_runtime": 166.4384,
+      "eval_samples_per_second": 161.723,
+      "eval_steps_per_second": 10.112,
+      "step": 10094
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.3856317093311314e-05,
+      "loss": 0.6937,
+      "step": 10100
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.3815028901734104e-05,
+      "loss": 0.692,
+      "step": 10110
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.3773740710156896e-05,
+      "loss": 0.7053,
+      "step": 10120
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.3732452518579686e-05,
+      "loss": 0.6046,
+      "step": 10130
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.3691164327002479e-05,
+      "loss": 0.6022,
+      "step": 10140
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.3649876135425268e-05,
+      "loss": 0.6027,
+      "step": 10150
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.360858794384806e-05,
+      "loss": 0.6424,
+      "step": 10160
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.3567299752270852e-05,
+      "loss": 0.6271,
+      "step": 10170
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.3526011560693641e-05,
+      "loss": 0.6264,
+      "step": 10180
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.3484723369116434e-05,
+      "loss": 0.6674,
+      "step": 10190
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.3443435177539224e-05,
+      "loss": 0.6816,
+      "step": 10200
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.3402146985962016e-05,
+      "loss": 0.6688,
+      "step": 10210
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.3360858794384806e-05,
+      "loss": 0.6307,
+      "step": 10220
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.3319570602807599e-05,
+      "loss": 0.6293,
+      "step": 10230
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.3278282411230388e-05,
+      "loss": 0.7323,
+      "step": 10240
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.323699421965318e-05,
+      "loss": 0.6245,
+      "step": 10250
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.3195706028075972e-05,
+      "loss": 0.6847,
+      "step": 10260
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.3154417836498761e-05,
+      "loss": 0.6803,
+      "step": 10270
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.3113129644921554e-05,
+      "loss": 0.5765,
+      "step": 10280
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.3071841453344344e-05,
+      "loss": 0.7276,
+      "step": 10290
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.3030553261767133e-05,
+      "loss": 0.6709,
+      "step": 10300
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.2989265070189926e-05,
+      "loss": 0.7282,
+      "step": 10310
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.2947976878612717e-05,
+      "loss": 0.7575,
+      "step": 10320
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.290668868703551e-05,
+      "loss": 0.6801,
+      "step": 10330
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.28654004954583e-05,
+      "loss": 0.6777,
+      "step": 10340
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.2824112303881092e-05,
+      "loss": 0.7004,
+      "step": 10350
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.2782824112303882e-05,
+      "loss": 0.7406,
+      "step": 10360
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.2741535920726671e-05,
+      "loss": 0.6308,
+      "step": 10370
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.2700247729149464e-05,
+      "loss": 0.6491,
+      "step": 10380
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.2658959537572255e-05,
+      "loss": 0.6228,
+      "step": 10390
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.2617671345995046e-05,
+      "loss": 0.7042,
+      "step": 10400
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.2576383154417837e-05,
+      "loss": 0.6618,
+      "step": 10410
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.253509496284063e-05,
+      "loss": 0.6241,
+      "step": 10420
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.249380677126342e-05,
+      "loss": 0.6329,
+      "step": 10430
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.245251857968621e-05,
+      "loss": 0.6857,
+      "step": 10440
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.2411230388109002e-05,
+      "loss": 0.6827,
+      "step": 10450
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.2369942196531791e-05,
+      "loss": 0.5959,
+      "step": 10460
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.2328654004954584e-05,
+      "loss": 0.671,
+      "step": 10470
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.2287365813377375e-05,
+      "loss": 0.6914,
+      "step": 10480
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.2246077621800166e-05,
+      "loss": 0.6575,
+      "step": 10490
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.2204789430222957e-05,
+      "loss": 0.7185,
+      "step": 10500
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.2163501238645748e-05,
+      "loss": 0.6612,
+      "step": 10510
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.212221304706854e-05,
+      "loss": 0.6569,
+      "step": 10520
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.2080924855491329e-05,
+      "loss": 0.6145,
+      "step": 10530
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.203963666391412e-05,
+      "loss": 0.704,
+      "step": 10540
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.1998348472336913e-05,
+      "loss": 0.6824,
+      "step": 10550
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.1957060280759704e-05,
+      "loss": 0.701,
+      "step": 10560
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.1915772089182495e-05,
+      "loss": 0.62,
+      "step": 10570
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.1874483897605286e-05,
+      "loss": 0.6429,
+      "step": 10580
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.1833195706028076e-05,
+      "loss": 0.6262,
+      "step": 10590
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.1791907514450867e-05,
+      "loss": 0.7207,
+      "step": 10600
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.1750619322873658e-05,
+      "loss": 0.6339,
+      "step": 10610
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.1709331131296449e-05,
+      "loss": 0.6222,
+      "step": 10620
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.1668042939719242e-05,
+      "loss": 0.6939,
+      "step": 10630
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.1626754748142033e-05,
+      "loss": 0.7472,
+      "step": 10640
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.1585466556564822e-05,
+      "loss": 0.624,
+      "step": 10650
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.1544178364987614e-05,
+      "loss": 0.6389,
+      "step": 10660
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.1502890173410405e-05,
+      "loss": 0.6456,
+      "step": 10670
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.1461601981833196e-05,
+      "loss": 0.6812,
+      "step": 10680
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.1420313790255987e-05,
+      "loss": 0.7531,
+      "step": 10690
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.1379025598678778e-05,
+      "loss": 0.6745,
+      "step": 10700
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.133773740710157e-05,
+      "loss": 0.6829,
+      "step": 10710
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.129644921552436e-05,
+      "loss": 0.7272,
+      "step": 10720
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.1255161023947151e-05,
+      "loss": 0.7304,
+      "step": 10730
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.1213872832369943e-05,
+      "loss": 0.6772,
+      "step": 10740
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.1172584640792734e-05,
+      "loss": 0.5812,
+      "step": 10750
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.1131296449215525e-05,
+      "loss": 0.7074,
+      "step": 10760
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.1090008257638316e-05,
+      "loss": 0.5895,
+      "step": 10770
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.1048720066061107e-05,
+      "loss": 0.6383,
+      "step": 10780
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.1007431874483898e-05,
+      "loss": 0.7039,
+      "step": 10790
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.096614368290669e-05,
+      "loss": 0.7018,
+      "step": 10800
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.092485549132948e-05,
+      "loss": 0.6026,
+      "step": 10810
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.0883567299752271e-05,
+      "loss": 0.618,
+      "step": 10820
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.0842279108175063e-05,
+      "loss": 0.6584,
+      "step": 10830
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.0800990916597854e-05,
+      "loss": 0.712,
+      "step": 10840
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.0759702725020645e-05,
+      "loss": 0.6934,
+      "step": 10850
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.0718414533443436e-05,
+      "loss": 0.7161,
+      "step": 10860
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.0677126341866227e-05,
+      "loss": 0.688,
+      "step": 10870
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.0635838150289018e-05,
+      "loss": 0.6298,
+      "step": 10880
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.059454995871181e-05,
+      "loss": 0.6871,
+      "step": 10890
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.05532617671346e-05,
+      "loss": 0.786,
+      "step": 10900
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.051197357555739e-05,
+      "loss": 0.6604,
+      "step": 10910
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.0470685383980183e-05,
+      "loss": 0.5722,
+      "step": 10920
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.0429397192402974e-05,
+      "loss": 0.6778,
+      "step": 10930
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.0388109000825765e-05,
+      "loss": 0.6601,
+      "step": 10940
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.0346820809248556e-05,
+      "loss": 0.718,
+      "step": 10950
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.0305532617671347e-05,
+      "loss": 0.7407,
+      "step": 10960
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.0264244426094137e-05,
+      "loss": 0.658,
+      "step": 10970
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.0222956234516928e-05,
+      "loss": 0.5704,
+      "step": 10980
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.0181668042939719e-05,
+      "loss": 0.6517,
+      "step": 10990
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.0140379851362512e-05,
+      "loss": 0.6814,
+      "step": 11000
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.0099091659785303e-05,
+      "loss": 0.7041,
+      "step": 11010
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.0057803468208094e-05,
+      "loss": 0.6566,
+      "step": 11020
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.0016515276630883e-05,
+      "loss": 0.6528,
+      "step": 11030
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 9.975227085053674e-06,
+      "loss": 0.6482,
+      "step": 11040
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 9.933938893476466e-06,
+      "loss": 0.6188,
+      "step": 11050
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.892650701899257e-06,
+      "loss": 0.7015,
+      "step": 11060
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.851362510322048e-06,
+      "loss": 0.6023,
+      "step": 11070
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 9.81007431874484e-06,
+      "loss": 0.6008,
+      "step": 11080
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.768786127167632e-06,
+      "loss": 0.5838,
+      "step": 11090
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.727497935590421e-06,
+      "loss": 0.7076,
+      "step": 11100
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.686209744013212e-06,
+      "loss": 0.6284,
+      "step": 11110
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 9.644921552436003e-06,
+      "loss": 0.6449,
+      "step": 11120
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 9.603633360858795e-06,
+      "loss": 0.5661,
+      "step": 11130
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 9.562345169281586e-06,
+      "loss": 0.7534,
+      "step": 11140
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 9.521056977704377e-06,
+      "loss": 0.6688,
+      "step": 11150
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.479768786127168e-06,
+      "loss": 0.584,
+      "step": 11160
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.438480594549959e-06,
+      "loss": 0.6462,
+      "step": 11170
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 9.39719240297275e-06,
+      "loss": 0.7479,
+      "step": 11180
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.355904211395541e-06,
+      "loss": 0.647,
+      "step": 11190
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.314616019818332e-06,
+      "loss": 0.6724,
+      "step": 11200
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.273327828241124e-06,
+      "loss": 0.7318,
+      "step": 11210
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 9.232039636663915e-06,
+      "loss": 0.637,
+      "step": 11220
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.190751445086706e-06,
+      "loss": 0.6005,
+      "step": 11230
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.149463253509497e-06,
+      "loss": 0.6096,
+      "step": 11240
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 9.108175061932288e-06,
+      "loss": 0.6284,
+      "step": 11250
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 9.066886870355079e-06,
+      "loss": 0.6579,
+      "step": 11260
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 9.02559867877787e-06,
+      "loss": 0.6509,
+      "step": 11270
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 8.984310487200661e-06,
+      "loss": 0.5788,
+      "step": 11280
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 8.94302229562345e-06,
+      "loss": 0.6669,
+      "step": 11290
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 8.901734104046244e-06,
+      "loss": 0.65,
+      "step": 11300
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 8.860445912469035e-06,
+      "loss": 0.6589,
+      "step": 11310
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 8.819157720891826e-06,
+      "loss": 0.5564,
+      "step": 11320
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 8.777869529314617e-06,
+      "loss": 0.6692,
+      "step": 11330
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 8.736581337737408e-06,
+      "loss": 0.6356,
+      "step": 11340
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 8.695293146160198e-06,
+      "loss": 0.6589,
+      "step": 11350
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 8.654004954582989e-06,
+      "loss": 0.6657,
+      "step": 11360
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 8.61271676300578e-06,
+      "loss": 0.6349,
+      "step": 11370
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 8.571428571428573e-06,
+      "loss": 0.6217,
+      "step": 11380
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.530140379851364e-06,
+      "loss": 0.6432,
+      "step": 11390
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.488852188274155e-06,
+      "loss": 0.6369,
+      "step": 11400
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.447563996696944e-06,
+      "loss": 0.6131,
+      "step": 11410
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 8.406275805119735e-06,
+      "loss": 0.67,
+      "step": 11420
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 8.364987613542527e-06,
+      "loss": 0.6781,
+      "step": 11430
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 8.323699421965318e-06,
+      "loss": 0.6087,
+      "step": 11440
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 8.28241123038811e-06,
+      "loss": 0.7239,
+      "step": 11450
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 8.241123038810902e-06,
+      "loss": 0.6677,
+      "step": 11460
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 8.199834847233693e-06,
+      "loss": 0.6151,
+      "step": 11470
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 8.158546655656482e-06,
+      "loss": 0.6126,
+      "step": 11480
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 8.117258464079273e-06,
+      "loss": 0.6492,
+      "step": 11490
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 8.075970272502064e-06,
+      "loss": 0.5863,
+      "step": 11500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 8.034682080924856e-06,
+      "loss": 0.5816,
+      "step": 11510
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 7.993393889347647e-06,
+      "loss": 0.6497,
+      "step": 11520
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 7.95210569777044e-06,
+      "loss": 0.623,
+      "step": 11530
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 7.910817506193229e-06,
+      "loss": 0.6746,
+      "step": 11540
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 7.86952931461602e-06,
+      "loss": 0.6464,
+      "step": 11550
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 7.828241123038811e-06,
+      "loss": 0.6464,
+      "step": 11560
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 7.786952931461602e-06,
+      "loss": 0.7262,
+      "step": 11570
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 7.745664739884393e-06,
+      "loss": 0.7009,
+      "step": 11580
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 7.704376548307184e-06,
+      "loss": 0.6737,
+      "step": 11590
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 7.663088356729976e-06,
+      "loss": 0.6538,
+      "step": 11600
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 7.621800165152766e-06,
+      "loss": 0.6234,
+      "step": 11610
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 7.580511973575558e-06,
+      "loss": 0.584,
+      "step": 11620
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 7.539223781998349e-06,
+      "loss": 0.6766,
+      "step": 11630
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 7.49793559042114e-06,
+      "loss": 0.6068,
+      "step": 11640
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 7.456647398843931e-06,
+      "loss": 0.6294,
+      "step": 11650
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 7.415359207266722e-06,
+      "loss": 0.6252,
+      "step": 11660
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 7.374071015689513e-06,
+      "loss": 0.6031,
+      "step": 11670
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 7.332782824112304e-06,
+      "loss": 0.6785,
+      "step": 11680
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 7.291494632535095e-06,
+      "loss": 0.6607,
+      "step": 11690
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 7.250206440957887e-06,
+      "loss": 0.73,
+      "step": 11700
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 7.208918249380678e-06,
+      "loss": 0.6378,
+      "step": 11710
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 7.167630057803469e-06,
+      "loss": 0.6217,
+      "step": 11720
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 7.126341866226259e-06,
+      "loss": 0.6099,
+      "step": 11730
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 7.0850536746490505e-06,
+      "loss": 0.6309,
+      "step": 11740
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 7.043765483071842e-06,
+      "loss": 0.5813,
+      "step": 11750
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 7.002477291494633e-06,
+      "loss": 0.6552,
+      "step": 11760
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 6.961189099917424e-06,
+      "loss": 0.5721,
+      "step": 11770
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 6.919900908340216e-06,
+      "loss": 0.7204,
+      "step": 11780
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 6.878612716763005e-06,
+      "loss": 0.648,
+      "step": 11790
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 6.837324525185796e-06,
+      "loss": 0.5631,
+      "step": 11800
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 6.796036333608588e-06,
+      "loss": 0.6259,
+      "step": 11810
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 6.7547481420313794e-06,
+      "loss": 0.6154,
+      "step": 11820
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.7134599504541706e-06,
+      "loss": 0.6494,
+      "step": 11830
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.672171758876962e-06,
+      "loss": 0.5928,
+      "step": 11840
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.630883567299754e-06,
+      "loss": 0.6874,
+      "step": 11850
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 6.589595375722543e-06,
+      "loss": 0.6273,
+      "step": 11860
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 6.548307184145334e-06,
+      "loss": 0.6155,
+      "step": 11870
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 6.507018992568126e-06,
+      "loss": 0.5436,
+      "step": 11880
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 6.465730800990917e-06,
+      "loss": 0.5748,
+      "step": 11890
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 6.424442609413708e-06,
+      "loss": 0.6767,
+      "step": 11900
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 6.3831544178364995e-06,
+      "loss": 0.6153,
+      "step": 11910
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 6.34186622625929e-06,
+      "loss": 0.6134,
+      "step": 11920
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 6.300578034682081e-06,
+      "loss": 0.6862,
+      "step": 11930
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 6.259289843104872e-06,
+      "loss": 0.6391,
+      "step": 11940
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 6.218001651527663e-06,
+      "loss": 0.7041,
+      "step": 11950
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 6.176713459950455e-06,
+      "loss": 0.7249,
+      "step": 11960
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 6.135425268373245e-06,
+      "loss": 0.5948,
+      "step": 11970
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 6.0941370767960365e-06,
+      "loss": 0.6403,
+      "step": 11980
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 6.052848885218828e-06,
+      "loss": 0.6428,
+      "step": 11990
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 6.011560693641619e-06,
+      "loss": 0.6733,
+      "step": 12000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 5.97027250206441e-06,
+      "loss": 0.639,
+      "step": 12010
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 5.928984310487201e-06,
+      "loss": 0.6142,
+      "step": 12020
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 5.887696118909992e-06,
+      "loss": 0.6362,
+      "step": 12030
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 5.846407927332783e-06,
+      "loss": 0.5982,
+      "step": 12040
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 5.805119735755574e-06,
+      "loss": 0.6699,
+      "step": 12050
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 5.763831544178365e-06,
+      "loss": 0.7005,
+      "step": 12060
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 5.722543352601157e-06,
+      "loss": 0.6051,
+      "step": 12070
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 5.681255161023948e-06,
+      "loss": 0.6817,
+      "step": 12080
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 5.639966969446738e-06,
+      "loss": 0.6245,
+      "step": 12090
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 5.598678777869529e-06,
+      "loss": 0.6462,
+      "step": 12100
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 5.557390586292321e-06,
+      "loss": 0.6718,
+      "step": 12110
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 5.516102394715112e-06,
+      "loss": 0.7305,
+      "step": 12120
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 5.4748142031379025e-06,
+      "loss": 0.6586,
+      "step": 12130
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 5.433526011560694e-06,
+      "loss": 0.7127,
+      "step": 12140
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 5.392237819983486e-06,
+      "loss": 0.6182,
+      "step": 12150
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 5.350949628406276e-06,
+      "loss": 0.6356,
+      "step": 12160
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 5.309661436829067e-06,
+      "loss": 0.5906,
+      "step": 12170
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 5.268373245251858e-06,
+      "loss": 0.6348,
+      "step": 12180
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 5.227085053674649e-06,
+      "loss": 0.591,
+      "step": 12190
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 5.18579686209744e-06,
+      "loss": 0.5745,
+      "step": 12200
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 5.1445086705202315e-06,
+      "loss": 0.6121,
+      "step": 12210
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 5.103220478943023e-06,
+      "loss": 0.6969,
+      "step": 12220
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 5.061932287365814e-06,
+      "loss": 0.718,
+      "step": 12230
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 5.020644095788605e-06,
+      "loss": 0.7024,
+      "step": 12240
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.979355904211395e-06,
+      "loss": 0.6118,
+      "step": 12250
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 4.938067712634187e-06,
+      "loss": 0.6547,
+      "step": 12260
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.896779521056978e-06,
+      "loss": 0.7086,
+      "step": 12270
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.8554913294797685e-06,
+      "loss": 0.615,
+      "step": 12280
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 4.81420313790256e-06,
+      "loss": 0.6193,
+      "step": 12290
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.772914946325352e-06,
+      "loss": 0.6487,
+      "step": 12300
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.731626754748143e-06,
+      "loss": 0.5845,
+      "step": 12310
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.690338563170933e-06,
+      "loss": 0.669,
+      "step": 12320
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 4.649050371593724e-06,
+      "loss": 0.6517,
+      "step": 12330
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.607762180016516e-06,
+      "loss": 0.6252,
+      "step": 12340
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.566473988439306e-06,
+      "loss": 0.7194,
+      "step": 12350
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 4.5251857968620975e-06,
+      "loss": 0.6455,
+      "step": 12360
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.483897605284889e-06,
+      "loss": 0.5814,
+      "step": 12370
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.44260941370768e-06,
+      "loss": 0.7044,
+      "step": 12380
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 4.401321222130471e-06,
+      "loss": 0.5711,
+      "step": 12390
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.360033030553262e-06,
+      "loss": 0.6398,
+      "step": 12400
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.318744838976053e-06,
+      "loss": 0.6932,
+      "step": 12410
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.277456647398844e-06,
+      "loss": 0.5871,
+      "step": 12420
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 4.236168455821635e-06,
+      "loss": 0.6411,
+      "step": 12430
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.194880264244426e-06,
+      "loss": 0.6153,
+      "step": 12440
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.1535920726672176e-06,
+      "loss": 0.6348,
+      "step": 12450
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 4.112303881090009e-06,
+      "loss": 0.6934,
+      "step": 12460
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.0710156895128e-06,
+      "loss": 0.6181,
+      "step": 12470
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 4.02972749793559e-06,
+      "loss": 0.6036,
+      "step": 12480
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.988439306358382e-06,
+      "loss": 0.6432,
+      "step": 12490
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.947151114781173e-06,
+      "loss": 0.5658,
+      "step": 12500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.9058629232039634e-06,
+      "loss": 0.5939,
+      "step": 12510
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.8645747316267546e-06,
+      "loss": 0.6759,
+      "step": 12520
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.8232865400495465e-06,
+      "loss": 0.5349,
+      "step": 12530
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.781998348472337e-06,
+      "loss": 0.6224,
+      "step": 12540
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.7407101568951284e-06,
+      "loss": 0.6048,
+      "step": 12550
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.6994219653179195e-06,
+      "loss": 0.5859,
+      "step": 12560
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.65813377374071e-06,
+      "loss": 0.6764,
+      "step": 12570
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.6168455821635013e-06,
+      "loss": 0.719,
+      "step": 12580
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.575557390586293e-06,
+      "loss": 0.7163,
+      "step": 12590
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.534269199009083e-06,
+      "loss": 0.6033,
+      "step": 12600
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4929810074318747e-06,
+      "loss": 0.6629,
+      "step": 12610
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4516928158546658e-06,
+      "loss": 0.6724,
+      "step": 12620
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.4104046242774565e-06,
+      "loss": 0.687,
+      "step": 12630
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.3691164327002476e-06,
+      "loss": 0.6423,
+      "step": 12640
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.327828241123039e-06,
+      "loss": 0.6265,
+      "step": 12650
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.2865400495458303e-06,
+      "loss": 0.6865,
+      "step": 12660
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.245251857968621e-06,
+      "loss": 0.6962,
+      "step": 12670
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.203963666391412e-06,
+      "loss": 0.6289,
+      "step": 12680
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.1626754748142036e-06,
+      "loss": 0.6613,
+      "step": 12690
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.1213872832369943e-06,
+      "loss": 0.5976,
+      "step": 12700
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.0800990916597855e-06,
+      "loss": 0.5983,
+      "step": 12710
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.0388109000825766e-06,
+      "loss": 0.5865,
+      "step": 12720
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 2.9975227085053677e-06,
+      "loss": 0.6748,
+      "step": 12730
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.956234516928159e-06,
+      "loss": 0.6116,
+      "step": 12740
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.9149463253509495e-06,
+      "loss": 0.7224,
+      "step": 12750
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 2.873658133773741e-06,
+      "loss": 0.6841,
+      "step": 12760
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.8323699421965318e-06,
+      "loss": 0.637,
+      "step": 12770
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.791081750619323e-06,
+      "loss": 0.6347,
+      "step": 12780
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.749793559042114e-06,
+      "loss": 0.6049,
+      "step": 12790
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 2.708505367464905e-06,
+      "loss": 0.6289,
+      "step": 12800
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.6672171758876962e-06,
+      "loss": 0.655,
+      "step": 12810
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.6259289843104874e-06,
+      "loss": 0.7168,
+      "step": 12820
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 2.5846407927332785e-06,
+      "loss": 0.6727,
+      "step": 12830
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.5433526011560696e-06,
+      "loss": 0.6576,
+      "step": 12840
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.5020644095788607e-06,
+      "loss": 0.647,
+      "step": 12850
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.4607762180016514e-06,
+      "loss": 0.5606,
+      "step": 12860
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 2.419488026424443e-06,
+      "loss": 0.6824,
+      "step": 12870
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.3781998348472337e-06,
+      "loss": 0.6617,
+      "step": 12880
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.3369116432700252e-06,
+      "loss": 0.6564,
+      "step": 12890
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 2.295623451692816e-06,
+      "loss": 0.6077,
+      "step": 12900
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.254335260115607e-06,
+      "loss": 0.6115,
+      "step": 12910
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.213047068538398e-06,
+      "loss": 0.7471,
+      "step": 12920
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 2.1717588769611893e-06,
+      "loss": 0.6362,
+      "step": 12930
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.13047068538398e-06,
+      "loss": 0.6719,
+      "step": 12940
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.0891824938067715e-06,
+      "loss": 0.6797,
+      "step": 12950
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.0478943022295622e-06,
+      "loss": 0.7031,
+      "step": 12960
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 2.0066061106523533e-06,
+      "loss": 0.676,
+      "step": 12970
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.9653179190751445e-06,
+      "loss": 0.5952,
+      "step": 12980
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.9240297274979356e-06,
+      "loss": 0.6634,
+      "step": 12990
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.882741535920727e-06,
+      "loss": 0.6021,
+      "step": 13000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.8414533443435178e-06,
+      "loss": 0.6647,
+      "step": 13010
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.8001651527663087e-06,
+      "loss": 0.7219,
+      "step": 13020
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.7588769611891e-06,
+      "loss": 0.5921,
+      "step": 13030
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.717588769611891e-06,
+      "loss": 0.5946,
+      "step": 13040
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.6763005780346821e-06,
+      "loss": 0.6199,
+      "step": 13050
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.6350123864574732e-06,
+      "loss": 0.6554,
+      "step": 13060
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.5937241948802644e-06,
+      "loss": 0.6812,
+      "step": 13070
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.5524360033030555e-06,
+      "loss": 0.6977,
+      "step": 13080
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.5111478117258466e-06,
+      "loss": 0.6438,
+      "step": 13090
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.4698596201486377e-06,
+      "loss": 0.6111,
+      "step": 13100
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.4285714285714286e-06,
+      "loss": 0.7156,
+      "step": 13110
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.3872832369942197e-06,
+      "loss": 0.6126,
+      "step": 13120
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.3459950454170109e-06,
+      "loss": 0.6866,
+      "step": 13130
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.3047068538398018e-06,
+      "loss": 0.6229,
+      "step": 13140
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.263418662262593e-06,
+      "loss": 0.6501,
+      "step": 13150
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.222130470685384e-06,
+      "loss": 0.6975,
+      "step": 13160
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.1808422791081751e-06,
+      "loss": 0.6253,
+      "step": 13170
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.139554087530966e-06,
+      "loss": 0.6167,
+      "step": 13180
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.0982658959537572e-06,
+      "loss": 0.7457,
+      "step": 13190
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.0569777043765483e-06,
+      "loss": 0.6742,
+      "step": 13200
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 1.0156895127993394e-06,
+      "loss": 0.6219,
+      "step": 13210
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 9.744013212221305e-07,
+      "loss": 0.6522,
+      "step": 13220
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 9.331131296449216e-07,
+      "loss": 0.6879,
+      "step": 13230
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 8.918249380677127e-07,
+      "loss": 0.6406,
+      "step": 13240
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 8.505367464905038e-07,
+      "loss": 0.7326,
+      "step": 13250
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 8.092485549132948e-07,
+      "loss": 0.6493,
+      "step": 13260
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 7.679603633360859e-07,
+      "loss": 0.7029,
+      "step": 13270
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 7.266721717588771e-07,
+      "loss": 0.6415,
+      "step": 13280
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 6.853839801816681e-07,
+      "loss": 0.558,
+      "step": 13290
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 6.440957886044592e-07,
+      "loss": 0.6842,
+      "step": 13300
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 6.028075970272502e-07,
+      "loss": 0.609,
+      "step": 13310
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 5.615194054500413e-07,
+      "loss": 0.698,
+      "step": 13320
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 5.202312138728324e-07,
+      "loss": 0.7018,
+      "step": 13330
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 4.789430222956235e-07,
+      "loss": 0.5928,
+      "step": 13340
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 4.3765483071841454e-07,
+      "loss": 0.6973,
+      "step": 13350
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.963666391412056e-07,
+      "loss": 0.5841,
+      "step": 13360
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.5507844756399673e-07,
+      "loss": 0.6611,
+      "step": 13370
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.137902559867878e-07,
+      "loss": 0.7326,
+      "step": 13380
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.7250206440957887e-07,
+      "loss": 0.6679,
+      "step": 13390
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 2.3121387283236997e-07,
+      "loss": 0.5928,
+      "step": 13400
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.8992568125516104e-07,
+      "loss": 0.6604,
+      "step": 13410
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.486374896779521e-07,
+      "loss": 0.6024,
+      "step": 13420
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.0734929810074319e-07,
+      "loss": 0.5838,
+      "step": 13430
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 6.606110652353427e-08,
+      "loss": 0.5949,
+      "step": 13440
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.477291494632535e-08,
+      "loss": 0.6493,
+      "step": 13450
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8621688895493554,
+      "eval_f1": 0.6977873555253621,
+      "eval_loss": 0.49737581610679626,
+      "eval_precision": 0.7789543746735181,
+      "eval_recall": 0.6581134959383467,
+      "eval_runtime": 165.8382,
+      "eval_samples_per_second": 162.309,
+      "eval_steps_per_second": 10.148,
+      "step": 13456
+    },
+    {
+      "epoch": 4.0,
+      "step": 13456,
+      "total_flos": 6.8284774551813e+19,
+      "train_loss": 1.060067500640039,
+      "train_runtime": 12677.4822,
+      "train_samples_per_second": 67.943,
+      "train_steps_per_second": 1.061
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 13456,
+  "num_train_epochs": 4,
+  "save_steps": 500,
+  "total_flos": 6.8284774551813e+19,
+  "trial_name": null,
+  "trial_params": null
+}