diff --git "a/checkpoint-1042/trainer_state.json" "b/checkpoint-1042/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-1042/trainer_state.json"
@@ -0,0 +1,6297 @@
+{
+  "best_metric": 0.5160893201828003,
+  "best_model_checkpoint": "./Einstein-v4-model/checkpoint-521",
+  "epoch": 1.0009606147934678,
+  "eval_steps": 521,
+  "global_step": 1042,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 1.1358,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "eval_loss": 1.2074627876281738,
+      "eval_runtime": 2270.2512,
+      "eval_samples_per_second": 1.197,
+      "eval_steps_per_second": 0.15,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 1.0958,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.5e-06,
+      "loss": 1.1535,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 1.0637,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.5e-06,
+      "loss": 0.9402,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3e-06,
+      "loss": 0.8635,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.5e-06,
+      "loss": 0.8675,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.7557,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.5e-06,
+      "loss": 0.791,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5e-06,
+      "loss": 0.8049,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999979246067429e-06,
+      "loss": 0.794,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999916984614294e-06,
+      "loss": 0.7392,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999813216674332e-06,
+      "loss": 0.7995,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.999667943970417e-06,
+      "loss": 0.7819,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9994811689145335e-06,
+      "loss": 0.7631,
+      "step": 15
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9992528946077344e-06,
+      "loss": 0.7677,
+      "step": 16
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9989831248400925e-06,
+      "loss": 0.742,
+      "step": 17
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.998671864090633e-06,
+      "loss": 0.7047,
+      "step": 18
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9983191175272635e-06,
+      "loss": 0.692,
+      "step": 19
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.997924891006689e-06,
+      "loss": 0.7237,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.997489191074308e-06,
+      "loss": 0.7048,
+      "step": 21
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.997012024964111e-06,
+      "loss": 0.7142,
+      "step": 22
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.996493400598555e-06,
+      "loss": 0.6877,
+      "step": 23
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995933326588439e-06,
+      "loss": 0.6681,
+      "step": 24
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995331812232752e-06,
+      "loss": 0.6409,
+      "step": 25
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994688867518523e-06,
+      "loss": 0.6709,
+      "step": 26
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9940045031206606e-06,
+      "loss": 0.655,
+      "step": 27
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.993278730401764e-06,
+      "loss": 0.6735,
+      "step": 28
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992511561411944e-06,
+      "loss": 0.7202,
+      "step": 29
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991703008888622e-06,
+      "loss": 0.7061,
+      "step": 30
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.99085308625631e-06,
+      "loss": 0.6744,
+      "step": 31
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9899618076264e-06,
+      "loss": 0.673,
+      "step": 32
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.98902918779692e-06,
+      "loss": 0.6492,
+      "step": 33
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.988055242252294e-06,
+      "loss": 0.6587,
+      "step": 34
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.987039987163081e-06,
+      "loss": 0.6684,
+      "step": 35
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.985983439385712e-06,
+      "loss": 0.6316,
+      "step": 36
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.984885616462202e-06,
+      "loss": 0.6634,
+      "step": 37
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.983746536619864e-06,
+      "loss": 0.6475,
+      "step": 38
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.982566218771011e-06,
+      "loss": 0.603,
+      "step": 39
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9813446825126296e-06,
+      "loss": 0.687,
+      "step": 40
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.980081948126066e-06,
+      "loss": 0.6423,
+      "step": 41
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.978778036576682e-06,
+      "loss": 0.6142,
+      "step": 42
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.977432969513514e-06,
+      "loss": 0.6496,
+      "step": 43
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.976046769268905e-06,
+      "loss": 0.6452,
+      "step": 44
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.974619458858142e-06,
+      "loss": 0.66,
+      "step": 45
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.973151061979065e-06,
+      "loss": 0.6364,
+      "step": 46
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.971641603011685e-06,
+      "loss": 0.5856,
+      "step": 47
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.970091107017768e-06,
+      "loss": 0.6662,
+      "step": 48
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.968499599740427e-06,
+      "loss": 0.6411,
+      "step": 49
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.966867107603688e-06,
+      "loss": 0.604,
+      "step": 50
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.965193657712057e-06,
+      "loss": 0.6326,
+      "step": 51
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.963479277850067e-06,
+      "loss": 0.7502,
+      "step": 52
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.961723996481817e-06,
+      "loss": 0.6184,
+      "step": 53
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.959927842750501e-06,
+      "loss": 0.667,
+      "step": 54
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.958090846477921e-06,
+      "loss": 0.6381,
+      "step": 55
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.956213038163995e-06,
+      "loss": 0.6413,
+      "step": 56
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.954294448986247e-06,
+      "loss": 0.6346,
+      "step": 57
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.952335110799295e-06,
+      "loss": 0.6128,
+      "step": 58
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9503350561343165e-06,
+      "loss": 0.6128,
+      "step": 59
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.948294318198511e-06,
+      "loss": 0.6328,
+      "step": 60
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.94621293087455e-06,
+      "loss": 0.6238,
+      "step": 61
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.944090928720009e-06,
+      "loss": 0.6206,
+      "step": 62
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.941928346966801e-06,
+      "loss": 0.6261,
+      "step": 63
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.939725221520586e-06,
+      "loss": 0.6082,
+      "step": 64
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.937481588960179e-06,
+      "loss": 0.6499,
+      "step": 65
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.935197486536937e-06,
+      "loss": 0.6235,
+      "step": 66
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.932872952174148e-06,
+      "loss": 0.6149,
+      "step": 67
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.930508024466395e-06,
+      "loss": 0.5835,
+      "step": 68
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.928102742678918e-06,
+      "loss": 0.6067,
+      "step": 69
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.925657146746961e-06,
+      "loss": 0.6046,
+      "step": 70
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9231712772751125e-06,
+      "loss": 0.5974,
+      "step": 71
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.920645175536624e-06,
+      "loss": 0.5931,
+      "step": 72
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9180788834727335e-06,
+      "loss": 0.5771,
+      "step": 73
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.915472443691962e-06,
+      "loss": 0.5965,
+      "step": 74
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.91282589946941e-06,
+      "loss": 0.5977,
+      "step": 75
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.910139294746038e-06,
+      "loss": 0.5848,
+      "step": 76
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.907412674127937e-06,
+      "loss": 0.5911,
+      "step": 77
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.904646082885586e-06,
+      "loss": 0.6085,
+      "step": 78
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.901839566953105e-06,
+      "loss": 0.5948,
+      "step": 79
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.898993172927488e-06,
+      "loss": 0.6027,
+      "step": 80
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.896106948067829e-06,
+      "loss": 0.5953,
+      "step": 81
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.893180940294542e-06,
+      "loss": 0.6232,
+      "step": 82
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.890215198188562e-06,
+      "loss": 0.5706,
+      "step": 83
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.887209770990537e-06,
+      "loss": 0.6225,
+      "step": 84
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.8841647086000145e-06,
+      "loss": 0.5969,
+      "step": 85
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.88108006157461e-06,
+      "loss": 0.5914,
+      "step": 86
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.877955881129169e-06,
+      "loss": 0.6087,
+      "step": 87
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.874792219134916e-06,
+      "loss": 0.618,
+      "step": 88
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.871589128118593e-06,
+      "loss": 0.6502,
+      "step": 89
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.868346661261586e-06,
+      "loss": 0.648,
+      "step": 90
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.865064872399048e-06,
+      "loss": 0.6616,
+      "step": 91
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.861743816018998e-06,
+      "loss": 0.5862,
+      "step": 92
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.85838354726142e-06,
+      "loss": 0.5644,
+      "step": 93
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.854984121917349e-06,
+      "loss": 0.5814,
+      "step": 94
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.851545596427938e-06,
+      "loss": 0.6027,
+      "step": 95
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.848068027883528e-06,
+      "loss": 0.5579,
+      "step": 96
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.844551474022699e-06,
+      "loss": 0.5995,
+      "step": 97
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.840995993231308e-06,
+      "loss": 0.5699,
+      "step": 98
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.83740164454152e-06,
+      "loss": 0.598,
+      "step": 99
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.833768487630834e-06,
+      "loss": 0.5892,
+      "step": 100
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.8300965828210835e-06,
+      "loss": 0.6006,
+      "step": 101
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.826385991077439e-06,
+      "loss": 0.616,
+      "step": 102
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.8226367740074e-06,
+      "loss": 0.5551,
+      "step": 103
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.8188489938597635e-06,
+      "loss": 0.6221,
+      "step": 104
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.815022713523597e-06,
+      "loss": 0.6124,
+      "step": 105
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.811157996527192e-06,
+      "loss": 0.6061,
+      "step": 106
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.807254907037007e-06,
+      "loss": 0.5755,
+      "step": 107
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.80331350985661e-06,
+      "loss": 0.6288,
+      "step": 108
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.799333870425594e-06,
+      "loss": 0.5737,
+      "step": 109
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.795316054818489e-06,
+      "loss": 0.6151,
+      "step": 110
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.79126012974368e-06,
+      "loss": 0.6037,
+      "step": 111
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.787166162542282e-06,
+      "loss": 0.6152,
+      "step": 112
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.783034221187028e-06,
+      "loss": 0.5692,
+      "step": 113
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.778864374281147e-06,
+      "loss": 0.5708,
+      "step": 114
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.774656691057214e-06,
+      "loss": 0.5656,
+      "step": 115
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.770411241376009e-06,
+      "loss": 0.6693,
+      "step": 116
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.766128095725352e-06,
+      "loss": 0.5772,
+      "step": 117
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.761807325218938e-06,
+      "loss": 0.5934,
+      "step": 118
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.757449001595149e-06,
+      "loss": 0.5912,
+      "step": 119
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.753053197215869e-06,
+      "loss": 0.5947,
+      "step": 120
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.7486199850652805e-06,
+      "loss": 0.599,
+      "step": 121
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.7441494387486535e-06,
+      "loss": 0.5785,
+      "step": 122
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.739641632491119e-06,
+      "loss": 0.602,
+      "step": 123
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.735096641136443e-06,
+      "loss": 0.5875,
+      "step": 124
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.730514540145783e-06,
+      "loss": 0.6053,
+      "step": 125
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.72589540559643e-06,
+      "loss": 0.5657,
+      "step": 126
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.721239314180549e-06,
+      "loss": 0.6151,
+      "step": 127
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.716546343203905e-06,
+      "loss": 0.5785,
+      "step": 128
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.711816570584583e-06,
+      "loss": 0.5716,
+      "step": 129
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.707050074851687e-06,
+      "loss": 0.5734,
+      "step": 130
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.7022469351440415e-06,
+      "loss": 0.5958,
+      "step": 131
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.697407231208878e-06,
+      "loss": 0.6023,
+      "step": 132
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.6925310434005064e-06,
+      "loss": 0.5966,
+      "step": 133
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.687618452678986e-06,
+      "loss": 0.5544,
+      "step": 134
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.682669540608778e-06,
+      "loss": 0.5701,
+      "step": 135
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.677684389357392e-06,
+      "loss": 0.5916,
+      "step": 136
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.672663081694022e-06,
+      "loss": 0.574,
+      "step": 137
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.667605700988174e-06,
+      "loss": 0.564,
+      "step": 138
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.662512331208277e-06,
+      "loss": 0.5684,
+      "step": 139
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.657383056920294e-06,
+      "loss": 0.587,
+      "step": 140
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.652217963286314e-06,
+      "loss": 0.6185,
+      "step": 141
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.647017136063143e-06,
+      "loss": 0.5602,
+      "step": 142
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.641780661600875e-06,
+      "loss": 0.5533,
+      "step": 143
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.636508626841458e-06,
+      "loss": 0.5889,
+      "step": 144
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.631201119317257e-06,
+      "loss": 0.5452,
+      "step": 145
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.625858227149595e-06,
+      "loss": 0.5664,
+      "step": 146
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.620480039047289e-06,
+      "loss": 0.5576,
+      "step": 147
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.615066644305183e-06,
+      "loss": 0.5901,
+      "step": 148
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.609618132802661e-06,
+      "loss": 0.5725,
+      "step": 149
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.604134595002154e-06,
+      "loss": 0.5779,
+      "step": 150
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.5986161219476425e-06,
+      "loss": 0.5737,
+      "step": 151
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.593062805263139e-06,
+      "loss": 0.5708,
+      "step": 152
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.5874747371511715e-06,
+      "loss": 0.5544,
+      "step": 153
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.581852010391253e-06,
+      "loss": 0.6194,
+      "step": 154
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.576194718338336e-06,
+      "loss": 0.5404,
+      "step": 155
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.570502954921266e-06,
+      "loss": 0.5863,
+      "step": 156
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.564776814641224e-06,
+      "loss": 0.5693,
+      "step": 157
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.559016392570152e-06,
+      "loss": 0.5793,
+      "step": 158
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.55322178434918e-06,
+      "loss": 0.5803,
+      "step": 159
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.547393086187033e-06,
+      "loss": 0.5699,
+      "step": 160
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.54153039485844e-06,
+      "loss": 0.5994,
+      "step": 161
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.5356338077025196e-06,
+      "loss": 0.5918,
+      "step": 162
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.5297034226211715e-06,
+      "loss": 0.5734,
+      "step": 163
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.523739338077443e-06,
+      "loss": 0.5965,
+      "step": 164
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.517741653093903e-06,
+      "loss": 0.5589,
+      "step": 165
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.511710467250991e-06,
+      "loss": 0.6116,
+      "step": 166
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.505645880685364e-06,
+      "loss": 0.5949,
+      "step": 167
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.499547994088242e-06,
+      "loss": 0.5775,
+      "step": 168
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.493416908703725e-06,
+      "loss": 0.5646,
+      "step": 169
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.48725272632712e-06,
+      "loss": 0.5606,
+      "step": 170
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.481055549303246e-06,
+      "loss": 0.5288,
+      "step": 171
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.4748254805247395e-06,
+      "loss": 0.5753,
+      "step": 172
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.4685626234303405e-06,
+      "loss": 0.5769,
+      "step": 173
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.462267082003183e-06,
+      "loss": 0.5468,
+      "step": 174
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.455938960769058e-06,
+      "loss": 0.5777,
+      "step": 175
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.4495783647946884e-06,
+      "loss": 0.5201,
+      "step": 176
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.443185399685978e-06,
+      "loss": 0.5659,
+      "step": 177
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.43676017158626e-06,
+      "loss": 0.5662,
+      "step": 178
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.430302787174535e-06,
+      "loss": 0.5522,
+      "step": 179
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.423813353663699e-06,
+      "loss": 0.5534,
+      "step": 180
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.417291978798765e-06,
+      "loss": 0.5715,
+      "step": 181
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.410738770855072e-06,
+      "loss": 0.5668,
+      "step": 182
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.404153838636489e-06,
+      "loss": 0.5986,
+      "step": 183
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.397537291473607e-06,
+      "loss": 0.5499,
+      "step": 184
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.390889239221926e-06,
+      "loss": 0.5658,
+      "step": 185
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.384209792260027e-06,
+      "loss": 0.562,
+      "step": 186
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.377499061487744e-06,
+      "loss": 0.5828,
+      "step": 187
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.370757158324322e-06,
+      "loss": 0.5731,
+      "step": 188
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.36398419470656e-06,
+      "loss": 0.5909,
+      "step": 189
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.357180283086966e-06,
+      "loss": 0.6445,
+      "step": 190
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.350345536431877e-06,
+      "loss": 0.6368,
+      "step": 191
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.343480068219589e-06,
+      "loss": 0.539,
+      "step": 192
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.336583992438475e-06,
+      "loss": 0.5724,
+      "step": 193
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.329657423585088e-06,
+      "loss": 0.5695,
+      "step": 194
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.32270047666226e-06,
+      "loss": 0.5642,
+      "step": 195
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.315713267177202e-06,
+      "loss": 0.5857,
+      "step": 196
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.308695911139569e-06,
+      "loss": 0.5642,
+      "step": 197
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.30164852505955e-06,
+      "loss": 0.5665,
+      "step": 198
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.2945712259459245e-06,
+      "loss": 0.5597,
+      "step": 199
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.287464131304125e-06,
+      "loss": 0.578,
+      "step": 200
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.280327359134281e-06,
+      "loss": 0.5987,
+      "step": 201
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.273161027929263e-06,
+      "loss": 0.5545,
+      "step": 202
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.265965256672716e-06,
+      "loss": 0.5727,
+      "step": 203
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.258740164837079e-06,
+      "loss": 0.5834,
+      "step": 204
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.251485872381608e-06,
+      "loss": 0.5859,
+      "step": 205
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.24420249975038e-06,
+      "loss": 0.5723,
+      "step": 206
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.2368901678702955e-06,
+      "loss": 0.5931,
+      "step": 207
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.2295489981490685e-06,
+      "loss": 0.5525,
+      "step": 208
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.222179112473211e-06,
+      "loss": 0.6104,
+      "step": 209
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.214780633206012e-06,
+      "loss": 0.5479,
+      "step": 210
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.207353683185503e-06,
+      "loss": 0.5445,
+      "step": 211
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.1998983857224215e-06,
+      "loss": 0.5784,
+      "step": 212
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.192414864598159e-06,
+      "loss": 0.581,
+      "step": 213
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.184903244062709e-06,
+      "loss": 0.5496,
+      "step": 214
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.177363648832605e-06,
+      "loss": 0.5548,
+      "step": 215
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.169796204088848e-06,
+      "loss": 0.5679,
+      "step": 216
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.162201035474829e-06,
+      "loss": 0.5759,
+      "step": 217
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.1545782690942416e-06,
+      "loss": 0.5702,
+      "step": 218
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.146928031508988e-06,
+      "loss": 0.5638,
+      "step": 219
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.139250449737081e-06,
+      "loss": 0.5371,
+      "step": 220
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.131545651250533e-06,
+      "loss": 0.5881,
+      "step": 221
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.123813763973237e-06,
+      "loss": 0.585,
+      "step": 222
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.116054916278848e-06,
+      "loss": 0.5776,
+      "step": 223
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.108269236988648e-06,
+      "loss": 0.5562,
+      "step": 224
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.100456855369406e-06,
+      "loss": 0.5479,
+      "step": 225
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.092617901131235e-06,
+      "loss": 0.5743,
+      "step": 226
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.084752504425439e-06,
+      "loss": 0.5396,
+      "step": 227
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.0768607958423455e-06,
+      "loss": 0.538,
+      "step": 228
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.068942906409148e-06,
+      "loss": 0.5909,
+      "step": 229
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.060998967587719e-06,
+      "loss": 0.5328,
+      "step": 230
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.053029111272435e-06,
+      "loss": 0.5569,
+      "step": 231
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.045033469787986e-06,
+      "loss": 0.5563,
+      "step": 232
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.037012175887174e-06,
+      "loss": 0.5309,
+      "step": 233
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.028965362748714e-06,
+      "loss": 0.5628,
+      "step": 234
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.0208931639750186e-06,
+      "loss": 0.5642,
+      "step": 235
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.012795713589984e-06,
+      "loss": 0.5642,
+      "step": 236
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.004673146036763e-06,
+      "loss": 0.586,
+      "step": 237
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.996525596175528e-06,
+      "loss": 0.6073,
+      "step": 238
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.988353199281242e-06,
+      "loss": 0.5875,
+      "step": 239
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.980156091041403e-06,
+      "loss": 0.5603,
+      "step": 240
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.971934407553797e-06,
+      "loss": 0.555,
+      "step": 241
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.9636882853242345e-06,
+      "loss": 0.5496,
+      "step": 242
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.955417861264289e-06,
+      "loss": 0.5365,
+      "step": 243
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 3.9471232726890185e-06,
+      "loss": 0.5795,
+      "step": 244
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.938804657314687e-06,
+      "loss": 0.5308,
+      "step": 245
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.9304621532564826e-06,
+      "loss": 0.5554,
+      "step": 246
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.922095899026219e-06,
+      "loss": 0.5288,
+      "step": 247
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.913706033530035e-06,
+      "loss": 0.5488,
+      "step": 248
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.905292696066095e-06,
+      "loss": 0.5531,
+      "step": 249
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.8968560263222675e-06,
+      "loss": 0.5709,
+      "step": 250
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.888396164373814e-06,
+      "loss": 0.5326,
+      "step": 251
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.879913250681059e-06,
+      "loss": 0.5464,
+      "step": 252
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.871407426087056e-06,
+      "loss": 0.5683,
+      "step": 253
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.862878831815253e-06,
+      "loss": 0.56,
+      "step": 254
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 3.8543276094671465e-06,
+      "loss": 0.5478,
+      "step": 255
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.845753901019931e-06,
+      "loss": 0.5644,
+      "step": 256
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.837157848824138e-06,
+      "loss": 0.5674,
+      "step": 257
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.82853959560128e-06,
+      "loss": 0.5495,
+      "step": 258
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.8198992844414736e-06,
+      "loss": 0.5718,
+      "step": 259
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.811237058801066e-06,
+      "loss": 0.5756,
+      "step": 260
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.802553062500256e-06,
+      "loss": 0.596,
+      "step": 261
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.7938474397207008e-06,
+      "loss": 0.5694,
+      "step": 262
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.7851203350031287e-06,
+      "loss": 0.558,
+      "step": 263
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.7763718932449322e-06,
+      "loss": 0.5275,
+      "step": 264
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 3.767602259697769e-06,
+      "loss": 0.5815,
+      "step": 265
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.7588115799651446e-06,
+      "loss": 0.5631,
+      "step": 266
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.7500000000000005e-06,
+      "loss": 0.5783,
+      "step": 267
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.741167666102284e-06,
+      "loss": 0.5381,
+      "step": 268
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.7323147249165255e-06,
+      "loss": 0.5974,
+      "step": 269
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.7234413234294e-06,
+      "loss": 0.549,
+      "step": 270
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.714547608967289e-06,
+      "loss": 0.5112,
+      "step": 271
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.7056337291938325e-06,
+      "loss": 0.5385,
+      "step": 272
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.6966998321074787e-06,
+      "loss": 0.5602,
+      "step": 273
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.6877460660390257e-06,
+      "loss": 0.54,
+      "step": 274
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 3.6787725796491593e-06,
+      "loss": 0.5598,
+      "step": 275
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6697795219259833e-06,
+      "loss": 0.5516,
+      "step": 276
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6607670421825507e-06,
+      "loss": 0.5524,
+      "step": 277
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.651735290054378e-06,
+      "loss": 0.5307,
+      "step": 278
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6426844154969654e-06,
+      "loss": 0.5568,
+      "step": 279
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6336145687833036e-06,
+      "loss": 0.5191,
+      "step": 280
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6245259005013843e-06,
+      "loss": 0.5517,
+      "step": 281
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6154185615516925e-06,
+      "loss": 0.5934,
+      "step": 282
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.6062927031447092e-06,
+      "loss": 0.5259,
+      "step": 283
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.597148476798393e-06,
+      "loss": 0.5782,
+      "step": 284
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.58798603433567e-06,
+      "loss": 0.5355,
+      "step": 285
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 3.57880552788191e-06,
+      "loss": 0.5316,
+      "step": 286
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.569607109862403e-06,
+      "loss": 0.5567,
+      "step": 287
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5603909329998275e-06,
+      "loss": 0.5309,
+      "step": 288
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5511571503117125e-06,
+      "loss": 0.5168,
+      "step": 289
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5419059151079026e-06,
+      "loss": 0.5368,
+      "step": 290
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5326373809880054e-06,
+      "loss": 0.5543,
+      "step": 291
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5233517018388473e-06,
+      "loss": 0.5354,
+      "step": 292
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.5140490318319153e-06,
+      "loss": 0.5535,
+      "step": 293
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.504729525420798e-06,
+      "loss": 0.57,
+      "step": 294
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.495393337338623e-06,
+      "loss": 0.5433,
+      "step": 295
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 3.4860406225954827e-06,
+      "loss": 0.5714,
+      "step": 296
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.4766715364758667e-06,
+      "loss": 0.5677,
+      "step": 297
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.4672862345360813e-06,
+      "loss": 0.552,
+      "step": 298
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.4578848726016646e-06,
+      "loss": 0.5066,
+      "step": 299
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.448467606764801e-06,
+      "loss": 0.5707,
+      "step": 300
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.439034593381733e-06,
+      "loss": 0.5426,
+      "step": 301
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.429585989070156e-06,
+      "loss": 0.5614,
+      "step": 302
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.4201219507066308e-06,
+      "loss": 0.5298,
+      "step": 303
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.410642635423967e-06,
+      "loss": 0.5611,
+      "step": 304
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.4011482006086214e-06,
+      "loss": 0.563,
+      "step": 305
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.3916388038980808e-06,
+      "loss": 0.5327,
+      "step": 306
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 3.382114603178249e-06,
+      "loss": 0.5681,
+      "step": 307
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.372575756580822e-06,
+      "loss": 0.5636,
+      "step": 308
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3630224224806616e-06,
+      "loss": 0.5329,
+      "step": 309
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3534547594931706e-06,
+      "loss": 0.5527,
+      "step": 310
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3438729264716535e-06,
+      "loss": 0.5577,
+      "step": 311
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3342770825046843e-06,
+      "loss": 0.5652,
+      "step": 312
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.324667386913462e-06,
+      "loss": 0.5612,
+      "step": 313
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3150439992491667e-06,
+      "loss": 0.5464,
+      "step": 314
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.3054070792903075e-06,
+      "loss": 0.5579,
+      "step": 315
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.295756787040076e-06,
+      "loss": 0.5784,
+      "step": 316
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 3.2860932827236822e-06,
+      "loss": 0.5562,
+      "step": 317
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.276416726785701e-06,
+      "loss": 0.5538,
+      "step": 318
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.2667272798874038e-06,
+      "loss": 0.5614,
+      "step": 319
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.257025102904092e-06,
+      "loss": 0.5297,
+      "step": 320
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.247310356922428e-06,
+      "loss": 0.5761,
+      "step": 321
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.2375832032377574e-06,
+      "loss": 0.5362,
+      "step": 322
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.2278438033514344e-06,
+      "loss": 0.5557,
+      "step": 323
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.2180923189681367e-06,
+      "loss": 0.5296,
+      "step": 324
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.208328911993185e-06,
+      "loss": 0.5449,
+      "step": 325
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.1985537445298497e-06,
+      "loss": 0.5743,
+      "step": 326
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 3.188766978876666e-06,
+      "loss": 0.5716,
+      "step": 327
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.1789687775247324e-06,
+      "loss": 0.5602,
+      "step": 328
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.169159303155017e-06,
+      "loss": 0.5533,
+      "step": 329
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.159338718635655e-06,
+      "loss": 0.5332,
+      "step": 330
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.1495071870192467e-06,
+      "loss": 0.5561,
+      "step": 331
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.1396648715401473e-06,
+      "loss": 0.5468,
+      "step": 332
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.129811935611757e-06,
+      "loss": 0.5361,
+      "step": 333
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.119948542823812e-06,
+      "loss": 0.552,
+      "step": 334
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.1100748569396614e-06,
+      "loss": 0.5647,
+      "step": 335
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.1001910418935562e-06,
+      "loss": 0.526,
+      "step": 336
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.090297261787919e-06,
+      "loss": 0.5432,
+      "step": 337
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 3.0803936808906277e-06,
+      "loss": 0.5759,
+      "step": 338
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.0704804636322818e-06,
+      "loss": 0.5221,
+      "step": 339
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.060557774603475e-06,
+      "loss": 0.5495,
+      "step": 340
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.050625778552063e-06,
+      "loss": 0.5076,
+      "step": 341
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.040684640380426e-06,
+      "loss": 0.5427,
+      "step": 342
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.0307345251427346e-06,
+      "loss": 0.5375,
+      "step": 343
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.0207755980422036e-06,
+      "loss": 0.5297,
+      "step": 344
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.010808024428356e-06,
+      "loss": 0.5755,
+      "step": 345
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 3.0008319697942712e-06,
+      "loss": 0.5255,
+      "step": 346
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.990847599773842e-06,
+      "loss": 0.5435,
+      "step": 347
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 2.980855080139021e-06,
+      "loss": 0.537,
+      "step": 348
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.970854576797073e-06,
+      "loss": 0.488,
+      "step": 349
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.9608462557878138e-06,
+      "loss": 0.5415,
+      "step": 350
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.9508302832808605e-06,
+      "loss": 0.5688,
+      "step": 351
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.940806825572866e-06,
+      "loss": 0.5425,
+      "step": 352
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.9307760490847636e-06,
+      "loss": 0.5372,
+      "step": 353
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.920738120359e-06,
+      "loss": 0.5119,
+      "step": 354
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.910693206056773e-06,
+      "loss": 0.5137,
+      "step": 355
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.90064147295526e-06,
+      "loss": 0.5451,
+      "step": 356
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8905830879448556e-06,
+      "loss": 0.5403,
+      "step": 357
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.880518218026394e-06,
+      "loss": 0.5258,
+      "step": 358
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 2.8704470303083808e-06,
+      "loss": 0.5453,
+      "step": 359
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8603696920042177e-06,
+      "loss": 0.5467,
+      "step": 360
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8502863704294236e-06,
+      "loss": 0.5318,
+      "step": 361
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8401972329988586e-06,
+      "loss": 0.5125,
+      "step": 362
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8301024472239465e-06,
+      "loss": 0.5532,
+      "step": 363
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.820002180709888e-06,
+      "loss": 0.5107,
+      "step": 364
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.8098966011528855e-06,
+      "loss": 0.5413,
+      "step": 365
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.7997858763373502e-06,
+      "loss": 0.5339,
+      "step": 366
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.7896701741331254e-06,
+      "loss": 0.5082,
+      "step": 367
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.7795496624926904e-06,
+      "loss": 0.5314,
+      "step": 368
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 2.769424509448379e-06,
+      "loss": 0.5678,
+      "step": 369
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.7592948831095863e-06,
+      "loss": 0.5524,
+      "step": 370
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.749160951659976e-06,
+      "loss": 0.5327,
+      "step": 371
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.7390228833546943e-06,
+      "loss": 0.5425,
+      "step": 372
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.7288808465175693e-06,
+      "loss": 0.5158,
+      "step": 373
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.7187350095383197e-06,
+      "loss": 0.5377,
+      "step": 374
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.7085855408697587e-06,
+      "loss": 0.5496,
+      "step": 375
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.6984326090249975e-06,
+      "loss": 0.5697,
+      "step": 376
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.6882763825746456e-06,
+      "loss": 0.5428,
+      "step": 377
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.6781170301440146e-06,
+      "loss": 0.5449,
+      "step": 378
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 2.6679547204103174e-06,
+      "loss": 0.5328,
+      "step": 379
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.6577896220998656e-06,
+      "loss": 0.5621,
+      "step": 380
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.647621903985272e-06,
+      "loss": 0.5381,
+      "step": 381
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.637451734882645e-06,
+      "loss": 0.5123,
+      "step": 382
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.627279283648788e-06,
+      "loss": 0.5305,
+      "step": 383
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.617104719178395e-06,
+      "loss": 0.5159,
+      "step": 384
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.606928210401245e-06,
+      "loss": 0.511,
+      "step": 385
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.5967499262794e-06,
+      "loss": 0.5218,
+      "step": 386
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.586570035804398e-06,
+      "loss": 0.5346,
+      "step": 387
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.576388707994447e-06,
+      "loss": 0.5061,
+      "step": 388
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.566206111891621e-06,
+      "loss": 0.5154,
+      "step": 389
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 2.5560224165590487e-06,
+      "loss": 0.5026,
+      "step": 390
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.545837791078112e-06,
+      "loss": 0.5114,
+      "step": 391
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.5356524045456348e-06,
+      "loss": 0.5823,
+      "step": 392
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.5254664260710777e-06,
+      "loss": 0.5521,
+      "step": 393
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.5152800247737293e-06,
+      "loss": 0.5173,
+      "step": 394
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.5050933697798975e-06,
+      "loss": 0.511,
+      "step": 395
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.4949066302201034e-06,
+      "loss": 0.5199,
+      "step": 396
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.4847199752262716e-06,
+      "loss": 0.5348,
+      "step": 397
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.474533573928923e-06,
+      "loss": 0.5209,
+      "step": 398
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.4643475954543657e-06,
+      "loss": 0.5404,
+      "step": 399
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 2.4541622089218893e-06,
+      "loss": 0.5664,
+      "step": 400
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.443977583440952e-06,
+      "loss": 0.5479,
+      "step": 401
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.43379388810838e-06,
+      "loss": 0.5462,
+      "step": 402
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.4236112920055534e-06,
+      "loss": 0.526,
+      "step": 403
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.413429964195603e-06,
+      "loss": 0.5471,
+      "step": 404
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.4032500737206014e-06,
+      "loss": 0.5213,
+      "step": 405
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.3930717895987562e-06,
+      "loss": 0.527,
+      "step": 406
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.3828952808216066e-06,
+      "loss": 0.5265,
+      "step": 407
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.3727207163512124e-06,
+      "loss": 0.5456,
+      "step": 408
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.362548265117355e-06,
+      "loss": 0.5279,
+      "step": 409
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.352378096014729e-06,
+      "loss": 0.5191,
+      "step": 410
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 2.3422103779001348e-06,
+      "loss": 0.5328,
+      "step": 411
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.3320452795896835e-06,
+      "loss": 0.4958,
+      "step": 412
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.3218829698559858e-06,
+      "loss": 0.5177,
+      "step": 413
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.3117236174253556e-06,
+      "loss": 0.5316,
+      "step": 414
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.3015673909750033e-06,
+      "loss": 0.551,
+      "step": 415
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.291414459130242e-06,
+      "loss": 0.5283,
+      "step": 416
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.281264990461681e-06,
+      "loss": 0.5527,
+      "step": 417
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.2711191534824315e-06,
+      "loss": 0.5226,
+      "step": 418
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.260977116645306e-06,
+      "loss": 0.5203,
+      "step": 419
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.2508390483400243e-06,
+      "loss": 0.5162,
+      "step": 420
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 2.240705116890415e-06,
+      "loss": 0.5277,
+      "step": 421
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.230575490551622e-06,
+      "loss": 0.5606,
+      "step": 422
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.2204503375073104e-06,
+      "loss": 0.4736,
+      "step": 423
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.2103298258668755e-06,
+      "loss": 0.5201,
+      "step": 424
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.20021412366265e-06,
+      "loss": 0.5639,
+      "step": 425
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.190103398847116e-06,
+      "loss": 0.5213,
+      "step": 426
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.179997819290113e-06,
+      "loss": 0.5327,
+      "step": 427
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.169897552776055e-06,
+      "loss": 0.523,
+      "step": 428
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.1598027670011427e-06,
+      "loss": 0.5363,
+      "step": 429
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.1497136295705776e-06,
+      "loss": 0.5076,
+      "step": 430
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.1396303079957835e-06,
+      "loss": 0.5405,
+      "step": 431
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 2.129552969691619e-06,
+      "loss": 0.5309,
+      "step": 432
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.119481781973606e-06,
+      "loss": 0.5369,
+      "step": 433
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.109416912055145e-06,
+      "loss": 0.5288,
+      "step": 434
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.09935852704474e-06,
+      "loss": 0.5454,
+      "step": 435
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.0893067939432275e-06,
+      "loss": 0.5314,
+      "step": 436
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.079261879641e-06,
+      "loss": 0.5423,
+      "step": 437
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.0692239509152368e-06,
+      "loss": 0.5085,
+      "step": 438
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.0591931744271343e-06,
+      "loss": 0.5283,
+      "step": 439
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.0491697167191403e-06,
+      "loss": 0.5111,
+      "step": 440
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.039153744212187e-06,
+      "loss": 0.5141,
+      "step": 441
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 2.029145423202928e-06,
+      "loss": 0.5005,
+      "step": 442
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.0191449198609796e-06,
+      "loss": 0.5045,
+      "step": 443
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 2.009152400226159e-06,
+      "loss": 0.5042,
+      "step": 444
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9991680302057296e-06,
+      "loss": 0.5325,
+      "step": 445
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9891919755716445e-06,
+      "loss": 0.5445,
+      "step": 446
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9792244019577973e-06,
+      "loss": 0.5313,
+      "step": 447
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9692654748572666e-06,
+      "loss": 0.551,
+      "step": 448
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.959315359619575e-06,
+      "loss": 0.5173,
+      "step": 449
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9493742214479384e-06,
+      "loss": 0.5518,
+      "step": 450
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9394422253965264e-06,
+      "loss": 0.5191,
+      "step": 451
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9295195363677195e-06,
+      "loss": 0.5343,
+      "step": 452
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9196063191093735e-06,
+      "loss": 0.5016,
+      "step": 453
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.909702738212082e-06,
+      "loss": 0.5533,
+      "step": 454
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8998089581064452e-06,
+      "loss": 0.5127,
+      "step": 455
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8899251430603385e-06,
+      "loss": 0.5506,
+      "step": 456
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8800514571761884e-06,
+      "loss": 0.5351,
+      "step": 457
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.870188064388243e-06,
+      "loss": 0.5102,
+      "step": 458
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.860335128459853e-06,
+      "loss": 0.5573,
+      "step": 459
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8504928129807537e-06,
+      "loss": 0.5432,
+      "step": 460
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8406612813643454e-06,
+      "loss": 0.5123,
+      "step": 461
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8308406968449837e-06,
+      "loss": 0.5576,
+      "step": 462
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8210312224752685e-06,
+      "loss": 0.5362,
+      "step": 463
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8112330211233345e-06,
+      "loss": 0.5051,
+      "step": 464
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8014462554701507e-06,
+      "loss": 0.5286,
+      "step": 465
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7916710880068162e-06,
+      "loss": 0.4955,
+      "step": 466
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7819076810318642e-06,
+      "loss": 0.5324,
+      "step": 467
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7721561966485664e-06,
+      "loss": 0.5224,
+      "step": 468
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7624167967622435e-06,
+      "loss": 0.5062,
+      "step": 469
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7526896430775725e-06,
+      "loss": 0.4976,
+      "step": 470
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7429748970959087e-06,
+      "loss": 0.496,
+      "step": 471
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7332727201125969e-06,
+      "loss": 0.5042,
+      "step": 472
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.7235832732142996e-06,
+      "loss": 0.4975,
+      "step": 473
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7139067172763182e-06,
+      "loss": 0.5525,
+      "step": 474
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.704243212959925e-06,
+      "loss": 0.5205,
+      "step": 475
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6945929207096937e-06,
+      "loss": 0.516,
+      "step": 476
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6849560007508348e-06,
+      "loss": 0.5112,
+      "step": 477
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.675332613086539e-06,
+      "loss": 0.5353,
+      "step": 478
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6657229174953165e-06,
+      "loss": 0.5069,
+      "step": 479
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.656127073528347e-06,
+      "loss": 0.5324,
+      "step": 480
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6465452405068305e-06,
+      "loss": 0.505,
+      "step": 481
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6369775775193386e-06,
+      "loss": 0.5413,
+      "step": 482
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6274242434191784e-06,
+      "loss": 0.5807,
+      "step": 483
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.6178853968217507e-06,
+      "loss": 0.5182,
+      "step": 484
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.6083611961019196e-06,
+      "loss": 0.4951,
+      "step": 485
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5988517993913794e-06,
+      "loss": 0.4969,
+      "step": 486
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5893573645760338e-06,
+      "loss": 0.5629,
+      "step": 487
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5798780492933696e-06,
+      "loss": 0.5023,
+      "step": 488
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5704140109298446e-06,
+      "loss": 0.5399,
+      "step": 489
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5609654066182682e-06,
+      "loss": 0.5239,
+      "step": 490
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5515323932351994e-06,
+      "loss": 0.5202,
+      "step": 491
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5421151273983358e-06,
+      "loss": 0.5458,
+      "step": 492
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5327137654639191e-06,
+      "loss": 0.5515,
+      "step": 493
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.5233284635241335e-06,
+      "loss": 0.5276,
+      "step": 494
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.5139593774045181e-06,
+      "loss": 0.5749,
+      "step": 495
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.5046066626613781e-06,
+      "loss": 0.5362,
+      "step": 496
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.4952704745792023e-06,
+      "loss": 0.5065,
+      "step": 497
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.485950968168086e-06,
+      "loss": 0.5313,
+      "step": 498
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.4766482981611538e-06,
+      "loss": 0.5272,
+      "step": 499
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.467362619011996e-06,
+      "loss": 0.5218,
+      "step": 500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.4580940848920985e-06,
+      "loss": 0.5059,
+      "step": 501
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.4488428496882883e-06,
+      "loss": 0.5005,
+      "step": 502
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.439609067000174e-06,
+      "loss": 0.5414,
+      "step": 503
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.4303928901375975e-06,
+      "loss": 0.4932,
+      "step": 504
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.4211944721180898e-06,
+      "loss": 0.5351,
+      "step": 505
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.4120139656643302e-06,
+      "loss": 0.5173,
+      "step": 506
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.4028515232016074e-06,
+      "loss": 0.5026,
+      "step": 507
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3937072968552914e-06,
+      "loss": 0.5421,
+      "step": 508
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3845814384483069e-06,
+      "loss": 0.5268,
+      "step": 509
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3754740994986163e-06,
+      "loss": 0.5461,
+      "step": 510
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3663854312166968e-06,
+      "loss": 0.5338,
+      "step": 511
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.357315584503036e-06,
+      "loss": 0.5478,
+      "step": 512
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.348264709945623e-06,
+      "loss": 0.5054,
+      "step": 513
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3392329578174495e-06,
+      "loss": 0.5017,
+      "step": 514
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.3302204780740169e-06,
+      "loss": 0.508,
+      "step": 515
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.3212274203508418e-06,
+      "loss": 0.5052,
+      "step": 516
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.3122539339609752e-06,
+      "loss": 0.4828,
+      "step": 517
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.3033001678925213e-06,
+      "loss": 0.5166,
+      "step": 518
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.2943662708061677e-06,
+      "loss": 0.5361,
+      "step": 519
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.2854523910327118e-06,
+      "loss": 0.5026,
+      "step": 520
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.2765586765706012e-06,
+      "loss": 0.5083,
+      "step": 521
+    },
+    {
+      "epoch": 0.5,
+      "eval_loss": 0.5160893201828003,
+      "eval_runtime": 2285.1688,
+      "eval_samples_per_second": 1.189,
+      "eval_steps_per_second": 0.149,
+      "step": 521
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.773300405821908e-06,
+      "loss": 0.5012,
+      "step": 522
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.768942802160411e-06,
+      "loss": 0.4993,
+      "step": 523
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.764579999039293e-06,
+      "loss": 0.5284,
+      "step": 524
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 3.760212014335024e-06,
+      "loss": 0.5014,
+      "step": 525
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7558388659453052e-06,
+      "loss": 0.5269,
+      "step": 526
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7514605717889974e-06,
+      "loss": 0.5435,
+      "step": 527
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7470771498060455e-06,
+      "loss": 0.5282,
+      "step": 528
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.742688617957405e-06,
+      "loss": 0.5227,
+      "step": 529
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7382949942249695e-06,
+      "loss": 0.4924,
+      "step": 530
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7338962966114956e-06,
+      "loss": 0.482,
+      "step": 531
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7294925431405306e-06,
+      "loss": 0.5172,
+      "step": 532
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7250837518563395e-06,
+      "loss": 0.527,
+      "step": 533
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.720669940823827e-06,
+      "loss": 0.5124,
+      "step": 534
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.716251128128466e-06,
+      "loss": 0.5255,
+      "step": 535
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 3.7118273318762275e-06,
+      "loss": 0.5645,
+      "step": 536
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.7073985701934983e-06,
+      "loss": 0.5431,
+      "step": 537
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.702964861227013e-06,
+      "loss": 0.5056,
+      "step": 538
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6985262231437764e-06,
+      "loss": 0.4891,
+      "step": 539
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.694082674130991e-06,
+      "loss": 0.5201,
+      "step": 540
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6896342323959814e-06,
+      "loss": 0.5286,
+      "step": 541
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6851809161661206e-06,
+      "loss": 0.5592,
+      "step": 542
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6807227436887536e-06,
+      "loss": 0.5125,
+      "step": 543
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6762597332311254e-06,
+      "loss": 0.5396,
+      "step": 544
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.671791903080303e-06,
+      "loss": 0.5476,
+      "step": 545
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 3.6673192715431016e-06,
+      "loss": 0.5423,
+      "step": 546
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.6628418569460122e-06,
+      "loss": 0.5085,
+      "step": 547
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.658359677635122e-06,
+      "loss": 0.5261,
+      "step": 548
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.653872751976043e-06,
+      "loss": 0.5195,
+      "step": 549
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.649381098353834e-06,
+      "loss": 0.5068,
+      "step": 550
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.6448847351729267e-06,
+      "loss": 0.5144,
+      "step": 551
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.6403836808570512e-06,
+      "loss": 0.5333,
+      "step": 552
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.6358779538491596e-06,
+      "loss": 0.5682,
+      "step": 553
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.631367572611348e-06,
+      "loss": 0.5531,
+      "step": 554
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.626852555624784e-06,
+      "loss": 0.5449,
+      "step": 555
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 3.6223329213896313e-06,
+      "loss": 0.5141,
+      "step": 556
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.6178086884249718e-06,
+      "loss": 0.5556,
+      "step": 557
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.613279875268731e-06,
+      "loss": 0.5268,
+      "step": 558
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.6087465004776008e-06,
+      "loss": 0.5416,
+      "step": 559
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.604208582626964e-06,
+      "loss": 0.5211,
+      "step": 560
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.5996661403108196e-06,
+      "loss": 0.5488,
+      "step": 561
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.5951191921417063e-06,
+      "loss": 0.502,
+      "step": 562
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.5905677567506227e-06,
+      "loss": 0.5176,
+      "step": 563
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.586011852786955e-06,
+      "loss": 0.5228,
+      "step": 564
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.5814514989184e-06,
+      "loss": 0.5118,
+      "step": 565
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.5768867138308872e-06,
+      "loss": 0.5171,
+      "step": 566
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 3.572317516228503e-06,
+      "loss": 0.5054,
+      "step": 567
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.5677439248334133e-06,
+      "loss": 0.5044,
+      "step": 568
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.563165958385789e-06,
+      "loss": 0.5083,
+      "step": 569
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.5585836356437266e-06,
+      "loss": 0.5344,
+      "step": 570
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.553996975383171e-06,
+      "loss": 0.5288,
+      "step": 571
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.5494059963978433e-06,
+      "loss": 0.5396,
+      "step": 572
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.544810717499158e-06,
+      "loss": 0.5564,
+      "step": 573
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.540211157516149e-06,
+      "loss": 0.5463,
+      "step": 574
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.5356073352953923e-06,
+      "loss": 0.54,
+      "step": 575
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.530999269700927e-06,
+      "loss": 0.4862,
+      "step": 576
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 3.5263869796141816e-06,
+      "loss": 0.5483,
+      "step": 577
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.521770483933891e-06,
+      "loss": 0.52,
+      "step": 578
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.517149801576026e-06,
+      "loss": 0.5496,
+      "step": 579
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.5125249514737093e-06,
+      "loss": 0.5217,
+      "step": 580
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.507895952577142e-06,
+      "loss": 0.5353,
+      "step": 581
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.503262823853527e-06,
+      "loss": 0.5181,
+      "step": 582
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.498625584286985e-06,
+      "loss": 0.4897,
+      "step": 583
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.493984252878483e-06,
+      "loss": 0.5392,
+      "step": 584
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.4893388486457534e-06,
+      "loss": 0.5448,
+      "step": 585
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.484689390623218e-06,
+      "loss": 0.5197,
+      "step": 586
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.4800358978619088e-06,
+      "loss": 0.5738,
+      "step": 587
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 3.4753783894293886e-06,
+      "loss": 0.5279,
+      "step": 588
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.470716884409675e-06,
+      "loss": 0.5063,
+      "step": 589
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.466051401903162e-06,
+      "loss": 0.5091,
+      "step": 590
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.461381961026542e-06,
+      "loss": 0.5596,
+      "step": 591
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.4567085809127247e-06,
+      "loss": 0.503,
+      "step": 592
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.452031280710762e-06,
+      "loss": 0.5231,
+      "step": 593
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.4473500795857674e-06,
+      "loss": 0.5416,
+      "step": 594
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.4426649967188385e-06,
+      "loss": 0.536,
+      "step": 595
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.4379760513069804e-06,
+      "loss": 0.5205,
+      "step": 596
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.433283262563022e-06,
+      "loss": 0.5162,
+      "step": 597
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 3.428586649715542e-06,
+      "loss": 0.5375,
+      "step": 598
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.423886232008785e-06,
+      "loss": 0.5171,
+      "step": 599
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.4191820287025916e-06,
+      "loss": 0.7379,
+      "step": 600
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.414474059072309e-06,
+      "loss": 0.5583,
+      "step": 601
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.4097623424087196e-06,
+      "loss": 0.5854,
+      "step": 602
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.4050468980179563e-06,
+      "loss": 0.5315,
+      "step": 603
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.4003277452214284e-06,
+      "loss": 0.5565,
+      "step": 604
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.3956049033557405e-06,
+      "loss": 0.5577,
+      "step": 605
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.3908783917726123e-06,
+      "loss": 0.575,
+      "step": 606
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.3861482298388006e-06,
+      "loss": 0.5496,
+      "step": 607
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 3.381414436936018e-06,
+      "loss": 0.5375,
+      "step": 608
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.3766770324608566e-06,
+      "loss": 0.5742,
+      "step": 609
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.3719360358247054e-06,
+      "loss": 0.6354,
+      "step": 610
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.367191466453674e-06,
+      "loss": 0.5275,
+      "step": 611
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.36244334378851e-06,
+      "loss": 0.55,
+      "step": 612
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.35769168728452e-06,
+      "loss": 0.5031,
+      "step": 613
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.3529365164114903e-06,
+      "loss": 0.5476,
+      "step": 614
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.34817785065361e-06,
+      "loss": 0.5319,
+      "step": 615
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.3434157095093846e-06,
+      "loss": 0.5428,
+      "step": 616
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.3386501124915626e-06,
+      "loss": 0.522,
+      "step": 617
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.333881079127052e-06,
+      "loss": 0.5892,
+      "step": 618
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 3.32910862895684e-06,
+      "loss": 0.572,
+      "step": 619
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.3243327815359168e-06,
+      "loss": 0.5633,
+      "step": 620
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.31955355643319e-06,
+      "loss": 0.5534,
+      "step": 621
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.314770973231408e-06,
+      "loss": 0.5528,
+      "step": 622
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.3099850515270804e-06,
+      "loss": 0.5198,
+      "step": 623
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.305195810930393e-06,
+      "loss": 0.524,
+      "step": 624
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.3004032710651346e-06,
+      "loss": 0.4849,
+      "step": 625
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.2956074515686105e-06,
+      "loss": 0.5266,
+      "step": 626
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.2908083720915636e-06,
+      "loss": 0.5003,
+      "step": 627
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.2860060522980945e-06,
+      "loss": 0.5299,
+      "step": 628
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.2812005118655837e-06,
+      "loss": 0.5339,
+      "step": 629
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.276391770484606e-06,
+      "loss": 0.5629,
+      "step": 630
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.271579847858851e-06,
+      "loss": 0.5071,
+      "step": 631
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.266764763705046e-06,
+      "loss": 0.5384,
+      "step": 632
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.26194653775287e-06,
+      "loss": 0.5422,
+      "step": 633
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.257125189744877e-06,
+      "loss": 0.53,
+      "step": 634
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.252300739436413e-06,
+      "loss": 0.5652,
+      "step": 635
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.247473206595536e-06,
+      "loss": 0.5423,
+      "step": 636
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.2426426110029347e-06,
+      "loss": 0.4923,
+      "step": 637
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.2378089724518464e-06,
+      "loss": 0.5195,
+      "step": 638
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.2329723107479766e-06,
+      "loss": 0.5019,
+      "step": 639
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.228132645709421e-06,
+      "loss": 0.5423,
+      "step": 640
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.2232899971665772e-06,
+      "loss": 0.5267,
+      "step": 641
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.218444384962071e-06,
+      "loss": 0.5098,
+      "step": 642
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.2135958289506697e-06,
+      "loss": 0.5343,
+      "step": 643
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.2087443489992043e-06,
+      "loss": 0.5408,
+      "step": 644
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.203889964986485e-06,
+      "loss": 0.5612,
+      "step": 645
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.1990326968032225e-06,
+      "loss": 0.5771,
+      "step": 646
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.1941725643519454e-06,
+      "loss": 0.5386,
+      "step": 647
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.189309587546917e-06,
+      "loss": 0.5382,
+      "step": 648
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.184443786314057e-06,
+      "loss": 0.5249,
+      "step": 649
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.1795751805908578e-06,
+      "loss": 0.5296,
+      "step": 650
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1747037903263023e-06,
+      "loss": 0.5535,
+      "step": 651
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.169829635480783e-06,
+      "loss": 0.5082,
+      "step": 652
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1649527360260207e-06,
+      "loss": 0.5104,
+      "step": 653
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.160073111944983e-06,
+      "loss": 0.5178,
+      "step": 654
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1551907832318003e-06,
+      "loss": 0.5098,
+      "step": 655
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.150305769891686e-06,
+      "loss": 0.527,
+      "step": 656
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.145418091940852e-06,
+      "loss": 0.5085,
+      "step": 657
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1405277694064306e-06,
+      "loss": 0.4871,
+      "step": 658
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.1356348223263897e-06,
+      "loss": 0.5192,
+      "step": 659
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.13073927074945e-06,
+      "loss": 0.5225,
+      "step": 660
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.125841134735005e-06,
+      "loss": 0.5145,
+      "step": 661
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.1209404343530374e-06,
+      "loss": 0.5393,
+      "step": 662
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.1160371896840376e-06,
+      "loss": 0.5251,
+      "step": 663
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.111131420818922e-06,
+      "loss": 0.5195,
+      "step": 664
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.1062231478589475e-06,
+      "loss": 0.5149,
+      "step": 665
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.1013123909156347e-06,
+      "loss": 0.4907,
+      "step": 666
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.0963991701106787e-06,
+      "loss": 0.5353,
+      "step": 667
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.091483505575873e-06,
+      "loss": 0.4889,
+      "step": 668
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.0865654174530233e-06,
+      "loss": 0.5521,
+      "step": 669
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.081644925893866e-06,
+      "loss": 0.5116,
+      "step": 670
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.0767220510599845e-06,
+      "loss": 0.575,
+      "step": 671
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0717968131227285e-06,
+      "loss": 0.5614,
+      "step": 672
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0668692322631304e-06,
+      "loss": 0.5349,
+      "step": 673
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.061939328671824e-06,
+      "loss": 0.54,
+      "step": 674
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0570071225489577e-06,
+      "loss": 0.4908,
+      "step": 675
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0520726341041165e-06,
+      "loss": 0.4967,
+      "step": 676
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.047135883556235e-06,
+      "loss": 0.5522,
+      "step": 677
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0421968911335196e-06,
+      "loss": 0.5822,
+      "step": 678
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.0372556770733617e-06,
+      "loss": 0.5083,
+      "step": 679
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.032312261622255e-06,
+      "loss": 0.5194,
+      "step": 680
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.027366665035713e-06,
+      "loss": 0.521,
+      "step": 681
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.0224189075781886e-06,
+      "loss": 0.5364,
+      "step": 682
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.0174690095229853e-06,
+      "loss": 0.5165,
+      "step": 683
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.012516991152181e-06,
+      "loss": 0.5542,
+      "step": 684
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.0075628727565404e-06,
+      "loss": 0.5462,
+      "step": 685
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.002606674635432e-06,
+      "loss": 0.5298,
+      "step": 686
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.9976484170967473e-06,
+      "loss": 0.5459,
+      "step": 687
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.9926881204568153e-06,
+      "loss": 0.501,
+      "step": 688
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.9877258050403214e-06,
+      "loss": 0.5296,
+      "step": 689
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.9827614911802205e-06,
+      "loss": 0.5221,
+      "step": 690
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.977795199217658e-06,
+      "loss": 0.5308,
+      "step": 691
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 2.972826949501884e-06,
+      "loss": 0.5087,
+      "step": 692
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.967856762390172e-06,
+      "loss": 0.5136,
+      "step": 693
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9628846582477305e-06,
+      "loss": 0.5034,
+      "step": 694
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9579106574476257e-06,
+      "loss": 0.4888,
+      "step": 695
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9529347803706943e-06,
+      "loss": 0.5097,
+      "step": 696
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.947957047405462e-06,
+      "loss": 0.5272,
+      "step": 697
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9429774789480576e-06,
+      "loss": 0.4999,
+      "step": 698
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.937996095402131e-06,
+      "loss": 0.5127,
+      "step": 699
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9330129171787704e-06,
+      "loss": 0.512,
+      "step": 700
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.9280279646964156e-06,
+      "loss": 0.5084,
+      "step": 701
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 2.923041258380779e-06,
+      "loss": 0.5346,
+      "step": 702
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.9180528186647585e-06,
+      "loss": 0.5239,
+      "step": 703
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.9130626659883537e-06,
+      "loss": 0.5313,
+      "step": 704
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.908070820798583e-06,
+      "loss": 0.5611,
+      "step": 705
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.9030773035493997e-06,
+      "loss": 0.5311,
+      "step": 706
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.8980821347016097e-06,
+      "loss": 0.498,
+      "step": 707
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.893085334722786e-06,
+      "loss": 0.5194,
+      "step": 708
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.8880869240871833e-06,
+      "loss": 0.514,
+      "step": 709
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.883086923275658e-06,
+      "loss": 0.5305,
+      "step": 710
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.878085352775581e-06,
+      "loss": 0.5134,
+      "step": 711
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.8730822330807556e-06,
+      "loss": 0.5037,
+      "step": 712
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 2.8680775846913323e-06,
+      "loss": 0.5445,
+      "step": 713
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8630714281137263e-06,
+      "loss": 0.5237,
+      "step": 714
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8580637838605315e-06,
+      "loss": 0.5337,
+      "step": 715
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.853054672450437e-06,
+      "loss": 0.5224,
+      "step": 716
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8480441144081465e-06,
+      "loss": 0.5362,
+      "step": 717
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8430321302642887e-06,
+      "loss": 0.5096,
+      "step": 718
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8380187405553355e-06,
+      "loss": 0.5224,
+      "step": 719
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8330039658235194e-06,
+      "loss": 0.4982,
+      "step": 720
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8279878266167467e-06,
+      "loss": 0.498,
+      "step": 721
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8229703434885165e-06,
+      "loss": 0.5308,
+      "step": 722
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 2.8179515369978327e-06,
+      "loss": 0.5387,
+      "step": 723
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.8129314277091224e-06,
+      "loss": 0.5193,
+      "step": 724
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.807910036192151e-06,
+      "loss": 0.4918,
+      "step": 725
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.8028873830219373e-06,
+      "loss": 0.525,
+      "step": 726
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.7978634887786702e-06,
+      "loss": 0.529,
+      "step": 727
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.7928383740476247e-06,
+      "loss": 0.509,
+      "step": 728
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.7878120594190742e-06,
+      "loss": 0.5237,
+      "step": 729
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.7827845654882112e-06,
+      "loss": 0.5205,
+      "step": 730
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.777755912855058e-06,
+      "loss": 0.5377,
+      "step": 731
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.7727261221243875e-06,
+      "loss": 0.5274,
+      "step": 732
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 2.767695213905634e-06,
+      "loss": 0.5341,
+      "step": 733
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.76266320881281e-06,
+      "loss": 0.5092,
+      "step": 734
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.757630127464425e-06,
+      "loss": 0.5218,
+      "step": 735
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7525959904833955e-06,
+      "loss": 0.5147,
+      "step": 736
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.747560818496966e-06,
+      "loss": 0.496,
+      "step": 737
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7425246321366205e-06,
+      "loss": 0.4896,
+      "step": 738
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7374874520379994e-06,
+      "loss": 0.5269,
+      "step": 739
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7324492988408146e-06,
+      "loss": 0.5061,
+      "step": 740
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7274101931887663e-06,
+      "loss": 0.5349,
+      "step": 741
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7223701557294574e-06,
+      "loss": 0.5359,
+      "step": 742
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.7173292071143075e-06,
+      "loss": 0.5119,
+      "step": 743
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 2.712287367998471e-06,
+      "loss": 0.4838,
+      "step": 744
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.707244659040749e-06,
+      "loss": 0.5297,
+      "step": 745
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.702201100903511e-06,
+      "loss": 0.4842,
+      "step": 746
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.697156714252601e-06,
+      "loss": 0.5243,
+      "step": 747
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.692111519757261e-06,
+      "loss": 0.5303,
+      "step": 748
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.6870655380900424e-06,
+      "loss": 0.512,
+      "step": 749
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.6820187899267203e-06,
+      "loss": 0.5073,
+      "step": 750
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.6769712959462143e-06,
+      "loss": 0.51,
+      "step": 751
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.671923076830496e-06,
+      "loss": 0.5088,
+      "step": 752
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.666874153264511e-06,
+      "loss": 0.5163,
+      "step": 753
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 2.6618245459360896e-06,
+      "loss": 0.4988,
+      "step": 754
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.6567742755358645e-06,
+      "loss": 0.5272,
+      "step": 755
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.651723362757186e-06,
+      "loss": 0.4955,
+      "step": 756
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.6466718282960354e-06,
+      "loss": 0.5197,
+      "step": 757
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.641619692850941e-06,
+      "loss": 0.5176,
+      "step": 758
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.6365669771228954e-06,
+      "loss": 0.5091,
+      "step": 759
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.631513701815267e-06,
+      "loss": 0.5419,
+      "step": 760
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.6264598876337188e-06,
+      "loss": 0.5124,
+      "step": 761
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.6214055552861213e-06,
+      "loss": 0.5086,
+      "step": 762
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.616350725482466e-06,
+      "loss": 0.5546,
+      "step": 763
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.611295418934786e-06,
+      "loss": 0.508,
+      "step": 764
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 2.606239656357066e-06,
+      "loss": 0.4877,
+      "step": 765
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.6011834584651597e-06,
+      "loss": 0.5011,
+      "step": 766
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5961268459767047e-06,
+      "loss": 0.5106,
+      "step": 767
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.591069839611036e-06,
+      "loss": 0.4771,
+      "step": 768
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.586012460089105e-06,
+      "loss": 0.4598,
+      "step": 769
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5809547281333904e-06,
+      "loss": 0.513,
+      "step": 770
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5758966644678145e-06,
+      "loss": 0.5362,
+      "step": 771
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.570838289817661e-06,
+      "loss": 0.4735,
+      "step": 772
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.5657796249094854e-06,
+      "loss": 0.5029,
+      "step": 773
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.560720690471033e-06,
+      "loss": 0.533,
+      "step": 774
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 2.555661507231155e-06,
+      "loss": 0.5391,
+      "step": 775
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5506020959197218e-06,
+      "loss": 0.49,
+      "step": 776
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.545542477267537e-06,
+      "loss": 0.511,
+      "step": 777
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5404826720062544e-06,
+      "loss": 0.4958,
+      "step": 778
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.535422700868292e-06,
+      "loss": 0.5112,
+      "step": 779
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5303625845867475e-06,
+      "loss": 0.5137,
+      "step": 780
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.525302343895315e-06,
+      "loss": 0.4968,
+      "step": 781
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5202419995281966e-06,
+      "loss": 0.4782,
+      "step": 782
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5151815722200203e-06,
+      "loss": 0.4964,
+      "step": 783
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5101210827057516e-06,
+      "loss": 0.4952,
+      "step": 784
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.5050605517206147e-06,
+      "loss": 0.4819,
+      "step": 785
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.5e-06,
+      "loss": 0.499,
+      "step": 786
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.494939448279386e-06,
+      "loss": 0.5327,
+      "step": 787
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4898789172942492e-06,
+      "loss": 0.5095,
+      "step": 788
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4848184277799805e-06,
+      "loss": 0.5092,
+      "step": 789
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4797580004718038e-06,
+      "loss": 0.4984,
+      "step": 790
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.474697656104686e-06,
+      "loss": 0.4933,
+      "step": 791
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4696374154132533e-06,
+      "loss": 0.5165,
+      "step": 792
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.464577299131709e-06,
+      "loss": 0.5206,
+      "step": 793
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4595173279937464e-06,
+      "loss": 0.5026,
+      "step": 794
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.454457522732464e-06,
+      "loss": 0.5272,
+      "step": 795
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.4493979040802786e-06,
+      "loss": 0.5178,
+      "step": 796
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4443384927688454e-06,
+      "loss": 0.5043,
+      "step": 797
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4392793095289677e-06,
+      "loss": 0.4816,
+      "step": 798
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.434220375090516e-06,
+      "loss": 0.5095,
+      "step": 799
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.42916171018234e-06,
+      "loss": 0.4979,
+      "step": 800
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4241033355321867e-06,
+      "loss": 0.509,
+      "step": 801
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.419045271866611e-06,
+      "loss": 0.511,
+      "step": 802
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.4139875399108957e-06,
+      "loss": 0.4958,
+      "step": 803
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.408930160388965e-06,
+      "loss": 0.4822,
+      "step": 804
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.403873154023297e-06,
+      "loss": 0.5199,
+      "step": 805
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.3988165415348416e-06,
+      "loss": 0.4976,
+      "step": 806
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.393760343642935e-06,
+      "loss": 0.4842,
+      "step": 807
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.388704581065215e-06,
+      "loss": 0.5084,
+      "step": 808
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.383649274517535e-06,
+      "loss": 0.5314,
+      "step": 809
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3785944447138804e-06,
+      "loss": 0.5048,
+      "step": 810
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.373540112366281e-06,
+      "loss": 0.5247,
+      "step": 811
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.368486298184733e-06,
+      "loss": 0.5083,
+      "step": 812
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.363433022877105e-06,
+      "loss": 0.5021,
+      "step": 813
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.358380307149059e-06,
+      "loss": 0.5061,
+      "step": 814
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.353328171703965e-06,
+      "loss": 0.5273,
+      "step": 815
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.348276637242814e-06,
+      "loss": 0.5222,
+      "step": 816
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3432257244641355e-06,
+      "loss": 0.4736,
+      "step": 817
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3381754540639108e-06,
+      "loss": 0.4966,
+      "step": 818
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3331258467354894e-06,
+      "loss": 0.5374,
+      "step": 819
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.328076923169504e-06,
+      "loss": 0.5206,
+      "step": 820
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.323028704053786e-06,
+      "loss": 0.5182,
+      "step": 821
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.31798121007328e-06,
+      "loss": 0.4952,
+      "step": 822
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3129344619099585e-06,
+      "loss": 0.4657,
+      "step": 823
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3078884802427394e-06,
+      "loss": 0.5101,
+      "step": 824
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3028432857474e-06,
+      "loss": 0.5021,
+      "step": 825
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.29779889909649e-06,
+      "loss": 0.5269,
+      "step": 826
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2927553409592513e-06,
+      "loss": 0.5228,
+      "step": 827
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2877126320015295e-06,
+      "loss": 0.5411,
+      "step": 828
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.282670792885693e-06,
+      "loss": 0.5205,
+      "step": 829
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2776298442705434e-06,
+      "loss": 0.5121,
+      "step": 830
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.272589806811234e-06,
+      "loss": 0.4685,
+      "step": 831
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.267550701159186e-06,
+      "loss": 0.4906,
+      "step": 832
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2625125479620014e-06,
+      "loss": 0.4959,
+      "step": 833
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.25747536786338e-06,
+      "loss": 0.5239,
+      "step": 834
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2524391815030346e-06,
+      "loss": 0.4844,
+      "step": 835
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.247404009516605e-06,
+      "loss": 0.4989,
+      "step": 836
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2423698725355754e-06,
+      "loss": 0.4755,
+      "step": 837
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.2373367911871904e-06,
+      "loss": 0.5006,
+      "step": 838
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.232304786094367e-06,
+      "loss": 0.5428,
+      "step": 839
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.227273877875613e-06,
+      "loss": 0.4605,
+      "step": 840
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.2222440871449426e-06,
+      "loss": 0.5078,
+      "step": 841
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.2172154345117896e-06,
+      "loss": 0.5262,
+      "step": 842
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.212187940580926e-06,
+      "loss": 0.4826,
+      "step": 843
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.207161625952376e-06,
+      "loss": 0.4923,
+      "step": 844
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.20213651122133e-06,
+      "loss": 0.5213,
+      "step": 845
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.1971126169780636e-06,
+      "loss": 0.5328,
+      "step": 846
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.19208996380785e-06,
+      "loss": 0.4996,
+      "step": 847
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 2.1870685722908784e-06,
+      "loss": 0.5155,
+      "step": 848
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1820484630021686e-06,
+      "loss": 0.5018,
+      "step": 849
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1770296565114847e-06,
+      "loss": 0.5046,
+      "step": 850
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.172012173383254e-06,
+      "loss": 0.5039,
+      "step": 851
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.166996034176482e-06,
+      "loss": 0.4967,
+      "step": 852
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1619812594446657e-06,
+      "loss": 0.4956,
+      "step": 853
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1569678697357126e-06,
+      "loss": 0.4927,
+      "step": 854
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1519558855918544e-06,
+      "loss": 0.5073,
+      "step": 855
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1469453275495634e-06,
+      "loss": 0.5174,
+      "step": 856
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.1419362161394698e-06,
+      "loss": 0.4996,
+      "step": 857
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 2.136928571886275e-06,
+      "loss": 0.5088,
+      "step": 858
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1319224153086677e-06,
+      "loss": 0.5473,
+      "step": 859
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.126917766919245e-06,
+      "loss": 0.4988,
+      "step": 860
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.121914647224419e-06,
+      "loss": 0.5084,
+      "step": 861
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1169130767243424e-06,
+      "loss": 0.5158,
+      "step": 862
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1119130759128166e-06,
+      "loss": 0.4717,
+      "step": 863
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1069146652772142e-06,
+      "loss": 0.4976,
+      "step": 864
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.1019178652983903e-06,
+      "loss": 0.527,
+      "step": 865
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0969226964506007e-06,
+      "loss": 0.5228,
+      "step": 866
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0919291792014176e-06,
+      "loss": 0.4958,
+      "step": 867
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0869373340116467e-06,
+      "loss": 0.5407,
+      "step": 868
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 2.0819471813352415e-06,
+      "loss": 0.5391,
+      "step": 869
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0769587416192212e-06,
+      "loss": 0.4819,
+      "step": 870
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0719720353035848e-06,
+      "loss": 0.4966,
+      "step": 871
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.066987082821231e-06,
+      "loss": 0.5225,
+      "step": 872
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.06200390459787e-06,
+      "loss": 0.4899,
+      "step": 873
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0570225210519433e-06,
+      "loss": 0.4696,
+      "step": 874
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0520429525945387e-06,
+      "loss": 0.4807,
+      "step": 875
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.047065219629306e-06,
+      "loss": 0.5001,
+      "step": 876
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0420893425523747e-06,
+      "loss": 0.4906,
+      "step": 877
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0371153417522703e-06,
+      "loss": 0.4888,
+      "step": 878
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 2.0321432376098286e-06,
+      "loss": 0.5089,
+      "step": 879
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.0271730504981165e-06,
+      "loss": 0.4711,
+      "step": 880
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.0222048007823424e-06,
+      "loss": 0.4886,
+      "step": 881
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.0172385088197804e-06,
+      "loss": 0.5886,
+      "step": 882
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.01227419495968e-06,
+      "loss": 0.5104,
+      "step": 883
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.007311879543185e-06,
+      "loss": 0.5049,
+      "step": 884
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 2.0023515829032535e-06,
+      "loss": 0.5236,
+      "step": 885
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9973933253645684e-06,
+      "loss": 0.5152,
+      "step": 886
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9924371272434604e-06,
+      "loss": 0.5075,
+      "step": 887
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9874830088478196e-06,
+      "loss": 0.5664,
+      "step": 888
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9825309904770155e-06,
+      "loss": 0.4783,
+      "step": 889
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9775810924218126e-06,
+      "loss": 0.507,
+      "step": 890
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9726333349642877e-06,
+      "loss": 0.5082,
+      "step": 891
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.967687738377746e-06,
+      "loss": 0.4941,
+      "step": 892
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.962744322926639e-06,
+      "loss": 0.4812,
+      "step": 893
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9578031088664812e-06,
+      "loss": 0.5064,
+      "step": 894
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.952864116443766e-06,
+      "loss": 0.5245,
+      "step": 895
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9479273658958852e-06,
+      "loss": 0.5347,
+      "step": 896
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9429928774510435e-06,
+      "loss": 0.4956,
+      "step": 897
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9380606713281773e-06,
+      "loss": 0.5046,
+      "step": 898
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9331307677368705e-06,
+      "loss": 0.5084,
+      "step": 899
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.928203186877273e-06,
+      "loss": 0.5355,
+      "step": 900
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9232779489400168e-06,
+      "loss": 0.4941,
+      "step": 901
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9183550741061354e-06,
+      "loss": 0.5205,
+      "step": 902
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.913434582546978e-06,
+      "loss": 0.4974,
+      "step": 903
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9085164944241275e-06,
+      "loss": 0.5284,
+      "step": 904
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9036008298893221e-06,
+      "loss": 0.4852,
+      "step": 905
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8986876090843668e-06,
+      "loss": 0.5125,
+      "step": 906
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8937768521410535e-06,
+      "loss": 0.5096,
+      "step": 907
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8888685791810784e-06,
+      "loss": 0.4974,
+      "step": 908
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8839628103159624e-06,
+      "loss": 0.4864,
+      "step": 909
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.8790595656469628e-06,
+      "loss": 0.4748,
+      "step": 910
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.874158865264995e-06,
+      "loss": 0.4799,
+      "step": 911
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.86926072925055e-06,
+      "loss": 0.5262,
+      "step": 912
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8643651776736102e-06,
+      "loss": 0.5242,
+      "step": 913
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8594722305935691e-06,
+      "loss": 0.499,
+      "step": 914
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8545819080591481e-06,
+      "loss": 0.5001,
+      "step": 915
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8496942301083142e-06,
+      "loss": 0.5448,
+      "step": 916
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8448092167681997e-06,
+      "loss": 0.5114,
+      "step": 917
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8399268880550174e-06,
+      "loss": 0.4528,
+      "step": 918
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.83504726397398e-06,
+      "loss": 0.4756,
+      "step": 919
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8301703645192178e-06,
+      "loss": 0.4987,
+      "step": 920
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.8252962096736987e-06,
+      "loss": 0.5011,
+      "step": 921
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8204248194091429e-06,
+      "loss": 0.5119,
+      "step": 922
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8155562136859433e-06,
+      "loss": 0.4583,
+      "step": 923
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8106904124530839e-06,
+      "loss": 0.5031,
+      "step": 924
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.8058274356480552e-06,
+      "loss": 0.4953,
+      "step": 925
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.800967303196778e-06,
+      "loss": 0.4917,
+      "step": 926
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.7961100350135158e-06,
+      "loss": 0.5185,
+      "step": 927
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.7912556510007967e-06,
+      "loss": 0.485,
+      "step": 928
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.7864041710493314e-06,
+      "loss": 0.4806,
+      "step": 929
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.7815556150379298e-06,
+      "loss": 0.523,
+      "step": 930
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.7767100028334236e-06,
+      "loss": 0.5369,
+      "step": 931
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.77186735429058e-06,
+      "loss": 0.5186,
+      "step": 932
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7670276892520238e-06,
+      "loss": 0.5126,
+      "step": 933
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7621910275481544e-06,
+      "loss": 0.4797,
+      "step": 934
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7573573889970657e-06,
+      "loss": 0.4785,
+      "step": 935
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7525267934044642e-06,
+      "loss": 0.502,
+      "step": 936
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7476992605635872e-06,
+      "loss": 0.4742,
+      "step": 937
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7428748102551237e-06,
+      "loss": 0.5254,
+      "step": 938
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7380534622471304e-06,
+      "loss": 0.5043,
+      "step": 939
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7332352362949546e-06,
+      "loss": 0.5013,
+      "step": 940
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7284201521411492e-06,
+      "loss": 0.5011,
+      "step": 941
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.7236082295153948e-06,
+      "loss": 0.4999,
+      "step": 942
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.7187994881344167e-06,
+      "loss": 0.5013,
+      "step": 943
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.7139939477019057e-06,
+      "loss": 0.4973,
+      "step": 944
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.709191627908438e-06,
+      "loss": 0.5167,
+      "step": 945
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.7043925484313911e-06,
+      "loss": 0.4975,
+      "step": 946
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.6995967289348664e-06,
+      "loss": 0.4787,
+      "step": 947
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.6948041890696076e-06,
+      "loss": 0.471,
+      "step": 948
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.690014948472921e-06,
+      "loss": 0.4808,
+      "step": 949
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.685229026768593e-06,
+      "loss": 0.5186,
+      "step": 950
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.6804464435668116e-06,
+      "loss": 0.5279,
+      "step": 951
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.6756672184640847e-06,
+      "loss": 0.5036,
+      "step": 952
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6708913710431607e-06,
+      "loss": 0.506,
+      "step": 953
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6661189208729492e-06,
+      "loss": 0.5154,
+      "step": 954
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6613498875084382e-06,
+      "loss": 0.5069,
+      "step": 955
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6565842904906154e-06,
+      "loss": 0.4657,
+      "step": 956
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6518221493463903e-06,
+      "loss": 0.4738,
+      "step": 957
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6470634835885097e-06,
+      "loss": 0.481,
+      "step": 958
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.642308312715481e-06,
+      "loss": 0.5073,
+      "step": 959
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6375566562114903e-06,
+      "loss": 0.4918,
+      "step": 960
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6328085335463257e-06,
+      "loss": 0.5002,
+      "step": 961
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.6280639641752944e-06,
+      "loss": 0.5037,
+      "step": 962
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.6233229675391439e-06,
+      "loss": 0.488,
+      "step": 963
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.6185855630639818e-06,
+      "loss": 0.4588,
+      "step": 964
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.6138517701611993e-06,
+      "loss": 0.5151,
+      "step": 965
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.6091216082273875e-06,
+      "loss": 0.4909,
+      "step": 966
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.6043950966442599e-06,
+      "loss": 0.5011,
+      "step": 967
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.5996722547785722e-06,
+      "loss": 0.4871,
+      "step": 968
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.5949531019820445e-06,
+      "loss": 0.5219,
+      "step": 969
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.5902376575912815e-06,
+      "loss": 0.4868,
+      "step": 970
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.5855259409276912e-06,
+      "loss": 0.5158,
+      "step": 971
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.580817971297409e-06,
+      "loss": 0.4802,
+      "step": 972
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.5761137679912152e-06,
+      "loss": 0.4952,
+      "step": 973
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5714133502844591e-06,
+      "loss": 0.4743,
+      "step": 974
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5667167374369786e-06,
+      "loss": 0.4861,
+      "step": 975
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.56202394869302e-06,
+      "loss": 0.499,
+      "step": 976
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.557335003281162e-06,
+      "loss": 0.5171,
+      "step": 977
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5526499204142332e-06,
+      "loss": 0.4934,
+      "step": 978
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.547968719289239e-06,
+      "loss": 0.4997,
+      "step": 979
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5432914190872757e-06,
+      "loss": 0.4582,
+      "step": 980
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5386180389734583e-06,
+      "loss": 0.5094,
+      "step": 981
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5339485980968383e-06,
+      "loss": 0.4918,
+      "step": 982
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5292831155903254e-06,
+      "loss": 0.5352,
+      "step": 983
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5246216105706124e-06,
+      "loss": 0.51,
+      "step": 984
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.519964102138092e-06,
+      "loss": 0.5113,
+      "step": 985
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5153106093767827e-06,
+      "loss": 0.461,
+      "step": 986
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5106611513542467e-06,
+      "loss": 0.4856,
+      "step": 987
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.506015747121518e-06,
+      "loss": 0.483,
+      "step": 988
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.5013744157130158e-06,
+      "loss": 0.5209,
+      "step": 989
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4967371761464738e-06,
+      "loss": 0.4929,
+      "step": 990
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4921040474228582e-06,
+      "loss": 0.4918,
+      "step": 991
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4874750485262917e-06,
+      "loss": 0.4743,
+      "step": 992
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4828501984239751e-06,
+      "loss": 0.4938,
+      "step": 993
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4782295160661103e-06,
+      "loss": 0.48,
+      "step": 994
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.47361302038582e-06,
+      "loss": 0.5003,
+      "step": 995
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.469000730299074e-06,
+      "loss": 0.5219,
+      "step": 996
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4643926647046086e-06,
+      "loss": 0.491,
+      "step": 997
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4597888424838519e-06,
+      "loss": 0.5006,
+      "step": 998
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4551892825008423e-06,
+      "loss": 0.477,
+      "step": 999
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.450594003602158e-06,
+      "loss": 0.5078,
+      "step": 1000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.44600302461683e-06,
+      "loss": 0.5425,
+      "step": 1001
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4414163643562755e-06,
+      "loss": 0.4932,
+      "step": 1002
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4368340416142123e-06,
+      "loss": 0.4973,
+      "step": 1003
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.4322560751665873e-06,
+      "loss": 0.4902,
+      "step": 1004
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4276824837714975e-06,
+      "loss": 0.5029,
+      "step": 1005
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4231132861691128e-06,
+      "loss": 0.5237,
+      "step": 1006
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4185485010816007e-06,
+      "loss": 0.4947,
+      "step": 1007
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4139881472130453e-06,
+      "loss": 0.5094,
+      "step": 1008
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4094322432493785e-06,
+      "loss": 0.4843,
+      "step": 1009
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.4048808078582943e-06,
+      "loss": 0.4625,
+      "step": 1010
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.40033385968918e-06,
+      "loss": 0.4828,
+      "step": 1011
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3957914173730366e-06,
+      "loss": 0.4858,
+      "step": 1012
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3912534995223996e-06,
+      "loss": 0.519,
+      "step": 1013
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.3867201247312697e-06,
+      "loss": 0.4983,
+      "step": 1014
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3821913115750284e-06,
+      "loss": 0.4919,
+      "step": 1015
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3776670786103685e-06,
+      "loss": 0.5207,
+      "step": 1016
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.373147444375217e-06,
+      "loss": 0.4724,
+      "step": 1017
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3686324273886531e-06,
+      "loss": 0.507,
+      "step": 1018
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3641220461508415e-06,
+      "loss": 0.4939,
+      "step": 1019
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.359616319142949e-06,
+      "loss": 0.514,
+      "step": 1020
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3551152648270733e-06,
+      "loss": 0.4999,
+      "step": 1021
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3506189016461674e-06,
+      "loss": 0.4874,
+      "step": 1022
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3461272480239576e-06,
+      "loss": 0.5405,
+      "step": 1023
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.341640322364878e-06,
+      "loss": 0.5018,
+      "step": 1024
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3371581430539884e-06,
+      "loss": 0.5408,
+      "step": 1025
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3326807284568984e-06,
+      "loss": 0.4818,
+      "step": 1026
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3282080969196982e-06,
+      "loss": 0.4731,
+      "step": 1027
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.323740266768875e-06,
+      "loss": 0.5141,
+      "step": 1028
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3192772563112464e-06,
+      "loss": 0.4853,
+      "step": 1029
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3148190838338804e-06,
+      "loss": 0.4832,
+      "step": 1030
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.310365767604019e-06,
+      "loss": 0.489,
+      "step": 1031
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3059173258690102e-06,
+      "loss": 0.4872,
+      "step": 1032
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.3014737768562246e-06,
+      "loss": 0.4798,
+      "step": 1033
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.2970351387729875e-06,
+      "loss": 0.507,
+      "step": 1034
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.2926014298065023e-06,
+      "loss": 0.4833,
+      "step": 1035
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2881726681237727e-06,
+      "loss": 0.5316,
+      "step": 1036
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2837488718715347e-06,
+      "loss": 0.5089,
+      "step": 1037
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2793300591761742e-06,
+      "loss": 0.4921,
+      "step": 1038
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2749162481436612e-06,
+      "loss": 0.4613,
+      "step": 1039
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.27050745685947e-06,
+      "loss": 0.4695,
+      "step": 1040
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2661037033885055e-06,
+      "loss": 0.4744,
+      "step": 1041
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2617050057750322e-06,
+      "loss": 0.4759,
+      "step": 1042
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 0.4901578426361084,
+      "eval_runtime": 2316.6043,
+      "eval_samples_per_second": 1.173,
+      "eval_steps_per_second": 0.147,
+      "step": 1042
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 1562,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 2,
+  "save_steps": 521,
+  "total_flos": 436137822781440.0,
+  "train_batch_size": 1,
+  "trial_name": null,
+  "trial_params": null
+}