diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,12019 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.9890601690701144,
+  "eval_steps": 500,
+  "global_step": 2000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.278688524590164e-07,
+      "loss": 1.1631,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.557377049180328e-07,
+      "loss": 1.8625,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9.836065573770493e-07,
+      "loss": 1.0719,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.3114754098360657e-06,
+      "loss": 0.9863,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.6393442622950819e-06,
+      "loss": 1.2453,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9672131147540985e-06,
+      "loss": 1.0458,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.295081967213115e-06,
+      "loss": 1.1643,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.6229508196721314e-06,
+      "loss": 0.9423,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.9508196721311478e-06,
+      "loss": 0.9232,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.2786885245901638e-06,
+      "loss": 1.1195,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.6065573770491806e-06,
+      "loss": 0.829,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.934426229508197e-06,
+      "loss": 1.1953,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.2622950819672135e-06,
+      "loss": 1.8621,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.59016393442623e-06,
+      "loss": 0.9418,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.918032786885246e-06,
+      "loss": 0.8859,
+      "step": 15
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.245901639344263e-06,
+      "loss": 1.1402,
+      "step": 16
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.573770491803278e-06,
+      "loss": 0.9034,
+      "step": 17
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5.9016393442622956e-06,
+      "loss": 1.4775,
+      "step": 18
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 6.229508196721312e-06,
+      "loss": 1.4377,
+      "step": 19
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 6.5573770491803276e-06,
+      "loss": 0.8942,
+      "step": 20
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 6.885245901639345e-06,
+      "loss": 0.9499,
+      "step": 21
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.213114754098361e-06,
+      "loss": 0.8495,
+      "step": 22
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.540983606557377e-06,
+      "loss": 1.5032,
+      "step": 23
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 7.868852459016394e-06,
+      "loss": 1.8468,
+      "step": 24
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 8.19672131147541e-06,
+      "loss": 1.7923,
+      "step": 25
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 8.524590163934427e-06,
+      "loss": 1.0237,
+      "step": 26
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 8.852459016393443e-06,
+      "loss": 0.7109,
+      "step": 27
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.18032786885246e-06,
+      "loss": 0.9663,
+      "step": 28
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.508196721311476e-06,
+      "loss": 1.0116,
+      "step": 29
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 9.836065573770493e-06,
+      "loss": 1.3598,
+      "step": 30
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.0163934426229509e-05,
+      "loss": 1.0879,
+      "step": 31
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.0491803278688525e-05,
+      "loss": 1.0015,
+      "step": 32
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.0819672131147544e-05,
+      "loss": 0.9813,
+      "step": 33
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.1147540983606557e-05,
+      "loss": 1.0592,
+      "step": 34
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.1475409836065575e-05,
+      "loss": 1.0062,
+      "step": 35
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.1803278688524591e-05,
+      "loss": 0.8501,
+      "step": 36
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.2131147540983608e-05,
+      "loss": 0.8943,
+      "step": 37
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.2459016393442624e-05,
+      "loss": 1.4784,
+      "step": 38
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.2786885245901642e-05,
+      "loss": 1.2013,
+      "step": 39
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.3114754098360655e-05,
+      "loss": 1.0282,
+      "step": 40
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.3442622950819673e-05,
+      "loss": 0.9437,
+      "step": 41
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.377049180327869e-05,
+      "loss": 1.6735,
+      "step": 42
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.4098360655737706e-05,
+      "loss": 0.7945,
+      "step": 43
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.4426229508196722e-05,
+      "loss": 1.3146,
+      "step": 44
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.4754098360655739e-05,
+      "loss": 1.0985,
+      "step": 45
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.5081967213114754e-05,
+      "loss": 0.6569,
+      "step": 46
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.5409836065573772e-05,
+      "loss": 0.8684,
+      "step": 47
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.5737704918032788e-05,
+      "loss": 1.2233,
+      "step": 48
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.6065573770491805e-05,
+      "loss": 0.7584,
+      "step": 49
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.639344262295082e-05,
+      "loss": 0.8788,
+      "step": 50
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.6721311475409837e-05,
+      "loss": 0.8581,
+      "step": 51
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.7049180327868854e-05,
+      "loss": 0.925,
+      "step": 52
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.737704918032787e-05,
+      "loss": 0.704,
+      "step": 53
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.7704918032786887e-05,
+      "loss": 0.9357,
+      "step": 54
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.8032786885245903e-05,
+      "loss": 1.0651,
+      "step": 55
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.836065573770492e-05,
+      "loss": 0.8416,
+      "step": 56
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.8688524590163936e-05,
+      "loss": 0.801,
+      "step": 57
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9016393442622952e-05,
+      "loss": 0.9371,
+      "step": 58
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.934426229508197e-05,
+      "loss": 0.8922,
+      "step": 59
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9672131147540985e-05,
+      "loss": 0.8713,
+      "step": 60
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 2e-05,
+      "loss": 0.8037,
+      "step": 61
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9999987008898457e-05,
+      "loss": 0.8481,
+      "step": 62
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9999948035627574e-05,
+      "loss": 0.9219,
+      "step": 63
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9999883080288618e-05,
+      "loss": 1.0054,
+      "step": 64
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9999792143050352e-05,
+      "loss": 1.3275,
+      "step": 65
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9999675224149054e-05,
+      "loss": 1.6285,
+      "step": 66
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9999532323888507e-05,
+      "loss": 1.0542,
+      "step": 67
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9999363442639997e-05,
+      "loss": 0.8898,
+      "step": 68
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.999916858084231e-05,
+      "loss": 0.9377,
+      "step": 69
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9998947739001743e-05,
+      "loss": 1.3403,
+      "step": 70
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.999870091769209e-05,
+      "loss": 0.9542,
+      "step": 71
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.999842811755465e-05,
+      "loss": 0.7805,
+      "step": 72
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9998129339298217e-05,
+      "loss": 0.9154,
+      "step": 73
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.999780458369908e-05,
+      "loss": 0.8465,
+      "step": 74
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.999745385160103e-05,
+      "loss": 0.8779,
+      "step": 75
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9997077143915345e-05,
+      "loss": 0.9247,
+      "step": 76
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.999667446162079e-05,
+      "loss": 1.4597,
+      "step": 77
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9996245805763628e-05,
+      "loss": 0.8224,
+      "step": 78
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9995791177457598e-05,
+      "loss": 0.7869,
+      "step": 79
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9995310577883928e-05,
+      "loss": 0.7324,
+      "step": 80
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.999480400829132e-05,
+      "loss": 0.8845,
+      "step": 81
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9994271469995953e-05,
+      "loss": 1.1086,
+      "step": 82
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9993712964381475e-05,
+      "loss": 0.9462,
+      "step": 83
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9993128492899012e-05,
+      "loss": 0.8624,
+      "step": 84
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.999251805706715e-05,
+      "loss": 0.8404,
+      "step": 85
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.999188165847193e-05,
+      "loss": 0.9484,
+      "step": 86
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9991219298766862e-05,
+      "loss": 1.0643,
+      "step": 87
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.99905309796729e-05,
+      "loss": 0.6939,
+      "step": 88
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9989816702978447e-05,
+      "loss": 0.7965,
+      "step": 89
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.998907647053935e-05,
+      "loss": 0.817,
+      "step": 90
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9988310284278904e-05,
+      "loss": 0.6953,
+      "step": 91
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9987518146187825e-05,
+      "loss": 1.1733,
+      "step": 92
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.998670005832426e-05,
+      "loss": 0.826,
+      "step": 93
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.998585602281378e-05,
+      "loss": 0.9648,
+      "step": 94
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9984986041849378e-05,
+      "loss": 0.9012,
+      "step": 95
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.998409011769146e-05,
+      "loss": 1.3652,
+      "step": 96
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9983168252667832e-05,
+      "loss": 1.3381,
+      "step": 97
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.99822204491737e-05,
+      "loss": 0.9126,
+      "step": 98
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9981246709671668e-05,
+      "loss": 0.9528,
+      "step": 99
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9980247036691723e-05,
+      "loss": 0.8905,
+      "step": 100
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.997922143283124e-05,
+      "loss": 0.9624,
+      "step": 101
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.997816990075496e-05,
+      "loss": 0.9044,
+      "step": 102
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9977092443194997e-05,
+      "loss": 0.8365,
+      "step": 103
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9975989062950828e-05,
+      "loss": 1.061,
+      "step": 104
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.997485976288927e-05,
+      "loss": 0.737,
+      "step": 105
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9973704545944494e-05,
+      "loss": 0.8608,
+      "step": 106
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9972523415118012e-05,
+      "loss": 0.8978,
+      "step": 107
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.997131637347866e-05,
+      "loss": 0.875,
+      "step": 108
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9970083424162598e-05,
+      "loss": 1.3394,
+      "step": 109
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.99688245703733e-05,
+      "loss": 1.3804,
+      "step": 110
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.996753981538155e-05,
+      "loss": 0.7778,
+      "step": 111
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9966229162525417e-05,
+      "loss": 0.9996,
+      "step": 112
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.996489261521027e-05,
+      "loss": 1.0724,
+      "step": 113
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9963530176908752e-05,
+      "loss": 1.2232,
+      "step": 114
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9962141851160778e-05,
+      "loss": 0.8759,
+      "step": 115
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.996072764157353e-05,
+      "loss": 1.0849,
+      "step": 116
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9959287551821425e-05,
+      "loss": 0.9536,
+      "step": 117
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9957821585646143e-05,
+      "loss": 1.3268,
+      "step": 118
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9956329746856583e-05,
+      "loss": 0.9561,
+      "step": 119
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9954812039328868e-05,
+      "loss": 0.8386,
+      "step": 120
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.995326846700634e-05,
+      "loss": 2.018,
+      "step": 121
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9951699033899544e-05,
+      "loss": 0.9314,
+      "step": 122
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9950103744086205e-05,
+      "loss": 0.9947,
+      "step": 123
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9948482601711245e-05,
+      "loss": 0.9588,
+      "step": 124
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.994683561098674e-05,
+      "loss": 1.3748,
+      "step": 125
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9945162776191946e-05,
+      "loss": 1.5894,
+      "step": 126
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9943464101673245e-05,
+      "loss": 1.2297,
+      "step": 127
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9941739591844173e-05,
+      "loss": 0.8366,
+      "step": 128
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9939989251185386e-05,
+      "loss": 0.817,
+      "step": 129
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9938213084244657e-05,
+      "loss": 0.7562,
+      "step": 130
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9936411095636855e-05,
+      "loss": 0.7552,
+      "step": 131
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.993458329004395e-05,
+      "loss": 0.8523,
+      "step": 132
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9932729672214975e-05,
+      "loss": 0.8466,
+      "step": 133
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.993085024696604e-05,
+      "loss": 0.914,
+      "step": 134
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.992894501918031e-05,
+      "loss": 0.7882,
+      "step": 135
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9927013993807985e-05,
+      "loss": 1.0892,
+      "step": 136
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.992505717586629e-05,
+      "loss": 0.8238,
+      "step": 137
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9923074570439476e-05,
+      "loss": 1.4146,
+      "step": 138
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.992106618267878e-05,
+      "loss": 0.9236,
+      "step": 139
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9919032017802445e-05,
+      "loss": 0.8065,
+      "step": 140
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9916972081095672e-05,
+      "loss": 0.9458,
+      "step": 141
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9914886377910637e-05,
+      "loss": 0.898,
+      "step": 142
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.991277491366645e-05,
+      "loss": 0.783,
+      "step": 143
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9910637693849166e-05,
+      "loss": 0.7972,
+      "step": 144
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9908474724011752e-05,
+      "loss": 1.448,
+      "step": 145
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9906286009774074e-05,
+      "loss": 0.9717,
+      "step": 146
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.99040715568229e-05,
+      "loss": 0.7907,
+      "step": 147
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9901831370911866e-05,
+      "loss": 0.9144,
+      "step": 148
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9899565457861463e-05,
+      "loss": 0.835,
+      "step": 149
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.989727382355904e-05,
+      "loss": 0.9172,
+      "step": 150
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9894956473958768e-05,
+      "loss": 0.9481,
+      "step": 151
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9892613415081623e-05,
+      "loss": 0.766,
+      "step": 152
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9890244653015397e-05,
+      "loss": 0.8229,
+      "step": 153
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.988785019391465e-05,
+      "loss": 0.7748,
+      "step": 154
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9885430044000715e-05,
+      "loss": 0.8486,
+      "step": 155
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.988298420956168e-05,
+      "loss": 0.8516,
+      "step": 156
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9880512696952356e-05,
+      "loss": 0.8075,
+      "step": 157
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9878015512594276e-05,
+      "loss": 0.6981,
+      "step": 158
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.987549266297568e-05,
+      "loss": 1.5784,
+      "step": 159
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.987294415465149e-05,
+      "loss": 0.7499,
+      "step": 160
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9870369994243283e-05,
+      "loss": 0.957,
+      "step": 161
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9867770188439302e-05,
+      "loss": 0.7369,
+      "step": 162
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.986514474399441e-05,
+      "loss": 0.8588,
+      "step": 163
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.986249366773009e-05,
+      "loss": 0.6999,
+      "step": 164
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.985981696653443e-05,
+      "loss": 0.7267,
+      "step": 165
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9857114647362082e-05,
+      "loss": 0.8667,
+      "step": 166
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9854386717234265e-05,
+      "loss": 0.7941,
+      "step": 167
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.985163318323875e-05,
+      "loss": 0.973,
+      "step": 168
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9848854052529822e-05,
+      "loss": 0.7735,
+      "step": 169
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.984604933232827e-05,
+      "loss": 0.8267,
+      "step": 170
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9843219029921387e-05,
+      "loss": 0.9171,
+      "step": 171
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9840363152662906e-05,
+      "loss": 1.4763,
+      "step": 172
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9837481707973037e-05,
+      "loss": 0.9691,
+      "step": 173
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9834574703338406e-05,
+      "loss": 0.9034,
+      "step": 174
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9831642146312043e-05,
+      "loss": 0.887,
+      "step": 175
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9828684044513392e-05,
+      "loss": 0.983,
+      "step": 176
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9825700405628245e-05,
+      "loss": 0.831,
+      "step": 177
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9822691237408747e-05,
+      "loss": 0.7617,
+      "step": 178
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9819656547673393e-05,
+      "loss": 1.2565,
+      "step": 179
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9816596344306965e-05,
+      "loss": 1.4362,
+      "step": 180
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.981351063526055e-05,
+      "loss": 1.2248,
+      "step": 181
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.98103994285515e-05,
+      "loss": 0.9034,
+      "step": 182
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9807262732263417e-05,
+      "loss": 0.7415,
+      "step": 183
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9804100554546127e-05,
+      "loss": 0.832,
+      "step": 184
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.980091290361566e-05,
+      "loss": 1.0681,
+      "step": 185
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.9797699787754243e-05,
+      "loss": 0.8454,
+      "step": 186
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9794461215310258e-05,
+      "loss": 0.7861,
+      "step": 187
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.979119719469822e-05,
+      "loss": 0.9102,
+      "step": 188
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9787907734398785e-05,
+      "loss": 0.71,
+      "step": 189
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9784592842958693e-05,
+      "loss": 0.9857,
+      "step": 190
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.978125252899076e-05,
+      "loss": 0.7978,
+      "step": 191
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.977788680117386e-05,
+      "loss": 0.829,
+      "step": 192
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.977449566825289e-05,
+      "loss": 0.7571,
+      "step": 193
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9771079139038765e-05,
+      "loss": 0.7473,
+      "step": 194
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.976763722240838e-05,
+      "loss": 0.7078,
+      "step": 195
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.9764169927304594e-05,
+      "loss": 0.7293,
+      "step": 196
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9760677262736202e-05,
+      "loss": 0.8904,
+      "step": 197
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9757159237777916e-05,
+      "loss": 0.7294,
+      "step": 198
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9753615861570338e-05,
+      "loss": 0.7273,
+      "step": 199
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.975004714331994e-05,
+      "loss": 0.6732,
+      "step": 200
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9746453092299042e-05,
+      "loss": 0.8208,
+      "step": 201
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.974283371784578e-05,
+      "loss": 0.6795,
+      "step": 202
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9739189029364083e-05,
+      "loss": 0.8884,
+      "step": 203
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9735519036323656e-05,
+      "loss": 0.7186,
+      "step": 204
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9731823748259953e-05,
+      "loss": 0.7591,
+      "step": 205
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.972810317477414e-05,
+      "loss": 0.7703,
+      "step": 206
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.972435732553309e-05,
+      "loss": 0.9294,
+      "step": 207
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9720586210269347e-05,
+      "loss": 0.6991,
+      "step": 208
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9716789838781095e-05,
+      "loss": 0.662,
+      "step": 209
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9712968220932144e-05,
+      "loss": 0.6357,
+      "step": 210
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.97091213666519e-05,
+      "loss": 0.6902,
+      "step": 211
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9705249285935344e-05,
+      "loss": 0.7826,
+      "step": 212
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9701351988842987e-05,
+      "loss": 0.6604,
+      "step": 213
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9697429485500862e-05,
+      "loss": 0.6934,
+      "step": 214
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9693481786100506e-05,
+      "loss": 0.7027,
+      "step": 215
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.9689508900898907e-05,
+      "loss": 0.7562,
+      "step": 216
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9685510840218497e-05,
+      "loss": 0.7664,
+      "step": 217
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.968148761444712e-05,
+      "loss": 0.5945,
+      "step": 218
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9677439234038004e-05,
+      "loss": 0.7876,
+      "step": 219
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9673365709509723e-05,
+      "loss": 0.6742,
+      "step": 220
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9669267051446208e-05,
+      "loss": 0.9253,
+      "step": 221
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9665143270496658e-05,
+      "loss": 0.6862,
+      "step": 222
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.9660994377375578e-05,
+      "loss": 0.7811,
+      "step": 223
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.96568203828627e-05,
+      "loss": 0.7367,
+      "step": 224
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.965262129780298e-05,
+      "loss": 0.68,
+      "step": 225
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.964839713310657e-05,
+      "loss": 0.7343,
+      "step": 226
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9644147899748784e-05,
+      "loss": 0.6716,
+      "step": 227
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.963987360877005e-05,
+      "loss": 0.6795,
+      "step": 228
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.963557427127594e-05,
+      "loss": 0.69,
+      "step": 229
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9631249898437066e-05,
+      "loss": 0.7206,
+      "step": 230
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9626900501489102e-05,
+      "loss": 0.679,
+      "step": 231
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9622526091732745e-05,
+      "loss": 0.7709,
+      "step": 232
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9618126680533672e-05,
+      "loss": 0.7466,
+      "step": 233
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9613702279322518e-05,
+      "loss": 0.6064,
+      "step": 234
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9609252899594863e-05,
+      "loss": 0.6615,
+      "step": 235
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.9604778552911167e-05,
+      "loss": 0.7465,
+      "step": 236
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.960027925089677e-05,
+      "loss": 0.7212,
+      "step": 237
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9595755005241853e-05,
+      "loss": 0.8214,
+      "step": 238
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.95912058277014e-05,
+      "loss": 0.6861,
+      "step": 239
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9586631730095175e-05,
+      "loss": 0.7779,
+      "step": 240
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.95820327243077e-05,
+      "loss": 0.686,
+      "step": 241
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9577408822288193e-05,
+      "loss": 0.679,
+      "step": 242
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9572760036050577e-05,
+      "loss": 0.7466,
+      "step": 243
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9568086377673422e-05,
+      "loss": 0.8037,
+      "step": 244
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.956338785929992e-05,
+      "loss": 0.6762,
+      "step": 245
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.9558664493137863e-05,
+      "loss": 0.8955,
+      "step": 246
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.955391629145959e-05,
+      "loss": 0.7347,
+      "step": 247
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9549143266601977e-05,
+      "loss": 0.7446,
+      "step": 248
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9544345430966398e-05,
+      "loss": 0.7133,
+      "step": 249
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9539522797018682e-05,
+      "loss": 0.7182,
+      "step": 250
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9534675377289094e-05,
+      "loss": 0.6604,
+      "step": 251
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9529803184372302e-05,
+      "loss": 0.6004,
+      "step": 252
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9524906230927338e-05,
+      "loss": 0.6171,
+      "step": 253
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.951998452967756e-05,
+      "loss": 0.6518,
+      "step": 254
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9515038093410633e-05,
+      "loss": 0.7039,
+      "step": 255
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.9510066934978496e-05,
+      "loss": 0.679,
+      "step": 256
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9505071067297304e-05,
+      "loss": 0.7146,
+      "step": 257
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.950005050334743e-05,
+      "loss": 0.6795,
+      "step": 258
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9495005256173398e-05,
+      "loss": 0.7601,
+      "step": 259
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9489935338883876e-05,
+      "loss": 0.641,
+      "step": 260
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9484840764651624e-05,
+      "loss": 0.6869,
+      "step": 261
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9479721546713472e-05,
+      "loss": 0.6668,
+      "step": 262
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.947457769837027e-05,
+      "loss": 0.6938,
+      "step": 263
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9469409232986876e-05,
+      "loss": 0.6747,
+      "step": 264
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.9464216163992097e-05,
+      "loss": 0.6927,
+      "step": 265
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.945899850487867e-05,
+      "loss": 0.7093,
+      "step": 266
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9453756269203225e-05,
+      "loss": 0.6612,
+      "step": 267
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9448489470586245e-05,
+      "loss": 0.6994,
+      "step": 268
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9443198122712036e-05,
+      "loss": 0.6974,
+      "step": 269
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9437882239328675e-05,
+      "loss": 0.7176,
+      "step": 270
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9432541834248014e-05,
+      "loss": 0.6262,
+      "step": 271
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9427176921345587e-05,
+      "loss": 0.7263,
+      "step": 272
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9421787514560625e-05,
+      "loss": 0.6594,
+      "step": 273
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9416373627896002e-05,
+      "loss": 0.6773,
+      "step": 274
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.9410935275418177e-05,
+      "loss": 0.7264,
+      "step": 275
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.940547247125719e-05,
+      "loss": 0.7352,
+      "step": 276
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9399985229606616e-05,
+      "loss": 0.6537,
+      "step": 277
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9394473564723515e-05,
+      "loss": 0.673,
+      "step": 278
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9388937490928402e-05,
+      "loss": 0.6809,
+      "step": 279
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.938337702260522e-05,
+      "loss": 0.6064,
+      "step": 280
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9377792174201295e-05,
+      "loss": 0.6943,
+      "step": 281
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9372182960227284e-05,
+      "loss": 0.7056,
+      "step": 282
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9366549395257167e-05,
+      "loss": 0.6422,
+      "step": 283
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9360891493928186e-05,
+      "loss": 0.7447,
+      "step": 284
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.9355209270940817e-05,
+      "loss": 0.7322,
+      "step": 285
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.934950274105872e-05,
+      "loss": 0.6957,
+      "step": 286
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9343771919108726e-05,
+      "loss": 0.6163,
+      "step": 287
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.933801681998077e-05,
+      "loss": 0.6324,
+      "step": 288
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.933223745862786e-05,
+      "loss": 0.7091,
+      "step": 289
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.932643385006606e-05,
+      "loss": 0.6288,
+      "step": 290
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9320606009374418e-05,
+      "loss": 0.7056,
+      "step": 291
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9314753951694952e-05,
+      "loss": 0.6282,
+      "step": 292
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9308877692232592e-05,
+      "loss": 0.6591,
+      "step": 293
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.930297724625516e-05,
+      "loss": 0.6914,
+      "step": 294
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9297052629093312e-05,
+      "loss": 0.6352,
+      "step": 295
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.929110385614051e-05,
+      "loss": 0.6771,
+      "step": 296
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9285130942852975e-05,
+      "loss": 0.7305,
+      "step": 297
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.927913390474965e-05,
+      "loss": 0.7441,
+      "step": 298
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9273112757412165e-05,
+      "loss": 0.6964,
+      "step": 299
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.926706751648479e-05,
+      "loss": 0.6472,
+      "step": 300
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9260998197674385e-05,
+      "loss": 0.6361,
+      "step": 301
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9254904816750376e-05,
+      "loss": 0.689,
+      "step": 302
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9248787389544722e-05,
+      "loss": 0.6399,
+      "step": 303
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9242645931951833e-05,
+      "loss": 0.6407,
+      "step": 304
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9236480459928573e-05,
+      "loss": 0.7198,
+      "step": 305
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9230290989494203e-05,
+      "loss": 0.6972,
+      "step": 306
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9224077536730323e-05,
+      "loss": 0.6848,
+      "step": 307
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9217840117780857e-05,
+      "loss": 0.6266,
+      "step": 308
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.921157874885199e-05,
+      "loss": 0.7203,
+      "step": 309
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9205293446212137e-05,
+      "loss": 0.6186,
+      "step": 310
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9198984226191905e-05,
+      "loss": 0.7296,
+      "step": 311
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9192651105184032e-05,
+      "loss": 0.6269,
+      "step": 312
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9186294099643367e-05,
+      "loss": 0.7291,
+      "step": 313
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.91799132260868e-05,
+      "loss": 0.7009,
+      "step": 314
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.917350850109326e-05,
+      "loss": 0.6599,
+      "step": 315
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.916707994130363e-05,
+      "loss": 0.6374,
+      "step": 316
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9160627563420714e-05,
+      "loss": 0.6662,
+      "step": 317
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.915415138420922e-05,
+      "loss": 0.6952,
+      "step": 318
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9147651420495696e-05,
+      "loss": 0.6496,
+      "step": 319
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9141127689168465e-05,
+      "loss": 0.6784,
+      "step": 320
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9134580207177625e-05,
+      "loss": 0.6576,
+      "step": 321
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9128008991534975e-05,
+      "loss": 0.6595,
+      "step": 322
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9121414059313985e-05,
+      "loss": 0.6332,
+      "step": 323
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9114795427649735e-05,
+      "loss": 0.6668,
+      "step": 324
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9108153113738896e-05,
+      "loss": 0.6232,
+      "step": 325
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9101487134839657e-05,
+      "loss": 0.6506,
+      "step": 326
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9094797508271702e-05,
+      "loss": 0.6441,
+      "step": 327
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9088084251416154e-05,
+      "loss": 0.6531,
+      "step": 328
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9081347381715535e-05,
+      "loss": 0.6683,
+      "step": 329
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9074586916673716e-05,
+      "loss": 0.6775,
+      "step": 330
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9067802873855875e-05,
+      "loss": 0.6849,
+      "step": 331
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9060995270888444e-05,
+      "loss": 0.6931,
+      "step": 332
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9054164125459084e-05,
+      "loss": 0.6496,
+      "step": 333
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.904730945531661e-05,
+      "loss": 0.66,
+      "step": 334
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9040431278270967e-05,
+      "loss": 0.6834,
+      "step": 335
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9033529612193177e-05,
+      "loss": 0.7538,
+      "step": 336
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9026604475015283e-05,
+      "loss": 0.5976,
+      "step": 337
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.901965588473032e-05,
+      "loss": 0.6549,
+      "step": 338
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.901268385939226e-05,
+      "loss": 0.6582,
+      "step": 339
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9005688417115954e-05,
+      "loss": 0.6617,
+      "step": 340
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.899866957607711e-05,
+      "loss": 0.6522,
+      "step": 341
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8991627354512213e-05,
+      "loss": 0.976,
+      "step": 342
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.898456177071852e-05,
+      "loss": 0.6434,
+      "step": 343
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8977472843053962e-05,
+      "loss": 0.676,
+      "step": 344
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8970360589937138e-05,
+      "loss": 0.7308,
+      "step": 345
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.8963225029847252e-05,
+      "loss": 0.5997,
+      "step": 346
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.895606618132406e-05,
+      "loss": 0.7292,
+      "step": 347
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8948884062967823e-05,
+      "loss": 0.6467,
+      "step": 348
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8941678693439272e-05,
+      "loss": 0.5916,
+      "step": 349
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8934450091459544e-05,
+      "loss": 0.6801,
+      "step": 350
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.892719827581014e-05,
+      "loss": 0.6193,
+      "step": 351
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8919923265332865e-05,
+      "loss": 0.7392,
+      "step": 352
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8912625078929814e-05,
+      "loss": 0.7519,
+      "step": 353
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8905303735563274e-05,
+      "loss": 0.6561,
+      "step": 354
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8897959254255715e-05,
+      "loss": 0.7031,
+      "step": 355
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.8890591654089705e-05,
+      "loss": 0.6356,
+      "step": 356
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8883200954207903e-05,
+      "loss": 0.6446,
+      "step": 357
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.887578717381297e-05,
+      "loss": 0.6155,
+      "step": 358
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.886835033216755e-05,
+      "loss": 0.7146,
+      "step": 359
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8860890448594182e-05,
+      "loss": 0.6414,
+      "step": 360
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.88534075424753e-05,
+      "loss": 0.6314,
+      "step": 361
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8845901633253132e-05,
+      "loss": 0.7255,
+      "step": 362
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8838372740429693e-05,
+      "loss": 0.6845,
+      "step": 363
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.88308208835667e-05,
+      "loss": 0.6271,
+      "step": 364
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.882324608228554e-05,
+      "loss": 0.6901,
+      "step": 365
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.881564835626722e-05,
+      "loss": 0.6465,
+      "step": 366
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.8808027725252297e-05,
+      "loss": 0.6494,
+      "step": 367
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8800384209040864e-05,
+      "loss": 0.6207,
+      "step": 368
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8792717827492446e-05,
+      "loss": 0.6646,
+      "step": 369
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8785028600526e-05,
+      "loss": 0.657,
+      "step": 370
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.877731654811983e-05,
+      "loss": 0.7176,
+      "step": 371
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.876958169031154e-05,
+      "loss": 0.6358,
+      "step": 372
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8761824047198003e-05,
+      "loss": 0.7008,
+      "step": 373
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8754043638935283e-05,
+      "loss": 0.6676,
+      "step": 374
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.8746240485738595e-05,
+      "loss": 0.6825,
+      "step": 375
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.873841460788225e-05,
+      "loss": 0.6394,
+      "step": 376
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.87305660256996e-05,
+      "loss": 0.6404,
+      "step": 377
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8722694759582992e-05,
+      "loss": 0.6467,
+      "step": 378
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.871480082998371e-05,
+      "loss": 0.6852,
+      "step": 379
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8706884257411925e-05,
+      "loss": 0.615,
+      "step": 380
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8698945062436633e-05,
+      "loss": 0.6319,
+      "step": 381
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.869098326568561e-05,
+      "loss": 0.6404,
+      "step": 382
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.868299888784536e-05,
+      "loss": 0.6251,
+      "step": 383
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.867499194966106e-05,
+      "loss": 0.6187,
+      "step": 384
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8666962471936498e-05,
+      "loss": 0.6202,
+      "step": 385
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.8658910475534016e-05,
+      "loss": 0.6843,
+      "step": 386
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.865083598137449e-05,
+      "loss": 0.593,
+      "step": 387
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.864273901043722e-05,
+      "loss": 0.6261,
+      "step": 388
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8634619583759933e-05,
+      "loss": 0.6289,
+      "step": 389
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.862647772243868e-05,
+      "loss": 0.7258,
+      "step": 390
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8618313447627815e-05,
+      "loss": 0.5747,
+      "step": 391
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.861012678053992e-05,
+      "loss": 0.6668,
+      "step": 392
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.860191774244576e-05,
+      "loss": 0.6911,
+      "step": 393
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8593686354674223e-05,
+      "loss": 0.6535,
+      "step": 394
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8585432638612272e-05,
+      "loss": 0.6716,
+      "step": 395
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.8577156615704877e-05,
+      "loss": 0.6456,
+      "step": 396
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.856885830745497e-05,
+      "loss": 0.6516,
+      "step": 397
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.856053773542338e-05,
+      "loss": 0.6124,
+      "step": 398
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8552194921228793e-05,
+      "loss": 0.6507,
+      "step": 399
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8543829886547674e-05,
+      "loss": 0.6223,
+      "step": 400
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8535442653114228e-05,
+      "loss": 0.6274,
+      "step": 401
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8527033242720328e-05,
+      "loss": 0.6839,
+      "step": 402
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8518601677215488e-05,
+      "loss": 0.675,
+      "step": 403
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.851014797850676e-05,
+      "loss": 0.657,
+      "step": 404
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.8501672168558726e-05,
+      "loss": 0.695,
+      "step": 405
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.84931742693934e-05,
+      "loss": 0.6864,
+      "step": 406
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.84846543030902e-05,
+      "loss": 0.6229,
+      "step": 407
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8476112291785876e-05,
+      "loss": 0.6588,
+      "step": 408
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8467548257674453e-05,
+      "loss": 0.6368,
+      "step": 409
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8458962223007177e-05,
+      "loss": 0.6543,
+      "step": 410
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.845035421009246e-05,
+      "loss": 0.6985,
+      "step": 411
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.844172424129582e-05,
+      "loss": 0.6677,
+      "step": 412
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.843307233903981e-05,
+      "loss": 0.7123,
+      "step": 413
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8424398525803983e-05,
+      "loss": 0.6168,
+      "step": 414
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8415702824124812e-05,
+      "loss": 0.7392,
+      "step": 415
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8406985256595652e-05,
+      "loss": 0.6636,
+      "step": 416
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.8398245845866657e-05,
+      "loss": 0.6216,
+      "step": 417
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8389484614644746e-05,
+      "loss": 0.6685,
+      "step": 418
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8380701585693526e-05,
+      "loss": 0.5938,
+      "step": 419
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8371896781833242e-05,
+      "loss": 0.6134,
+      "step": 420
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8363070225940714e-05,
+      "loss": 0.637,
+      "step": 421
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8354221940949282e-05,
+      "loss": 0.8702,
+      "step": 422
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8345351949848735e-05,
+      "loss": 0.7037,
+      "step": 423
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8336460275685267e-05,
+      "loss": 0.6907,
+      "step": 424
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8327546941561403e-05,
+      "loss": 0.7027,
+      "step": 425
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8318611970635953e-05,
+      "loss": 0.6431,
+      "step": 426
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.8309655386123935e-05,
+      "loss": 0.7247,
+      "step": 427
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8300677211296532e-05,
+      "loss": 0.6558,
+      "step": 428
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8291677469481025e-05,
+      "loss": 0.7563,
+      "step": 429
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8282656184060713e-05,
+      "loss": 0.6238,
+      "step": 430
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.82736133784749e-05,
+      "loss": 0.6946,
+      "step": 431
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.826454907621877e-05,
+      "loss": 0.6227,
+      "step": 432
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.825546330084339e-05,
+      "loss": 0.6576,
+      "step": 433
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8246356075955594e-05,
+      "loss": 0.6126,
+      "step": 434
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8237227425217976e-05,
+      "loss": 0.7517,
+      "step": 435
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8228077372348765e-05,
+      "loss": 0.6504,
+      "step": 436
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.8218905941121824e-05,
+      "loss": 0.6934,
+      "step": 437
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8209713155366546e-05,
+      "loss": 0.614,
+      "step": 438
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.820049903896782e-05,
+      "loss": 0.6809,
+      "step": 439
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8191263615865942e-05,
+      "loss": 0.6117,
+      "step": 440
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8182006910056582e-05,
+      "loss": 0.62,
+      "step": 441
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8172728945590702e-05,
+      "loss": 0.7094,
+      "step": 442
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8163429746574493e-05,
+      "loss": 0.6945,
+      "step": 443
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8154109337169326e-05,
+      "loss": 0.6693,
+      "step": 444
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8144767741591676e-05,
+      "loss": 0.5903,
+      "step": 445
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.8135404984113067e-05,
+      "loss": 0.643,
+      "step": 446
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.812602108906001e-05,
+      "loss": 0.6264,
+      "step": 447
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8116616080813925e-05,
+      "loss": 0.6724,
+      "step": 448
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8107189983811094e-05,
+      "loss": 0.6087,
+      "step": 449
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8097742822542605e-05,
+      "loss": 0.6531,
+      "step": 450
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.808827462155425e-05,
+      "loss": 0.5679,
+      "step": 451
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8078785405446516e-05,
+      "loss": 0.7,
+      "step": 452
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8069275198874464e-05,
+      "loss": 0.6416,
+      "step": 453
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8059744026547713e-05,
+      "loss": 0.6474,
+      "step": 454
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.8050191913230345e-05,
+      "loss": 0.6297,
+      "step": 455
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.804061888374086e-05,
+      "loss": 0.6923,
+      "step": 456
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.803102496295209e-05,
+      "loss": 0.6188,
+      "step": 457
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.802141017579116e-05,
+      "loss": 0.5695,
+      "step": 458
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.8011774547239403e-05,
+      "loss": 0.6514,
+      "step": 459
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.8002118102332308e-05,
+      "loss": 0.6491,
+      "step": 460
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7992440866159443e-05,
+      "loss": 0.68,
+      "step": 461
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.79827428638644e-05,
+      "loss": 0.6495,
+      "step": 462
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.797302412064473e-05,
+      "loss": 0.6987,
+      "step": 463
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.796328466175186e-05,
+      "loss": 0.6891,
+      "step": 464
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.795352451249106e-05,
+      "loss": 0.6765,
+      "step": 465
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7943743698221337e-05,
+      "loss": 0.6371,
+      "step": 466
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.7933942244355415e-05,
+      "loss": 0.6979,
+      "step": 467
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.792412017635962e-05,
+      "loss": 0.6372,
+      "step": 468
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.791427751975385e-05,
+      "loss": 0.6674,
+      "step": 469
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7904414300111497e-05,
+      "loss": 0.6552,
+      "step": 470
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.789453054305938e-05,
+      "loss": 0.6788,
+      "step": 471
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7884626274277674e-05,
+      "loss": 0.7142,
+      "step": 472
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7874701519499855e-05,
+      "loss": 0.5597,
+      "step": 473
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.786475630451262e-05,
+      "loss": 0.6239,
+      "step": 474
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.7854790655155827e-05,
+      "loss": 0.6677,
+      "step": 475
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.784480459732243e-05,
+      "loss": 0.6527,
+      "step": 476
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.783479815695841e-05,
+      "loss": 0.6342,
+      "step": 477
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.78247713600627e-05,
+      "loss": 0.6251,
+      "step": 478
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.781472423268713e-05,
+      "loss": 0.6716,
+      "step": 479
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.780465680093635e-05,
+      "loss": 0.6483,
+      "step": 480
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7794569090967763e-05,
+      "loss": 0.6613,
+      "step": 481
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7784461128991465e-05,
+      "loss": 0.64,
+      "step": 482
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7774332941270165e-05,
+      "loss": 0.6742,
+      "step": 483
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.776418455411913e-05,
+      "loss": 0.6902,
+      "step": 484
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7754015993906103e-05,
+      "loss": 0.6196,
+      "step": 485
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.774382728705124e-05,
+      "loss": 0.6905,
+      "step": 486
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.7733618460027055e-05,
+      "loss": 0.6993,
+      "step": 487
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7723389539358323e-05,
+      "loss": 0.5793,
+      "step": 488
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7713140551622032e-05,
+      "loss": 0.627,
+      "step": 489
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7702871523447317e-05,
+      "loss": 0.6394,
+      "step": 490
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7692582481515372e-05,
+      "loss": 0.6372,
+      "step": 491
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.768227345255939e-05,
+      "loss": 0.6464,
+      "step": 492
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.767194446336451e-05,
+      "loss": 0.6816,
+      "step": 493
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7661595540767714e-05,
+      "loss": 0.6807,
+      "step": 494
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7651226711657786e-05,
+      "loss": 0.661,
+      "step": 495
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7640838002975223e-05,
+      "loss": 0.5681,
+      "step": 496
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.7630429441712184e-05,
+      "loss": 0.5885,
+      "step": 497
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.76200010549124e-05,
+      "loss": 0.6487,
+      "step": 498
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7609552869671126e-05,
+      "loss": 0.6879,
+      "step": 499
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7599084913135038e-05,
+      "loss": 0.6085,
+      "step": 500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7588597212502204e-05,
+      "loss": 0.6497,
+      "step": 501
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.757808979502197e-05,
+      "loss": 0.6377,
+      "step": 502
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.756756268799493e-05,
+      "loss": 0.6577,
+      "step": 503
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7557015918772822e-05,
+      "loss": 0.59,
+      "step": 504
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.754644951475848e-05,
+      "loss": 0.6151,
+      "step": 505
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7535863503405743e-05,
+      "loss": 0.6877,
+      "step": 506
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.7525257912219412e-05,
+      "loss": 0.6664,
+      "step": 507
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.751463276875514e-05,
+      "loss": 0.6925,
+      "step": 508
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.750398810061939e-05,
+      "loss": 0.7353,
+      "step": 509
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7493323935469364e-05,
+      "loss": 0.6527,
+      "step": 510
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7482640301012904e-05,
+      "loss": 0.6653,
+      "step": 511
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7471937225008447e-05,
+      "loss": 0.6105,
+      "step": 512
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7461214735264947e-05,
+      "loss": 0.663,
+      "step": 513
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.745047285964179e-05,
+      "loss": 0.6139,
+      "step": 514
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7439711626048738e-05,
+      "loss": 0.7019,
+      "step": 515
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.7428931062445844e-05,
+      "loss": 0.5709,
+      "step": 516
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.741813119684339e-05,
+      "loss": 0.6271,
+      "step": 517
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7407312057301807e-05,
+      "loss": 0.6192,
+      "step": 518
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7396473671931597e-05,
+      "loss": 0.5847,
+      "step": 519
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.738561606889328e-05,
+      "loss": 0.6313,
+      "step": 520
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.73747392763973e-05,
+      "loss": 0.6231,
+      "step": 521
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7363843322703956e-05,
+      "loss": 0.5879,
+      "step": 522
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7352928236123337e-05,
+      "loss": 0.6387,
+      "step": 523
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7341994045015245e-05,
+      "loss": 0.6781,
+      "step": 524
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.7331040777789122e-05,
+      "loss": 0.6501,
+      "step": 525
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.732006846290396e-05,
+      "loss": 0.6286,
+      "step": 526
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.730907712886825e-05,
+      "loss": 0.6595,
+      "step": 527
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7298066804239904e-05,
+      "loss": 0.604,
+      "step": 528
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7287037517626174e-05,
+      "loss": 0.6222,
+      "step": 529
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7275989297683575e-05,
+      "loss": 0.6315,
+      "step": 530
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.726492217311781e-05,
+      "loss": 0.6157,
+      "step": 531
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.725383617268371e-05,
+      "loss": 0.6616,
+      "step": 532
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7242731325185152e-05,
+      "loss": 0.5931,
+      "step": 533
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7231607659474972e-05,
+      "loss": 0.658,
+      "step": 534
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7220465204454902e-05,
+      "loss": 0.6186,
+      "step": 535
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7209303989075497e-05,
+      "loss": 0.5955,
+      "step": 536
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.7198124042336055e-05,
+      "loss": 0.6567,
+      "step": 537
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7186925393284538e-05,
+      "loss": 0.6332,
+      "step": 538
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7175708071017503e-05,
+      "loss": 0.628,
+      "step": 539
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7164472104680026e-05,
+      "loss": 0.6144,
+      "step": 540
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.715321752346563e-05,
+      "loss": 0.664,
+      "step": 541
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7141944356616185e-05,
+      "loss": 0.6983,
+      "step": 542
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7130652633421864e-05,
+      "loss": 0.6177,
+      "step": 543
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7119342383221055e-05,
+      "loss": 0.6634,
+      "step": 544
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7108013635400283e-05,
+      "loss": 0.6421,
+      "step": 545
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7096666419394122e-05,
+      "loss": 0.6453,
+      "step": 546
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.7085300764685147e-05,
+      "loss": 0.6599,
+      "step": 547
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.707391670080383e-05,
+      "loss": 0.7101,
+      "step": 548
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.7062514257328474e-05,
+      "loss": 0.6757,
+      "step": 549
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.705109346388514e-05,
+      "loss": 0.6949,
+      "step": 550
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.703965435014757e-05,
+      "loss": 0.6671,
+      "step": 551
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.7028196945837097e-05,
+      "loss": 0.6966,
+      "step": 552
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.7016721280722585e-05,
+      "loss": 0.6081,
+      "step": 553
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.7005227384620336e-05,
+      "loss": 0.6207,
+      "step": 554
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.699371528739403e-05,
+      "loss": 0.569,
+      "step": 555
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.698218501895463e-05,
+      "loss": 0.6214,
+      "step": 556
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6970636609260308e-05,
+      "loss": 0.6936,
+      "step": 557
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.6959070088316384e-05,
+      "loss": 0.6793,
+      "step": 558
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6947485486175223e-05,
+      "loss": 0.587,
+      "step": 559
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6935882832936177e-05,
+      "loss": 0.6241,
+      "step": 560
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6924262158745493e-05,
+      "loss": 0.6018,
+      "step": 561
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.691262349379624e-05,
+      "loss": 0.6667,
+      "step": 562
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6900966868328236e-05,
+      "loss": 0.666,
+      "step": 563
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.688929231262797e-05,
+      "loss": 0.6603,
+      "step": 564
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6877599857028495e-05,
+      "loss": 0.6142,
+      "step": 565
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.6865889531909397e-05,
+      "loss": 0.5954,
+      "step": 566
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.685416136769668e-05,
+      "loss": 0.6811,
+      "step": 567
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.684241539486269e-05,
+      "loss": 0.6814,
+      "step": 568
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.683065164392606e-05,
+      "loss": 0.7413,
+      "step": 569
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.681887014545161e-05,
+      "loss": 0.6803,
+      "step": 570
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.680707093005026e-05,
+      "loss": 0.707,
+      "step": 571
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.679525402837897e-05,
+      "loss": 0.6282,
+      "step": 572
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6783419471140665e-05,
+      "loss": 0.6586,
+      "step": 573
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6771567289084122e-05,
+      "loss": 0.64,
+      "step": 574
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6759697513003926e-05,
+      "loss": 0.6282,
+      "step": 575
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.674781017374037e-05,
+      "loss": 0.5877,
+      "step": 576
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6735905302179377e-05,
+      "loss": 0.6247,
+      "step": 577
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.6723982929252428e-05,
+      "loss": 0.6917,
+      "step": 578
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6712043085936473e-05,
+      "loss": 0.6647,
+      "step": 579
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6700085803253858e-05,
+      "loss": 0.6931,
+      "step": 580
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.668811111227224e-05,
+      "loss": 0.6597,
+      "step": 581
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6676119044104494e-05,
+      "loss": 0.627,
+      "step": 582
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.666410962990867e-05,
+      "loss": 0.5954,
+      "step": 583
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6652082900887858e-05,
+      "loss": 0.6516,
+      "step": 584
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.664003888829016e-05,
+      "loss": 0.6377,
+      "step": 585
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6627977623408566e-05,
+      "loss": 0.6542,
+      "step": 586
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.6615899137580904e-05,
+      "loss": 0.6716,
+      "step": 587
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.660380346218974e-05,
+      "loss": 0.606,
+      "step": 588
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6591690628662305e-05,
+      "loss": 0.616,
+      "step": 589
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.657956066847041e-05,
+      "loss": 0.6054,
+      "step": 590
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6567413613130354e-05,
+      "loss": 0.612,
+      "step": 591
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6555249494202875e-05,
+      "loss": 0.654,
+      "step": 592
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.654306834329303e-05,
+      "loss": 0.7106,
+      "step": 593
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6530870192050134e-05,
+      "loss": 0.6601,
+      "step": 594
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6518655072167666e-05,
+      "loss": 0.6028,
+      "step": 595
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.65064230153832e-05,
+      "loss": 0.6381,
+      "step": 596
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.649417405347832e-05,
+      "loss": 0.6052,
+      "step": 597
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.6481908218278524e-05,
+      "loss": 0.6177,
+      "step": 598
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6469625541653152e-05,
+      "loss": 0.668,
+      "step": 599
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6457326055515305e-05,
+      "loss": 0.6135,
+      "step": 600
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.644500979182176e-05,
+      "loss": 0.6258,
+      "step": 601
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6432676782572883e-05,
+      "loss": 0.6496,
+      "step": 602
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6420327059812547e-05,
+      "loss": 0.5972,
+      "step": 603
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6407960655628055e-05,
+      "loss": 0.6349,
+      "step": 604
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6395577602150043e-05,
+      "loss": 0.6827,
+      "step": 605
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.6383177931552422e-05,
+      "loss": 0.6728,
+      "step": 606
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.637076167605226e-05,
+      "loss": 0.7084,
+      "step": 607
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.635832886790973e-05,
+      "loss": 0.7076,
+      "step": 608
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6345879539428e-05,
+      "loss": 0.597,
+      "step": 609
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6333413722953175e-05,
+      "loss": 0.6083,
+      "step": 610
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.632093145087419e-05,
+      "loss": 0.5898,
+      "step": 611
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.630843275562274e-05,
+      "loss": 0.622,
+      "step": 612
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.629591766967318e-05,
+      "loss": 0.5837,
+      "step": 613
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6283386225542467e-05,
+      "loss": 0.6169,
+      "step": 614
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6270838455790056e-05,
+      "loss": 0.6189,
+      "step": 615
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6258274393017814e-05,
+      "loss": 0.6919,
+      "step": 616
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.6245694069869945e-05,
+      "loss": 0.6176,
+      "step": 617
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.62330975190329e-05,
+      "loss": 0.6348,
+      "step": 618
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.622048477323529e-05,
+      "loss": 0.6459,
+      "step": 619
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.620785586524781e-05,
+      "loss": 0.6236,
+      "step": 620
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.6195210827883146e-05,
+      "loss": 0.6287,
+      "step": 621
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.6182549693995893e-05,
+      "loss": 0.6218,
+      "step": 622
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.6169872496482462e-05,
+      "loss": 0.6691,
+      "step": 623
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.6157179268281007e-05,
+      "loss": 0.653,
+      "step": 624
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.614447004237133e-05,
+      "loss": 0.6649,
+      "step": 625
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.61317448517748e-05,
+      "loss": 0.6055,
+      "step": 626
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.611900372955427e-05,
+      "loss": 0.7133,
+      "step": 627
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.6106246708813973e-05,
+      "loss": 0.661,
+      "step": 628
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6093473822699467e-05,
+      "loss": 0.6101,
+      "step": 629
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.608068510439752e-05,
+      "loss": 0.6209,
+      "step": 630
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6067880587136047e-05,
+      "loss": 0.6485,
+      "step": 631
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6055060304184e-05,
+      "loss": 0.6972,
+      "step": 632
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6042224288851292e-05,
+      "loss": 0.6392,
+      "step": 633
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6029372574488732e-05,
+      "loss": 0.5941,
+      "step": 634
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.6016505194487894e-05,
+      "loss": 0.6168,
+      "step": 635
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.600362218228107e-05,
+      "loss": 0.6962,
+      "step": 636
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5990723571341167e-05,
+      "loss": 0.6139,
+      "step": 637
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.5977809395181615e-05,
+      "loss": 0.6416,
+      "step": 638
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5964879687356286e-05,
+      "loss": 0.5989,
+      "step": 639
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.595193448145941e-05,
+      "loss": 0.6464,
+      "step": 640
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5938973811125493e-05,
+      "loss": 0.6293,
+      "step": 641
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.59259977100292e-05,
+      "loss": 0.6367,
+      "step": 642
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.591300621188531e-05,
+      "loss": 0.6843,
+      "step": 643
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.589999935044859e-05,
+      "loss": 0.6632,
+      "step": 644
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5886977159513732e-05,
+      "loss": 0.6747,
+      "step": 645
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5873939672915263e-05,
+      "loss": 0.6435,
+      "step": 646
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.586088692452744e-05,
+      "loss": 0.6339,
+      "step": 647
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.5847818948264183e-05,
+      "loss": 0.6781,
+      "step": 648
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5834735778078968e-05,
+      "loss": 0.6314,
+      "step": 649
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5821637447964757e-05,
+      "loss": 0.6327,
+      "step": 650
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.58085239919539e-05,
+      "loss": 0.5932,
+      "step": 651
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5795395444118037e-05,
+      "loss": 0.6264,
+      "step": 652
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5782251838568034e-05,
+      "loss": 0.574,
+      "step": 653
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5769093209453876e-05,
+      "loss": 0.6047,
+      "step": 654
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5755919590964572e-05,
+      "loss": 0.6573,
+      "step": 655
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.5742731017328087e-05,
+      "loss": 0.6067,
+      "step": 656
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.572952752281125e-05,
+      "loss": 0.7343,
+      "step": 657
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.571630914171964e-05,
+      "loss": 0.598,
+      "step": 658
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5703075908397523e-05,
+      "loss": 0.6209,
+      "step": 659
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5689827857227755e-05,
+      "loss": 0.6603,
+      "step": 660
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5676565022631696e-05,
+      "loss": 0.6373,
+      "step": 661
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5663287439069114e-05,
+      "loss": 0.6221,
+      "step": 662
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5649995141038088e-05,
+      "loss": 0.6442,
+      "step": 663
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.563668816307494e-05,
+      "loss": 0.6095,
+      "step": 664
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.562336653975413e-05,
+      "loss": 0.625,
+      "step": 665
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.5610030305688173e-05,
+      "loss": 0.6606,
+      "step": 666
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.559667949552754e-05,
+      "loss": 0.684,
+      "step": 667
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.558331414396058e-05,
+      "loss": 0.6011,
+      "step": 668
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.556993428571342e-05,
+      "loss": 0.6515,
+      "step": 669
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5556539955549878e-05,
+      "loss": 0.6045,
+      "step": 670
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5543131188271374e-05,
+      "loss": 0.6265,
+      "step": 671
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5529708018716846e-05,
+      "loss": 0.6261,
+      "step": 672
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5516270481762634e-05,
+      "loss": 0.6576,
+      "step": 673
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.550281861232243e-05,
+      "loss": 0.6596,
+      "step": 674
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.548935244534715e-05,
+      "loss": 0.6664,
+      "step": 675
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.547587201582486e-05,
+      "loss": 0.6487,
+      "step": 676
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5462377358780693e-05,
+      "loss": 0.5812,
+      "step": 677
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.5448868509276736e-05,
+      "loss": 0.6703,
+      "step": 678
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5435345502411956e-05,
+      "loss": 0.6343,
+      "step": 679
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5421808373322105e-05,
+      "loss": 0.6391,
+      "step": 680
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5408257157179627e-05,
+      "loss": 0.6565,
+      "step": 681
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5394691889193564e-05,
+      "loss": 0.6323,
+      "step": 682
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.538111260460948e-05,
+      "loss": 0.6661,
+      "step": 683
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.536751933870934e-05,
+      "loss": 0.654,
+      "step": 684
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5353912126811435e-05,
+      "loss": 0.6332,
+      "step": 685
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.534029100427032e-05,
+      "loss": 0.6235,
+      "step": 686
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.5326656006476658e-05,
+      "loss": 0.5774,
+      "step": 687
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.531300716885718e-05,
+      "loss": 0.6491,
+      "step": 688
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5299344526874576e-05,
+      "loss": 0.6699,
+      "step": 689
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5285668116027397e-05,
+      "loss": 0.645,
+      "step": 690
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5271977971849973e-05,
+      "loss": 0.6511,
+      "step": 691
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5258274129912311e-05,
+      "loss": 0.6154,
+      "step": 692
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5244556625820015e-05,
+      "loss": 0.5958,
+      "step": 693
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5230825495214184e-05,
+      "loss": 0.6602,
+      "step": 694
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5217080773771315e-05,
+      "loss": 0.6737,
+      "step": 695
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5203322497203228e-05,
+      "loss": 0.647,
+      "step": 696
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5189550701256955e-05,
+      "loss": 0.6099,
+      "step": 697
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.5175765421714652e-05,
+      "loss": 0.622,
+      "step": 698
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5161966694393516e-05,
+      "loss": 0.5852,
+      "step": 699
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5148154555145684e-05,
+      "loss": 0.654,
+      "step": 700
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.513432903985813e-05,
+      "loss": 0.6009,
+      "step": 701
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.512049018445259e-05,
+      "loss": 0.6546,
+      "step": 702
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5106638024885465e-05,
+      "loss": 0.6528,
+      "step": 703
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5092772597147707e-05,
+      "loss": 0.597,
+      "step": 704
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.507889393726476e-05,
+      "loss": 0.5966,
+      "step": 705
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5065002081296443e-05,
+      "loss": 0.6459,
+      "step": 706
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.5051097065336846e-05,
+      "loss": 0.6364,
+      "step": 707
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.503717892551427e-05,
+      "loss": 0.7106,
+      "step": 708
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5023247697991114e-05,
+      "loss": 0.6433,
+      "step": 709
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.5009303418963772e-05,
+      "loss": 0.7182,
+      "step": 710
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4995346124662551e-05,
+      "loss": 0.5917,
+      "step": 711
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4981375851351579e-05,
+      "loss": 0.6753,
+      "step": 712
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4967392635328702e-05,
+      "loss": 0.6453,
+      "step": 713
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4953396512925398e-05,
+      "loss": 0.6367,
+      "step": 714
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4939387520506675e-05,
+      "loss": 0.6132,
+      "step": 715
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.492536569447098e-05,
+      "loss": 0.6269,
+      "step": 716
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4911331071250113e-05,
+      "loss": 0.6404,
+      "step": 717
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.4897283687309107e-05,
+      "loss": 0.6296,
+      "step": 718
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4883223579146167e-05,
+      "loss": 0.6437,
+      "step": 719
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4869150783292552e-05,
+      "loss": 0.6409,
+      "step": 720
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4855065336312482e-05,
+      "loss": 0.5544,
+      "step": 721
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4840967274803055e-05,
+      "loss": 0.6783,
+      "step": 722
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.482685663539414e-05,
+      "loss": 0.6072,
+      "step": 723
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4812733454748283e-05,
+      "loss": 0.6093,
+      "step": 724
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4798597769560623e-05,
+      "loss": 0.5691,
+      "step": 725
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4784449616558785e-05,
+      "loss": 0.6241,
+      "step": 726
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.4770289032502785e-05,
+      "loss": 0.6191,
+      "step": 727
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.475611605418494e-05,
+      "loss": 0.8272,
+      "step": 728
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4741930718429772e-05,
+      "loss": 0.5563,
+      "step": 729
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4727733062093905e-05,
+      "loss": 0.6515,
+      "step": 730
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4713523122065981e-05,
+      "loss": 0.6024,
+      "step": 731
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4699300935266557e-05,
+      "loss": 0.6459,
+      "step": 732
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4685066538648e-05,
+      "loss": 0.6279,
+      "step": 733
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4670819969194416e-05,
+      "loss": 0.637,
+      "step": 734
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4656561263921529e-05,
+      "loss": 0.6378,
+      "step": 735
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.464229045987659e-05,
+      "loss": 0.5813,
+      "step": 736
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4628007594138307e-05,
+      "loss": 0.6014,
+      "step": 737
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4613712703816696e-05,
+      "loss": 0.5909,
+      "step": 738
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.4599405826053039e-05,
+      "loss": 0.6687,
+      "step": 739
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4585086998019757e-05,
+      "loss": 0.6203,
+      "step": 740
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4570756256920318e-05,
+      "loss": 0.5934,
+      "step": 741
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4556413639989142e-05,
+      "loss": 0.6561,
+      "step": 742
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4542059184491513e-05,
+      "loss": 0.59,
+      "step": 743
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4527692927723465e-05,
+      "loss": 0.69,
+      "step": 744
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4513314907011698e-05,
+      "loss": 0.6064,
+      "step": 745
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.449892515971348e-05,
+      "loss": 0.6644,
+      "step": 746
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4484523723216542e-05,
+      "loss": 0.6059,
+      "step": 747
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.447011063493899e-05,
+      "loss": 0.6174,
+      "step": 748
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.4455685932329204e-05,
+      "loss": 0.636,
+      "step": 749
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4441249652865737e-05,
+      "loss": 0.5573,
+      "step": 750
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4426801834057224e-05,
+      "loss": 0.6531,
+      "step": 751
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4412342513442283e-05,
+      "loss": 0.5889,
+      "step": 752
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.439787172858941e-05,
+      "loss": 0.606,
+      "step": 753
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4383389517096899e-05,
+      "loss": 0.6039,
+      "step": 754
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4368895916592723e-05,
+      "loss": 0.619,
+      "step": 755
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4354390964734444e-05,
+      "loss": 0.6378,
+      "step": 756
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.433987469920913e-05,
+      "loss": 0.5943,
+      "step": 757
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4325347157733232e-05,
+      "loss": 0.6004,
+      "step": 758
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.4310808378052506e-05,
+      "loss": 0.6814,
+      "step": 759
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4296258397941905e-05,
+      "loss": 0.6315,
+      "step": 760
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4281697255205478e-05,
+      "loss": 0.6236,
+      "step": 761
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4267124987676288e-05,
+      "loss": 0.6036,
+      "step": 762
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4252541633216295e-05,
+      "loss": 0.6046,
+      "step": 763
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4237947229716262e-05,
+      "loss": 0.5763,
+      "step": 764
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4223341815095672e-05,
+      "loss": 0.5621,
+      "step": 765
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4208725427302606e-05,
+      "loss": 0.5938,
+      "step": 766
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4194098104313656e-05,
+      "loss": 0.6015,
+      "step": 767
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4179459884133836e-05,
+      "loss": 0.6045,
+      "step": 768
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.4164810804796464e-05,
+      "loss": 0.6367,
+      "step": 769
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4150150904363072e-05,
+      "loss": 0.6264,
+      "step": 770
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.413548022092332e-05,
+      "loss": 0.6353,
+      "step": 771
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4120798792594868e-05,
+      "loss": 0.5893,
+      "step": 772
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4106106657523301e-05,
+      "loss": 0.6197,
+      "step": 773
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.409140385388203e-05,
+      "loss": 0.5622,
+      "step": 774
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.407669041987217e-05,
+      "loss": 0.6142,
+      "step": 775
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4061966393722468e-05,
+      "loss": 0.6106,
+      "step": 776
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4047231813689193e-05,
+      "loss": 0.6018,
+      "step": 777
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4032486718056016e-05,
+      "loss": 0.596,
+      "step": 778
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.4017731145133955e-05,
+      "loss": 0.6388,
+      "step": 779
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.4002965133261238e-05,
+      "loss": 0.6808,
+      "step": 780
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3988188720803213e-05,
+      "loss": 0.6211,
+      "step": 781
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.397340194615226e-05,
+      "loss": 0.6149,
+      "step": 782
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3958604847727673e-05,
+      "loss": 0.5746,
+      "step": 783
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3943797463975575e-05,
+      "loss": 0.6059,
+      "step": 784
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3928979833368813e-05,
+      "loss": 0.6003,
+      "step": 785
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3914151994406852e-05,
+      "loss": 0.6174,
+      "step": 786
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3899313985615687e-05,
+      "loss": 0.6488,
+      "step": 787
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3884465845547734e-05,
+      "loss": 0.6069,
+      "step": 788
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.3869607612781733e-05,
+      "loss": 0.6358,
+      "step": 789
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.385473932592264e-05,
+      "loss": 0.6356,
+      "step": 790
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.383986102360155e-05,
+      "loss": 0.5956,
+      "step": 791
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3824972744475557e-05,
+      "loss": 0.6346,
+      "step": 792
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3810074527227703e-05,
+      "loss": 0.6317,
+      "step": 793
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3795166410566834e-05,
+      "loss": 0.6343,
+      "step": 794
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3780248433227517e-05,
+      "loss": 0.5908,
+      "step": 795
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.376532063396995e-05,
+      "loss": 0.6344,
+      "step": 796
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3750383051579842e-05,
+      "loss": 0.5779,
+      "step": 797
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.3735435724868323e-05,
+      "loss": 0.6194,
+      "step": 798
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.372047869267184e-05,
+      "loss": 0.6273,
+      "step": 799
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.370551199385206e-05,
+      "loss": 0.6385,
+      "step": 800
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3690535667295759e-05,
+      "loss": 0.6242,
+      "step": 801
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3675549751914736e-05,
+      "loss": 0.5952,
+      "step": 802
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3660554286645704e-05,
+      "loss": 0.6523,
+      "step": 803
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.364554931045018e-05,
+      "loss": 0.7132,
+      "step": 804
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.36305348623144e-05,
+      "loss": 0.6442,
+      "step": 805
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.361551098124921e-05,
+      "loss": 0.616,
+      "step": 806
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3600477706289956e-05,
+      "loss": 0.582,
+      "step": 807
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3585435076496407e-05,
+      "loss": 0.6146,
+      "step": 808
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.3570383130952627e-05,
+      "loss": 0.6244,
+      "step": 809
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3555321908766882e-05,
+      "loss": 0.6546,
+      "step": 810
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3540251449071552e-05,
+      "loss": 0.6114,
+      "step": 811
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3525171791023002e-05,
+      "loss": 0.6234,
+      "step": 812
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3510082973801515e-05,
+      "loss": 0.6471,
+      "step": 813
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.349498503661116e-05,
+      "loss": 0.6366,
+      "step": 814
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3479878018679702e-05,
+      "loss": 0.619,
+      "step": 815
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3464761959258506e-05,
+      "loss": 0.5983,
+      "step": 816
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3449636897622416e-05,
+      "loss": 0.6182,
+      "step": 817
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3434502873069683e-05,
+      "loss": 0.6611,
+      "step": 818
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.3419359924921833e-05,
+      "loss": 0.612,
+      "step": 819
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3404208092523584e-05,
+      "loss": 0.6394,
+      "step": 820
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.338904741524273e-05,
+      "loss": 0.6003,
+      "step": 821
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3373877932470057e-05,
+      "loss": 0.6585,
+      "step": 822
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3358699683619216e-05,
+      "loss": 0.6685,
+      "step": 823
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3343512708126642e-05,
+      "loss": 0.6073,
+      "step": 824
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3328317045451455e-05,
+      "loss": 0.6581,
+      "step": 825
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.331311273507532e-05,
+      "loss": 0.6696,
+      "step": 826
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3297899816502391e-05,
+      "loss": 0.6221,
+      "step": 827
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.3282678329259185e-05,
+      "loss": 0.6457,
+      "step": 828
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.326744831289447e-05,
+      "loss": 0.6109,
+      "step": 829
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3252209806979193e-05,
+      "loss": 0.5569,
+      "step": 830
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3236962851106346e-05,
+      "loss": 0.6697,
+      "step": 831
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.322170748489088e-05,
+      "loss": 0.6458,
+      "step": 832
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.32064437479696e-05,
+      "loss": 0.6235,
+      "step": 833
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3191171680001048e-05,
+      "loss": 0.5603,
+      "step": 834
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3175891320665431e-05,
+      "loss": 0.6584,
+      "step": 835
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3160602709664482e-05,
+      "loss": 0.5955,
+      "step": 836
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3145305886721389e-05,
+      "loss": 0.637,
+      "step": 837
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3130000891580661e-05,
+      "loss": 0.607,
+      "step": 838
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3114687764008048e-05,
+      "loss": 0.6458,
+      "step": 839
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.309936654379043e-05,
+      "loss": 0.6418,
+      "step": 840
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3084037270735714e-05,
+      "loss": 0.5792,
+      "step": 841
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3068699984672728e-05,
+      "loss": 0.57,
+      "step": 842
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3053354725451119e-05,
+      "loss": 0.5929,
+      "step": 843
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3038001532941249e-05,
+      "loss": 0.5656,
+      "step": 844
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3022640447034104e-05,
+      "loss": 0.5648,
+      "step": 845
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3007271507641156e-05,
+      "loss": 0.5659,
+      "step": 846
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2991894754694303e-05,
+      "loss": 0.5866,
+      "step": 847
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.297651022814574e-05,
+      "loss": 0.5759,
+      "step": 848
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2961117967967844e-05,
+      "loss": 0.6466,
+      "step": 849
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2945718014153116e-05,
+      "loss": 0.61,
+      "step": 850
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2930310406714011e-05,
+      "loss": 0.6612,
+      "step": 851
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2914895185682899e-05,
+      "loss": 0.6192,
+      "step": 852
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2899472391111916e-05,
+      "loss": 0.6024,
+      "step": 853
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2884042063072881e-05,
+      "loss": 0.6221,
+      "step": 854
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2868604241657187e-05,
+      "loss": 0.6554,
+      "step": 855
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2853158966975693e-05,
+      "loss": 0.6397,
+      "step": 856
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2837706279158626e-05,
+      "loss": 0.6314,
+      "step": 857
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2822246218355475e-05,
+      "loss": 0.6293,
+      "step": 858
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.280677882473488e-05,
+      "loss": 0.6527,
+      "step": 859
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2791304138484539e-05,
+      "loss": 0.6089,
+      "step": 860
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2775822199811097e-05,
+      "loss": 0.6274,
+      "step": 861
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2760333048940043e-05,
+      "loss": 0.5853,
+      "step": 862
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.27448367261156e-05,
+      "loss": 0.6028,
+      "step": 863
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.272933327160063e-05,
+      "loss": 0.6786,
+      "step": 864
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2713822725676526e-05,
+      "loss": 0.6135,
+      "step": 865
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2698305128643099e-05,
+      "loss": 0.6188,
+      "step": 866
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2682780520818488e-05,
+      "loss": 0.6596,
+      "step": 867
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.266724894253904e-05,
+      "loss": 0.6224,
+      "step": 868
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.2651710434159223e-05,
+      "loss": 0.6583,
+      "step": 869
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2636165036051505e-05,
+      "loss": 0.6162,
+      "step": 870
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2620612788606246e-05,
+      "loss": 0.658,
+      "step": 871
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.260505373223162e-05,
+      "loss": 0.5569,
+      "step": 872
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2589487907353484e-05,
+      "loss": 0.5709,
+      "step": 873
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2573915354415274e-05,
+      "loss": 0.5685,
+      "step": 874
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2558336113877916e-05,
+      "loss": 0.6361,
+      "step": 875
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2542750226219711e-05,
+      "loss": 0.6597,
+      "step": 876
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2527157731936228e-05,
+      "loss": 0.5965,
+      "step": 877
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2511558671540198e-05,
+      "loss": 0.5527,
+      "step": 878
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.2495953085561426e-05,
+      "loss": 0.6598,
+      "step": 879
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2480341014546653e-05,
+      "loss": 0.6458,
+      "step": 880
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2464722499059481e-05,
+      "loss": 0.6243,
+      "step": 881
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2449097579680261e-05,
+      "loss": 0.6272,
+      "step": 882
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2433466297005964e-05,
+      "loss": 0.6277,
+      "step": 883
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.241782869165012e-05,
+      "loss": 0.6145,
+      "step": 884
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2402184804242658e-05,
+      "loss": 0.6369,
+      "step": 885
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2386534675429858e-05,
+      "loss": 0.5982,
+      "step": 886
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2370878345874195e-05,
+      "loss": 0.5905,
+      "step": 887
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.235521585625426e-05,
+      "loss": 0.5931,
+      "step": 888
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.2339547247264658e-05,
+      "loss": 0.593,
+      "step": 889
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2323872559615884e-05,
+      "loss": 0.6239,
+      "step": 890
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.230819183403423e-05,
+      "loss": 0.5766,
+      "step": 891
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2292505111261674e-05,
+      "loss": 0.6276,
+      "step": 892
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.227681243205578e-05,
+      "loss": 0.6529,
+      "step": 893
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2261113837189587e-05,
+      "loss": 0.5941,
+      "step": 894
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2245409367451498e-05,
+      "loss": 0.6162,
+      "step": 895
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2229699063645191e-05,
+      "loss": 0.6013,
+      "step": 896
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2213982966589492e-05,
+      "loss": 0.6211,
+      "step": 897
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2198261117118287e-05,
+      "loss": 0.6658,
+      "step": 898
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.2182533556080402e-05,
+      "loss": 0.5758,
+      "step": 899
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2166800324339505e-05,
+      "loss": 0.6558,
+      "step": 900
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2151061462774006e-05,
+      "loss": 0.5785,
+      "step": 901
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.213531701227692e-05,
+      "loss": 0.5656,
+      "step": 902
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2119567013755811e-05,
+      "loss": 0.643,
+      "step": 903
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2103811508132642e-05,
+      "loss": 0.6632,
+      "step": 904
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.208805053634368e-05,
+      "loss": 0.6232,
+      "step": 905
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2072284139339414e-05,
+      "loss": 0.6222,
+      "step": 906
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2056512358084408e-05,
+      "loss": 0.6544,
+      "step": 907
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.204073523355723e-05,
+      "loss": 0.6373,
+      "step": 908
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.2024952806750321e-05,
+      "loss": 0.5505,
+      "step": 909
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.2009165118669905e-05,
+      "loss": 0.6273,
+      "step": 910
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1993372210335875e-05,
+      "loss": 0.5647,
+      "step": 911
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.197757412278168e-05,
+      "loss": 0.5965,
+      "step": 912
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1961770897054242e-05,
+      "loss": 0.5739,
+      "step": 913
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1945962574213814e-05,
+      "loss": 0.5656,
+      "step": 914
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1930149195333904e-05,
+      "loss": 0.6217,
+      "step": 915
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.191433080150116e-05,
+      "loss": 0.6263,
+      "step": 916
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1898507433815244e-05,
+      "loss": 0.6491,
+      "step": 917
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1882679133388763e-05,
+      "loss": 0.6359,
+      "step": 918
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1866845941347118e-05,
+      "loss": 0.5871,
+      "step": 919
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.1851007898828432e-05,
+      "loss": 0.6018,
+      "step": 920
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1835165046983436e-05,
+      "loss": 0.6192,
+      "step": 921
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1819317426975343e-05,
+      "loss": 0.6123,
+      "step": 922
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1803465079979762e-05,
+      "loss": 0.6608,
+      "step": 923
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1787608047184583e-05,
+      "loss": 0.6352,
+      "step": 924
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1771746369789872e-05,
+      "loss": 0.6021,
+      "step": 925
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1755880089007761e-05,
+      "loss": 0.6079,
+      "step": 926
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1740009246062343e-05,
+      "loss": 0.6216,
+      "step": 927
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1724133882189562e-05,
+      "loss": 0.6276,
+      "step": 928
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.1708254038637115e-05,
+      "loss": 0.604,
+      "step": 929
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.169236975666433e-05,
+      "loss": 0.5523,
+      "step": 930
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1676481077542072e-05,
+      "loss": 0.6034,
+      "step": 931
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.166058804255263e-05,
+      "loss": 0.6243,
+      "step": 932
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1644690692989616e-05,
+      "loss": 0.5911,
+      "step": 933
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1628789070157836e-05,
+      "loss": 0.6007,
+      "step": 934
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1612883215373221e-05,
+      "loss": 0.5507,
+      "step": 935
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1596973169962675e-05,
+      "loss": 0.644,
+      "step": 936
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1581058975264003e-05,
+      "loss": 0.6084,
+      "step": 937
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1565140672625799e-05,
+      "loss": 0.6591,
+      "step": 938
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1549218303407305e-05,
+      "loss": 0.5793,
+      "step": 939
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.1533291908978356e-05,
+      "loss": 0.5984,
+      "step": 940
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1517361530719233e-05,
+      "loss": 0.6161,
+      "step": 941
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1501427210020559e-05,
+      "loss": 0.5447,
+      "step": 942
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1485488988283217e-05,
+      "loss": 0.6139,
+      "step": 943
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1469546906918219e-05,
+      "loss": 0.5928,
+      "step": 944
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1453601007346599e-05,
+      "loss": 0.6124,
+      "step": 945
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1437651330999324e-05,
+      "loss": 0.6111,
+      "step": 946
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1421697919317161e-05,
+      "loss": 0.5999,
+      "step": 947
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1405740813750593e-05,
+      "loss": 0.6364,
+      "step": 948
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1389780055759689e-05,
+      "loss": 0.569,
+      "step": 949
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.1373815686814025e-05,
+      "loss": 0.5304,
+      "step": 950
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1357847748392539e-05,
+      "loss": 0.6998,
+      "step": 951
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1341876281983457e-05,
+      "loss": 0.5927,
+      "step": 952
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1325901329084167e-05,
+      "loss": 0.6094,
+      "step": 953
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1309922931201114e-05,
+      "loss": 0.6459,
+      "step": 954
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1293941129849701e-05,
+      "loss": 0.6598,
+      "step": 955
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1277955966554165e-05,
+      "loss": 0.5724,
+      "step": 956
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.126196748284748e-05,
+      "loss": 0.5728,
+      "step": 957
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1245975720271257e-05,
+      "loss": 0.6222,
+      "step": 958
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1229980720375609e-05,
+      "loss": 0.6267,
+      "step": 959
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1213982524719074e-05,
+      "loss": 0.5778,
+      "step": 960
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1197981174868488e-05,
+      "loss": 0.5592,
+      "step": 961
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1181976712398885e-05,
+      "loss": 0.633,
+      "step": 962
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1165969178893384e-05,
+      "loss": 0.636,
+      "step": 963
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.114995861594308e-05,
+      "loss": 0.6022,
+      "step": 964
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1133945065146947e-05,
+      "loss": 0.5528,
+      "step": 965
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1117928568111715e-05,
+      "loss": 0.5927,
+      "step": 966
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1101909166451777e-05,
+      "loss": 0.5654,
+      "step": 967
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.108588690178906e-05,
+      "loss": 0.5545,
+      "step": 968
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.1069861815752944e-05,
+      "loss": 0.5902,
+      "step": 969
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.105383394998013e-05,
+      "loss": 0.6311,
+      "step": 970
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.1037803346114541e-05,
+      "loss": 0.5778,
+      "step": 971
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.1021770045807228e-05,
+      "loss": 0.655,
+      "step": 972
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.1005734090716227e-05,
+      "loss": 0.645,
+      "step": 973
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.0989695522506486e-05,
+      "loss": 0.6269,
+      "step": 974
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.097365438284974e-05,
+      "loss": 0.6069,
+      "step": 975
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.09576107134244e-05,
+      "loss": 0.613,
+      "step": 976
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.0941564555915455e-05,
+      "loss": 0.6226,
+      "step": 977
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.0925515952014363e-05,
+      "loss": 0.5708,
+      "step": 978
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.0909464943418926e-05,
+      "loss": 0.6813,
+      "step": 979
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.0893411571833203e-05,
+      "loss": 0.6205,
+      "step": 980
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0877355878967391e-05,
+      "loss": 0.6243,
+      "step": 981
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0861297906537715e-05,
+      "loss": 0.6294,
+      "step": 982
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0845237696266326e-05,
+      "loss": 0.5664,
+      "step": 983
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0829175289881188e-05,
+      "loss": 0.5712,
+      "step": 984
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0813110729115975e-05,
+      "loss": 0.7026,
+      "step": 985
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0797044055709949e-05,
+      "loss": 0.6169,
+      "step": 986
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.078097531140787e-05,
+      "loss": 0.6167,
+      "step": 987
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.0764904537959875e-05,
+      "loss": 0.6129,
+      "step": 988
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.074883177712138e-05,
+      "loss": 0.5771,
+      "step": 989
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.073275707065295e-05,
+      "loss": 0.6294,
+      "step": 990
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0716680460320217e-05,
+      "loss": 0.6451,
+      "step": 991
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0700601987893759e-05,
+      "loss": 0.6639,
+      "step": 992
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0684521695148986e-05,
+      "loss": 0.6375,
+      "step": 993
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0668439623866043e-05,
+      "loss": 0.6193,
+      "step": 994
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0652355815829694e-05,
+      "loss": 0.6088,
+      "step": 995
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0636270312829215e-05,
+      "loss": 0.6115,
+      "step": 996
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0620183156658288e-05,
+      "loss": 0.6357,
+      "step": 997
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0604094389114887e-05,
+      "loss": 0.6447,
+      "step": 998
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0588004052001177e-05,
+      "loss": 0.6226,
+      "step": 999
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0571912187123399e-05,
+      "loss": 0.6733,
+      "step": 1000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0555818836291759e-05,
+      "loss": 0.6605,
+      "step": 1001
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.053972404132033e-05,
+      "loss": 0.5505,
+      "step": 1002
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0523627844026935e-05,
+      "loss": 0.587,
+      "step": 1003
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0507530286233042e-05,
+      "loss": 0.5886,
+      "step": 1004
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0491431409763654e-05,
+      "loss": 0.5763,
+      "step": 1005
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0475331256447195e-05,
+      "loss": 0.6181,
+      "step": 1006
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0459229868115412e-05,
+      "loss": 0.5827,
+      "step": 1007
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0443127286603256e-05,
+      "loss": 0.5975,
+      "step": 1008
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.0427023553748792e-05,
+      "loss": 0.6164,
+      "step": 1009
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.041091871139305e-05,
+      "loss": 0.5616,
+      "step": 1010
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0394812801379972e-05,
+      "loss": 0.5532,
+      "step": 1011
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0378705865556255e-05,
+      "loss": 0.5873,
+      "step": 1012
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0362597945771264e-05,
+      "loss": 0.6133,
+      "step": 1013
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0346489083876928e-05,
+      "loss": 0.5386,
+      "step": 1014
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0330379321727617e-05,
+      "loss": 0.5457,
+      "step": 1015
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0314268701180043e-05,
+      "loss": 0.5681,
+      "step": 1016
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0298157264093146e-05,
+      "loss": 0.5558,
+      "step": 1017
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0282045052327994e-05,
+      "loss": 0.5985,
+      "step": 1018
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0265932107747656e-05,
+      "loss": 0.5263,
+      "step": 1019
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.0249818472217115e-05,
+      "loss": 0.5071,
+      "step": 1020
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0233704187603143e-05,
+      "loss": 0.5088,
+      "step": 1021
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0217589295774208e-05,
+      "loss": 0.5521,
+      "step": 1022
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0201473838600346e-05,
+      "loss": 0.5484,
+      "step": 1023
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0185357857953064e-05,
+      "loss": 0.6462,
+      "step": 1024
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0169241395705229e-05,
+      "loss": 0.5859,
+      "step": 1025
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0153124493730967e-05,
+      "loss": 0.566,
+      "step": 1026
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.013700719390553e-05,
+      "loss": 0.5128,
+      "step": 1027
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0120889538105223e-05,
+      "loss": 0.5201,
+      "step": 1028
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0104771568207266e-05,
+      "loss": 0.5337,
+      "step": 1029
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.0088653326089685e-05,
+      "loss": 0.5802,
+      "step": 1030
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0072534853631236e-05,
+      "loss": 0.5381,
+      "step": 1031
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0056416192711256e-05,
+      "loss": 0.5426,
+      "step": 1032
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0040297385209581e-05,
+      "loss": 0.4954,
+      "step": 1033
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0024178473006418e-05,
+      "loss": 0.5174,
+      "step": 1034
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.0008059497982258e-05,
+      "loss": 0.5123,
+      "step": 1035
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.991940502017745e-06,
+      "loss": 0.5067,
+      "step": 1036
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.975821526993584e-06,
+      "loss": 0.5187,
+      "step": 1037
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.95970261479042e-06,
+      "loss": 0.526,
+      "step": 1038
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.943583807288746e-06,
+      "loss": 0.5253,
+      "step": 1039
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 9.927465146368766e-06,
+      "loss": 0.4935,
+      "step": 1040
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.911346673910318e-06,
+      "loss": 0.4796,
+      "step": 1041
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.895228431792739e-06,
+      "loss": 0.4967,
+      "step": 1042
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.879110461894778e-06,
+      "loss": 0.4785,
+      "step": 1043
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.862992806094473e-06,
+      "loss": 0.5386,
+      "step": 1044
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.846875506269038e-06,
+      "loss": 0.5061,
+      "step": 1045
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.830758604294773e-06,
+      "loss": 0.5057,
+      "step": 1046
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.814642142046938e-06,
+      "loss": 0.4852,
+      "step": 1047
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.798526161399657e-06,
+      "loss": 0.4942,
+      "step": 1048
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.782410704225793e-06,
+      "loss": 0.4849,
+      "step": 1049
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 9.76629581239686e-06,
+      "loss": 0.4403,
+      "step": 1050
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.750181527782892e-06,
+      "loss": 0.4734,
+      "step": 1051
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.734067892252349e-06,
+      "loss": 0.4184,
+      "step": 1052
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.71795494767201e-06,
+      "loss": 0.4972,
+      "step": 1053
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.701842735906855e-06,
+      "loss": 0.4304,
+      "step": 1054
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.685731298819957e-06,
+      "loss": 0.4736,
+      "step": 1055
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.669620678272386e-06,
+      "loss": 0.4434,
+      "step": 1056
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.653510916123074e-06,
+      "loss": 0.47,
+      "step": 1057
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.637402054228739e-06,
+      "loss": 0.456,
+      "step": 1058
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.621294134443747e-06,
+      "loss": 0.4444,
+      "step": 1059
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 9.60518719862003e-06,
+      "loss": 0.4903,
+      "step": 1060
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.589081288606952e-06,
+      "loss": 0.4554,
+      "step": 1061
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.572976446251215e-06,
+      "loss": 0.4465,
+      "step": 1062
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.556872713396746e-06,
+      "loss": 0.4435,
+      "step": 1063
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.54077013188459e-06,
+      "loss": 0.4831,
+      "step": 1064
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.524668743552807e-06,
+      "loss": 0.4435,
+      "step": 1065
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.508568590236349e-06,
+      "loss": 0.4574,
+      "step": 1066
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.492469713766961e-06,
+      "loss": 0.4344,
+      "step": 1067
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.47637215597307e-06,
+      "loss": 0.4691,
+      "step": 1068
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.460275958679674e-06,
+      "loss": 0.4686,
+      "step": 1069
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 9.444181163708245e-06,
+      "loss": 0.5084,
+      "step": 1070
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.428087812876604e-06,
+      "loss": 0.5401,
+      "step": 1071
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.411995947998823e-06,
+      "loss": 0.4631,
+      "step": 1072
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.395905610885116e-06,
+      "loss": 0.4554,
+      "step": 1073
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.379816843341715e-06,
+      "loss": 0.4259,
+      "step": 1074
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.363729687170787e-06,
+      "loss": 0.5428,
+      "step": 1075
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.34764418417031e-06,
+      "loss": 0.4416,
+      "step": 1076
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.331560376133958e-06,
+      "loss": 0.4454,
+      "step": 1077
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.315478304851017e-06,
+      "loss": 0.4477,
+      "step": 1078
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.299398012106246e-06,
+      "loss": 0.4558,
+      "step": 1079
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 9.283319539679787e-06,
+      "loss": 0.4508,
+      "step": 1080
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.267242929347052e-06,
+      "loss": 0.4171,
+      "step": 1081
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.251168222878624e-06,
+      "loss": 0.4752,
+      "step": 1082
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.235095462040125e-06,
+      "loss": 0.4687,
+      "step": 1083
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.219024688592136e-06,
+      "loss": 0.4312,
+      "step": 1084
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.202955944290058e-06,
+      "loss": 0.4642,
+      "step": 1085
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.18688927088403e-06,
+      "loss": 0.4324,
+      "step": 1086
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.170824710118814e-06,
+      "loss": 0.4864,
+      "step": 1087
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.154762303733676e-06,
+      "loss": 0.4702,
+      "step": 1088
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.138702093462286e-06,
+      "loss": 0.4753,
+      "step": 1089
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 9.122644121032614e-06,
+      "loss": 0.5504,
+      "step": 1090
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.106588428166799e-06,
+      "loss": 0.5089,
+      "step": 1091
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.090535056581075e-06,
+      "loss": 0.4596,
+      "step": 1092
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.074484047985638e-06,
+      "loss": 0.4191,
+      "step": 1093
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.058435444084543e-06,
+      "loss": 0.4464,
+      "step": 1094
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.042389286575603e-06,
+      "loss": 0.4733,
+      "step": 1095
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.026345617150265e-06,
+      "loss": 0.4598,
+      "step": 1096
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 9.01030447749352e-06,
+      "loss": 0.4918,
+      "step": 1097
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 8.994265909283776e-06,
+      "loss": 0.4966,
+      "step": 1098
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 8.978229954192775e-06,
+      "loss": 0.4567,
+      "step": 1099
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 8.962196653885459e-06,
+      "loss": 0.4587,
+      "step": 1100
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 8.946166050019875e-06,
+      "loss": 0.6759,
+      "step": 1101
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.93013818424706e-06,
+      "loss": 0.5431,
+      "step": 1102
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.914113098210941e-06,
+      "loss": 0.4956,
+      "step": 1103
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.898090833548226e-06,
+      "loss": 0.421,
+      "step": 1104
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.882071431888286e-06,
+      "loss": 0.472,
+      "step": 1105
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.866054934853053e-06,
+      "loss": 0.4715,
+      "step": 1106
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.850041384056924e-06,
+      "loss": 0.4357,
+      "step": 1107
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.83403082110662e-06,
+      "loss": 0.4281,
+      "step": 1108
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.818023287601117e-06,
+      "loss": 0.4434,
+      "step": 1109
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.802018825131513e-06,
+      "loss": 0.4632,
+      "step": 1110
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 8.786017475280927e-06,
+      "loss": 0.4511,
+      "step": 1111
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.770019279624398e-06,
+      "loss": 0.4479,
+      "step": 1112
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.754024279728748e-06,
+      "loss": 0.4924,
+      "step": 1113
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.738032517152523e-06,
+      "loss": 0.4408,
+      "step": 1114
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.72204403344584e-06,
+      "loss": 0.5394,
+      "step": 1115
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.7060588701503e-06,
+      "loss": 0.5221,
+      "step": 1116
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.690077068798886e-06,
+      "loss": 0.3811,
+      "step": 1117
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.674098670915838e-06,
+      "loss": 0.4497,
+      "step": 1118
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.658123718016548e-06,
+      "loss": 0.4778,
+      "step": 1119
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.642152251607465e-06,
+      "loss": 0.5056,
+      "step": 1120
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 8.626184313185979e-06,
+      "loss": 0.4521,
+      "step": 1121
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.610219944240313e-06,
+      "loss": 0.4424,
+      "step": 1122
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.594259186249409e-06,
+      "loss": 0.4574,
+      "step": 1123
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.578302080682844e-06,
+      "loss": 0.5127,
+      "step": 1124
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.56234866900068e-06,
+      "loss": 0.4494,
+      "step": 1125
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.546398992653403e-06,
+      "loss": 0.5,
+      "step": 1126
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.530453093081784e-06,
+      "loss": 0.4865,
+      "step": 1127
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.514511011716783e-06,
+      "loss": 0.4455,
+      "step": 1128
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.498572789979446e-06,
+      "loss": 0.4628,
+      "step": 1129
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.482638469280772e-06,
+      "loss": 0.4743,
+      "step": 1130
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 8.466708091021645e-06,
+      "loss": 0.4697,
+      "step": 1131
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.450781696592696e-06,
+      "loss": 0.4999,
+      "step": 1132
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.434859327374205e-06,
+      "loss": 0.4862,
+      "step": 1133
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.418941024735997e-06,
+      "loss": 0.4133,
+      "step": 1134
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.403026830037331e-06,
+      "loss": 0.4194,
+      "step": 1135
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.387116784626785e-06,
+      "loss": 0.4846,
+      "step": 1136
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.371210929842166e-06,
+      "loss": 0.4354,
+      "step": 1137
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.355309307010386e-06,
+      "loss": 0.4425,
+      "step": 1138
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.33941195744737e-06,
+      "loss": 0.4387,
+      "step": 1139
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.32351892245793e-06,
+      "loss": 0.4361,
+      "step": 1140
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 8.307630243335676e-06,
+      "loss": 0.4334,
+      "step": 1141
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.29174596136289e-06,
+      "loss": 0.4852,
+      "step": 1142
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.27586611781044e-06,
+      "loss": 0.465,
+      "step": 1143
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.259990753937662e-06,
+      "loss": 0.4546,
+      "step": 1144
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.24411991099224e-06,
+      "loss": 0.4274,
+      "step": 1145
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.228253630210133e-06,
+      "loss": 0.4261,
+      "step": 1146
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.212391952815419e-06,
+      "loss": 0.4052,
+      "step": 1147
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.196534920020241e-06,
+      "loss": 0.4525,
+      "step": 1148
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.18068257302466e-06,
+      "loss": 0.4151,
+      "step": 1149
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.164834953016566e-06,
+      "loss": 0.4807,
+      "step": 1150
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 8.148992101171566e-06,
+      "loss": 0.543,
+      "step": 1151
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.133154058652887e-06,
+      "loss": 0.4785,
+      "step": 1152
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.117320866611242e-06,
+      "loss": 0.3805,
+      "step": 1153
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.101492566184757e-06,
+      "loss": 0.4612,
+      "step": 1154
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.085669198498842e-06,
+      "loss": 0.4322,
+      "step": 1155
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.069850804666096e-06,
+      "loss": 0.4129,
+      "step": 1156
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.054037425786189e-06,
+      "loss": 0.456,
+      "step": 1157
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.038229102945763e-06,
+      "loss": 0.4281,
+      "step": 1158
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.022425877218321e-06,
+      "loss": 0.4815,
+      "step": 1159
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 8.00662778966413e-06,
+      "loss": 0.4849,
+      "step": 1160
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.990834881330098e-06,
+      "loss": 0.4722,
+      "step": 1161
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.97504719324968e-06,
+      "loss": 0.4204,
+      "step": 1162
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.959264766442775e-06,
+      "loss": 0.4397,
+      "step": 1163
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.943487641915595e-06,
+      "loss": 0.4367,
+      "step": 1164
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.927715860660589e-06,
+      "loss": 0.4946,
+      "step": 1165
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.911949463656322e-06,
+      "loss": 0.5255,
+      "step": 1166
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.896188491867361e-06,
+      "loss": 0.5271,
+      "step": 1167
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.88043298624419e-06,
+      "loss": 0.45,
+      "step": 1168
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.864682987723082e-06,
+      "loss": 0.462,
+      "step": 1169
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.848938537226001e-06,
+      "loss": 0.3939,
+      "step": 1170
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.833199675660496e-06,
+      "loss": 0.4968,
+      "step": 1171
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.8174664439196e-06,
+      "loss": 0.4905,
+      "step": 1172
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.801738882881715e-06,
+      "loss": 0.4115,
+      "step": 1173
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.78601703341051e-06,
+      "loss": 0.4309,
+      "step": 1174
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.770300936354814e-06,
+      "loss": 0.4585,
+      "step": 1175
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.754590632548506e-06,
+      "loss": 0.4238,
+      "step": 1176
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.738886162810417e-06,
+      "loss": 0.4687,
+      "step": 1177
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.723187567944223e-06,
+      "loss": 0.4527,
+      "step": 1178
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.70749488873833e-06,
+      "loss": 0.4223,
+      "step": 1179
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.691808165965776e-06,
+      "loss": 0.4414,
+      "step": 1180
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.67612744038412e-06,
+      "loss": 0.4325,
+      "step": 1181
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.660452752735345e-06,
+      "loss": 0.4703,
+      "step": 1182
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.644784143745743e-06,
+      "loss": 0.3923,
+      "step": 1183
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.629121654125808e-06,
+      "loss": 0.4444,
+      "step": 1184
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.613465324570144e-06,
+      "loss": 0.601,
+      "step": 1185
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.597815195757343e-06,
+      "loss": 0.5431,
+      "step": 1186
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.582171308349886e-06,
+      "loss": 0.4524,
+      "step": 1187
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.566533702994038e-06,
+      "loss": 0.4335,
+      "step": 1188
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.550902420319742e-06,
+      "loss": 0.4856,
+      "step": 1189
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.53527750094052e-06,
+      "loss": 0.4863,
+      "step": 1190
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.519658985453351e-06,
+      "loss": 0.48,
+      "step": 1191
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.5040469144385795e-06,
+      "loss": 0.4425,
+      "step": 1192
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.488441328459806e-06,
+      "loss": 0.4289,
+      "step": 1193
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.472842268063776e-06,
+      "loss": 0.4446,
+      "step": 1194
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.457249773780292e-06,
+      "loss": 0.4995,
+      "step": 1195
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.441663886122086e-06,
+      "loss": 0.4633,
+      "step": 1196
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.4260846455847314e-06,
+      "loss": 0.4416,
+      "step": 1197
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.41051209264652e-06,
+      "loss": 0.4072,
+      "step": 1198
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.394946267768381e-06,
+      "loss": 0.4066,
+      "step": 1199
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.379387211393756e-06,
+      "loss": 0.4205,
+      "step": 1200
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 7.363834963948499e-06,
+      "loss": 0.4264,
+      "step": 1201
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.348289565840777e-06,
+      "loss": 0.4995,
+      "step": 1202
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.332751057460961e-06,
+      "loss": 0.4385,
+      "step": 1203
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.317219479181517e-06,
+      "loss": 0.4439,
+      "step": 1204
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.3016948713569046e-06,
+      "loss": 0.4277,
+      "step": 1205
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.286177274323476e-06,
+      "loss": 0.4873,
+      "step": 1206
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.27066672839937e-06,
+      "loss": 0.4533,
+      "step": 1207
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.255163273884401e-06,
+      "loss": 0.4733,
+      "step": 1208
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.23966695105996e-06,
+      "loss": 0.5394,
+      "step": 1209
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.224177800188906e-06,
+      "loss": 0.4487,
+      "step": 1210
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 7.208695861515462e-06,
+      "loss": 0.4575,
+      "step": 1211
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.193221175265124e-06,
+      "loss": 0.522,
+      "step": 1212
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.1777537816445295e-06,
+      "loss": 0.4806,
+      "step": 1213
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.162293720841378e-06,
+      "loss": 0.443,
+      "step": 1214
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.146841033024312e-06,
+      "loss": 0.4517,
+      "step": 1215
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.131395758342815e-06,
+      "loss": 0.3867,
+      "step": 1216
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.11595793692712e-06,
+      "loss": 0.4784,
+      "step": 1217
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.1005276088880855e-06,
+      "loss": 0.5366,
+      "step": 1218
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.085104814317101e-06,
+      "loss": 0.4215,
+      "step": 1219
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.069689593285991e-06,
+      "loss": 0.4384,
+      "step": 1220
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 7.0542819858468895e-06,
+      "loss": 0.4139,
+      "step": 1221
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 7.038882032032157e-06,
+      "loss": 0.4231,
+      "step": 1222
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 7.023489771854264e-06,
+      "loss": 0.4137,
+      "step": 1223
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 7.008105245305699e-06,
+      "loss": 0.4484,
+      "step": 1224
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.992728492358847e-06,
+      "loss": 0.4556,
+      "step": 1225
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.977359552965903e-06,
+      "loss": 0.4686,
+      "step": 1226
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.961998467058753e-06,
+      "loss": 0.4852,
+      "step": 1227
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.9466452745488835e-06,
+      "loss": 0.4396,
+      "step": 1228
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.931300015327274e-06,
+      "loss": 0.4397,
+      "step": 1229
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.915962729264287e-06,
+      "loss": 0.4521,
+      "step": 1230
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.900633456209574e-06,
+      "loss": 0.42,
+      "step": 1231
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.8853122359919565e-06,
+      "loss": 0.4065,
+      "step": 1232
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.869999108419343e-06,
+      "loss": 0.3824,
+      "step": 1233
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.854694113278614e-06,
+      "loss": 0.4728,
+      "step": 1234
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.8393972903355185e-06,
+      "loss": 0.4246,
+      "step": 1235
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.8241086793345715e-06,
+      "loss": 0.3887,
+      "step": 1236
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.808828319998957e-06,
+      "loss": 0.4785,
+      "step": 1237
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.7935562520304065e-06,
+      "loss": 0.4461,
+      "step": 1238
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.7782925151091224e-06,
+      "loss": 0.4633,
+      "step": 1239
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.763037148893656e-06,
+      "loss": 0.419,
+      "step": 1240
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.747790193020808e-06,
+      "loss": 0.4209,
+      "step": 1241
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.7325516871055305e-06,
+      "loss": 0.3817,
+      "step": 1242
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.717321670740822e-06,
+      "loss": 0.4239,
+      "step": 1243
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.702100183497613e-06,
+      "loss": 0.4402,
+      "step": 1244
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.686887264924682e-06,
+      "loss": 0.4443,
+      "step": 1245
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.6716829545485485e-06,
+      "loss": 0.4481,
+      "step": 1246
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.656487291873357e-06,
+      "loss": 0.4729,
+      "step": 1247
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.64130031638079e-06,
+      "loss": 0.4976,
+      "step": 1248
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.62612206752995e-06,
+      "loss": 0.4698,
+      "step": 1249
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.610952584757273e-06,
+      "loss": 0.4696,
+      "step": 1250
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.59579190747642e-06,
+      "loss": 0.4466,
+      "step": 1251
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.580640075078169e-06,
+      "loss": 0.4938,
+      "step": 1252
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.565497126930318e-06,
+      "loss": 0.4727,
+      "step": 1253
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.550363102377588e-06,
+      "loss": 0.4173,
+      "step": 1254
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.535238040741498e-06,
+      "loss": 0.3832,
+      "step": 1255
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.5201219813203e-06,
+      "loss": 0.411,
+      "step": 1256
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.505014963388843e-06,
+      "loss": 0.4297,
+      "step": 1257
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.4899170261984844e-06,
+      "loss": 0.4091,
+      "step": 1258
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.474828208976998e-06,
+      "loss": 0.347,
+      "step": 1259
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.459748550928454e-06,
+      "loss": 0.4239,
+      "step": 1260
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.444678091233122e-06,
+      "loss": 0.4469,
+      "step": 1261
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.429616869047376e-06,
+      "loss": 0.4475,
+      "step": 1262
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.4145649235035944e-06,
+      "loss": 0.4572,
+      "step": 1263
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.3995222937100455e-06,
+      "loss": 0.4116,
+      "step": 1264
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.384489018750797e-06,
+      "loss": 0.4882,
+      "step": 1265
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.369465137685604e-06,
+      "loss": 0.4142,
+      "step": 1266
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.354450689549823e-06,
+      "loss": 0.4881,
+      "step": 1267
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.339445713354299e-06,
+      "loss": 0.4133,
+      "step": 1268
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.324450248085265e-06,
+      "loss": 0.4634,
+      "step": 1269
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.309464332704243e-06,
+      "loss": 0.4565,
+      "step": 1270
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.294488006147947e-06,
+      "loss": 0.4402,
+      "step": 1271
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.279521307328163e-06,
+      "loss": 0.4717,
+      "step": 1272
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.264564275131679e-06,
+      "loss": 0.3883,
+      "step": 1273
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.249616948420161e-06,
+      "loss": 0.4883,
+      "step": 1274
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.23467936603005e-06,
+      "loss": 0.5146,
+      "step": 1275
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.219751566772483e-06,
+      "loss": 0.382,
+      "step": 1276
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.20483358943317e-06,
+      "loss": 0.5157,
+      "step": 1277
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.1899254727723e-06,
+      "loss": 0.5106,
+      "step": 1278
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.175027255524446e-06,
+      "loss": 0.3965,
+      "step": 1279
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.160138976398456e-06,
+      "loss": 0.4031,
+      "step": 1280
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.145260674077363e-06,
+      "loss": 0.4331,
+      "step": 1281
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.130392387218274e-06,
+      "loss": 0.4464,
+      "step": 1282
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.115534154452269e-06,
+      "loss": 0.4845,
+      "step": 1283
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.100686014384315e-06,
+      "loss": 0.4564,
+      "step": 1284
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.08584800559315e-06,
+      "loss": 0.4286,
+      "step": 1285
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.07102016663119e-06,
+      "loss": 0.4857,
+      "step": 1286
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.056202536024425e-06,
+      "loss": 0.4496,
+      "step": 1287
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.041395152272331e-06,
+      "loss": 0.4527,
+      "step": 1288
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.026598053847743e-06,
+      "loss": 0.5159,
+      "step": 1289
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.011811279196788e-06,
+      "loss": 0.5407,
+      "step": 1290
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 5.997034866738765e-06,
+      "loss": 0.4741,
+      "step": 1291
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 5.982268854866045e-06,
+      "loss": 0.4186,
+      "step": 1292
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.967513281943984e-06,
+      "loss": 0.4449,
+      "step": 1293
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.952768186310813e-06,
+      "loss": 0.3943,
+      "step": 1294
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.938033606277534e-06,
+      "loss": 0.4286,
+      "step": 1295
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.923309580127832e-06,
+      "loss": 0.4185,
+      "step": 1296
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.908596146117971e-06,
+      "loss": 0.4808,
+      "step": 1297
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.8938933424766985e-06,
+      "loss": 0.4532,
+      "step": 1298
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.879201207405136e-06,
+      "loss": 0.4193,
+      "step": 1299
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.864519779076685e-06,
+      "loss": 0.4765,
+      "step": 1300
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.84984909563693e-06,
+      "loss": 0.4474,
+      "step": 1301
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 5.8351891952035415e-06,
+      "loss": 0.4204,
+      "step": 1302
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.820540115866164e-06,
+      "loss": 0.4322,
+      "step": 1303
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.805901895686344e-06,
+      "loss": 0.4613,
+      "step": 1304
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.791274572697401e-06,
+      "loss": 0.3923,
+      "step": 1305
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.776658184904334e-06,
+      "loss": 0.4037,
+      "step": 1306
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.7620527702837415e-06,
+      "loss": 0.4529,
+      "step": 1307
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.747458366783707e-06,
+      "loss": 0.4309,
+      "step": 1308
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.732875012323712e-06,
+      "loss": 0.4296,
+      "step": 1309
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.718302744794522e-06,
+      "loss": 0.398,
+      "step": 1310
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.7037416020581014e-06,
+      "loss": 0.4507,
+      "step": 1311
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 5.689191621947495e-06,
+      "loss": 0.4555,
+      "step": 1312
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.67465284226677e-06,
+      "loss": 0.4147,
+      "step": 1313
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.660125300790873e-06,
+      "loss": 0.3885,
+      "step": 1314
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.645609035265558e-06,
+      "loss": 0.4455,
+      "step": 1315
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.631104083407285e-06,
+      "loss": 0.4349,
+      "step": 1316
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.616610482903102e-06,
+      "loss": 0.4201,
+      "step": 1317
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.60212827141059e-06,
+      "loss": 0.4361,
+      "step": 1318
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.58765748655772e-06,
+      "loss": 0.4301,
+      "step": 1319
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.5731981659427785e-06,
+      "loss": 0.4615,
+      "step": 1320
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.558750347134265e-06,
+      "loss": 0.4035,
+      "step": 1321
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 5.544314067670798e-06,
+      "loss": 0.4512,
+      "step": 1322
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.529889365061012e-06,
+      "loss": 0.4578,
+      "step": 1323
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.5154762767834605e-06,
+      "loss": 0.4028,
+      "step": 1324
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.501074840286523e-06,
+      "loss": 0.3962,
+      "step": 1325
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.4866850929883045e-06,
+      "loss": 0.4219,
+      "step": 1326
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.472307072276539e-06,
+      "loss": 0.4342,
+      "step": 1327
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.457940815508491e-06,
+      "loss": 0.4261,
+      "step": 1328
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.443586360010859e-06,
+      "loss": 0.4099,
+      "step": 1329
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.429243743079686e-06,
+      "loss": 0.4589,
+      "step": 1330
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.414913001980246e-06,
+      "loss": 0.4234,
+      "step": 1331
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 5.400594173946963e-06,
+      "loss": 0.439,
+      "step": 1332
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.3862872961833065e-06,
+      "loss": 0.3922,
+      "step": 1333
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.3719924058616975e-06,
+      "loss": 0.3905,
+      "step": 1334
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.35770954012341e-06,
+      "loss": 0.4081,
+      "step": 1335
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.343438736078475e-06,
+      "loss": 0.4215,
+      "step": 1336
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.329180030805584e-06,
+      "loss": 0.4213,
+      "step": 1337
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.314933461352e-06,
+      "loss": 0.3818,
+      "step": 1338
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.30069906473345e-06,
+      "loss": 0.3387,
+      "step": 1339
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.286476877934023e-06,
+      "loss": 0.3873,
+      "step": 1340
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.2722669379061e-06,
+      "loss": 0.4715,
+      "step": 1341
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 5.258069281570231e-06,
+      "loss": 0.4677,
+      "step": 1342
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.243883945815063e-06,
+      "loss": 0.4222,
+      "step": 1343
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.2297109674972166e-06,
+      "loss": 0.4828,
+      "step": 1344
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.215550383441221e-06,
+      "loss": 0.443,
+      "step": 1345
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.201402230439381e-06,
+      "loss": 0.4672,
+      "step": 1346
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.187266545251719e-06,
+      "loss": 0.447,
+      "step": 1347
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.173143364605864e-06,
+      "loss": 0.6682,
+      "step": 1348
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.159032725196946e-06,
+      "loss": 0.4944,
+      "step": 1349
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.144934663687523e-06,
+      "loss": 0.3803,
+      "step": 1350
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.1308492167074545e-06,
+      "loss": 0.4179,
+      "step": 1351
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 5.116776420853834e-06,
+      "loss": 0.437,
+      "step": 1352
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.102716312690895e-06,
+      "loss": 0.4987,
+      "step": 1353
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.088668928749891e-06,
+      "loss": 0.3789,
+      "step": 1354
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.074634305529021e-06,
+      "loss": 0.4428,
+      "step": 1355
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.060612479493328e-06,
+      "loss": 0.4766,
+      "step": 1356
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.046603487074605e-06,
+      "loss": 0.4213,
+      "step": 1357
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.032607364671301e-06,
+      "loss": 0.3667,
+      "step": 1358
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.0186241486484245e-06,
+      "loss": 0.4147,
+      "step": 1359
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 5.004653875337452e-06,
+      "loss": 0.3781,
+      "step": 1360
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.990696581036231e-06,
+      "loss": 0.4185,
+      "step": 1361
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.976752302008888e-06,
+      "loss": 0.4101,
+      "step": 1362
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.962821074485731e-06,
+      "loss": 0.3671,
+      "step": 1363
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.948902934663158e-06,
+      "loss": 0.4203,
+      "step": 1364
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.934997918703564e-06,
+      "loss": 0.4502,
+      "step": 1365
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.921106062735241e-06,
+      "loss": 0.3939,
+      "step": 1366
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.907227402852296e-06,
+      "loss": 0.3998,
+      "step": 1367
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.89336197511454e-06,
+      "loss": 0.4318,
+      "step": 1368
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.879509815547413e-06,
+      "loss": 0.3551,
+      "step": 1369
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.865670960141874e-06,
+      "loss": 0.4126,
+      "step": 1370
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.851845444854321e-06,
+      "loss": 0.4198,
+      "step": 1371
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.8380333056064825e-06,
+      "loss": 0.4359,
+      "step": 1372
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.824234578285352e-06,
+      "loss": 0.433,
+      "step": 1373
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.810449298743051e-06,
+      "loss": 0.4012,
+      "step": 1374
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.796677502796776e-06,
+      "loss": 0.4082,
+      "step": 1375
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.782919226228685e-06,
+      "loss": 0.4195,
+      "step": 1376
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.769174504785818e-06,
+      "loss": 0.4044,
+      "step": 1377
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.7554433741799854e-06,
+      "loss": 0.4191,
+      "step": 1378
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.741725870087693e-06,
+      "loss": 0.4289,
+      "step": 1379
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.728022028150033e-06,
+      "loss": 0.4506,
+      "step": 1380
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.7143318839726035e-06,
+      "loss": 0.4235,
+      "step": 1381
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.700655473125425e-06,
+      "loss": 0.4138,
+      "step": 1382
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.686992831142819e-06,
+      "loss": 0.4137,
+      "step": 1383
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.673343993523347e-06,
+      "loss": 0.3904,
+      "step": 1384
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.659708995729685e-06,
+      "loss": 0.4351,
+      "step": 1385
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.646087873188563e-06,
+      "loss": 0.4152,
+      "step": 1386
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.6324806612906654e-06,
+      "loss": 0.3547,
+      "step": 1387
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.618887395390523e-06,
+      "loss": 0.383,
+      "step": 1388
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.605308110806436e-06,
+      "loss": 0.4666,
+      "step": 1389
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.591742842820379e-06,
+      "loss": 0.4357,
+      "step": 1390
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.578191626677897e-06,
+      "loss": 0.4093,
+      "step": 1391
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.564654497588047e-06,
+      "loss": 0.4491,
+      "step": 1392
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.551131490723267e-06,
+      "loss": 0.4566,
+      "step": 1393
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.537622641219309e-06,
+      "loss": 0.422,
+      "step": 1394
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.5241279841751405e-06,
+      "loss": 0.4501,
+      "step": 1395
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.510647554652854e-06,
+      "loss": 0.382,
+      "step": 1396
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.497181387677574e-06,
+      "loss": 0.3973,
+      "step": 1397
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.483729518237369e-06,
+      "loss": 0.4333,
+      "step": 1398
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.47029198128316e-06,
+      "loss": 0.4235,
+      "step": 1399
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.456868811728629e-06,
+      "loss": 0.3611,
+      "step": 1400
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.443460044450125e-06,
+      "loss": 0.4176,
+      "step": 1401
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.4300657142865835e-06,
+      "loss": 0.3964,
+      "step": 1402
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.416685856039423e-06,
+      "loss": 0.4177,
+      "step": 1403
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.403320504472463e-06,
+      "loss": 0.3893,
+      "step": 1404
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.389969694311831e-06,
+      "loss": 0.4372,
+      "step": 1405
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.3766334602458695e-06,
+      "loss": 0.4166,
+      "step": 1406
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.3633118369250645e-06,
+      "loss": 0.4073,
+      "step": 1407
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.350004858961917e-06,
+      "loss": 0.3875,
+      "step": 1408
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.336712560930891e-06,
+      "loss": 0.4031,
+      "step": 1409
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.323434977368306e-06,
+      "loss": 0.3649,
+      "step": 1410
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.310172142772243e-06,
+      "loss": 0.4079,
+      "step": 1411
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.296924091602478e-06,
+      "loss": 0.3782,
+      "step": 1412
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.283690858280366e-06,
+      "loss": 0.4052,
+      "step": 1413
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.270472477188755e-06,
+      "loss": 0.3942,
+      "step": 1414
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.257268982671912e-06,
+      "loss": 0.4215,
+      "step": 1415
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.244080409035431e-06,
+      "loss": 0.3743,
+      "step": 1416
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.230906790546128e-06,
+      "loss": 0.4581,
+      "step": 1417
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.217748161431969e-06,
+      "loss": 0.4189,
+      "step": 1418
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.204604555881967e-06,
+      "loss": 0.4407,
+      "step": 1419
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.191476008046103e-06,
+      "loss": 0.3669,
+      "step": 1420
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.1783625520352435e-06,
+      "loss": 0.383,
+      "step": 1421
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.165264221921033e-06,
+      "loss": 0.4078,
+      "step": 1422
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.15218105173582e-06,
+      "loss": 0.4346,
+      "step": 1423
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.139113075472565e-06,
+      "loss": 0.4068,
+      "step": 1424
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.126060327084739e-06,
+      "loss": 0.3872,
+      "step": 1425
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.113022840486268e-06,
+      "loss": 0.4293,
+      "step": 1426
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.100000649551413e-06,
+      "loss": 0.4859,
+      "step": 1427
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.086993788114694e-06,
+      "loss": 0.5384,
+      "step": 1428
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.074002289970801e-06,
+      "loss": 0.3733,
+      "step": 1429
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.061026188874509e-06,
+      "loss": 0.4293,
+      "step": 1430
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.048065518540589e-06,
+      "loss": 0.3839,
+      "step": 1431
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.035120312643718e-06,
+      "loss": 0.4572,
+      "step": 1432
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.02219060481839e-06,
+      "loss": 0.5494,
+      "step": 1433
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.009276428658836e-06,
+      "loss": 0.4731,
+      "step": 1434
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.996377817718932e-06,
+      "loss": 0.4338,
+      "step": 1435
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.983494805512109e-06,
+      "loss": 0.4206,
+      "step": 1436
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.970627425511272e-06,
+      "loss": 0.4375,
+      "step": 1437
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.9577757111487095e-06,
+      "loss": 0.4494,
+      "step": 1438
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.944939695816005e-06,
+      "loss": 0.4143,
+      "step": 1439
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.932119412863952e-06,
+      "loss": 0.4882,
+      "step": 1440
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.9193148956024795e-06,
+      "loss": 0.4838,
+      "step": 1441
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 3.906526177300536e-06,
+      "loss": 0.4216,
+      "step": 1442
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.893753291186031e-06,
+      "loss": 0.4707,
+      "step": 1443
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.8809962704457375e-06,
+      "loss": 0.378,
+      "step": 1444
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.868255148225199e-06,
+      "loss": 0.4112,
+      "step": 1445
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.855529957628671e-06,
+      "loss": 0.4056,
+      "step": 1446
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.842820731718997e-06,
+      "loss": 0.4148,
+      "step": 1447
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.830127503517541e-06,
+      "loss": 0.3965,
+      "step": 1448
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.81745030600411e-06,
+      "loss": 0.4471,
+      "step": 1449
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.8047891721168517e-06,
+      "loss": 0.4257,
+      "step": 1450
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.7921441347521893e-06,
+      "loss": 0.351,
+      "step": 1451
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 3.779515226764714e-06,
+      "loss": 0.4552,
+      "step": 1452
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.766902480967106e-06,
+      "loss": 0.4194,
+      "step": 1453
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.75430593013006e-06,
+      "loss": 0.4381,
+      "step": 1454
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.7417256069821872e-06,
+      "loss": 0.4185,
+      "step": 1455
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.729161544209945e-06,
+      "loss": 0.4467,
+      "step": 1456
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.7166137744575324e-06,
+      "loss": 0.3922,
+      "step": 1457
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.704082330326826e-06,
+      "loss": 0.4088,
+      "step": 1458
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.6915672443772644e-06,
+      "loss": 0.4264,
+      "step": 1459
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.6790685491258104e-06,
+      "loss": 0.4519,
+      "step": 1460
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.666586277046825e-06,
+      "loss": 0.3527,
+      "step": 1461
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.6541204605719992e-06,
+      "loss": 0.4036,
+      "step": 1462
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 3.6416711320902722e-06,
+      "loss": 0.4331,
+      "step": 1463
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.62923832394774e-06,
+      "loss": 0.4224,
+      "step": 1464
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.616822068447581e-06,
+      "loss": 0.3773,
+      "step": 1465
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.604422397849958e-06,
+      "loss": 0.3991,
+      "step": 1466
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.592039344371949e-06,
+      "loss": 0.4457,
+      "step": 1467
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.579672940187455e-06,
+      "loss": 0.4314,
+      "step": 1468
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.56732321742712e-06,
+      "loss": 0.4106,
+      "step": 1469
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.554990208178242e-06,
+      "loss": 0.4725,
+      "step": 1470
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.5426739444846967e-06,
+      "loss": 0.4547,
+      "step": 1471
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.5303744583468515e-06,
+      "loss": 0.4111,
+      "step": 1472
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 3.5180917817214798e-06,
+      "loss": 0.4242,
+      "step": 1473
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.5058259465216828e-06,
+      "loss": 0.4458,
+      "step": 1474
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.493576984616801e-06,
+      "loss": 0.4277,
+      "step": 1475
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.4813449278323374e-06,
+      "loss": 0.4247,
+      "step": 1476
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.46912980794987e-06,
+      "loss": 0.4239,
+      "step": 1477
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.456931656706972e-06,
+      "loss": 0.3874,
+      "step": 1478
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.444750505797123e-06,
+      "loss": 0.3985,
+      "step": 1479
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.4325863868696453e-06,
+      "loss": 0.3705,
+      "step": 1480
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.420439331529597e-06,
+      "loss": 0.4043,
+      "step": 1481
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.408309371337699e-06,
+      "loss": 0.4601,
+      "step": 1482
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 3.3961965378102635e-06,
+      "loss": 0.4226,
+      "step": 1483
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.384100862419096e-06,
+      "loss": 0.4397,
+      "step": 1484
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.372022376591435e-06,
+      "loss": 0.4365,
+      "step": 1485
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.3599611117098463e-06,
+      "loss": 0.4348,
+      "step": 1486
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.3479170991121455e-06,
+      "loss": 0.4159,
+      "step": 1487
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.3358903700913357e-06,
+      "loss": 0.4254,
+      "step": 1488
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.3238809558955054e-06,
+      "loss": 0.4367,
+      "step": 1489
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.311888887727763e-06,
+      "loss": 0.3994,
+      "step": 1490
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.2999141967461435e-06,
+      "loss": 0.3789,
+      "step": 1491
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.2879569140635324e-06,
+      "loss": 0.4485,
+      "step": 1492
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 3.276017070747579e-06,
+      "loss": 0.46,
+      "step": 1493
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.2640946978206266e-06,
+      "loss": 0.403,
+      "step": 1494
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.252189826259634e-06,
+      "loss": 0.4251,
+      "step": 1495
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.2403024869960765e-06,
+      "loss": 0.423,
+      "step": 1496
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.22843271091588e-06,
+      "loss": 0.4406,
+      "step": 1497
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.2165805288593377e-06,
+      "loss": 0.4234,
+      "step": 1498
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.2047459716210306e-06,
+      "loss": 0.408,
+      "step": 1499
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.192929069949744e-06,
+      "loss": 0.424,
+      "step": 1500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.1811298545483937e-06,
+      "loss": 0.3563,
+      "step": 1501
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.16934835607394e-06,
+      "loss": 0.4295,
+      "step": 1502
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 3.1575846051373117e-06,
+      "loss": 0.4023,
+      "step": 1503
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.145838632303325e-06,
+      "loss": 0.4371,
+      "step": 1504
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.1341104680906055e-06,
+      "loss": 0.4269,
+      "step": 1505
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.122400142971507e-06,
+      "loss": 0.3805,
+      "step": 1506
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.1107076873720344e-06,
+      "loss": 0.3622,
+      "step": 1507
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.0990331316717635e-06,
+      "loss": 0.3704,
+      "step": 1508
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.087376506203763e-06,
+      "loss": 0.3553,
+      "step": 1509
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.0757378412545114e-06,
+      "loss": 0.3825,
+      "step": 1510
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.064117167063827e-06,
+      "loss": 0.4002,
+      "step": 1511
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.0525145138247793e-06,
+      "loss": 0.4351,
+      "step": 1512
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 3.040929911683619e-06,
+      "loss": 0.4098,
+      "step": 1513
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.0293633907396903e-06,
+      "loss": 0.4543,
+      "step": 1514
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.017814981045374e-06,
+      "loss": 0.4124,
+      "step": 1515
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 3.006284712605971e-06,
+      "loss": 0.4123,
+      "step": 1516
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.994772615379665e-06,
+      "loss": 0.4105,
+      "step": 1517
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.983278719277418e-06,
+      "loss": 0.3972,
+      "step": 1518
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.971803054162903e-06,
+      "loss": 0.4142,
+      "step": 1519
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.9603456498524342e-06,
+      "loss": 0.4307,
+      "step": 1520
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.948906536114864e-06,
+      "loss": 0.3872,
+      "step": 1521
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.937485742671532e-06,
+      "loss": 0.3959,
+      "step": 1522
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.926083299196174e-06,
+      "loss": 0.465,
+      "step": 1523
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.914699235314855e-06,
+      "loss": 0.3882,
+      "step": 1524
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.903333580605878e-06,
+      "loss": 0.4345,
+      "step": 1525
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.8919863645997227e-06,
+      "loss": 0.3806,
+      "step": 1526
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.880657616778948e-06,
+      "loss": 0.4074,
+      "step": 1527
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.8693473665781367e-06,
+      "loss": 0.3848,
+      "step": 1528
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.858055643383818e-06,
+      "loss": 0.4056,
+      "step": 1529
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.846782476534373e-06,
+      "loss": 0.411,
+      "step": 1530
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.835527895319973e-06,
+      "loss": 0.4615,
+      "step": 1531
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.824291928982501e-06,
+      "loss": 0.4257,
+      "step": 1532
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.813074606715465e-06,
+      "loss": 0.3549,
+      "step": 1533
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.8018759576639478e-06,
+      "loss": 0.429,
+      "step": 1534
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.790696010924505e-06,
+      "loss": 0.394,
+      "step": 1535
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7795347955451004e-06,
+      "loss": 0.3725,
+      "step": 1536
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7683923405250315e-06,
+      "loss": 0.4298,
+      "step": 1537
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.757268674814849e-06,
+      "loss": 0.4425,
+      "step": 1538
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7461638273162895e-06,
+      "loss": 0.4309,
+      "step": 1539
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.735077826882192e-06,
+      "loss": 0.3683,
+      "step": 1540
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.724010702316429e-06,
+      "loss": 0.4057,
+      "step": 1541
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7129624823738267e-06,
+      "loss": 0.4663,
+      "step": 1542
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.7019331957600958e-06,
+      "loss": 0.4449,
+      "step": 1543
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6909228711317526e-06,
+      "loss": 0.4235,
+      "step": 1544
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6799315370960454e-06,
+      "loss": 0.3336,
+      "step": 1545
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6689592222108827e-06,
+      "loss": 0.3715,
+      "step": 1546
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6580059549847546e-06,
+      "loss": 0.4626,
+      "step": 1547
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6470717638766607e-06,
+      "loss": 0.3586,
+      "step": 1548
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6361566772960466e-06,
+      "loss": 0.3947,
+      "step": 1549
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.625260723602703e-06,
+      "loss": 0.4712,
+      "step": 1550
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.614383931106722e-06,
+      "loss": 0.4215,
+      "step": 1551
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.6035263280684055e-06,
+      "loss": 0.4123,
+      "step": 1552
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5926879426981943e-06,
+      "loss": 0.4329,
+      "step": 1553
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5818688031566132e-06,
+      "loss": 0.4774,
+      "step": 1554
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5710689375541596e-06,
+      "loss": 0.4757,
+      "step": 1555
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5602883739512675e-06,
+      "loss": 0.399,
+      "step": 1556
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5495271403582146e-06,
+      "loss": 0.4433,
+      "step": 1557
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5387852647350553e-06,
+      "loss": 0.4027,
+      "step": 1558
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.5280627749915544e-06,
+      "loss": 0.4577,
+      "step": 1559
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.517359698987102e-06,
+      "loss": 0.4261,
+      "step": 1560
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.506676064530641e-06,
+      "loss": 0.3354,
+      "step": 1561
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.496011899380609e-06,
+      "loss": 0.4003,
+      "step": 1562
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.485367231244863e-06,
+      "loss": 0.4171,
+      "step": 1563
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4747420877805905e-06,
+      "loss": 0.4299,
+      "step": 1564
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4641364965942572e-06,
+      "loss": 0.4129,
+      "step": 1565
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.453550485241526e-06,
+      "loss": 0.4238,
+      "step": 1566
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.442984081227181e-06,
+      "loss": 0.3878,
+      "step": 1567
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4324373120050738e-06,
+      "loss": 0.4472,
+      "step": 1568
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.421910204978033e-06,
+      "loss": 0.398,
+      "step": 1569
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.411402787497801e-06,
+      "loss": 0.4246,
+      "step": 1570
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.400915086864967e-06,
+      "loss": 0.4614,
+      "step": 1571
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.390447130328878e-06,
+      "loss": 0.4215,
+      "step": 1572
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.3799989450876005e-06,
+      "loss": 0.3916,
+      "step": 1573
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.369570558287819e-06,
+      "loss": 0.4531,
+      "step": 1574
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3591619970247803e-06,
+      "loss": 0.4677,
+      "step": 1575
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3487732883422186e-06,
+      "loss": 0.4801,
+      "step": 1576
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3384044592322875e-06,
+      "loss": 0.426,
+      "step": 1577
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3280555366354906e-06,
+      "loss": 0.4049,
+      "step": 1578
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3177265474406084e-06,
+      "loss": 0.4584,
+      "step": 1579
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3074175184846303e-06,
+      "loss": 0.4135,
+      "step": 1580
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.2971284765526847e-06,
+      "loss": 0.4577,
+      "step": 1581
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.286859448377966e-06,
+      "loss": 0.4299,
+      "step": 1582
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.276610460641682e-06,
+      "loss": 0.42,
+      "step": 1583
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.2663815399729495e-06,
+      "loss": 0.3987,
+      "step": 1584
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.2561727129487622e-06,
+      "loss": 0.4673,
+      "step": 1585
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.245984006093902e-06,
+      "loss": 0.4219,
+      "step": 1586
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.2358154458808713e-06,
+      "loss": 0.4034,
+      "step": 1587
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.2256670587298378e-06,
+      "loss": 0.43,
+      "step": 1588
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.215538871008538e-06,
+      "loss": 0.4309,
+      "step": 1589
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.205430909032239e-06,
+      "loss": 0.3674,
+      "step": 1590
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.195343199063653e-06,
+      "loss": 0.3717,
+      "step": 1591
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.185275767312869e-06,
+      "loss": 0.4351,
+      "step": 1592
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.175228639937299e-06,
+      "loss": 0.3968,
+      "step": 1593
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.1652018430415923e-06,
+      "loss": 0.4552,
+      "step": 1594
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.1551954026775723e-06,
+      "loss": 0.4301,
+      "step": 1595
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.145209344844177e-06,
+      "loss": 0.4121,
+      "step": 1596
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.1352436954873825e-06,
+      "loss": 0.4006,
+      "step": 1597
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.1252984805001465e-06,
+      "loss": 0.4343,
+      "step": 1598
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.115373725722326e-06,
+      "loss": 0.4547,
+      "step": 1599
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.1054694569406243e-06,
+      "loss": 0.4235,
+      "step": 1600
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.095585699888504e-06,
+      "loss": 0.4522,
+      "step": 1601
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.0857224802461516e-06,
+      "loss": 0.3813,
+      "step": 1602
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.075879823640382e-06,
+      "loss": 0.3938,
+      "step": 1603
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.066057755644587e-06,
+      "loss": 0.4117,
+      "step": 1604
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.056256301778664e-06,
+      "loss": 0.447,
+      "step": 1605
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.046475487508943e-06,
+      "loss": 0.4229,
+      "step": 1606
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.0367153382481407e-06,
+      "loss": 0.3651,
+      "step": 1607
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.026975879355273e-06,
+      "loss": 0.4046,
+      "step": 1608
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.0172571361356007e-06,
+      "loss": 0.368,
+      "step": 1609
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.0075591338405586e-06,
+      "loss": 0.4077,
+      "step": 1610
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.997881897667695e-06,
+      "loss": 0.4161,
+      "step": 1611
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.9882254527605995e-06,
+      "loss": 0.4501,
+      "step": 1612
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.978589824208843e-06,
+      "loss": 0.4884,
+      "step": 1613
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9689750370479134e-06,
+      "loss": 0.4173,
+      "step": 1614
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9593811162591446e-06,
+      "loss": 0.3664,
+      "step": 1615
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9498080867696568e-06,
+      "loss": 0.416,
+      "step": 1616
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9402559734522895e-06,
+      "loss": 0.4258,
+      "step": 1617
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.930724801125539e-06,
+      "loss": 0.3957,
+      "step": 1618
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.921214594553488e-06,
+      "loss": 0.4217,
+      "step": 1619
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.91172537844575e-06,
+      "loss": 0.3908,
+      "step": 1620
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9022571774573995e-06,
+      "loss": 0.4552,
+      "step": 1621
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.8928100161889062e-06,
+      "loss": 0.393,
+      "step": 1622
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.8833839191860803e-06,
+      "loss": 0.4295,
+      "step": 1623
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8739789109399954e-06,
+      "loss": 0.3818,
+      "step": 1624
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8645950158869353e-06,
+      "loss": 0.4454,
+      "step": 1625
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8552322584083249e-06,
+      "loss": 0.4449,
+      "step": 1626
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.845890662830675e-06,
+      "loss": 0.4192,
+      "step": 1627
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8365702534255103e-06,
+      "loss": 0.3922,
+      "step": 1628
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8272710544093019e-06,
+      "loss": 0.3795,
+      "step": 1629
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8179930899434207e-06,
+      "loss": 0.4471,
+      "step": 1630
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.8087363841340588e-06,
+      "loss": 0.3972,
+      "step": 1631
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.7995009610321833e-06,
+      "loss": 0.4345,
+      "step": 1632
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.7902868446334555e-06,
+      "loss": 0.4326,
+      "step": 1633
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7810940588781811e-06,
+      "loss": 0.4103,
+      "step": 1634
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.771922627651238e-06,
+      "loss": 0.4264,
+      "step": 1635
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.762772574782027e-06,
+      "loss": 0.4327,
+      "step": 1636
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7536439240444037e-06,
+      "loss": 0.4268,
+      "step": 1637
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7445366991566126e-06,
+      "loss": 0.3864,
+      "step": 1638
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7354509237812334e-06,
+      "loss": 0.4465,
+      "step": 1639
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7263866215251034e-06,
+      "loss": 0.4102,
+      "step": 1640
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7173438159392863e-06,
+      "loss": 0.4742,
+      "step": 1641
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.7083225305189777e-06,
+      "loss": 0.4043,
+      "step": 1642
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.699322788703468e-06,
+      "loss": 0.3951,
+      "step": 1643
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.690344613876066e-06,
+      "loss": 0.3938,
+      "step": 1644
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6813880293640505e-06,
+      "loss": 0.3878,
+      "step": 1645
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.672453058438599e-06,
+      "loss": 0.393,
+      "step": 1646
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6635397243147366e-06,
+      "loss": 0.5156,
+      "step": 1647
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6546480501512674e-06,
+      "loss": 0.4314,
+      "step": 1648
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.64577805905072e-06,
+      "loss": 0.3952,
+      "step": 1649
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6369297740592872e-06,
+      "loss": 0.4036,
+      "step": 1650
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.62810321816676e-06,
+      "loss": 0.3635,
+      "step": 1651
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.6192984143064771e-06,
+      "loss": 0.4383,
+      "step": 1652
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.610515385355258e-06,
+      "loss": 0.4462,
+      "step": 1653
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.601754154133347e-06,
+      "loss": 0.3936,
+      "step": 1654
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.593014743404353e-06,
+      "loss": 0.4637,
+      "step": 1655
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5842971758751913e-06,
+      "loss": 0.4519,
+      "step": 1656
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5756014741960213e-06,
+      "loss": 0.4327,
+      "step": 1657
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5669276609601925e-06,
+      "loss": 0.3949,
+      "step": 1658
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.558275758704183e-06,
+      "loss": 0.4259,
+      "step": 1659
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.54964578990754e-06,
+      "loss": 0.4188,
+      "step": 1660
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.541037776992822e-06,
+      "loss": 0.418,
+      "step": 1661
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5324517423255503e-06,
+      "loss": 0.4514,
+      "step": 1662
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5238877082141268e-06,
+      "loss": 0.4014,
+      "step": 1663
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.5153456969098013e-06,
+      "loss": 0.3741,
+      "step": 1664
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.5068257306065991e-06,
+      "loss": 0.4268,
+      "step": 1665
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.498327831441274e-06,
+      "loss": 0.4001,
+      "step": 1666
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4898520214932388e-06,
+      "loss": 0.445,
+      "step": 1667
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4813983227845164e-06,
+      "loss": 0.4074,
+      "step": 1668
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4729667572796735e-06,
+      "loss": 0.4337,
+      "step": 1669
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4645573468857754e-06,
+      "loss": 0.412,
+      "step": 1670
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4561701134523288e-06,
+      "loss": 0.3682,
+      "step": 1671
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4478050787712094e-06,
+      "loss": 0.4196,
+      "step": 1672
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.4394622645766232e-06,
+      "loss": 0.4367,
+      "step": 1673
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.431141692545036e-06,
+      "loss": 0.4157,
+      "step": 1674
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.4228433842951251e-06,
+      "loss": 0.4,
+      "step": 1675
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.4145673613877298e-06,
+      "loss": 0.4462,
+      "step": 1676
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.4063136453257787e-06,
+      "loss": 0.4257,
+      "step": 1677
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.398082257554243e-06,
+      "loss": 0.4134,
+      "step": 1678
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.389873219460085e-06,
+      "loss": 0.3996,
+      "step": 1679
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.3816865523721867e-06,
+      "loss": 0.3794,
+      "step": 1680
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.373522277561321e-06,
+      "loss": 0.4278,
+      "step": 1681
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.3653804162400686e-06,
+      "loss": 0.3806,
+      "step": 1682
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.3572609895627786e-06,
+      "loss": 0.393,
+      "step": 1683
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.349164018625513e-06,
+      "loss": 0.4187,
+      "step": 1684
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3410895244659828e-06,
+      "loss": 0.4172,
+      "step": 1685
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3330375280635054e-06,
+      "loss": 0.4424,
+      "step": 1686
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3250080503389396e-06,
+      "loss": 0.4068,
+      "step": 1687
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3170011121546388e-06,
+      "loss": 0.4133,
+      "step": 1688
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3090167343143911e-06,
+      "loss": 0.4315,
+      "step": 1689
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.3010549375633697e-06,
+      "loss": 0.4035,
+      "step": 1690
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.2931157425880781e-06,
+      "loss": 0.4547,
+      "step": 1691
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.2851991700162914e-06,
+      "loss": 0.3627,
+      "step": 1692
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.2773052404170106e-06,
+      "loss": 0.4197,
+      "step": 1693
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.2694339743004037e-06,
+      "loss": 0.3916,
+      "step": 1694
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2615853921177512e-06,
+      "loss": 0.3983,
+      "step": 1695
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2537595142614078e-06,
+      "loss": 0.36,
+      "step": 1696
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2459563610647186e-06,
+      "loss": 0.4108,
+      "step": 1697
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2381759528019988e-06,
+      "loss": 0.406,
+      "step": 1698
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2304183096884626e-06,
+      "loss": 0.4029,
+      "step": 1699
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2226834518801746e-06,
+      "loss": 0.4446,
+      "step": 1700
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.214971399474002e-06,
+      "loss": 0.4296,
+      "step": 1701
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.2072821725075567e-06,
+      "loss": 0.4039,
+      "step": 1702
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.199615790959141e-06,
+      "loss": 0.4387,
+      "step": 1703
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.1919722747477024e-06,
+      "loss": 0.4182,
+      "step": 1704
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.184351643732784e-06,
+      "loss": 0.4049,
+      "step": 1705
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1767539177144616e-06,
+      "loss": 0.3915,
+      "step": 1706
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1691791164333054e-06,
+      "loss": 0.4263,
+      "step": 1707
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1616272595703114e-06,
+      "loss": 0.4193,
+      "step": 1708
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1540983667468686e-06,
+      "loss": 0.434,
+      "step": 1709
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1465924575247022e-06,
+      "loss": 0.3768,
+      "step": 1710
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1391095514058182e-06,
+      "loss": 0.4126,
+      "step": 1711
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.131649667832453e-06,
+      "loss": 0.4245,
+      "step": 1712
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1242128261870311e-06,
+      "loss": 0.4378,
+      "step": 1713
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.1167990457920985e-06,
+      "loss": 0.4006,
+      "step": 1714
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.1094083459102966e-06,
+      "loss": 0.4273,
+      "step": 1715
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.1020407457442905e-06,
+      "loss": 0.3455,
+      "step": 1716
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0946962644367265e-06,
+      "loss": 0.4709,
+      "step": 1717
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0873749210701868e-06,
+      "loss": 0.4943,
+      "step": 1718
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0800767346671347e-06,
+      "loss": 0.4557,
+      "step": 1719
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0728017241898648e-06,
+      "loss": 0.3414,
+      "step": 1720
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0655499085404587e-06,
+      "loss": 0.4201,
+      "step": 1721
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.05832130656073e-06,
+      "loss": 0.4245,
+      "step": 1722
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.0511159370321789e-06,
+      "loss": 0.421,
+      "step": 1723
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.043933818675944e-06,
+      "loss": 0.4152,
+      "step": 1724
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0367749701527508e-06,
+      "loss": 0.4252,
+      "step": 1725
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0296394100628648e-06,
+      "loss": 0.4145,
+      "step": 1726
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0225271569460426e-06,
+      "loss": 0.4427,
+      "step": 1727
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0154382292814846e-06,
+      "loss": 0.4319,
+      "step": 1728
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.008372645487785e-06,
+      "loss": 0.4014,
+      "step": 1729
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.0013304239228938e-06,
+      "loss": 0.4022,
+      "step": 1730
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.943115828840477e-07,
+      "loss": 0.4164,
+      "step": 1731
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.873161406077435e-07,
+      "loss": 0.4665,
+      "step": 1732
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.803441152696824e-07,
+      "loss": 0.4893,
+      "step": 1733
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 9.733955249847183e-07,
+      "loss": 0.5233,
+      "step": 1734
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.66470387806826e-07,
+      "loss": 0.3745,
+      "step": 1735
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.595687217290362e-07,
+      "loss": 0.399,
+      "step": 1736
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.526905446833934e-07,
+      "loss": 0.4313,
+      "step": 1737
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.458358745409202e-07,
+      "loss": 0.3941,
+      "step": 1738
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.390047291115567e-07,
+      "loss": 0.5249,
+      "step": 1739
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.321971261441287e-07,
+      "loss": 0.5003,
+      "step": 1740
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.254130833262876e-07,
+      "loss": 0.3672,
+      "step": 1741
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.186526182844669e-07,
+      "loss": 0.4503,
+      "step": 1742
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.119157485838459e-07,
+      "loss": 0.4171,
+      "step": 1743
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 9.052024917282987e-07,
+      "loss": 0.4422,
+      "step": 1744
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.985128651603437e-07,
+      "loss": 0.4272,
+      "step": 1745
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.918468862611051e-07,
+      "loss": 0.3973,
+      "step": 1746
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.852045723502667e-07,
+      "loss": 0.4454,
+      "step": 1747
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.785859406860176e-07,
+      "loss": 0.3968,
+      "step": 1748
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.719910084650262e-07,
+      "loss": 0.3769,
+      "step": 1749
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.654197928223773e-07,
+      "loss": 0.4152,
+      "step": 1750
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.588723108315377e-07,
+      "loss": 0.4385,
+      "step": 1751
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.523485795043073e-07,
+      "loss": 0.3897,
+      "step": 1752
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.458486157907786e-07,
+      "loss": 0.4034,
+      "step": 1753
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 8.393724365792866e-07,
+      "loss": 0.4438,
+      "step": 1754
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.329200586963748e-07,
+      "loss": 0.4125,
+      "step": 1755
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.264914989067407e-07,
+      "loss": 0.395,
+      "step": 1756
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.20086773913199e-07,
+      "loss": 0.4147,
+      "step": 1757
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.137059003566372e-07,
+      "loss": 0.413,
+      "step": 1758
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.073488948159691e-07,
+      "loss": 0.3923,
+      "step": 1759
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 8.01015773808097e-07,
+      "loss": 0.427,
+      "step": 1760
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.94706553787864e-07,
+      "loss": 0.4168,
+      "step": 1761
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.884212511480139e-07,
+      "loss": 0.3666,
+      "step": 1762
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.821598822191468e-07,
+      "loss": 0.4001,
+      "step": 1763
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 7.759224632696793e-07,
+      "loss": 0.4285,
+      "step": 1764
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.697090105057991e-07,
+      "loss": 0.4237,
+      "step": 1765
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.635195400714279e-07,
+      "loss": 0.3727,
+      "step": 1766
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.573540680481705e-07,
+      "loss": 0.3825,
+      "step": 1767
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.512126104552809e-07,
+      "loss": 0.4473,
+      "step": 1768
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.450951832496233e-07,
+      "loss": 0.3716,
+      "step": 1769
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.390018023256196e-07,
+      "loss": 0.4208,
+      "step": 1770
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.32932483515214e-07,
+      "loss": 0.414,
+      "step": 1771
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.268872425878348e-07,
+      "loss": 0.4233,
+      "step": 1772
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.208660952503488e-07,
+      "loss": 0.4298,
+      "step": 1773
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 7.148690571470251e-07,
+      "loss": 0.3739,
+      "step": 1774
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.088961438594922e-07,
+      "loss": 0.429,
+      "step": 1775
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 7.029473709066892e-07,
+      "loss": 0.4304,
+      "step": 1776
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.970227537448415e-07,
+      "loss": 0.4122,
+      "step": 1777
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.911223077674079e-07,
+      "loss": 0.4515,
+      "step": 1778
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.852460483050494e-07,
+      "loss": 0.4311,
+      "step": 1779
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.793939906255831e-07,
+      "loss": 0.3644,
+      "step": 1780
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.735661499339441e-07,
+      "loss": 0.4018,
+      "step": 1781
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.677625413721433e-07,
+      "loss": 0.4104,
+      "step": 1782
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.619831800192355e-07,
+      "loss": 0.4103,
+      "step": 1783
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 6.562280808912768e-07,
+      "loss": 0.4216,
+      "step": 1784
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.504972589412806e-07,
+      "loss": 0.4133,
+      "step": 1785
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.447907290591859e-07,
+      "loss": 0.4507,
+      "step": 1786
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.391085060718149e-07,
+      "loss": 0.3958,
+      "step": 1787
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.334506047428346e-07,
+      "loss": 0.3868,
+      "step": 1788
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.278170397727179e-07,
+      "loss": 0.4689,
+      "step": 1789
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.222078257987085e-07,
+      "loss": 0.4259,
+      "step": 1790
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.166229773947796e-07,
+      "loss": 0.404,
+      "step": 1791
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.110625090715994e-07,
+      "loss": 0.3952,
+      "step": 1792
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.055264352764878e-07,
+      "loss": 0.3922,
+      "step": 1793
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 6.000147703933845e-07,
+      "loss": 0.409,
+      "step": 1794
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.945275287428099e-07,
+      "loss": 0.4693,
+      "step": 1795
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.890647245818259e-07,
+      "loss": 0.4072,
+      "step": 1796
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.836263721040014e-07,
+      "loss": 0.4519,
+      "step": 1797
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.782124854393745e-07,
+      "loss": 0.4268,
+      "step": 1798
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.728230786544153e-07,
+      "loss": 0.406,
+      "step": 1799
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.674581657519906e-07,
+      "loss": 0.3909,
+      "step": 1800
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.621177606713257e-07,
+      "loss": 0.4146,
+      "step": 1801
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.568018772879691e-07,
+      "loss": 0.429,
+      "step": 1802
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.515105294137546e-07,
+      "loss": 0.3798,
+      "step": 1803
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 5.46243730796776e-07,
+      "loss": 0.4169,
+      "step": 1804
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.410014951213316e-07,
+      "loss": 0.4352,
+      "step": 1805
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.357838360079059e-07,
+      "loss": 0.44,
+      "step": 1806
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.305907670131249e-07,
+      "loss": 0.399,
+      "step": 1807
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.254223016297289e-07,
+      "loss": 0.4008,
+      "step": 1808
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.202784532865302e-07,
+      "loss": 0.4693,
+      "step": 1809
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.15159235348377e-07,
+      "loss": 0.3937,
+      "step": 1810
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.100646611161264e-07,
+      "loss": 0.4296,
+      "step": 1811
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 5.049947438266023e-07,
+      "loss": 0.4094,
+      "step": 1812
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.99949496652572e-07,
+      "loss": 0.4356,
+      "step": 1813
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.949289327026952e-07,
+      "loss": 0.4738,
+      "step": 1814
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.899330650215062e-07,
+      "loss": 0.4332,
+      "step": 1815
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.849619065893673e-07,
+      "loss": 0.4374,
+      "step": 1816
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.800154703224424e-07,
+      "loss": 0.4581,
+      "step": 1817
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.7509376907266533e-07,
+      "loss": 0.4332,
+      "step": 1818
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.7019681562769816e-07,
+      "loss": 0.407,
+      "step": 1819
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.6532462271090763e-07,
+      "loss": 0.4488,
+      "step": 1820
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.6047720298132205e-07,
+      "loss": 0.4136,
+      "step": 1821
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.556545690336045e-07,
+      "loss": 0.3715,
+      "step": 1822
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.5085673339802407e-07,
+      "loss": 0.398,
+      "step": 1823
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.460837085404113e-07,
+      "loss": 0.3887,
+      "step": 1824
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.413355068621394e-07,
+      "loss": 0.4453,
+      "step": 1825
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.3661214070008006e-07,
+      "loss": 0.4426,
+      "step": 1826
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.319136223265796e-07,
+      "loss": 0.4101,
+      "step": 1827
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.272399639494251e-07,
+      "loss": 0.4348,
+      "step": 1828
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.225911777118097e-07,
+      "loss": 0.4306,
+      "step": 1829
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.179672756923037e-07,
+      "loss": 0.4273,
+      "step": 1830
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.133682699048247e-07,
+      "loss": 0.4176,
+      "step": 1831
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.087941722986022e-07,
+      "loss": 0.4283,
+      "step": 1832
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.0424499475814836e-07,
+      "loss": 0.4328,
+      "step": 1833
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.9972074910323066e-07,
+      "loss": 0.4124,
+      "step": 1834
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.9522144708883493e-07,
+      "loss": 0.4686,
+      "step": 1835
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.9074710040513887e-07,
+      "loss": 0.4194,
+      "step": 1836
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.862977206774798e-07,
+      "loss": 0.434,
+      "step": 1837
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.8187331946633154e-07,
+      "loss": 0.3925,
+      "step": 1838
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.7747390826725736e-07,
+      "loss": 0.3995,
+      "step": 1839
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.730994985108993e-07,
+      "loss": 0.4118,
+      "step": 1840
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.687501015629369e-07,
+      "loss": 0.4033,
+      "step": 1841
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.6442572872406155e-07,
+      "loss": 0.4642,
+      "step": 1842
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.601263912299491e-07,
+      "loss": 0.4304,
+      "step": 1843
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.5585210025122166e-07,
+      "loss": 0.4238,
+      "step": 1844
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.5160286689343126e-07,
+      "loss": 0.3883,
+      "step": 1845
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.4737870219702207e-07,
+      "loss": 0.4413,
+      "step": 1846
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.431796171373025e-07,
+      "loss": 0.4022,
+      "step": 1847
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.390056226244243e-07,
+      "loss": 0.4196,
+      "step": 1848
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.3485672950334447e-07,
+      "loss": 0.4275,
+      "step": 1849
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.3073294855379803e-07,
+      "loss": 0.4182,
+      "step": 1850
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.266342904902764e-07,
+      "loss": 0.4444,
+      "step": 1851
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.22560765962e-07,
+      "loss": 0.4487,
+      "step": 1852
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.1851238555288046e-07,
+      "loss": 0.4218,
+      "step": 1853
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.1448915978150365e-07,
+      "loss": 0.4217,
+      "step": 1854
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.1049109910109453e-07,
+      "loss": 0.3923,
+      "step": 1855
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.06518213899496e-07,
+      "loss": 0.4503,
+      "step": 1856
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.025705144991398e-07,
+      "loss": 0.4388,
+      "step": 1857
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.986480111570178e-07,
+      "loss": 0.3703,
+      "step": 1858
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.947507140646588e-07,
+      "loss": 0.4014,
+      "step": 1859
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.908786333480995e-07,
+      "loss": 0.4158,
+      "step": 1860
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.8703177906785675e-07,
+      "loss": 0.4001,
+      "step": 1861
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.832101612189064e-07,
+      "loss": 0.4517,
+      "step": 1862
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.794137897306548e-07,
+      "loss": 0.4167,
+      "step": 1863
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.756426744669105e-07,
+      "loss": 0.4421,
+      "step": 1864
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 2.7189682522586135e-07,
+      "loss": 0.4287,
+      "step": 1865
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.681762517400499e-07,
+      "loss": 0.4487,
+      "step": 1866
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.644809636763446e-07,
+      "loss": 0.4113,
+      "step": 1867
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.6081097063591855e-07,
+      "loss": 0.4303,
+      "step": 1868
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.57166282154222e-07,
+      "loss": 0.4129,
+      "step": 1869
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.5354690770096004e-07,
+      "loss": 0.4333,
+      "step": 1870
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.499528566800613e-07,
+      "loss": 0.4321,
+      "step": 1871
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.4638413842966725e-07,
+      "loss": 0.4611,
+      "step": 1872
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.4284076222208964e-07,
+      "loss": 0.4593,
+      "step": 1873
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.393227372638018e-07,
+      "loss": 0.4183,
+      "step": 1874
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 2.3583007269540882e-07,
+      "loss": 0.4801,
+      "step": 1875
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.323627775916204e-07,
+      "loss": 0.4098,
+      "step": 1876
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.28920860961237e-07,
+      "loss": 0.4224,
+      "step": 1877
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.2550433174711283e-07,
+      "loss": 0.3989,
+      "step": 1878
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.221131988261438e-07,
+      "loss": 0.4619,
+      "step": 1879
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.1874747100924188e-07,
+      "loss": 0.4716,
+      "step": 1880
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.1540715704130745e-07,
+      "loss": 0.4662,
+      "step": 1881
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.1209226560121477e-07,
+      "loss": 0.4681,
+      "step": 1882
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.0880280530178098e-07,
+      "loss": 0.3667,
+      "step": 1883
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.055387846897472e-07,
+      "loss": 0.4084,
+      "step": 1884
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 2.0230021224575846e-07,
+      "loss": 0.4257,
+      "step": 1885
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.9908709638434053e-07,
+      "loss": 0.4271,
+      "step": 1886
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.9589944545387652e-07,
+      "loss": 0.4333,
+      "step": 1887
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.9273726773658464e-07,
+      "loss": 0.4739,
+      "step": 1888
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8960057144850163e-07,
+      "loss": 0.4329,
+      "step": 1889
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8648936473945057e-07,
+      "loss": 0.4165,
+      "step": 1890
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8340365569303742e-07,
+      "loss": 0.4458,
+      "step": 1891
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.803434523266101e-07,
+      "loss": 0.4275,
+      "step": 1892
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.7730876259125396e-07,
+      "loss": 0.4342,
+      "step": 1893
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.742995943717607e-07,
+      "loss": 0.4406,
+      "step": 1894
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.7131595548661063e-07,
+      "loss": 0.3921,
+      "step": 1895
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.6835785368795598e-07,
+      "loss": 0.4436,
+      "step": 1896
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.654252966615977e-07,
+      "loss": 0.3591,
+      "step": 1897
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.6251829202696524e-07,
+      "loss": 0.4667,
+      "step": 1898
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5963684733709462e-07,
+      "loss": 0.4801,
+      "step": 1899
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5678097007861715e-07,
+      "loss": 0.479,
+      "step": 1900
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.539506676717284e-07,
+      "loss": 0.4634,
+      "step": 1901
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.511459474701793e-07,
+      "loss": 0.4211,
+      "step": 1902
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4836681676124954e-07,
+      "loss": 0.4284,
+      "step": 1903
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4561328276573415e-07,
+      "loss": 0.4645,
+      "step": 1904
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.4288535263792148e-07,
+      "loss": 0.3599,
+      "step": 1905
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.4018303346557295e-07,
+      "loss": 0.4592,
+      "step": 1906
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.3750633226990996e-07,
+      "loss": 0.4193,
+      "step": 1907
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.3485525600559378e-07,
+      "loss": 0.4526,
+      "step": 1908
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.3222981156070126e-07,
+      "loss": 0.4943,
+      "step": 1909
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.2963000575671792e-07,
+      "loss": 0.4589,
+      "step": 1910
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.2705584534851268e-07,
+      "loss": 0.4149,
+      "step": 1911
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.2450733702431884e-07,
+      "loss": 0.4419,
+      "step": 1912
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.219844874057241e-07,
+      "loss": 0.4058,
+      "step": 1913
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.1948730304764622e-07,
+      "loss": 0.3808,
+      "step": 1914
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.1701579043832179e-07,
+      "loss": 0.3885,
+      "step": 1915
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.1456995599928522e-07,
+      "loss": 0.4662,
+      "step": 1916
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.1214980608535209e-07,
+      "loss": 0.4368,
+      "step": 1917
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.0975534698460577e-07,
+      "loss": 0.3539,
+      "step": 1918
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.073865849183786e-07,
+      "loss": 0.4123,
+      "step": 1919
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.0504352604123413e-07,
+      "loss": 0.4394,
+      "step": 1920
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.0272617644095928e-07,
+      "loss": 0.4522,
+      "step": 1921
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.0043454213853665e-07,
+      "loss": 0.4574,
+      "step": 1922
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.816862908813784e-08,
+      "loss": 0.4396,
+      "step": 1923
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.592844317710238e-08,
+      "loss": 0.4324,
+      "step": 1924
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 9.371399022592765e-08,
+      "loss": 0.445,
+      "step": 1925
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 9.152527598825123e-08,
+      "loss": 0.4121,
+      "step": 1926
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.93623061508353e-08,
+      "loss": 0.4433,
+      "step": 1927
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.722508633355109e-08,
+      "loss": 0.4401,
+      "step": 1928
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.511362208936447e-08,
+      "loss": 0.4224,
+      "step": 1929
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.302791890432815e-08,
+      "loss": 0.4306,
+      "step": 1930
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 8.096798219755731e-08,
+      "loss": 0.4107,
+      "step": 1931
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.893381732122063e-08,
+      "loss": 0.4567,
+      "step": 1932
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.692542956052706e-08,
+      "loss": 0.426,
+      "step": 1933
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.494282413371135e-08,
+      "loss": 0.4088,
+      "step": 1934
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 7.298600619201735e-08,
+      "loss": 0.4629,
+      "step": 1935
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 7.105498081969142e-08,
+      "loss": 0.4153,
+      "step": 1936
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.914975303396021e-08,
+      "loss": 0.3911,
+      "step": 1937
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.727032778502729e-08,
+      "loss": 0.4286,
+      "step": 1938
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.541670995605321e-08,
+      "loss": 0.4342,
+      "step": 1939
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.358890436314547e-08,
+      "loss": 0.4025,
+      "step": 1940
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.178691575534412e-08,
+      "loss": 0.4329,
+      "step": 1941
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 6.001074881461511e-08,
+      "loss": 0.4373,
+      "step": 1942
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.826040815582912e-08,
+      "loss": 0.4208,
+      "step": 1943
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.653589832675943e-08,
+      "loss": 0.3861,
+      "step": 1944
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 5.483722380805967e-08,
+      "loss": 0.4193,
+      "step": 1945
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.31643890132616e-08,
+      "loss": 0.4287,
+      "step": 1946
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 5.151739828875846e-08,
+      "loss": 0.4071,
+      "step": 1947
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.98962559137961e-08,
+      "loss": 0.4064,
+      "step": 1948
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.830096610045854e-08,
+      "loss": 0.4186,
+      "step": 1949
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.673153299365907e-08,
+      "loss": 0.4798,
+      "step": 1950
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.5187960671133626e-08,
+      "loss": 0.3861,
+      "step": 1951
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.367025314342077e-08,
+      "loss": 0.4104,
+      "step": 1952
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.21784143538595e-08,
+      "loss": 0.4535,
+      "step": 1953
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.071244817857589e-08,
+      "loss": 0.441,
+      "step": 1954
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.9272358426473146e-08,
+      "loss": 0.436,
+      "step": 1955
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.7858148839221565e-08,
+      "loss": 0.4062,
+      "step": 1956
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.646982309124969e-08,
+      "loss": 0.4799,
+      "step": 1957
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.510738478973208e-08,
+      "loss": 0.4585,
+      "step": 1958
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.3770837474584874e-08,
+      "loss": 0.4864,
+      "step": 1959
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.2460184618452464e-08,
+      "loss": 0.4516,
+      "step": 1960
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 3.117542962669973e-08,
+      "loss": 0.4382,
+      "step": 1961
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.991657583740315e-08,
+      "loss": 0.4013,
+      "step": 1962
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.8683626521341934e-08,
+      "loss": 0.4508,
+      "step": 1963
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.747658488199023e-08,
+      "loss": 0.461,
+      "step": 1964
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.6295454055508263e-08,
+      "loss": 0.4273,
+      "step": 1965
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.5140237110733433e-08,
+      "loss": 0.4016,
+      "step": 1966
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.4010937049174788e-08,
+      "loss": 0.3793,
+      "step": 1967
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.2907556805001895e-08,
+      "loss": 0.4418,
+      "step": 1968
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.1830099245040427e-08,
+      "loss": 0.3802,
+      "step": 1969
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.0778567168761032e-08,
+      "loss": 0.4189,
+      "step": 1970
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.975296330827825e-08,
+      "loss": 0.4188,
+      "step": 1971
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.875329032833495e-08,
+      "loss": 0.4489,
+      "step": 1972
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.777955082630234e-08,
+      "loss": 0.4489,
+      "step": 1973
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.683174733216997e-08,
+      "loss": 0.4434,
+      "step": 1974
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.5909882308540182e-08,
+      "loss": 0.4098,
+      "step": 1975
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.5013958150621455e-08,
+      "loss": 0.3922,
+      "step": 1976
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.4143977186221735e-08,
+      "loss": 0.4721,
+      "step": 1977
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3299941675743999e-08,
+      "loss": 0.4666,
+      "step": 1978
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.248185381217848e-08,
+      "loss": 0.4483,
+      "step": 1979
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.1689715721097117e-08,
+      "loss": 0.4192,
+      "step": 1980
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.0923529460649119e-08,
+      "loss": 0.4573,
+      "step": 1981
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.0183297021555405e-08,
+      "loss": 0.4715,
+      "step": 1982
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 9.469020327103063e-09,
+      "loss": 0.4233,
+      "step": 1983
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 8.780701233139789e-09,
+      "loss": 0.4787,
+      "step": 1984
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 8.118341528071671e-09,
+      "loss": 0.4704,
+      "step": 1985
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 7.481942932853204e-09,
+      "loss": 0.4177,
+      "step": 1986
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.8715071009894944e-09,
+      "loss": 0.4958,
+      "step": 1987
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 6.2870356185273924e-09,
+      "loss": 0.4805,
+      "step": 1988
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.728530004051047e-09,
+      "loss": 0.4165,
+      "step": 1989
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 5.195991708681902e-09,
+      "loss": 0.4895,
+      "step": 1990
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.68942211607315e-09,
+      "loss": 0.4932,
+      "step": 1991
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.208822542401958e-09,
+      "loss": 0.5008,
+      "step": 1992
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.754194236373909e-09,
+      "loss": 0.4242,
+      "step": 1993
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 3.325538379211901e-09,
+      "loss": 0.4235,
+      "step": 1994
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.9228560846583655e-09,
+      "loss": 0.4523,
+      "step": 1995
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.546148398971937e-09,
+      "loss": 0.4823,
+      "step": 1996
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.1954163009219043e-09,
+      "loss": 0.3952,
+      "step": 1997
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.870660701785987e-09,
+      "loss": 0.5206,
+      "step": 1998
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.5718824453525572e-09,
+      "loss": 0.4393,
+      "step": 1999
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.299082307912869e-09,
+      "loss": 0.3969,
+      "step": 2000
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 2010,
+  "num_train_epochs": 2,
+  "save_steps": 100,
+  "total_flos": 1440382941904896.0,
+  "trial_name": null,
+  "trial_params": null
+}