diff --git "a/checkpoint-2803/trainer_state.json" "b/checkpoint-2803/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-2803/trainer_state.json"
@@ -0,0 +1,16843 @@
+{
+  "best_metric": 0.5148146748542786,
+  "best_model_checkpoint": "./medium-article-titles-engagement/checkpoint-2803",
+  "epoch": 1.0,
+  "global_step": 2803,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 0.7159,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.7009,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3e-06,
+      "loss": 0.7291,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.7201,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 5e-06,
+      "loss": 0.7007,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6e-06,
+      "loss": 0.7034,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 7.000000000000001e-06,
+      "loss": 0.74,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.6726,
+      "step": 8
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 9e-06,
+      "loss": 0.7118,
+      "step": 9
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1e-05,
+      "loss": 0.7025,
+      "step": 10
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.1000000000000001e-05,
+      "loss": 0.7034,
+      "step": 11
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.2e-05,
+      "loss": 0.6956,
+      "step": 12
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.3000000000000001e-05,
+      "loss": 0.6789,
+      "step": 13
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 0.6891,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.5e-05,
+      "loss": 0.6971,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 0.7049,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.7000000000000003e-05,
+      "loss": 0.6963,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.8e-05,
+      "loss": 0.6857,
+      "step": 18
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.9e-05,
+      "loss": 0.7105,
+      "step": 19
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2e-05,
+      "loss": 0.6904,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.1e-05,
+      "loss": 0.694,
+      "step": 21
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 0.7006,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.3000000000000003e-05,
+      "loss": 0.6932,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.4e-05,
+      "loss": 0.6898,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.5e-05,
+      "loss": 0.6834,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 0.694,
+      "step": 26
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.7000000000000002e-05,
+      "loss": 0.6915,
+      "step": 27
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 0.6742,
+      "step": 28
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 2.9e-05,
+      "loss": 0.6847,
+      "step": 29
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3e-05,
+      "loss": 0.6809,
+      "step": 30
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.1e-05,
+      "loss": 0.7081,
+      "step": 31
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 0.6871,
+      "step": 32
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.3e-05,
+      "loss": 0.6796,
+      "step": 33
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 0.6808,
+      "step": 34
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.5e-05,
+      "loss": 0.706,
+      "step": 35
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.6e-05,
+      "loss": 0.6854,
+      "step": 36
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.7e-05,
+      "loss": 0.691,
+      "step": 37
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.8e-05,
+      "loss": 0.6791,
+      "step": 38
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.9000000000000006e-05,
+      "loss": 0.6871,
+      "step": 39
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4e-05,
+      "loss": 0.7009,
+      "step": 40
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.1e-05,
+      "loss": 0.6879,
+      "step": 41
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.2e-05,
+      "loss": 0.6903,
+      "step": 42
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.3e-05,
+      "loss": 0.7009,
+      "step": 43
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 0.6622,
+      "step": 44
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.5e-05,
+      "loss": 0.6796,
+      "step": 45
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 0.6778,
+      "step": 46
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.7e-05,
+      "loss": 0.7009,
+      "step": 47
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.8e-05,
+      "loss": 0.6946,
+      "step": 48
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9e-05,
+      "loss": 0.6546,
+      "step": 49
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 5e-05,
+      "loss": 0.6836,
+      "step": 50
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9996419620479776e-05,
+      "loss": 0.6608,
+      "step": 51
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.999283924095954e-05,
+      "loss": 0.702,
+      "step": 52
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9989258861439315e-05,
+      "loss": 0.7556,
+      "step": 53
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.998567848191908e-05,
+      "loss": 0.6576,
+      "step": 54
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9982098102398855e-05,
+      "loss": 0.6991,
+      "step": 55
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.997851772287863e-05,
+      "loss": 0.6654,
+      "step": 56
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9974937343358394e-05,
+      "loss": 0.7093,
+      "step": 57
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.997135696383817e-05,
+      "loss": 0.7182,
+      "step": 58
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.996777658431794e-05,
+      "loss": 0.6627,
+      "step": 59
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.996419620479771e-05,
+      "loss": 0.6947,
+      "step": 60
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.996061582527748e-05,
+      "loss": 0.677,
+      "step": 61
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9957035445757254e-05,
+      "loss": 0.652,
+      "step": 62
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.995345506623703e-05,
+      "loss": 0.631,
+      "step": 63
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.99498746867168e-05,
+      "loss": 0.6458,
+      "step": 64
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9946294307196566e-05,
+      "loss": 0.6994,
+      "step": 65
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.994271392767634e-05,
+      "loss": 0.6767,
+      "step": 66
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9939133548156106e-05,
+      "loss": 0.6782,
+      "step": 67
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.993555316863588e-05,
+      "loss": 0.6532,
+      "step": 68
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.993197278911565e-05,
+      "loss": 0.721,
+      "step": 69
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.992839240959542e-05,
+      "loss": 0.6353,
+      "step": 70
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992481203007519e-05,
+      "loss": 0.6671,
+      "step": 71
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.992123165055496e-05,
+      "loss": 0.693,
+      "step": 72
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991765127103473e-05,
+      "loss": 0.6639,
+      "step": 73
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9914070891514505e-05,
+      "loss": 0.6729,
+      "step": 74
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.991049051199427e-05,
+      "loss": 0.6678,
+      "step": 75
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9906910132474044e-05,
+      "loss": 0.6627,
+      "step": 76
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.990332975295382e-05,
+      "loss": 0.7264,
+      "step": 77
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9899749373433584e-05,
+      "loss": 0.7034,
+      "step": 78
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.989616899391336e-05,
+      "loss": 0.7133,
+      "step": 79
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9892588614393123e-05,
+      "loss": 0.6636,
+      "step": 80
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.98890082348729e-05,
+      "loss": 0.6578,
+      "step": 81
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.988542785535267e-05,
+      "loss": 0.6693,
+      "step": 82
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9881847475832436e-05,
+      "loss": 0.6566,
+      "step": 83
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.987826709631221e-05,
+      "loss": 0.7203,
+      "step": 84
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.987468671679198e-05,
+      "loss": 0.6635,
+      "step": 85
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9871106337271756e-05,
+      "loss": 0.6184,
+      "step": 86
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.986752595775153e-05,
+      "loss": 0.6922,
+      "step": 87
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9863945578231295e-05,
+      "loss": 0.6691,
+      "step": 88
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.986036519871107e-05,
+      "loss": 0.6873,
+      "step": 89
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9856784819190835e-05,
+      "loss": 0.6924,
+      "step": 90
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.985320443967061e-05,
+      "loss": 0.6754,
+      "step": 91
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.984962406015038e-05,
+      "loss": 0.658,
+      "step": 92
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.984604368063015e-05,
+      "loss": 0.6457,
+      "step": 93
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.984246330110992e-05,
+      "loss": 0.6697,
+      "step": 94
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9838882921589694e-05,
+      "loss": 0.6244,
+      "step": 95
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.983530254206946e-05,
+      "loss": 0.6171,
+      "step": 96
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9831722162549234e-05,
+      "loss": 0.7583,
+      "step": 97
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.9828141783029e-05,
+      "loss": 0.7512,
+      "step": 98
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9824561403508773e-05,
+      "loss": 0.6467,
+      "step": 99
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9820981023988547e-05,
+      "loss": 0.7834,
+      "step": 100
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.981740064446831e-05,
+      "loss": 0.6662,
+      "step": 101
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9813820264948086e-05,
+      "loss": 0.693,
+      "step": 102
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.981023988542786e-05,
+      "loss": 0.6854,
+      "step": 103
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9806659505907626e-05,
+      "loss": 0.6908,
+      "step": 104
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.98030791263874e-05,
+      "loss": 0.6454,
+      "step": 105
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9799498746867165e-05,
+      "loss": 0.7064,
+      "step": 106
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.979591836734694e-05,
+      "loss": 0.6956,
+      "step": 107
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.979233798782671e-05,
+      "loss": 0.6842,
+      "step": 108
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9788757608306485e-05,
+      "loss": 0.6716,
+      "step": 109
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.978517722878626e-05,
+      "loss": 0.6738,
+      "step": 110
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9781596849266025e-05,
+      "loss": 0.6509,
+      "step": 111
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.97780164697458e-05,
+      "loss": 0.6746,
+      "step": 112
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.977443609022557e-05,
+      "loss": 0.6497,
+      "step": 113
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.977085571070534e-05,
+      "loss": 0.652,
+      "step": 114
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.976727533118511e-05,
+      "loss": 0.6651,
+      "step": 115
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.976369495166488e-05,
+      "loss": 0.6341,
+      "step": 116
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.976011457214465e-05,
+      "loss": 0.698,
+      "step": 117
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.975653419262442e-05,
+      "loss": 0.6979,
+      "step": 118
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.975295381310419e-05,
+      "loss": 0.671,
+      "step": 119
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.974937343358396e-05,
+      "loss": 0.6906,
+      "step": 120
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9745793054063736e-05,
+      "loss": 0.6825,
+      "step": 121
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.97422126745435e-05,
+      "loss": 0.6585,
+      "step": 122
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9738632295023276e-05,
+      "loss": 0.6658,
+      "step": 123
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.973505191550304e-05,
+      "loss": 0.6184,
+      "step": 124
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9731471535982815e-05,
+      "loss": 0.6479,
+      "step": 125
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.972789115646259e-05,
+      "loss": 0.7496,
+      "step": 126
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9724310776942355e-05,
+      "loss": 0.604,
+      "step": 127
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.972073039742213e-05,
+      "loss": 0.7436,
+      "step": 128
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9717150017901894e-05,
+      "loss": 0.6447,
+      "step": 129
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.971356963838167e-05,
+      "loss": 0.7154,
+      "step": 130
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.970998925886144e-05,
+      "loss": 0.6345,
+      "step": 131
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9706408879341214e-05,
+      "loss": 0.6996,
+      "step": 132
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.970282849982099e-05,
+      "loss": 0.6765,
+      "step": 133
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9699248120300754e-05,
+      "loss": 0.6099,
+      "step": 134
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.969566774078053e-05,
+      "loss": 0.6562,
+      "step": 135
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.96920873612603e-05,
+      "loss": 0.6532,
+      "step": 136
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9688506981740066e-05,
+      "loss": 0.6839,
+      "step": 137
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.968492660221984e-05,
+      "loss": 0.6499,
+      "step": 138
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.968134622269961e-05,
+      "loss": 0.6182,
+      "step": 139
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.967776584317938e-05,
+      "loss": 0.682,
+      "step": 140
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.967418546365915e-05,
+      "loss": 0.7618,
+      "step": 141
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.967060508413892e-05,
+      "loss": 0.7105,
+      "step": 142
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.966702470461869e-05,
+      "loss": 0.6784,
+      "step": 143
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9663444325098465e-05,
+      "loss": 0.7058,
+      "step": 144
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.965986394557823e-05,
+      "loss": 0.6746,
+      "step": 145
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9656283566058005e-05,
+      "loss": 0.6963,
+      "step": 146
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.965270318653777e-05,
+      "loss": 0.6886,
+      "step": 147
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9649122807017544e-05,
+      "loss": 0.6601,
+      "step": 148
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.964554242749732e-05,
+      "loss": 0.6388,
+      "step": 149
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9641962047977084e-05,
+      "loss": 0.7187,
+      "step": 150
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.963838166845686e-05,
+      "loss": 0.6441,
+      "step": 151
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.963480128893663e-05,
+      "loss": 0.635,
+      "step": 152
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.96312209094164e-05,
+      "loss": 0.6421,
+      "step": 153
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.962764052989617e-05,
+      "loss": 0.6535,
+      "step": 154
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9624060150375936e-05,
+      "loss": 0.6718,
+      "step": 155
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9620479770855716e-05,
+      "loss": 0.6475,
+      "step": 156
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.961689939133549e-05,
+      "loss": 0.6324,
+      "step": 157
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9613319011815256e-05,
+      "loss": 0.6297,
+      "step": 158
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.960973863229503e-05,
+      "loss": 0.6823,
+      "step": 159
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9606158252774796e-05,
+      "loss": 0.7251,
+      "step": 160
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.960257787325457e-05,
+      "loss": 0.7437,
+      "step": 161
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.959899749373434e-05,
+      "loss": 0.6722,
+      "step": 162
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.959541711421411e-05,
+      "loss": 0.6506,
+      "step": 163
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.959183673469388e-05,
+      "loss": 0.7161,
+      "step": 164
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.958825635517365e-05,
+      "loss": 0.6613,
+      "step": 165
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.958467597565342e-05,
+      "loss": 0.6492,
+      "step": 166
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9581095596133194e-05,
+      "loss": 0.6708,
+      "step": 167
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.957751521661296e-05,
+      "loss": 0.75,
+      "step": 168
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9573934837092734e-05,
+      "loss": 0.6443,
+      "step": 169
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.957035445757251e-05,
+      "loss": 0.6741,
+      "step": 170
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9566774078052273e-05,
+      "loss": 0.6885,
+      "step": 171
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.956319369853205e-05,
+      "loss": 0.6739,
+      "step": 172
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.955961331901181e-05,
+      "loss": 0.6247,
+      "step": 173
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9556032939491586e-05,
+      "loss": 0.704,
+      "step": 174
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.955245255997136e-05,
+      "loss": 0.6588,
+      "step": 175
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9548872180451126e-05,
+      "loss": 0.7004,
+      "step": 176
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.95452918009309e-05,
+      "loss": 0.6415,
+      "step": 177
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.954171142141067e-05,
+      "loss": 0.6693,
+      "step": 178
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9538131041890445e-05,
+      "loss": 0.672,
+      "step": 179
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.953455066237022e-05,
+      "loss": 0.635,
+      "step": 180
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.9530970282849985e-05,
+      "loss": 0.6926,
+      "step": 181
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.952738990332976e-05,
+      "loss": 0.6558,
+      "step": 182
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9523809523809525e-05,
+      "loss": 0.6828,
+      "step": 183
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.95202291442893e-05,
+      "loss": 0.7209,
+      "step": 184
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.951664876476907e-05,
+      "loss": 0.6179,
+      "step": 185
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.951306838524884e-05,
+      "loss": 0.7027,
+      "step": 186
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.950948800572861e-05,
+      "loss": 0.6311,
+      "step": 187
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9505907626208384e-05,
+      "loss": 0.6565,
+      "step": 188
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.950232724668815e-05,
+      "loss": 0.7102,
+      "step": 189
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9498746867167923e-05,
+      "loss": 0.6111,
+      "step": 190
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.949516648764769e-05,
+      "loss": 0.6164,
+      "step": 191
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.949158610812746e-05,
+      "loss": 0.6465,
+      "step": 192
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9488005728607236e-05,
+      "loss": 0.6344,
+      "step": 193
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9484425349087e-05,
+      "loss": 0.6496,
+      "step": 194
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9480844969566776e-05,
+      "loss": 0.6155,
+      "step": 195
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.947726459004655e-05,
+      "loss": 0.6761,
+      "step": 196
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9473684210526315e-05,
+      "loss": 0.6198,
+      "step": 197
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.947010383100609e-05,
+      "loss": 0.7867,
+      "step": 198
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9466523451485855e-05,
+      "loss": 0.6036,
+      "step": 199
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.946294307196563e-05,
+      "loss": 0.5905,
+      "step": 200
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.94593626924454e-05,
+      "loss": 0.7537,
+      "step": 201
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9455782312925175e-05,
+      "loss": 0.6287,
+      "step": 202
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.945220193340495e-05,
+      "loss": 0.6808,
+      "step": 203
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9448621553884714e-05,
+      "loss": 0.7303,
+      "step": 204
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.944504117436449e-05,
+      "loss": 0.6118,
+      "step": 205
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.944146079484426e-05,
+      "loss": 0.6998,
+      "step": 206
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.943788041532403e-05,
+      "loss": 0.6566,
+      "step": 207
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.94343000358038e-05,
+      "loss": 0.6808,
+      "step": 208
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.9430719656283567e-05,
+      "loss": 0.613,
+      "step": 209
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.942713927676334e-05,
+      "loss": 0.6322,
+      "step": 210
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.942355889724311e-05,
+      "loss": 0.6834,
+      "step": 211
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.941997851772288e-05,
+      "loss": 0.6896,
+      "step": 212
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.941639813820265e-05,
+      "loss": 0.6805,
+      "step": 213
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9412817758682426e-05,
+      "loss": 0.6569,
+      "step": 214
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.940923737916219e-05,
+      "loss": 0.6865,
+      "step": 215
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9405656999641965e-05,
+      "loss": 0.6639,
+      "step": 216
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.940207662012173e-05,
+      "loss": 0.6121,
+      "step": 217
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9398496240601505e-05,
+      "loss": 0.6661,
+      "step": 218
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.939491586108128e-05,
+      "loss": 0.6537,
+      "step": 219
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9391335481561044e-05,
+      "loss": 0.6549,
+      "step": 220
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.938775510204082e-05,
+      "loss": 0.6223,
+      "step": 221
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.938417472252059e-05,
+      "loss": 0.6496,
+      "step": 222
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.938059434300036e-05,
+      "loss": 0.6532,
+      "step": 223
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.937701396348013e-05,
+      "loss": 0.6497,
+      "step": 224
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.93734335839599e-05,
+      "loss": 0.6836,
+      "step": 225
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.936985320443968e-05,
+      "loss": 0.6646,
+      "step": 226
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.936627282491944e-05,
+      "loss": 0.6123,
+      "step": 227
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9362692445399216e-05,
+      "loss": 0.695,
+      "step": 228
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.935911206587899e-05,
+      "loss": 0.576,
+      "step": 229
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9355531686358756e-05,
+      "loss": 0.624,
+      "step": 230
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.935195130683853e-05,
+      "loss": 0.7198,
+      "step": 231
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.93483709273183e-05,
+      "loss": 0.6552,
+      "step": 232
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.934479054779807e-05,
+      "loss": 0.688,
+      "step": 233
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.934121016827784e-05,
+      "loss": 0.6205,
+      "step": 234
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.933762978875761e-05,
+      "loss": 0.6495,
+      "step": 235
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.933404940923738e-05,
+      "loss": 0.6722,
+      "step": 236
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9330469029717155e-05,
+      "loss": 0.6437,
+      "step": 237
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.932688865019692e-05,
+      "loss": 0.6648,
+      "step": 238
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9323308270676694e-05,
+      "loss": 0.629,
+      "step": 239
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.931972789115647e-05,
+      "loss": 0.7004,
+      "step": 240
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9316147511636234e-05,
+      "loss": 0.6584,
+      "step": 241
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.931256713211601e-05,
+      "loss": 0.6705,
+      "step": 242
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9308986752595774e-05,
+      "loss": 0.6621,
+      "step": 243
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.930540637307555e-05,
+      "loss": 0.6223,
+      "step": 244
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.930182599355532e-05,
+      "loss": 0.688,
+      "step": 245
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9298245614035086e-05,
+      "loss": 0.5823,
+      "step": 246
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.929466523451486e-05,
+      "loss": 0.6132,
+      "step": 247
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9291084854994626e-05,
+      "loss": 0.683,
+      "step": 248
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9287504475474406e-05,
+      "loss": 0.6604,
+      "step": 249
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.928392409595418e-05,
+      "loss": 0.6685,
+      "step": 250
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9280343716433946e-05,
+      "loss": 0.6335,
+      "step": 251
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.927676333691372e-05,
+      "loss": 0.7343,
+      "step": 252
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9273182957393485e-05,
+      "loss": 0.6402,
+      "step": 253
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.926960257787326e-05,
+      "loss": 0.6632,
+      "step": 254
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.926602219835303e-05,
+      "loss": 0.6252,
+      "step": 255
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.92624418188328e-05,
+      "loss": 0.6656,
+      "step": 256
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.925886143931257e-05,
+      "loss": 0.6923,
+      "step": 257
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.925528105979234e-05,
+      "loss": 0.657,
+      "step": 258
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.925170068027211e-05,
+      "loss": 0.672,
+      "step": 259
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9248120300751884e-05,
+      "loss": 0.6562,
+      "step": 260
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.924453992123165e-05,
+      "loss": 0.6776,
+      "step": 261
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9240959541711423e-05,
+      "loss": 0.6536,
+      "step": 262
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.92373791621912e-05,
+      "loss": 0.6506,
+      "step": 263
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.923379878267096e-05,
+      "loss": 0.6681,
+      "step": 264
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9230218403150736e-05,
+      "loss": 0.6001,
+      "step": 265
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.92266380236305e-05,
+      "loss": 0.6295,
+      "step": 266
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9223057644110276e-05,
+      "loss": 0.6362,
+      "step": 267
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.921947726459005e-05,
+      "loss": 0.6425,
+      "step": 268
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9215896885069815e-05,
+      "loss": 0.6809,
+      "step": 269
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.921231650554959e-05,
+      "loss": 0.6225,
+      "step": 270
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.920873612602936e-05,
+      "loss": 0.6077,
+      "step": 271
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9205155746509135e-05,
+      "loss": 0.6758,
+      "step": 272
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.920157536698891e-05,
+      "loss": 0.7356,
+      "step": 273
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9197994987468675e-05,
+      "loss": 0.6248,
+      "step": 274
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.919441460794845e-05,
+      "loss": 0.61,
+      "step": 275
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9190834228428214e-05,
+      "loss": 0.6932,
+      "step": 276
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.918725384890799e-05,
+      "loss": 0.5455,
+      "step": 277
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.918367346938776e-05,
+      "loss": 0.7632,
+      "step": 278
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.918009308986753e-05,
+      "loss": 0.7385,
+      "step": 279
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.91765127103473e-05,
+      "loss": 0.6461,
+      "step": 280
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9172932330827073e-05,
+      "loss": 0.6154,
+      "step": 281
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.916935195130684e-05,
+      "loss": 0.6747,
+      "step": 282
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.916577157178661e-05,
+      "loss": 0.6463,
+      "step": 283
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.916219119226638e-05,
+      "loss": 0.5522,
+      "step": 284
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.915861081274615e-05,
+      "loss": 0.7031,
+      "step": 285
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9155030433225926e-05,
+      "loss": 0.6883,
+      "step": 286
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.915145005370569e-05,
+      "loss": 0.6796,
+      "step": 287
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9147869674185465e-05,
+      "loss": 0.6623,
+      "step": 288
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.914428929466524e-05,
+      "loss": 0.6417,
+      "step": 289
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9140708915145005e-05,
+      "loss": 0.6574,
+      "step": 290
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.913712853562478e-05,
+      "loss": 0.6741,
+      "step": 291
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9133548156104545e-05,
+      "loss": 0.6445,
+      "step": 292
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.912996777658432e-05,
+      "loss": 0.6071,
+      "step": 293
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.912638739706409e-05,
+      "loss": 0.6877,
+      "step": 294
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.912280701754386e-05,
+      "loss": 0.6561,
+      "step": 295
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.911922663802364e-05,
+      "loss": 0.6447,
+      "step": 296
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9115646258503404e-05,
+      "loss": 0.6917,
+      "step": 297
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.911206587898318e-05,
+      "loss": 0.6759,
+      "step": 298
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.910848549946295e-05,
+      "loss": 0.6906,
+      "step": 299
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9104905119942717e-05,
+      "loss": 0.671,
+      "step": 300
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.910132474042249e-05,
+      "loss": 0.6472,
+      "step": 301
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9097744360902256e-05,
+      "loss": 0.6306,
+      "step": 302
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.909416398138203e-05,
+      "loss": 0.6318,
+      "step": 303
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.90905836018618e-05,
+      "loss": 0.7244,
+      "step": 304
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.908700322234157e-05,
+      "loss": 0.6813,
+      "step": 305
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.908342284282134e-05,
+      "loss": 0.7184,
+      "step": 306
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9079842463301115e-05,
+      "loss": 0.6831,
+      "step": 307
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.907626208378088e-05,
+      "loss": 0.6572,
+      "step": 308
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9072681704260655e-05,
+      "loss": 0.6687,
+      "step": 309
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.906910132474042e-05,
+      "loss": 0.6577,
+      "step": 310
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9065520945220194e-05,
+      "loss": 0.7001,
+      "step": 311
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.906194056569997e-05,
+      "loss": 0.6641,
+      "step": 312
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9058360186179734e-05,
+      "loss": 0.6668,
+      "step": 313
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.905477980665951e-05,
+      "loss": 0.6482,
+      "step": 314
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.905119942713928e-05,
+      "loss": 0.6939,
+      "step": 315
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.904761904761905e-05,
+      "loss": 0.636,
+      "step": 316
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.904403866809882e-05,
+      "loss": 0.685,
+      "step": 317
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9040458288578586e-05,
+      "loss": 0.6946,
+      "step": 318
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9036877909058366e-05,
+      "loss": 0.6496,
+      "step": 319
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.903329752953813e-05,
+      "loss": 0.6315,
+      "step": 320
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9029717150017906e-05,
+      "loss": 0.7018,
+      "step": 321
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.902613677049768e-05,
+      "loss": 0.6714,
+      "step": 322
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9022556390977446e-05,
+      "loss": 0.6552,
+      "step": 323
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.901897601145722e-05,
+      "loss": 0.6131,
+      "step": 324
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.901539563193699e-05,
+      "loss": 0.61,
+      "step": 325
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.901181525241676e-05,
+      "loss": 0.6432,
+      "step": 326
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.900823487289653e-05,
+      "loss": 0.687,
+      "step": 327
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.90046544933763e-05,
+      "loss": 0.6628,
+      "step": 328
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.900107411385607e-05,
+      "loss": 0.5912,
+      "step": 329
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8997493734335844e-05,
+      "loss": 0.7056,
+      "step": 330
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.899391335481561e-05,
+      "loss": 0.6474,
+      "step": 331
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8990332975295384e-05,
+      "loss": 0.6698,
+      "step": 332
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.898675259577516e-05,
+      "loss": 0.7126,
+      "step": 333
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8983172216254924e-05,
+      "loss": 0.6747,
+      "step": 334
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.89795918367347e-05,
+      "loss": 0.6181,
+      "step": 335
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.897601145721446e-05,
+      "loss": 0.6507,
+      "step": 336
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8972431077694236e-05,
+      "loss": 0.6724,
+      "step": 337
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.896885069817401e-05,
+      "loss": 0.6737,
+      "step": 338
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8965270318653776e-05,
+      "loss": 0.6476,
+      "step": 339
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.896168993913355e-05,
+      "loss": 0.6422,
+      "step": 340
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8958109559613316e-05,
+      "loss": 0.6096,
+      "step": 341
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8954529180093096e-05,
+      "loss": 0.5987,
+      "step": 342
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.895094880057287e-05,
+      "loss": 0.6941,
+      "step": 343
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8947368421052635e-05,
+      "loss": 0.6435,
+      "step": 344
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.894378804153241e-05,
+      "loss": 0.6444,
+      "step": 345
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8940207662012175e-05,
+      "loss": 0.6439,
+      "step": 346
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.893662728249195e-05,
+      "loss": 0.6361,
+      "step": 347
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.893304690297172e-05,
+      "loss": 0.6791,
+      "step": 348
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.892946652345149e-05,
+      "loss": 0.6774,
+      "step": 349
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.892588614393126e-05,
+      "loss": 0.6584,
+      "step": 350
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8922305764411034e-05,
+      "loss": 0.5579,
+      "step": 351
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.89187253848908e-05,
+      "loss": 0.68,
+      "step": 352
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8915145005370573e-05,
+      "loss": 0.6111,
+      "step": 353
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.891156462585034e-05,
+      "loss": 0.7083,
+      "step": 354
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.890798424633011e-05,
+      "loss": 0.5948,
+      "step": 355
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8904403866809886e-05,
+      "loss": 0.6101,
+      "step": 356
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.890082348728965e-05,
+      "loss": 0.659,
+      "step": 357
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8897243107769426e-05,
+      "loss": 0.5919,
+      "step": 358
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.889366272824919e-05,
+      "loss": 0.6367,
+      "step": 359
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8890082348728965e-05,
+      "loss": 0.5942,
+      "step": 360
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.888650196920874e-05,
+      "loss": 0.7403,
+      "step": 361
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8882921589688505e-05,
+      "loss": 0.6494,
+      "step": 362
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.887934121016828e-05,
+      "loss": 0.6883,
+      "step": 363
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.887576083064805e-05,
+      "loss": 0.7156,
+      "step": 364
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.887218045112782e-05,
+      "loss": 0.6999,
+      "step": 365
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.88686000716076e-05,
+      "loss": 0.621,
+      "step": 366
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8865019692087364e-05,
+      "loss": 0.6363,
+      "step": 367
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.886143931256714e-05,
+      "loss": 0.736,
+      "step": 368
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.8857858933046904e-05,
+      "loss": 0.6141,
+      "step": 369
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.885427855352668e-05,
+      "loss": 0.634,
+      "step": 370
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.885069817400645e-05,
+      "loss": 0.6934,
+      "step": 371
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.884711779448622e-05,
+      "loss": 0.6729,
+      "step": 372
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.884353741496599e-05,
+      "loss": 0.6844,
+      "step": 373
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.883995703544576e-05,
+      "loss": 0.6387,
+      "step": 374
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.883637665592553e-05,
+      "loss": 0.6455,
+      "step": 375
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.88327962764053e-05,
+      "loss": 0.6749,
+      "step": 376
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.882921589688507e-05,
+      "loss": 0.6524,
+      "step": 377
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.882563551736484e-05,
+      "loss": 0.6688,
+      "step": 378
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8822055137844615e-05,
+      "loss": 0.6294,
+      "step": 379
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.881847475832438e-05,
+      "loss": 0.6947,
+      "step": 380
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8814894378804155e-05,
+      "loss": 0.6572,
+      "step": 381
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.881131399928393e-05,
+      "loss": 0.6082,
+      "step": 382
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8807733619763695e-05,
+      "loss": 0.6239,
+      "step": 383
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.880415324024347e-05,
+      "loss": 0.6248,
+      "step": 384
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8800572860723234e-05,
+      "loss": 0.6568,
+      "step": 385
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.879699248120301e-05,
+      "loss": 0.6226,
+      "step": 386
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.879341210168278e-05,
+      "loss": 0.6218,
+      "step": 387
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.878983172216255e-05,
+      "loss": 0.7277,
+      "step": 388
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.878625134264233e-05,
+      "loss": 0.6453,
+      "step": 389
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.878267096312209e-05,
+      "loss": 0.6878,
+      "step": 390
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8779090583601867e-05,
+      "loss": 0.6557,
+      "step": 391
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.877551020408164e-05,
+      "loss": 0.6571,
+      "step": 392
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8771929824561406e-05,
+      "loss": 0.7255,
+      "step": 393
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.876834944504118e-05,
+      "loss": 0.6093,
+      "step": 394
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8764769065520946e-05,
+      "loss": 0.6427,
+      "step": 395
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.876118868600072e-05,
+      "loss": 0.6775,
+      "step": 396
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.875760830648049e-05,
+      "loss": 0.6224,
+      "step": 397
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.875402792696026e-05,
+      "loss": 0.6647,
+      "step": 398
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.875044754744003e-05,
+      "loss": 0.6915,
+      "step": 399
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8746867167919805e-05,
+      "loss": 0.6612,
+      "step": 400
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.874328678839957e-05,
+      "loss": 0.6733,
+      "step": 401
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8739706408879344e-05,
+      "loss": 0.704,
+      "step": 402
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.873612602935911e-05,
+      "loss": 0.6456,
+      "step": 403
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8732545649838884e-05,
+      "loss": 0.6645,
+      "step": 404
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.872896527031866e-05,
+      "loss": 0.6658,
+      "step": 405
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.8725384890798424e-05,
+      "loss": 0.5848,
+      "step": 406
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.87218045112782e-05,
+      "loss": 0.6454,
+      "step": 407
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.871822413175797e-05,
+      "loss": 0.6447,
+      "step": 408
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8714643752237736e-05,
+      "loss": 0.6263,
+      "step": 409
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.871106337271751e-05,
+      "loss": 0.6417,
+      "step": 410
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8707482993197276e-05,
+      "loss": 0.6794,
+      "step": 411
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8703902613677056e-05,
+      "loss": 0.6687,
+      "step": 412
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.870032223415682e-05,
+      "loss": 0.6588,
+      "step": 413
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8696741854636596e-05,
+      "loss": 0.6687,
+      "step": 414
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.869316147511637e-05,
+      "loss": 0.6341,
+      "step": 415
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8689581095596135e-05,
+      "loss": 0.6101,
+      "step": 416
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.868600071607591e-05,
+      "loss": 0.5921,
+      "step": 417
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.868242033655568e-05,
+      "loss": 0.6727,
+      "step": 418
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.867883995703545e-05,
+      "loss": 0.6331,
+      "step": 419
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.867525957751522e-05,
+      "loss": 0.7366,
+      "step": 420
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.867167919799499e-05,
+      "loss": 0.6447,
+      "step": 421
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.866809881847476e-05,
+      "loss": 0.6373,
+      "step": 422
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8664518438954534e-05,
+      "loss": 0.716,
+      "step": 423
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.86609380594343e-05,
+      "loss": 0.6327,
+      "step": 424
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8657357679914074e-05,
+      "loss": 0.6562,
+      "step": 425
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.865377730039385e-05,
+      "loss": 0.6641,
+      "step": 426
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.865019692087361e-05,
+      "loss": 0.6815,
+      "step": 427
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8646616541353386e-05,
+      "loss": 0.6255,
+      "step": 428
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.864303616183315e-05,
+      "loss": 0.6015,
+      "step": 429
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8639455782312926e-05,
+      "loss": 0.6325,
+      "step": 430
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.86358754027927e-05,
+      "loss": 0.7066,
+      "step": 431
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8632295023272466e-05,
+      "loss": 0.6477,
+      "step": 432
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.862871464375224e-05,
+      "loss": 0.6505,
+      "step": 433
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.8625134264232005e-05,
+      "loss": 0.6186,
+      "step": 434
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.862155388471178e-05,
+      "loss": 0.6616,
+      "step": 435
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.861797350519156e-05,
+      "loss": 0.5488,
+      "step": 436
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8614393125671325e-05,
+      "loss": 0.6712,
+      "step": 437
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.86108127461511e-05,
+      "loss": 0.6749,
+      "step": 438
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8607232366630864e-05,
+      "loss": 0.5795,
+      "step": 439
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.860365198711064e-05,
+      "loss": 0.653,
+      "step": 440
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.860007160759041e-05,
+      "loss": 0.7009,
+      "step": 441
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.859649122807018e-05,
+      "loss": 0.6194,
+      "step": 442
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.859291084854995e-05,
+      "loss": 0.6678,
+      "step": 443
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8589330469029724e-05,
+      "loss": 0.647,
+      "step": 444
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.858575008950949e-05,
+      "loss": 0.6413,
+      "step": 445
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.858216970998926e-05,
+      "loss": 0.6327,
+      "step": 446
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.857858933046903e-05,
+      "loss": 0.6109,
+      "step": 447
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.85750089509488e-05,
+      "loss": 0.6762,
+      "step": 448
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8571428571428576e-05,
+      "loss": 0.7022,
+      "step": 449
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.856784819190834e-05,
+      "loss": 0.6419,
+      "step": 450
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8564267812388115e-05,
+      "loss": 0.6153,
+      "step": 451
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.856068743286788e-05,
+      "loss": 0.6892,
+      "step": 452
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8557107053347655e-05,
+      "loss": 0.6952,
+      "step": 453
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.855352667382743e-05,
+      "loss": 0.583,
+      "step": 454
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8549946294307195e-05,
+      "loss": 0.6318,
+      "step": 455
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.854636591478697e-05,
+      "loss": 0.6183,
+      "step": 456
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.854278553526674e-05,
+      "loss": 0.5793,
+      "step": 457
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.853920515574651e-05,
+      "loss": 0.6385,
+      "step": 458
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.853562477622629e-05,
+      "loss": 0.664,
+      "step": 459
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.8532044396706054e-05,
+      "loss": 0.632,
+      "step": 460
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.852846401718583e-05,
+      "loss": 0.6475,
+      "step": 461
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.85248836376656e-05,
+      "loss": 0.6592,
+      "step": 462
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.852130325814537e-05,
+      "loss": 0.6733,
+      "step": 463
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.851772287862514e-05,
+      "loss": 0.6162,
+      "step": 464
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8514142499104906e-05,
+      "loss": 0.618,
+      "step": 465
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.851056211958468e-05,
+      "loss": 0.6883,
+      "step": 466
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.850698174006445e-05,
+      "loss": 0.6399,
+      "step": 467
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.850340136054422e-05,
+      "loss": 0.6409,
+      "step": 468
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.849982098102399e-05,
+      "loss": 0.6838,
+      "step": 469
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.849624060150376e-05,
+      "loss": 0.6646,
+      "step": 470
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.849266022198353e-05,
+      "loss": 0.5508,
+      "step": 471
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8489079842463305e-05,
+      "loss": 0.5931,
+      "step": 472
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.848549946294307e-05,
+      "loss": 0.66,
+      "step": 473
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8481919083422845e-05,
+      "loss": 0.6226,
+      "step": 474
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.847833870390262e-05,
+      "loss": 0.6174,
+      "step": 475
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8474758324382384e-05,
+      "loss": 0.6376,
+      "step": 476
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.847117794486216e-05,
+      "loss": 0.5678,
+      "step": 477
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8467597565341924e-05,
+      "loss": 0.687,
+      "step": 478
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.84640171858217e-05,
+      "loss": 0.6057,
+      "step": 479
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.846043680630147e-05,
+      "loss": 0.6648,
+      "step": 480
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8456856426781237e-05,
+      "loss": 0.6361,
+      "step": 481
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.845327604726101e-05,
+      "loss": 0.5769,
+      "step": 482
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.844969566774078e-05,
+      "loss": 0.5918,
+      "step": 483
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8446115288220556e-05,
+      "loss": 0.6066,
+      "step": 484
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.844253490870033e-05,
+      "loss": 0.6267,
+      "step": 485
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8438954529180096e-05,
+      "loss": 0.6467,
+      "step": 486
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.843537414965987e-05,
+      "loss": 0.6491,
+      "step": 487
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.8431793770139635e-05,
+      "loss": 0.8145,
+      "step": 488
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.842821339061941e-05,
+      "loss": 0.742,
+      "step": 489
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.842463301109918e-05,
+      "loss": 0.6979,
+      "step": 490
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.842105263157895e-05,
+      "loss": 0.6173,
+      "step": 491
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.841747225205872e-05,
+      "loss": 0.5638,
+      "step": 492
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8413891872538494e-05,
+      "loss": 0.7286,
+      "step": 493
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.841031149301826e-05,
+      "loss": 0.687,
+      "step": 494
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8406731113498034e-05,
+      "loss": 0.6512,
+      "step": 495
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.84031507339778e-05,
+      "loss": 0.6249,
+      "step": 496
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8399570354457574e-05,
+      "loss": 0.6255,
+      "step": 497
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.839598997493735e-05,
+      "loss": 0.6909,
+      "step": 498
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.839240959541711e-05,
+      "loss": 0.6348,
+      "step": 499
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8388829215896886e-05,
+      "loss": 0.6658,
+      "step": 500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.838524883637666e-05,
+      "loss": 0.644,
+      "step": 501
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8381668456856426e-05,
+      "loss": 0.6356,
+      "step": 502
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.83780880773362e-05,
+      "loss": 0.6128,
+      "step": 503
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8374507697815966e-05,
+      "loss": 0.6952,
+      "step": 504
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.837092731829574e-05,
+      "loss": 0.6279,
+      "step": 505
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.836734693877551e-05,
+      "loss": 0.6199,
+      "step": 506
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8363766559255285e-05,
+      "loss": 0.624,
+      "step": 507
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.836018617973506e-05,
+      "loss": 0.5719,
+      "step": 508
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8356605800214825e-05,
+      "loss": 0.6383,
+      "step": 509
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.83530254206946e-05,
+      "loss": 0.7009,
+      "step": 510
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.834944504117437e-05,
+      "loss": 0.6177,
+      "step": 511
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.834586466165414e-05,
+      "loss": 0.6602,
+      "step": 512
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.834228428213391e-05,
+      "loss": 0.6886,
+      "step": 513
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.833870390261368e-05,
+      "loss": 0.6373,
+      "step": 514
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.833512352309345e-05,
+      "loss": 0.6096,
+      "step": 515
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.8331543143573224e-05,
+      "loss": 0.6428,
+      "step": 516
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.832796276405299e-05,
+      "loss": 0.624,
+      "step": 517
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.832438238453276e-05,
+      "loss": 0.5812,
+      "step": 518
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8320802005012536e-05,
+      "loss": 0.6127,
+      "step": 519
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.83172216254923e-05,
+      "loss": 0.5992,
+      "step": 520
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8313641245972076e-05,
+      "loss": 0.6678,
+      "step": 521
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.831006086645184e-05,
+      "loss": 0.6842,
+      "step": 522
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8306480486931616e-05,
+      "loss": 0.7032,
+      "step": 523
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.830290010741139e-05,
+      "loss": 0.6641,
+      "step": 524
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8299319727891155e-05,
+      "loss": 0.7118,
+      "step": 525
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.829573934837093e-05,
+      "loss": 0.7305,
+      "step": 526
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8292158968850695e-05,
+      "loss": 0.6732,
+      "step": 527
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.828857858933047e-05,
+      "loss": 0.6677,
+      "step": 528
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.828499820981025e-05,
+      "loss": 0.6551,
+      "step": 529
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8281417830290014e-05,
+      "loss": 0.6346,
+      "step": 530
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.827783745076979e-05,
+      "loss": 0.6542,
+      "step": 531
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8274257071249554e-05,
+      "loss": 0.6681,
+      "step": 532
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.827067669172933e-05,
+      "loss": 0.7053,
+      "step": 533
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.82670963122091e-05,
+      "loss": 0.6796,
+      "step": 534
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.826351593268887e-05,
+      "loss": 0.6098,
+      "step": 535
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.825993555316864e-05,
+      "loss": 0.6273,
+      "step": 536
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.825635517364841e-05,
+      "loss": 0.645,
+      "step": 537
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.825277479412818e-05,
+      "loss": 0.6251,
+      "step": 538
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.824919441460795e-05,
+      "loss": 0.6119,
+      "step": 539
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.824561403508772e-05,
+      "loss": 0.6612,
+      "step": 540
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.824203365556749e-05,
+      "loss": 0.6999,
+      "step": 541
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8238453276047265e-05,
+      "loss": 0.5999,
+      "step": 542
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.823487289652703e-05,
+      "loss": 0.6121,
+      "step": 543
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8231292517006805e-05,
+      "loss": 0.7019,
+      "step": 544
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.822771213748657e-05,
+      "loss": 0.6551,
+      "step": 545
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.8224131757966345e-05,
+      "loss": 0.6495,
+      "step": 546
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.822055137844612e-05,
+      "loss": 0.6392,
+      "step": 547
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8216970998925884e-05,
+      "loss": 0.6866,
+      "step": 548
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.821339061940566e-05,
+      "loss": 0.5955,
+      "step": 549
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.820981023988543e-05,
+      "loss": 0.6209,
+      "step": 550
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.82062298603652e-05,
+      "loss": 0.6364,
+      "step": 551
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.820264948084497e-05,
+      "loss": 0.6199,
+      "step": 552
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8199069101324743e-05,
+      "loss": 0.7362,
+      "step": 553
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.819548872180452e-05,
+      "loss": 0.6252,
+      "step": 554
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.819190834228429e-05,
+      "loss": 0.6603,
+      "step": 555
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8188327962764056e-05,
+      "loss": 0.6274,
+      "step": 556
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.818474758324383e-05,
+      "loss": 0.6248,
+      "step": 557
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8181167203723596e-05,
+      "loss": 0.6941,
+      "step": 558
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.817758682420337e-05,
+      "loss": 0.6473,
+      "step": 559
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.817400644468314e-05,
+      "loss": 0.6254,
+      "step": 560
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.817042606516291e-05,
+      "loss": 0.616,
+      "step": 561
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.816684568564268e-05,
+      "loss": 0.6774,
+      "step": 562
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.816326530612245e-05,
+      "loss": 0.6736,
+      "step": 563
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.815968492660222e-05,
+      "loss": 0.6353,
+      "step": 564
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8156104547081995e-05,
+      "loss": 0.6226,
+      "step": 565
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.815252416756176e-05,
+      "loss": 0.7304,
+      "step": 566
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8148943788041534e-05,
+      "loss": 0.608,
+      "step": 567
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.814536340852131e-05,
+      "loss": 0.5938,
+      "step": 568
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8141783029001074e-05,
+      "loss": 0.6176,
+      "step": 569
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.813820264948085e-05,
+      "loss": 0.6421,
+      "step": 570
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8134622269960613e-05,
+      "loss": 0.621,
+      "step": 571
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8131041890440387e-05,
+      "loss": 0.6653,
+      "step": 572
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.812746151092016e-05,
+      "loss": 0.6567,
+      "step": 573
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.8123881131399926e-05,
+      "loss": 0.6493,
+      "step": 574
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.81203007518797e-05,
+      "loss": 0.6662,
+      "step": 575
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.811672037235947e-05,
+      "loss": 0.6493,
+      "step": 576
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8113139992839246e-05,
+      "loss": 0.6751,
+      "step": 577
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.810955961331902e-05,
+      "loss": 0.6376,
+      "step": 578
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8105979233798785e-05,
+      "loss": 0.6608,
+      "step": 579
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.810239885427856e-05,
+      "loss": 0.6628,
+      "step": 580
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8098818474758325e-05,
+      "loss": 0.5823,
+      "step": 581
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.80952380952381e-05,
+      "loss": 0.6506,
+      "step": 582
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.809165771571787e-05,
+      "loss": 0.6328,
+      "step": 583
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.808807733619764e-05,
+      "loss": 0.6182,
+      "step": 584
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.808449695667741e-05,
+      "loss": 0.6496,
+      "step": 585
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8080916577157184e-05,
+      "loss": 0.6087,
+      "step": 586
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.807733619763695e-05,
+      "loss": 0.6368,
+      "step": 587
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8073755818116724e-05,
+      "loss": 0.737,
+      "step": 588
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.807017543859649e-05,
+      "loss": 0.6798,
+      "step": 589
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.806659505907626e-05,
+      "loss": 0.635,
+      "step": 590
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8063014679556036e-05,
+      "loss": 0.7139,
+      "step": 591
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.80594343000358e-05,
+      "loss": 0.644,
+      "step": 592
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8055853920515576e-05,
+      "loss": 0.6121,
+      "step": 593
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.805227354099535e-05,
+      "loss": 0.6477,
+      "step": 594
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8048693161475116e-05,
+      "loss": 0.6887,
+      "step": 595
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.804511278195489e-05,
+      "loss": 0.6153,
+      "step": 596
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8041532402434655e-05,
+      "loss": 0.5794,
+      "step": 597
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.803795202291443e-05,
+      "loss": 0.662,
+      "step": 598
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.80343716433942e-05,
+      "loss": 0.6223,
+      "step": 599
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8030791263873975e-05,
+      "loss": 0.6186,
+      "step": 600
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.802721088435375e-05,
+      "loss": 0.6053,
+      "step": 601
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.8023630504833514e-05,
+      "loss": 0.658,
+      "step": 602
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.802005012531329e-05,
+      "loss": 0.6329,
+      "step": 603
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.801646974579306e-05,
+      "loss": 0.5941,
+      "step": 604
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.801288936627283e-05,
+      "loss": 0.6245,
+      "step": 605
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.80093089867526e-05,
+      "loss": 0.6253,
+      "step": 606
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.800572860723237e-05,
+      "loss": 0.7008,
+      "step": 607
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.800214822771214e-05,
+      "loss": 0.629,
+      "step": 608
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.799856784819191e-05,
+      "loss": 0.6127,
+      "step": 609
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.799498746867168e-05,
+      "loss": 0.6018,
+      "step": 610
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.799140708915145e-05,
+      "loss": 0.6164,
+      "step": 611
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7987826709631226e-05,
+      "loss": 0.6795,
+      "step": 612
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.798424633011099e-05,
+      "loss": 0.6771,
+      "step": 613
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7980665950590766e-05,
+      "loss": 0.5993,
+      "step": 614
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.797708557107053e-05,
+      "loss": 0.5903,
+      "step": 615
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7973505191550305e-05,
+      "loss": 0.6103,
+      "step": 616
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.796992481203008e-05,
+      "loss": 0.6272,
+      "step": 617
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7966344432509845e-05,
+      "loss": 0.744,
+      "step": 618
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.796276405298962e-05,
+      "loss": 0.7213,
+      "step": 619
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.795918367346939e-05,
+      "loss": 0.654,
+      "step": 620
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.795560329394916e-05,
+      "loss": 0.5945,
+      "step": 621
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.795202291442893e-05,
+      "loss": 0.6277,
+      "step": 622
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7948442534908704e-05,
+      "loss": 0.6289,
+      "step": 623
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.794486215538848e-05,
+      "loss": 0.6487,
+      "step": 624
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7941281775868244e-05,
+      "loss": 0.6325,
+      "step": 625
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.793770139634802e-05,
+      "loss": 0.7204,
+      "step": 626
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.793412101682779e-05,
+      "loss": 0.6242,
+      "step": 627
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.7930540637307556e-05,
+      "loss": 0.6503,
+      "step": 628
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.792696025778733e-05,
+      "loss": 0.6597,
+      "step": 629
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.79233798782671e-05,
+      "loss": 0.6289,
+      "step": 630
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.791979949874687e-05,
+      "loss": 0.6229,
+      "step": 631
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.791621911922664e-05,
+      "loss": 0.6277,
+      "step": 632
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.791263873970641e-05,
+      "loss": 0.643,
+      "step": 633
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.790905836018618e-05,
+      "loss": 0.5702,
+      "step": 634
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7905477980665955e-05,
+      "loss": 0.6031,
+      "step": 635
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.790189760114572e-05,
+      "loss": 0.6371,
+      "step": 636
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7898317221625495e-05,
+      "loss": 0.7021,
+      "step": 637
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.789473684210526e-05,
+      "loss": 0.6573,
+      "step": 638
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7891156462585034e-05,
+      "loss": 0.6184,
+      "step": 639
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.788757608306481e-05,
+      "loss": 0.6831,
+      "step": 640
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7883995703544574e-05,
+      "loss": 0.5919,
+      "step": 641
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.788041532402435e-05,
+      "loss": 0.6523,
+      "step": 642
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.787683494450412e-05,
+      "loss": 0.6552,
+      "step": 643
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.787325456498389e-05,
+      "loss": 0.6059,
+      "step": 644
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.786967418546366e-05,
+      "loss": 0.6209,
+      "step": 645
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.786609380594343e-05,
+      "loss": 0.6632,
+      "step": 646
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7862513426423206e-05,
+      "loss": 0.6659,
+      "step": 647
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.785893304690298e-05,
+      "loss": 0.6899,
+      "step": 648
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7855352667382746e-05,
+      "loss": 0.5999,
+      "step": 649
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.785177228786252e-05,
+      "loss": 0.5852,
+      "step": 650
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7848191908342285e-05,
+      "loss": 0.6576,
+      "step": 651
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.784461152882206e-05,
+      "loss": 0.7359,
+      "step": 652
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.784103114930183e-05,
+      "loss": 0.5948,
+      "step": 653
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.78374507697816e-05,
+      "loss": 0.6502,
+      "step": 654
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.783387039026137e-05,
+      "loss": 0.6524,
+      "step": 655
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.783029001074114e-05,
+      "loss": 0.5936,
+      "step": 656
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.782670963122091e-05,
+      "loss": 0.6206,
+      "step": 657
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.7823129251700684e-05,
+      "loss": 0.6518,
+      "step": 658
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.781954887218045e-05,
+      "loss": 0.6619,
+      "step": 659
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7815968492660224e-05,
+      "loss": 0.6513,
+      "step": 660
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.781238811314e-05,
+      "loss": 0.6366,
+      "step": 661
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7808807733619763e-05,
+      "loss": 0.5907,
+      "step": 662
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7805227354099537e-05,
+      "loss": 0.6417,
+      "step": 663
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.78016469745793e-05,
+      "loss": 0.6137,
+      "step": 664
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7798066595059076e-05,
+      "loss": 0.5948,
+      "step": 665
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.779448621553885e-05,
+      "loss": 0.6987,
+      "step": 666
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7790905836018616e-05,
+      "loss": 0.6289,
+      "step": 667
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.778732545649839e-05,
+      "loss": 0.6208,
+      "step": 668
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.778374507697816e-05,
+      "loss": 0.6273,
+      "step": 669
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7780164697457935e-05,
+      "loss": 0.5621,
+      "step": 670
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.777658431793771e-05,
+      "loss": 0.6657,
+      "step": 671
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7773003938417475e-05,
+      "loss": 0.579,
+      "step": 672
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.776942355889725e-05,
+      "loss": 0.582,
+      "step": 673
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7765843179377015e-05,
+      "loss": 0.6313,
+      "step": 674
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.776226279985679e-05,
+      "loss": 0.674,
+      "step": 675
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.775868242033656e-05,
+      "loss": 0.7313,
+      "step": 676
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.775510204081633e-05,
+      "loss": 0.6226,
+      "step": 677
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.77515216612961e-05,
+      "loss": 0.4999,
+      "step": 678
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7747941281775874e-05,
+      "loss": 0.6622,
+      "step": 679
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.774436090225564e-05,
+      "loss": 0.6635,
+      "step": 680
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.774078052273541e-05,
+      "loss": 0.714,
+      "step": 681
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.773720014321518e-05,
+      "loss": 0.68,
+      "step": 682
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.773361976369495e-05,
+      "loss": 0.602,
+      "step": 683
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7730039384174726e-05,
+      "loss": 0.5681,
+      "step": 684
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.772645900465449e-05,
+      "loss": 0.684,
+      "step": 685
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.7722878625134266e-05,
+      "loss": 0.589,
+      "step": 686
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.771929824561404e-05,
+      "loss": 0.6831,
+      "step": 687
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7715717866093805e-05,
+      "loss": 0.6739,
+      "step": 688
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.771213748657358e-05,
+      "loss": 0.6169,
+      "step": 689
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7708557107053345e-05,
+      "loss": 0.676,
+      "step": 690
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.770497672753312e-05,
+      "loss": 0.6173,
+      "step": 691
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.770139634801289e-05,
+      "loss": 0.6763,
+      "step": 692
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7697815968492664e-05,
+      "loss": 0.6784,
+      "step": 693
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.769423558897244e-05,
+      "loss": 0.6773,
+      "step": 694
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7690655209452204e-05,
+      "loss": 0.5804,
+      "step": 695
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.768707482993198e-05,
+      "loss": 0.6445,
+      "step": 696
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.768349445041175e-05,
+      "loss": 0.6043,
+      "step": 697
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.767991407089152e-05,
+      "loss": 0.7136,
+      "step": 698
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.767633369137129e-05,
+      "loss": 0.6129,
+      "step": 699
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7672753311851056e-05,
+      "loss": 0.6159,
+      "step": 700
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.766917293233083e-05,
+      "loss": 0.6054,
+      "step": 701
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.76655925528106e-05,
+      "loss": 0.6427,
+      "step": 702
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.766201217329037e-05,
+      "loss": 0.641,
+      "step": 703
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.765843179377014e-05,
+      "loss": 0.6576,
+      "step": 704
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7654851414249916e-05,
+      "loss": 0.6144,
+      "step": 705
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.765127103472968e-05,
+      "loss": 0.5989,
+      "step": 706
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7647690655209455e-05,
+      "loss": 0.632,
+      "step": 707
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.764411027568922e-05,
+      "loss": 0.6109,
+      "step": 708
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7640529896168995e-05,
+      "loss": 0.5963,
+      "step": 709
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.763694951664877e-05,
+      "loss": 0.5481,
+      "step": 710
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.7633369137128534e-05,
+      "loss": 0.6597,
+      "step": 711
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.762978875760831e-05,
+      "loss": 0.6289,
+      "step": 712
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.762620837808808e-05,
+      "loss": 0.585,
+      "step": 713
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.762262799856785e-05,
+      "loss": 0.6017,
+      "step": 714
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.761904761904762e-05,
+      "loss": 0.6541,
+      "step": 715
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7615467239527394e-05,
+      "loss": 0.5871,
+      "step": 716
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.761188686000717e-05,
+      "loss": 0.6162,
+      "step": 717
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.760830648048693e-05,
+      "loss": 0.6705,
+      "step": 718
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7604726100966706e-05,
+      "loss": 0.688,
+      "step": 719
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.760114572144648e-05,
+      "loss": 0.6886,
+      "step": 720
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7597565341926246e-05,
+      "loss": 0.6604,
+      "step": 721
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.759398496240602e-05,
+      "loss": 0.6039,
+      "step": 722
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.759040458288579e-05,
+      "loss": 0.6018,
+      "step": 723
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.758682420336556e-05,
+      "loss": 0.6566,
+      "step": 724
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.758324382384533e-05,
+      "loss": 0.6245,
+      "step": 725
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.75796634443251e-05,
+      "loss": 0.5896,
+      "step": 726
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.757608306480487e-05,
+      "loss": 0.615,
+      "step": 727
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7572502685284645e-05,
+      "loss": 0.5735,
+      "step": 728
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.756892230576441e-05,
+      "loss": 0.6692,
+      "step": 729
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7565341926244184e-05,
+      "loss": 0.5977,
+      "step": 730
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.756176154672396e-05,
+      "loss": 0.6393,
+      "step": 731
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7558181167203724e-05,
+      "loss": 0.624,
+      "step": 732
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.75546007876835e-05,
+      "loss": 0.6192,
+      "step": 733
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7551020408163263e-05,
+      "loss": 0.6107,
+      "step": 734
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.754744002864304e-05,
+      "loss": 0.6727,
+      "step": 735
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.754385964912281e-05,
+      "loss": 0.6203,
+      "step": 736
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7540279269602576e-05,
+      "loss": 0.589,
+      "step": 737
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.753669889008235e-05,
+      "loss": 0.5663,
+      "step": 738
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.753311851056212e-05,
+      "loss": 0.6419,
+      "step": 739
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7529538131041896e-05,
+      "loss": 0.7377,
+      "step": 740
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.752595775152167e-05,
+      "loss": 0.6478,
+      "step": 741
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.7522377372001435e-05,
+      "loss": 0.6434,
+      "step": 742
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.751879699248121e-05,
+      "loss": 0.5502,
+      "step": 743
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7515216612960975e-05,
+      "loss": 0.5807,
+      "step": 744
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.751163623344075e-05,
+      "loss": 0.6415,
+      "step": 745
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.750805585392052e-05,
+      "loss": 0.6201,
+      "step": 746
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.750447547440029e-05,
+      "loss": 0.6885,
+      "step": 747
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.750089509488006e-05,
+      "loss": 0.6618,
+      "step": 748
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.749731471535983e-05,
+      "loss": 0.629,
+      "step": 749
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.74937343358396e-05,
+      "loss": 0.6916,
+      "step": 750
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7490153956319374e-05,
+      "loss": 0.6061,
+      "step": 751
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.748657357679914e-05,
+      "loss": 0.6416,
+      "step": 752
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7482993197278913e-05,
+      "loss": 0.639,
+      "step": 753
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7479412817758687e-05,
+      "loss": 0.6176,
+      "step": 754
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.747583243823845e-05,
+      "loss": 0.686,
+      "step": 755
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7472252058718226e-05,
+      "loss": 0.5925,
+      "step": 756
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.746867167919799e-05,
+      "loss": 0.6369,
+      "step": 757
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7465091299677766e-05,
+      "loss": 0.6552,
+      "step": 758
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.746151092015754e-05,
+      "loss": 0.5725,
+      "step": 759
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7457930540637305e-05,
+      "loss": 0.5793,
+      "step": 760
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.745435016111708e-05,
+      "loss": 0.651,
+      "step": 761
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.745076978159685e-05,
+      "loss": 0.5542,
+      "step": 762
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7447189402076625e-05,
+      "loss": 0.5755,
+      "step": 763
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.74436090225564e-05,
+      "loss": 0.5698,
+      "step": 764
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7440028643036165e-05,
+      "loss": 0.5871,
+      "step": 765
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.743644826351594e-05,
+      "loss": 0.5948,
+      "step": 766
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.7432867883995704e-05,
+      "loss": 0.5909,
+      "step": 767
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.742928750447548e-05,
+      "loss": 0.571,
+      "step": 768
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.742570712495525e-05,
+      "loss": 0.6713,
+      "step": 769
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.742212674543502e-05,
+      "loss": 0.6979,
+      "step": 770
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.741854636591479e-05,
+      "loss": 0.6522,
+      "step": 771
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.741496598639456e-05,
+      "loss": 0.5555,
+      "step": 772
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.741138560687433e-05,
+      "loss": 0.6757,
+      "step": 773
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.74078052273541e-05,
+      "loss": 0.6042,
+      "step": 774
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.740422484783387e-05,
+      "loss": 0.6209,
+      "step": 775
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.740064446831364e-05,
+      "loss": 0.6783,
+      "step": 776
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7397064088793416e-05,
+      "loss": 0.6464,
+      "step": 777
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.739348370927318e-05,
+      "loss": 0.5771,
+      "step": 778
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7389903329752955e-05,
+      "loss": 0.6112,
+      "step": 779
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.738632295023273e-05,
+      "loss": 0.5757,
+      "step": 780
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7382742570712495e-05,
+      "loss": 0.5478,
+      "step": 781
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.737916219119227e-05,
+      "loss": 0.6216,
+      "step": 782
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7375581811672034e-05,
+      "loss": 0.6566,
+      "step": 783
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.737200143215181e-05,
+      "loss": 0.5559,
+      "step": 784
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.736842105263158e-05,
+      "loss": 0.5706,
+      "step": 785
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7364840673111354e-05,
+      "loss": 0.7015,
+      "step": 786
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.736126029359113e-05,
+      "loss": 0.5659,
+      "step": 787
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7357679914070894e-05,
+      "loss": 0.6617,
+      "step": 788
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.735409953455067e-05,
+      "loss": 0.5842,
+      "step": 789
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.735051915503044e-05,
+      "loss": 0.5757,
+      "step": 790
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7346938775510206e-05,
+      "loss": 0.5771,
+      "step": 791
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.734335839598998e-05,
+      "loss": 0.6813,
+      "step": 792
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7339778016469746e-05,
+      "loss": 0.5735,
+      "step": 793
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.733619763694952e-05,
+      "loss": 0.5787,
+      "step": 794
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.733261725742929e-05,
+      "loss": 0.5939,
+      "step": 795
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.732903687790906e-05,
+      "loss": 0.5919,
+      "step": 796
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.732545649838883e-05,
+      "loss": 0.759,
+      "step": 797
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.7321876118868605e-05,
+      "loss": 0.5852,
+      "step": 798
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.731829573934837e-05,
+      "loss": 0.691,
+      "step": 799
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7314715359828145e-05,
+      "loss": 0.7036,
+      "step": 800
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.731113498030791e-05,
+      "loss": 0.6744,
+      "step": 801
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7307554600787684e-05,
+      "loss": 0.5779,
+      "step": 802
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.730397422126746e-05,
+      "loss": 0.616,
+      "step": 803
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7300393841747224e-05,
+      "loss": 0.6257,
+      "step": 804
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7296813462227e-05,
+      "loss": 0.5997,
+      "step": 805
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.729323308270677e-05,
+      "loss": 0.6384,
+      "step": 806
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.728965270318654e-05,
+      "loss": 0.5652,
+      "step": 807
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.728607232366631e-05,
+      "loss": 0.6217,
+      "step": 808
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.728249194414608e-05,
+      "loss": 0.6238,
+      "step": 809
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7278911564625856e-05,
+      "loss": 0.6617,
+      "step": 810
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.727533118510562e-05,
+      "loss": 0.5764,
+      "step": 811
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7271750805585396e-05,
+      "loss": 0.5569,
+      "step": 812
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.726817042606517e-05,
+      "loss": 0.6364,
+      "step": 813
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7264590046544936e-05,
+      "loss": 0.6016,
+      "step": 814
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.726100966702471e-05,
+      "loss": 0.684,
+      "step": 815
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.725742928750448e-05,
+      "loss": 0.5959,
+      "step": 816
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.725384890798425e-05,
+      "loss": 0.6257,
+      "step": 817
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.725026852846402e-05,
+      "loss": 0.61,
+      "step": 818
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.724668814894379e-05,
+      "loss": 0.6271,
+      "step": 819
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.724310776942356e-05,
+      "loss": 0.5828,
+      "step": 820
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7239527389903334e-05,
+      "loss": 0.6316,
+      "step": 821
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.72359470103831e-05,
+      "loss": 0.6191,
+      "step": 822
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7232366630862874e-05,
+      "loss": 0.584,
+      "step": 823
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.722878625134265e-05,
+      "loss": 0.6755,
+      "step": 824
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.7225205871822413e-05,
+      "loss": 0.6218,
+      "step": 825
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.722162549230219e-05,
+      "loss": 0.5374,
+      "step": 826
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.721804511278195e-05,
+      "loss": 0.621,
+      "step": 827
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7214464733261726e-05,
+      "loss": 0.6616,
+      "step": 828
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.72108843537415e-05,
+      "loss": 0.5874,
+      "step": 829
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7207303974221266e-05,
+      "loss": 0.6147,
+      "step": 830
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.720372359470104e-05,
+      "loss": 0.5938,
+      "step": 831
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7200143215180805e-05,
+      "loss": 0.6514,
+      "step": 832
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7196562835660585e-05,
+      "loss": 0.6204,
+      "step": 833
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.719298245614036e-05,
+      "loss": 0.5867,
+      "step": 834
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7189402076620125e-05,
+      "loss": 0.642,
+      "step": 835
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.71858216970999e-05,
+      "loss": 0.5799,
+      "step": 836
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7182241317579665e-05,
+      "loss": 0.67,
+      "step": 837
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.717866093805944e-05,
+      "loss": 0.5356,
+      "step": 838
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.717508055853921e-05,
+      "loss": 0.566,
+      "step": 839
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.717150017901898e-05,
+      "loss": 0.6405,
+      "step": 840
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.716791979949875e-05,
+      "loss": 0.6777,
+      "step": 841
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7164339419978524e-05,
+      "loss": 0.6366,
+      "step": 842
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.716075904045829e-05,
+      "loss": 0.6348,
+      "step": 843
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7157178660938063e-05,
+      "loss": 0.6344,
+      "step": 844
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.715359828141783e-05,
+      "loss": 0.7312,
+      "step": 845
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.71500179018976e-05,
+      "loss": 0.6219,
+      "step": 846
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7146437522377376e-05,
+      "loss": 0.552,
+      "step": 847
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.714285714285714e-05,
+      "loss": 0.6205,
+      "step": 848
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7139276763336916e-05,
+      "loss": 0.6674,
+      "step": 849
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.713569638381668e-05,
+      "loss": 0.5688,
+      "step": 850
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7132116004296455e-05,
+      "loss": 0.6034,
+      "step": 851
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.712853562477623e-05,
+      "loss": 0.5472,
+      "step": 852
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.7124955245255995e-05,
+      "loss": 0.6654,
+      "step": 853
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.712137486573577e-05,
+      "loss": 0.6103,
+      "step": 854
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.711779448621554e-05,
+      "loss": 0.6437,
+      "step": 855
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7114214106695315e-05,
+      "loss": 0.5895,
+      "step": 856
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.711063372717509e-05,
+      "loss": 0.6347,
+      "step": 857
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7107053347654854e-05,
+      "loss": 0.6135,
+      "step": 858
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.710347296813463e-05,
+      "loss": 0.5635,
+      "step": 859
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7099892588614394e-05,
+      "loss": 0.5908,
+      "step": 860
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.709631220909417e-05,
+      "loss": 0.5871,
+      "step": 861
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.709273182957394e-05,
+      "loss": 0.6856,
+      "step": 862
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7089151450053707e-05,
+      "loss": 0.5981,
+      "step": 863
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.708557107053348e-05,
+      "loss": 0.6354,
+      "step": 864
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.708199069101325e-05,
+      "loss": 0.5895,
+      "step": 865
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.707841031149302e-05,
+      "loss": 0.6859,
+      "step": 866
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.707482993197279e-05,
+      "loss": 0.5745,
+      "step": 867
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.707124955245256e-05,
+      "loss": 0.6336,
+      "step": 868
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.706766917293233e-05,
+      "loss": 0.6705,
+      "step": 869
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7064088793412105e-05,
+      "loss": 0.6259,
+      "step": 870
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.706050841389187e-05,
+      "loss": 0.6278,
+      "step": 871
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7056928034371645e-05,
+      "loss": 0.602,
+      "step": 872
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.705334765485142e-05,
+      "loss": 0.6464,
+      "step": 873
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7049767275331184e-05,
+      "loss": 0.7884,
+      "step": 874
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.704618689581096e-05,
+      "loss": 0.5779,
+      "step": 875
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7042606516290724e-05,
+      "loss": 0.6103,
+      "step": 876
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.70390261367705e-05,
+      "loss": 0.6722,
+      "step": 877
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.703544575725027e-05,
+      "loss": 0.6517,
+      "step": 878
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.703186537773004e-05,
+      "loss": 0.5835,
+      "step": 879
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.702828499820982e-05,
+      "loss": 0.5853,
+      "step": 880
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.702470461868958e-05,
+      "loss": 0.5948,
+      "step": 881
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.7021124239169356e-05,
+      "loss": 0.5676,
+      "step": 882
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.701754385964913e-05,
+      "loss": 0.6713,
+      "step": 883
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.7013963480128896e-05,
+      "loss": 0.5899,
+      "step": 884
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.701038310060867e-05,
+      "loss": 0.5555,
+      "step": 885
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.7006802721088436e-05,
+      "loss": 0.5958,
+      "step": 886
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.700322234156821e-05,
+      "loss": 0.6801,
+      "step": 887
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.699964196204798e-05,
+      "loss": 0.6851,
+      "step": 888
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.699606158252775e-05,
+      "loss": 0.6117,
+      "step": 889
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.699248120300752e-05,
+      "loss": 0.621,
+      "step": 890
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6988900823487295e-05,
+      "loss": 0.5166,
+      "step": 891
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.698532044396706e-05,
+      "loss": 0.6238,
+      "step": 892
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6981740064446834e-05,
+      "loss": 0.5591,
+      "step": 893
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.69781596849266e-05,
+      "loss": 0.6256,
+      "step": 894
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6974579305406374e-05,
+      "loss": 0.6114,
+      "step": 895
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.697099892588615e-05,
+      "loss": 0.6451,
+      "step": 896
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6967418546365914e-05,
+      "loss": 0.6534,
+      "step": 897
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.696383816684569e-05,
+      "loss": 0.7154,
+      "step": 898
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.696025778732546e-05,
+      "loss": 0.6499,
+      "step": 899
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6956677407805226e-05,
+      "loss": 0.6366,
+      "step": 900
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6953097028285e-05,
+      "loss": 0.6129,
+      "step": 901
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6949516648764766e-05,
+      "loss": 0.5906,
+      "step": 902
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6945936269244546e-05,
+      "loss": 0.6005,
+      "step": 903
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.694235588972431e-05,
+      "loss": 0.7187,
+      "step": 904
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6938775510204086e-05,
+      "loss": 0.638,
+      "step": 905
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.693519513068386e-05,
+      "loss": 0.5951,
+      "step": 906
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.6931614751163625e-05,
+      "loss": 0.6234,
+      "step": 907
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.69280343716434e-05,
+      "loss": 0.6947,
+      "step": 908
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.692445399212317e-05,
+      "loss": 0.5582,
+      "step": 909
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.692087361260294e-05,
+      "loss": 0.5085,
+      "step": 910
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.691729323308271e-05,
+      "loss": 0.6483,
+      "step": 911
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.691371285356248e-05,
+      "loss": 0.5519,
+      "step": 912
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.691013247404225e-05,
+      "loss": 0.7038,
+      "step": 913
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6906552094522024e-05,
+      "loss": 0.5783,
+      "step": 914
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.690297171500179e-05,
+      "loss": 0.5942,
+      "step": 915
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6899391335481563e-05,
+      "loss": 0.6228,
+      "step": 916
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.689581095596134e-05,
+      "loss": 0.6308,
+      "step": 917
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.68922305764411e-05,
+      "loss": 0.6745,
+      "step": 918
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6888650196920876e-05,
+      "loss": 0.6317,
+      "step": 919
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.688506981740064e-05,
+      "loss": 0.6484,
+      "step": 920
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6881489437880416e-05,
+      "loss": 0.676,
+      "step": 921
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.687790905836019e-05,
+      "loss": 0.6246,
+      "step": 922
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6874328678839955e-05,
+      "loss": 0.7367,
+      "step": 923
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.687074829931973e-05,
+      "loss": 0.6958,
+      "step": 924
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6867167919799495e-05,
+      "loss": 0.6646,
+      "step": 925
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6863587540279275e-05,
+      "loss": 0.6246,
+      "step": 926
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.686000716075905e-05,
+      "loss": 0.6119,
+      "step": 927
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6856426781238815e-05,
+      "loss": 0.6369,
+      "step": 928
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.685284640171859e-05,
+      "loss": 0.6242,
+      "step": 929
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6849266022198354e-05,
+      "loss": 0.5706,
+      "step": 930
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.684568564267813e-05,
+      "loss": 0.6378,
+      "step": 931
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.68421052631579e-05,
+      "loss": 0.5909,
+      "step": 932
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.683852488363767e-05,
+      "loss": 0.6158,
+      "step": 933
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.683494450411744e-05,
+      "loss": 0.649,
+      "step": 934
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.6831364124597213e-05,
+      "loss": 0.6368,
+      "step": 935
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.682778374507698e-05,
+      "loss": 0.589,
+      "step": 936
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.682420336555675e-05,
+      "loss": 0.6531,
+      "step": 937
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.682062298603652e-05,
+      "loss": 0.6511,
+      "step": 938
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.681704260651629e-05,
+      "loss": 0.5871,
+      "step": 939
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6813462226996066e-05,
+      "loss": 0.5874,
+      "step": 940
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.680988184747583e-05,
+      "loss": 0.5381,
+      "step": 941
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6806301467955605e-05,
+      "loss": 0.6121,
+      "step": 942
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.680272108843537e-05,
+      "loss": 0.662,
+      "step": 943
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6799140708915145e-05,
+      "loss": 0.5993,
+      "step": 944
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.679556032939492e-05,
+      "loss": 0.6021,
+      "step": 945
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6791979949874685e-05,
+      "loss": 0.5491,
+      "step": 946
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.678839957035446e-05,
+      "loss": 0.519,
+      "step": 947
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.678481919083423e-05,
+      "loss": 0.6508,
+      "step": 948
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6781238811314e-05,
+      "loss": 0.6183,
+      "step": 949
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.677765843179378e-05,
+      "loss": 0.5984,
+      "step": 950
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6774078052273544e-05,
+      "loss": 0.6919,
+      "step": 951
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.677049767275332e-05,
+      "loss": 0.5597,
+      "step": 952
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.676691729323309e-05,
+      "loss": 0.558,
+      "step": 953
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6763336913712857e-05,
+      "loss": 0.6319,
+      "step": 954
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.675975653419263e-05,
+      "loss": 0.6414,
+      "step": 955
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6756176154672396e-05,
+      "loss": 0.7025,
+      "step": 956
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.675259577515217e-05,
+      "loss": 0.599,
+      "step": 957
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.674901539563194e-05,
+      "loss": 0.603,
+      "step": 958
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.674543501611171e-05,
+      "loss": 0.59,
+      "step": 959
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.674185463659148e-05,
+      "loss": 0.5935,
+      "step": 960
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.673827425707125e-05,
+      "loss": 0.6459,
+      "step": 961
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.673469387755102e-05,
+      "loss": 0.61,
+      "step": 962
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6731113498030795e-05,
+      "loss": 0.6742,
+      "step": 963
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.672753311851056e-05,
+      "loss": 0.4971,
+      "step": 964
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6723952738990334e-05,
+      "loss": 0.677,
+      "step": 965
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.672037235947011e-05,
+      "loss": 0.6509,
+      "step": 966
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.6716791979949874e-05,
+      "loss": 0.6588,
+      "step": 967
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.671321160042965e-05,
+      "loss": 0.6436,
+      "step": 968
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6709631220909414e-05,
+      "loss": 0.6011,
+      "step": 969
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.670605084138919e-05,
+      "loss": 0.5969,
+      "step": 970
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.670247046186896e-05,
+      "loss": 0.5635,
+      "step": 971
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6698890082348726e-05,
+      "loss": 0.5857,
+      "step": 972
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6695309702828506e-05,
+      "loss": 0.5805,
+      "step": 973
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.669172932330827e-05,
+      "loss": 0.5955,
+      "step": 974
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6688148943788046e-05,
+      "loss": 0.6117,
+      "step": 975
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.668456856426782e-05,
+      "loss": 0.5805,
+      "step": 976
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6680988184747586e-05,
+      "loss": 0.6374,
+      "step": 977
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.667740780522736e-05,
+      "loss": 0.5811,
+      "step": 978
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6673827425707125e-05,
+      "loss": 0.6176,
+      "step": 979
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.66702470461869e-05,
+      "loss": 0.592,
+      "step": 980
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.666666666666667e-05,
+      "loss": 0.5797,
+      "step": 981
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.666308628714644e-05,
+      "loss": 0.5918,
+      "step": 982
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.665950590762621e-05,
+      "loss": 0.5987,
+      "step": 983
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6655925528105984e-05,
+      "loss": 0.5861,
+      "step": 984
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.665234514858575e-05,
+      "loss": 0.6632,
+      "step": 985
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6648764769065524e-05,
+      "loss": 0.6804,
+      "step": 986
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.664518438954529e-05,
+      "loss": 0.6425,
+      "step": 987
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6641604010025064e-05,
+      "loss": 0.6847,
+      "step": 988
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.663802363050484e-05,
+      "loss": 0.6121,
+      "step": 989
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.66344432509846e-05,
+      "loss": 0.6,
+      "step": 990
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6630862871464376e-05,
+      "loss": 0.6774,
+      "step": 991
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.662728249194415e-05,
+      "loss": 0.6269,
+      "step": 992
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6623702112423916e-05,
+      "loss": 0.5897,
+      "step": 993
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.662012173290369e-05,
+      "loss": 0.6187,
+      "step": 994
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.6616541353383456e-05,
+      "loss": 0.5873,
+      "step": 995
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6612960973863236e-05,
+      "loss": 0.638,
+      "step": 996
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6609380594343e-05,
+      "loss": 0.5715,
+      "step": 997
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6605800214822775e-05,
+      "loss": 0.6211,
+      "step": 998
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.660221983530255e-05,
+      "loss": 0.577,
+      "step": 999
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6598639455782315e-05,
+      "loss": 0.5802,
+      "step": 1000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.659505907626209e-05,
+      "loss": 0.596,
+      "step": 1001
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.659147869674186e-05,
+      "loss": 0.5693,
+      "step": 1002
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.658789831722163e-05,
+      "loss": 0.5805,
+      "step": 1003
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.65843179377014e-05,
+      "loss": 0.5854,
+      "step": 1004
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.658073755818117e-05,
+      "loss": 0.6109,
+      "step": 1005
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.657715717866094e-05,
+      "loss": 0.5919,
+      "step": 1006
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6573576799140714e-05,
+      "loss": 0.6611,
+      "step": 1007
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.656999641962048e-05,
+      "loss": 0.6238,
+      "step": 1008
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.656641604010025e-05,
+      "loss": 0.691,
+      "step": 1009
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6562835660580026e-05,
+      "loss": 0.6763,
+      "step": 1010
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.655925528105979e-05,
+      "loss": 0.6107,
+      "step": 1011
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6555674901539566e-05,
+      "loss": 0.6566,
+      "step": 1012
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.655209452201933e-05,
+      "loss": 0.5724,
+      "step": 1013
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6548514142499105e-05,
+      "loss": 0.6358,
+      "step": 1014
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.654493376297888e-05,
+      "loss": 0.6135,
+      "step": 1015
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6541353383458645e-05,
+      "loss": 0.6206,
+      "step": 1016
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.653777300393842e-05,
+      "loss": 0.6334,
+      "step": 1017
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6534192624418185e-05,
+      "loss": 0.6253,
+      "step": 1018
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.653061224489796e-05,
+      "loss": 0.6363,
+      "step": 1019
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.652703186537774e-05,
+      "loss": 0.655,
+      "step": 1020
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6523451485857504e-05,
+      "loss": 0.5597,
+      "step": 1021
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.651987110633728e-05,
+      "loss": 0.613,
+      "step": 1022
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.6516290726817044e-05,
+      "loss": 0.6009,
+      "step": 1023
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.651271034729682e-05,
+      "loss": 0.6324,
+      "step": 1024
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.650912996777659e-05,
+      "loss": 0.5835,
+      "step": 1025
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.650554958825636e-05,
+      "loss": 0.5916,
+      "step": 1026
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.650196920873613e-05,
+      "loss": 0.6254,
+      "step": 1027
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.64983888292159e-05,
+      "loss": 0.6446,
+      "step": 1028
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.649480844969567e-05,
+      "loss": 0.639,
+      "step": 1029
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.649122807017544e-05,
+      "loss": 0.6257,
+      "step": 1030
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.648764769065521e-05,
+      "loss": 0.7095,
+      "step": 1031
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.648406731113498e-05,
+      "loss": 0.6922,
+      "step": 1032
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6480486931614755e-05,
+      "loss": 0.6045,
+      "step": 1033
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.647690655209452e-05,
+      "loss": 0.6134,
+      "step": 1034
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6473326172574295e-05,
+      "loss": 0.6234,
+      "step": 1035
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.646974579305406e-05,
+      "loss": 0.6726,
+      "step": 1036
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6466165413533835e-05,
+      "loss": 0.6133,
+      "step": 1037
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.646258503401361e-05,
+      "loss": 0.6138,
+      "step": 1038
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6459004654493374e-05,
+      "loss": 0.6091,
+      "step": 1039
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.645542427497315e-05,
+      "loss": 0.536,
+      "step": 1040
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.645184389545292e-05,
+      "loss": 0.5595,
+      "step": 1041
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.644826351593269e-05,
+      "loss": 0.633,
+      "step": 1042
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.644468313641247e-05,
+      "loss": 0.5593,
+      "step": 1043
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.644110275689223e-05,
+      "loss": 0.5435,
+      "step": 1044
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6437522377372007e-05,
+      "loss": 0.6041,
+      "step": 1045
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.643394199785178e-05,
+      "loss": 0.6958,
+      "step": 1046
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6430361618331546e-05,
+      "loss": 0.5786,
+      "step": 1047
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.642678123881132e-05,
+      "loss": 0.6889,
+      "step": 1048
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.6423200859291086e-05,
+      "loss": 0.5802,
+      "step": 1049
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.641962047977086e-05,
+      "loss": 0.5233,
+      "step": 1050
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.641604010025063e-05,
+      "loss": 0.6665,
+      "step": 1051
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.64124597207304e-05,
+      "loss": 0.6309,
+      "step": 1052
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.640887934121017e-05,
+      "loss": 0.697,
+      "step": 1053
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.640529896168994e-05,
+      "loss": 0.5385,
+      "step": 1054
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.640171858216971e-05,
+      "loss": 0.6051,
+      "step": 1055
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6398138202649484e-05,
+      "loss": 0.5907,
+      "step": 1056
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.639455782312925e-05,
+      "loss": 0.5226,
+      "step": 1057
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6390977443609024e-05,
+      "loss": 0.6127,
+      "step": 1058
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.63873970640888e-05,
+      "loss": 0.7291,
+      "step": 1059
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6383816684568564e-05,
+      "loss": 0.4886,
+      "step": 1060
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.638023630504834e-05,
+      "loss": 0.6304,
+      "step": 1061
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.63766559255281e-05,
+      "loss": 0.5866,
+      "step": 1062
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6373075546007876e-05,
+      "loss": 0.6125,
+      "step": 1063
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.636949516648765e-05,
+      "loss": 0.6053,
+      "step": 1064
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6365914786967416e-05,
+      "loss": 0.6521,
+      "step": 1065
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6362334407447196e-05,
+      "loss": 0.6062,
+      "step": 1066
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.635875402792696e-05,
+      "loss": 0.5406,
+      "step": 1067
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6355173648406736e-05,
+      "loss": 0.5386,
+      "step": 1068
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.635159326888651e-05,
+      "loss": 0.6663,
+      "step": 1069
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6348012889366275e-05,
+      "loss": 0.5704,
+      "step": 1070
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.634443250984605e-05,
+      "loss": 0.612,
+      "step": 1071
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6340852130325815e-05,
+      "loss": 0.6014,
+      "step": 1072
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.633727175080559e-05,
+      "loss": 0.5785,
+      "step": 1073
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.633369137128536e-05,
+      "loss": 0.6531,
+      "step": 1074
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.633011099176513e-05,
+      "loss": 0.5799,
+      "step": 1075
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.63265306122449e-05,
+      "loss": 0.6021,
+      "step": 1076
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6322950232724674e-05,
+      "loss": 0.6322,
+      "step": 1077
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.631936985320444e-05,
+      "loss": 0.5358,
+      "step": 1078
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.6315789473684214e-05,
+      "loss": 0.5858,
+      "step": 1079
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.631220909416398e-05,
+      "loss": 0.6852,
+      "step": 1080
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.630862871464375e-05,
+      "loss": 0.5826,
+      "step": 1081
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6305048335123526e-05,
+      "loss": 0.6315,
+      "step": 1082
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.630146795560329e-05,
+      "loss": 0.5886,
+      "step": 1083
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6297887576083066e-05,
+      "loss": 0.6191,
+      "step": 1084
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.629430719656284e-05,
+      "loss": 0.5792,
+      "step": 1085
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6290726817042606e-05,
+      "loss": 0.6413,
+      "step": 1086
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.628714643752238e-05,
+      "loss": 0.5461,
+      "step": 1087
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6283566058002145e-05,
+      "loss": 0.6258,
+      "step": 1088
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.627998567848192e-05,
+      "loss": 0.6321,
+      "step": 1089
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.627640529896169e-05,
+      "loss": 0.5986,
+      "step": 1090
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6272824919441465e-05,
+      "loss": 0.6726,
+      "step": 1091
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.626924453992124e-05,
+      "loss": 0.6255,
+      "step": 1092
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6265664160401004e-05,
+      "loss": 0.6877,
+      "step": 1093
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.626208378088078e-05,
+      "loss": 0.6733,
+      "step": 1094
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.625850340136055e-05,
+      "loss": 0.5915,
+      "step": 1095
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.625492302184032e-05,
+      "loss": 0.5628,
+      "step": 1096
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.625134264232009e-05,
+      "loss": 0.6341,
+      "step": 1097
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.624776226279986e-05,
+      "loss": 0.6154,
+      "step": 1098
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.624418188327963e-05,
+      "loss": 0.6416,
+      "step": 1099
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.62406015037594e-05,
+      "loss": 0.5294,
+      "step": 1100
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.623702112423917e-05,
+      "loss": 0.6237,
+      "step": 1101
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.623344074471894e-05,
+      "loss": 0.6366,
+      "step": 1102
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6229860365198716e-05,
+      "loss": 0.5905,
+      "step": 1103
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.622627998567848e-05,
+      "loss": 0.6122,
+      "step": 1104
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6222699606158255e-05,
+      "loss": 0.561,
+      "step": 1105
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.621911922663802e-05,
+      "loss": 0.5946,
+      "step": 1106
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.6215538847117795e-05,
+      "loss": 0.6495,
+      "step": 1107
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.621195846759757e-05,
+      "loss": 0.5825,
+      "step": 1108
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6208378088077335e-05,
+      "loss": 0.6402,
+      "step": 1109
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.620479770855711e-05,
+      "loss": 0.5731,
+      "step": 1110
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.620121732903688e-05,
+      "loss": 0.6238,
+      "step": 1111
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.619763694951665e-05,
+      "loss": 0.5702,
+      "step": 1112
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.619405656999643e-05,
+      "loss": 0.6366,
+      "step": 1113
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6190476190476194e-05,
+      "loss": 0.6204,
+      "step": 1114
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.618689581095597e-05,
+      "loss": 0.5949,
+      "step": 1115
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6183315431435733e-05,
+      "loss": 0.6061,
+      "step": 1116
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.617973505191551e-05,
+      "loss": 0.5727,
+      "step": 1117
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.617615467239528e-05,
+      "loss": 0.5692,
+      "step": 1118
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6172574292875046e-05,
+      "loss": 0.6131,
+      "step": 1119
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.616899391335482e-05,
+      "loss": 0.5821,
+      "step": 1120
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.616541353383459e-05,
+      "loss": 0.6537,
+      "step": 1121
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.616183315431436e-05,
+      "loss": 0.6501,
+      "step": 1122
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.615825277479413e-05,
+      "loss": 0.533,
+      "step": 1123
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.61546723952739e-05,
+      "loss": 0.5931,
+      "step": 1124
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.615109201575367e-05,
+      "loss": 0.5532,
+      "step": 1125
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6147511636233445e-05,
+      "loss": 0.5964,
+      "step": 1126
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.614393125671321e-05,
+      "loss": 0.6437,
+      "step": 1127
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6140350877192985e-05,
+      "loss": 0.6301,
+      "step": 1128
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.613677049767275e-05,
+      "loss": 0.627,
+      "step": 1129
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6133190118152524e-05,
+      "loss": 0.6636,
+      "step": 1130
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.61296097386323e-05,
+      "loss": 0.6019,
+      "step": 1131
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6126029359112064e-05,
+      "loss": 0.6136,
+      "step": 1132
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.612244897959184e-05,
+      "loss": 0.6426,
+      "step": 1133
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.611886860007161e-05,
+      "loss": 0.575,
+      "step": 1134
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.6115288220551377e-05,
+      "loss": 0.6652,
+      "step": 1135
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6111707841031157e-05,
+      "loss": 0.644,
+      "step": 1136
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.610812746151092e-05,
+      "loss": 0.578,
+      "step": 1137
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6104547081990696e-05,
+      "loss": 0.545,
+      "step": 1138
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.610096670247047e-05,
+      "loss": 0.582,
+      "step": 1139
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6097386322950236e-05,
+      "loss": 0.6065,
+      "step": 1140
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.609380594343001e-05,
+      "loss": 0.5412,
+      "step": 1141
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6090225563909775e-05,
+      "loss": 0.6027,
+      "step": 1142
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.608664518438955e-05,
+      "loss": 0.6968,
+      "step": 1143
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.608306480486932e-05,
+      "loss": 0.6591,
+      "step": 1144
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.607948442534909e-05,
+      "loss": 0.5979,
+      "step": 1145
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.607590404582886e-05,
+      "loss": 0.5719,
+      "step": 1146
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.607232366630863e-05,
+      "loss": 0.5839,
+      "step": 1147
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.60687432867884e-05,
+      "loss": 0.5571,
+      "step": 1148
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6065162907268174e-05,
+      "loss": 0.6108,
+      "step": 1149
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.606158252774794e-05,
+      "loss": 0.6502,
+      "step": 1150
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6058002148227714e-05,
+      "loss": 0.5885,
+      "step": 1151
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.605442176870749e-05,
+      "loss": 0.5603,
+      "step": 1152
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.605084138918725e-05,
+      "loss": 0.567,
+      "step": 1153
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6047261009667026e-05,
+      "loss": 0.6602,
+      "step": 1154
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.604368063014679e-05,
+      "loss": 0.6226,
+      "step": 1155
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6040100250626566e-05,
+      "loss": 0.588,
+      "step": 1156
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.603651987110634e-05,
+      "loss": 0.631,
+      "step": 1157
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6032939491586106e-05,
+      "loss": 0.6639,
+      "step": 1158
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.602935911206588e-05,
+      "loss": 0.617,
+      "step": 1159
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.602577873254565e-05,
+      "loss": 0.566,
+      "step": 1160
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6022198353025425e-05,
+      "loss": 0.6075,
+      "step": 1161
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.60186179735052e-05,
+      "loss": 0.5346,
+      "step": 1162
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.6015037593984965e-05,
+      "loss": 0.756,
+      "step": 1163
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.601145721446474e-05,
+      "loss": 0.5488,
+      "step": 1164
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.6007876834944504e-05,
+      "loss": 0.6731,
+      "step": 1165
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.600429645542428e-05,
+      "loss": 0.5547,
+      "step": 1166
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.600071607590405e-05,
+      "loss": 0.5873,
+      "step": 1167
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.599713569638382e-05,
+      "loss": 0.544,
+      "step": 1168
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.599355531686359e-05,
+      "loss": 0.6154,
+      "step": 1169
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5989974937343364e-05,
+      "loss": 0.5926,
+      "step": 1170
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.598639455782313e-05,
+      "loss": 0.5547,
+      "step": 1171
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.59828141783029e-05,
+      "loss": 0.5935,
+      "step": 1172
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.597923379878267e-05,
+      "loss": 0.7053,
+      "step": 1173
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.597565341926244e-05,
+      "loss": 0.5968,
+      "step": 1174
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5972073039742216e-05,
+      "loss": 0.5802,
+      "step": 1175
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.596849266022198e-05,
+      "loss": 0.6311,
+      "step": 1176
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5964912280701756e-05,
+      "loss": 0.6786,
+      "step": 1177
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.596133190118153e-05,
+      "loss": 0.623,
+      "step": 1178
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5957751521661295e-05,
+      "loss": 0.5511,
+      "step": 1179
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.595417114214107e-05,
+      "loss": 0.5798,
+      "step": 1180
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5950590762620835e-05,
+      "loss": 0.6675,
+      "step": 1181
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.594701038310061e-05,
+      "loss": 0.7011,
+      "step": 1182
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.594343000358038e-05,
+      "loss": 0.6311,
+      "step": 1183
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5939849624060154e-05,
+      "loss": 0.6911,
+      "step": 1184
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.593626924453993e-05,
+      "loss": 0.6358,
+      "step": 1185
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5932688865019694e-05,
+      "loss": 0.6025,
+      "step": 1186
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.592910848549947e-05,
+      "loss": 0.5972,
+      "step": 1187
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.592552810597924e-05,
+      "loss": 0.608,
+      "step": 1188
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.592194772645901e-05,
+      "loss": 0.575,
+      "step": 1189
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.591836734693878e-05,
+      "loss": 0.7249,
+      "step": 1190
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.5914786967418546e-05,
+      "loss": 0.5996,
+      "step": 1191
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.591120658789832e-05,
+      "loss": 0.5665,
+      "step": 1192
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.590762620837809e-05,
+      "loss": 0.583,
+      "step": 1193
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.590404582885786e-05,
+      "loss": 0.6985,
+      "step": 1194
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.590046544933763e-05,
+      "loss": 0.6615,
+      "step": 1195
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5896885069817405e-05,
+      "loss": 0.6313,
+      "step": 1196
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.589330469029717e-05,
+      "loss": 0.6556,
+      "step": 1197
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5889724310776945e-05,
+      "loss": 0.5848,
+      "step": 1198
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.588614393125671e-05,
+      "loss": 0.5918,
+      "step": 1199
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5882563551736485e-05,
+      "loss": 0.6167,
+      "step": 1200
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.587898317221626e-05,
+      "loss": 0.6604,
+      "step": 1201
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5875402792696024e-05,
+      "loss": 0.6393,
+      "step": 1202
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.58718224131758e-05,
+      "loss": 0.673,
+      "step": 1203
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.586824203365557e-05,
+      "loss": 0.5803,
+      "step": 1204
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.586466165413534e-05,
+      "loss": 0.6283,
+      "step": 1205
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.586108127461512e-05,
+      "loss": 0.5731,
+      "step": 1206
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5857500895094883e-05,
+      "loss": 0.5175,
+      "step": 1207
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.585392051557466e-05,
+      "loss": 0.5852,
+      "step": 1208
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.585034013605442e-05,
+      "loss": 0.6266,
+      "step": 1209
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5846759756534196e-05,
+      "loss": 0.6389,
+      "step": 1210
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.584317937701397e-05,
+      "loss": 0.5959,
+      "step": 1211
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5839598997493736e-05,
+      "loss": 0.6002,
+      "step": 1212
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.583601861797351e-05,
+      "loss": 0.6004,
+      "step": 1213
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.583243823845328e-05,
+      "loss": 0.5752,
+      "step": 1214
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.582885785893305e-05,
+      "loss": 0.6275,
+      "step": 1215
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.582527747941282e-05,
+      "loss": 0.5978,
+      "step": 1216
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.582169709989259e-05,
+      "loss": 0.6065,
+      "step": 1217
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.581811672037236e-05,
+      "loss": 0.5409,
+      "step": 1218
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.5814536340852135e-05,
+      "loss": 0.6012,
+      "step": 1219
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.58109559613319e-05,
+      "loss": 0.7826,
+      "step": 1220
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5807375581811674e-05,
+      "loss": 0.6438,
+      "step": 1221
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.580379520229145e-05,
+      "loss": 0.6654,
+      "step": 1222
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5800214822771214e-05,
+      "loss": 0.5663,
+      "step": 1223
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.579663444325099e-05,
+      "loss": 0.6455,
+      "step": 1224
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5793054063730753e-05,
+      "loss": 0.6404,
+      "step": 1225
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5789473684210527e-05,
+      "loss": 0.53,
+      "step": 1226
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.57858933046903e-05,
+      "loss": 0.5531,
+      "step": 1227
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5782312925170066e-05,
+      "loss": 0.5585,
+      "step": 1228
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.577873254564984e-05,
+      "loss": 0.6588,
+      "step": 1229
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.577515216612961e-05,
+      "loss": 0.6389,
+      "step": 1230
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5771571786609386e-05,
+      "loss": 0.5763,
+      "step": 1231
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.576799140708916e-05,
+      "loss": 0.6936,
+      "step": 1232
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5764411027568925e-05,
+      "loss": 0.4588,
+      "step": 1233
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.57608306480487e-05,
+      "loss": 0.6626,
+      "step": 1234
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5757250268528465e-05,
+      "loss": 0.5558,
+      "step": 1235
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.575366988900824e-05,
+      "loss": 0.5934,
+      "step": 1236
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.575008950948801e-05,
+      "loss": 0.6543,
+      "step": 1237
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.574650912996778e-05,
+      "loss": 0.5989,
+      "step": 1238
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.574292875044755e-05,
+      "loss": 0.5251,
+      "step": 1239
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5739348370927324e-05,
+      "loss": 0.7058,
+      "step": 1240
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.573576799140709e-05,
+      "loss": 0.643,
+      "step": 1241
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5732187611886864e-05,
+      "loss": 0.5821,
+      "step": 1242
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.572860723236663e-05,
+      "loss": 0.5795,
+      "step": 1243
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.57250268528464e-05,
+      "loss": 0.6112,
+      "step": 1244
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5721446473326176e-05,
+      "loss": 0.605,
+      "step": 1245
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.571786609380594e-05,
+      "loss": 0.621,
+      "step": 1246
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.5714285714285716e-05,
+      "loss": 0.4887,
+      "step": 1247
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.571070533476548e-05,
+      "loss": 0.698,
+      "step": 1248
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5707124955245256e-05,
+      "loss": 0.5479,
+      "step": 1249
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.570354457572503e-05,
+      "loss": 0.6316,
+      "step": 1250
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5699964196204795e-05,
+      "loss": 0.6238,
+      "step": 1251
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.569638381668457e-05,
+      "loss": 0.5282,
+      "step": 1252
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.569280343716434e-05,
+      "loss": 0.6507,
+      "step": 1253
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5689223057644115e-05,
+      "loss": 0.6992,
+      "step": 1254
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.568564267812389e-05,
+      "loss": 0.655,
+      "step": 1255
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5682062298603654e-05,
+      "loss": 0.5074,
+      "step": 1256
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.567848191908343e-05,
+      "loss": 0.5345,
+      "step": 1257
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5674901539563194e-05,
+      "loss": 0.6028,
+      "step": 1258
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.567132116004297e-05,
+      "loss": 0.6915,
+      "step": 1259
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.566774078052274e-05,
+      "loss": 0.6796,
+      "step": 1260
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.566416040100251e-05,
+      "loss": 0.5023,
+      "step": 1261
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.566058002148228e-05,
+      "loss": 0.6313,
+      "step": 1262
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.565699964196205e-05,
+      "loss": 0.6047,
+      "step": 1263
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.565341926244182e-05,
+      "loss": 0.6533,
+      "step": 1264
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.564983888292159e-05,
+      "loss": 0.6586,
+      "step": 1265
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.564625850340136e-05,
+      "loss": 0.5987,
+      "step": 1266
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.564267812388113e-05,
+      "loss": 0.5772,
+      "step": 1267
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5639097744360906e-05,
+      "loss": 0.651,
+      "step": 1268
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.563551736484067e-05,
+      "loss": 0.6039,
+      "step": 1269
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5631936985320445e-05,
+      "loss": 0.6249,
+      "step": 1270
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.562835660580022e-05,
+      "loss": 0.5785,
+      "step": 1271
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5624776226279985e-05,
+      "loss": 0.5673,
+      "step": 1272
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.562119584675976e-05,
+      "loss": 0.6881,
+      "step": 1273
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.5617615467239524e-05,
+      "loss": 0.6203,
+      "step": 1274
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.56140350877193e-05,
+      "loss": 0.5912,
+      "step": 1275
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.561045470819907e-05,
+      "loss": 0.6112,
+      "step": 1276
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5606874328678844e-05,
+      "loss": 0.6251,
+      "step": 1277
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.560329394915862e-05,
+      "loss": 0.7052,
+      "step": 1278
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5599713569638384e-05,
+      "loss": 0.6772,
+      "step": 1279
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.559613319011816e-05,
+      "loss": 0.6207,
+      "step": 1280
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.559255281059793e-05,
+      "loss": 0.5513,
+      "step": 1281
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5588972431077696e-05,
+      "loss": 0.5603,
+      "step": 1282
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.558539205155747e-05,
+      "loss": 0.623,
+      "step": 1283
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5581811672037236e-05,
+      "loss": 0.5386,
+      "step": 1284
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.557823129251701e-05,
+      "loss": 0.6579,
+      "step": 1285
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.557465091299678e-05,
+      "loss": 0.648,
+      "step": 1286
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.557107053347655e-05,
+      "loss": 0.604,
+      "step": 1287
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.556749015395632e-05,
+      "loss": 0.6753,
+      "step": 1288
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5563909774436095e-05,
+      "loss": 0.5936,
+      "step": 1289
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.556032939491586e-05,
+      "loss": 0.6524,
+      "step": 1290
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5556749015395635e-05,
+      "loss": 0.6395,
+      "step": 1291
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.55531686358754e-05,
+      "loss": 0.5559,
+      "step": 1292
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5549588256355174e-05,
+      "loss": 0.6252,
+      "step": 1293
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.554600787683495e-05,
+      "loss": 0.6472,
+      "step": 1294
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5542427497314714e-05,
+      "loss": 0.5291,
+      "step": 1295
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.553884711779449e-05,
+      "loss": 0.5875,
+      "step": 1296
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.553526673827426e-05,
+      "loss": 0.6038,
+      "step": 1297
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.553168635875403e-05,
+      "loss": 0.5771,
+      "step": 1298
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.55281059792338e-05,
+      "loss": 0.6105,
+      "step": 1299
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.552452559971357e-05,
+      "loss": 0.6352,
+      "step": 1300
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5520945220193346e-05,
+      "loss": 0.5363,
+      "step": 1301
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.551736484067311e-05,
+      "loss": 0.6191,
+      "step": 1302
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.5513784461152886e-05,
+      "loss": 0.4904,
+      "step": 1303
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.551020408163266e-05,
+      "loss": 0.5796,
+      "step": 1304
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5506623702112425e-05,
+      "loss": 0.6163,
+      "step": 1305
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.55030433225922e-05,
+      "loss": 0.6488,
+      "step": 1306
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.549946294307197e-05,
+      "loss": 0.7144,
+      "step": 1307
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.549588256355174e-05,
+      "loss": 0.6226,
+      "step": 1308
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.549230218403151e-05,
+      "loss": 0.5813,
+      "step": 1309
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.548872180451128e-05,
+      "loss": 0.5756,
+      "step": 1310
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.548514142499105e-05,
+      "loss": 0.581,
+      "step": 1311
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5481561045470824e-05,
+      "loss": 0.5377,
+      "step": 1312
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.547798066595059e-05,
+      "loss": 0.6405,
+      "step": 1313
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5474400286430364e-05,
+      "loss": 0.5658,
+      "step": 1314
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.547081990691014e-05,
+      "loss": 0.6155,
+      "step": 1315
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5467239527389903e-05,
+      "loss": 0.6922,
+      "step": 1316
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5463659147869677e-05,
+      "loss": 0.5488,
+      "step": 1317
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.546007876834944e-05,
+      "loss": 0.5357,
+      "step": 1318
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5456498388829216e-05,
+      "loss": 0.561,
+      "step": 1319
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.545291800930899e-05,
+      "loss": 0.6186,
+      "step": 1320
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5449337629788756e-05,
+      "loss": 0.5704,
+      "step": 1321
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.544575725026853e-05,
+      "loss": 0.5193,
+      "step": 1322
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.54421768707483e-05,
+      "loss": 0.622,
+      "step": 1323
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5438596491228075e-05,
+      "loss": 0.5216,
+      "step": 1324
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.543501611170785e-05,
+      "loss": 0.5669,
+      "step": 1325
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5431435732187615e-05,
+      "loss": 0.6909,
+      "step": 1326
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.542785535266739e-05,
+      "loss": 0.6022,
+      "step": 1327
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.5424274973147155e-05,
+      "loss": 0.5354,
+      "step": 1328
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.542069459362693e-05,
+      "loss": 0.588,
+      "step": 1329
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.54171142141067e-05,
+      "loss": 0.6173,
+      "step": 1330
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.541353383458647e-05,
+      "loss": 0.6279,
+      "step": 1331
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.540995345506624e-05,
+      "loss": 0.667,
+      "step": 1332
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5406373075546014e-05,
+      "loss": 0.469,
+      "step": 1333
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.540279269602578e-05,
+      "loss": 0.5679,
+      "step": 1334
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.539921231650555e-05,
+      "loss": 0.5864,
+      "step": 1335
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.539563193698532e-05,
+      "loss": 0.5953,
+      "step": 1336
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.539205155746509e-05,
+      "loss": 0.6446,
+      "step": 1337
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5388471177944866e-05,
+      "loss": 0.5452,
+      "step": 1338
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.538489079842463e-05,
+      "loss": 0.4833,
+      "step": 1339
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5381310418904406e-05,
+      "loss": 0.7119,
+      "step": 1340
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.537773003938417e-05,
+      "loss": 0.6352,
+      "step": 1341
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5374149659863945e-05,
+      "loss": 0.6251,
+      "step": 1342
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.537056928034372e-05,
+      "loss": 0.4452,
+      "step": 1343
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5366988900823485e-05,
+      "loss": 0.5991,
+      "step": 1344
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.536340852130326e-05,
+      "loss": 0.6204,
+      "step": 1345
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.535982814178303e-05,
+      "loss": 0.5764,
+      "step": 1346
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5356247762262804e-05,
+      "loss": 0.685,
+      "step": 1347
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.535266738274258e-05,
+      "loss": 0.605,
+      "step": 1348
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5349087003222344e-05,
+      "loss": 0.5017,
+      "step": 1349
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.534550662370212e-05,
+      "loss": 0.6132,
+      "step": 1350
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.534192624418189e-05,
+      "loss": 0.6392,
+      "step": 1351
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.533834586466166e-05,
+      "loss": 0.6204,
+      "step": 1352
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.533476548514143e-05,
+      "loss": 0.6816,
+      "step": 1353
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.5331185105621196e-05,
+      "loss": 0.5694,
+      "step": 1354
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.532760472610097e-05,
+      "loss": 0.6299,
+      "step": 1355
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.532402434658074e-05,
+      "loss": 0.6828,
+      "step": 1356
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.532044396706051e-05,
+      "loss": 0.6593,
+      "step": 1357
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.531686358754028e-05,
+      "loss": 0.6334,
+      "step": 1358
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.531328320802005e-05,
+      "loss": 0.6353,
+      "step": 1359
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.530970282849982e-05,
+      "loss": 0.5285,
+      "step": 1360
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5306122448979595e-05,
+      "loss": 0.6161,
+      "step": 1361
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.530254206945936e-05,
+      "loss": 0.5342,
+      "step": 1362
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5298961689939135e-05,
+      "loss": 0.6492,
+      "step": 1363
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.529538131041891e-05,
+      "loss": 0.5478,
+      "step": 1364
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5291800930898674e-05,
+      "loss": 0.6477,
+      "step": 1365
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.528822055137845e-05,
+      "loss": 0.5803,
+      "step": 1366
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5284640171858214e-05,
+      "loss": 0.5768,
+      "step": 1367
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.528105979233799e-05,
+      "loss": 0.6276,
+      "step": 1368
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.527747941281776e-05,
+      "loss": 0.5352,
+      "step": 1369
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5273899033297534e-05,
+      "loss": 0.6251,
+      "step": 1370
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.527031865377731e-05,
+      "loss": 0.6439,
+      "step": 1371
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.526673827425707e-05,
+      "loss": 0.592,
+      "step": 1372
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5263157894736846e-05,
+      "loss": 0.579,
+      "step": 1373
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.525957751521662e-05,
+      "loss": 0.6963,
+      "step": 1374
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5255997135696386e-05,
+      "loss": 0.6745,
+      "step": 1375
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.525241675617616e-05,
+      "loss": 0.7231,
+      "step": 1376
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5248836376655926e-05,
+      "loss": 0.5597,
+      "step": 1377
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.52452559971357e-05,
+      "loss": 0.6008,
+      "step": 1378
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.524167561761547e-05,
+      "loss": 0.6338,
+      "step": 1379
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.523809523809524e-05,
+      "loss": 0.5436,
+      "step": 1380
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.523451485857501e-05,
+      "loss": 0.5605,
+      "step": 1381
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5230934479054785e-05,
+      "loss": 0.594,
+      "step": 1382
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.522735409953455e-05,
+      "loss": 0.5767,
+      "step": 1383
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5223773720014324e-05,
+      "loss": 0.5586,
+      "step": 1384
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.522019334049409e-05,
+      "loss": 0.5544,
+      "step": 1385
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.5216612960973864e-05,
+      "loss": 0.5581,
+      "step": 1386
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.521303258145364e-05,
+      "loss": 0.6487,
+      "step": 1387
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5209452201933403e-05,
+      "loss": 0.5819,
+      "step": 1388
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.520587182241318e-05,
+      "loss": 0.5502,
+      "step": 1389
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.520229144289295e-05,
+      "loss": 0.5956,
+      "step": 1390
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5198711063372716e-05,
+      "loss": 0.6002,
+      "step": 1391
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.519513068385249e-05,
+      "loss": 0.6072,
+      "step": 1392
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.519155030433226e-05,
+      "loss": 0.6331,
+      "step": 1393
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5187969924812036e-05,
+      "loss": 0.5829,
+      "step": 1394
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.51843895452918e-05,
+      "loss": 0.4877,
+      "step": 1395
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5180809165771575e-05,
+      "loss": 0.5958,
+      "step": 1396
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.517722878625135e-05,
+      "loss": 0.6928,
+      "step": 1397
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5173648406731115e-05,
+      "loss": 0.4927,
+      "step": 1398
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.517006802721089e-05,
+      "loss": 0.7108,
+      "step": 1399
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.516648764769066e-05,
+      "loss": 0.5632,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.516290726817043e-05,
+      "loss": 0.5975,
+      "step": 1401
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.51593268886502e-05,
+      "loss": 0.6039,
+      "step": 1402
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.515574650912997e-05,
+      "loss": 0.6618,
+      "step": 1403
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.515216612960974e-05,
+      "loss": 0.4991,
+      "step": 1404
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5148585750089514e-05,
+      "loss": 0.5985,
+      "step": 1405
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.514500537056928e-05,
+      "loss": 0.594,
+      "step": 1406
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5141424991049053e-05,
+      "loss": 0.7162,
+      "step": 1407
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5137844611528827e-05,
+      "loss": 0.652,
+      "step": 1408
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.513426423200859e-05,
+      "loss": 0.6115,
+      "step": 1409
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5130683852488366e-05,
+      "loss": 0.5318,
+      "step": 1410
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.512710347296813e-05,
+      "loss": 0.6125,
+      "step": 1411
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5123523093447906e-05,
+      "loss": 0.5787,
+      "step": 1412
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.511994271392768e-05,
+      "loss": 0.6564,
+      "step": 1413
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.5116362334407445e-05,
+      "loss": 0.6142,
+      "step": 1414
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.511278195488722e-05,
+      "loss": 0.6241,
+      "step": 1415
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5109201575366985e-05,
+      "loss": 0.6926,
+      "step": 1416
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5105621195846765e-05,
+      "loss": 0.6123,
+      "step": 1417
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.510204081632654e-05,
+      "loss": 0.5987,
+      "step": 1418
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5098460436806305e-05,
+      "loss": 0.6282,
+      "step": 1419
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.509488005728608e-05,
+      "loss": 0.56,
+      "step": 1420
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5091299677765844e-05,
+      "loss": 0.5879,
+      "step": 1421
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.508771929824562e-05,
+      "loss": 0.6116,
+      "step": 1422
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.508413891872539e-05,
+      "loss": 0.5266,
+      "step": 1423
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.508055853920516e-05,
+      "loss": 0.6385,
+      "step": 1424
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.507697815968493e-05,
+      "loss": 0.5186,
+      "step": 1425
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.50733977801647e-05,
+      "loss": 0.626,
+      "step": 1426
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.506981740064447e-05,
+      "loss": 0.6803,
+      "step": 1427
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.506623702112424e-05,
+      "loss": 0.6777,
+      "step": 1428
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.506265664160401e-05,
+      "loss": 0.5614,
+      "step": 1429
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.505907626208378e-05,
+      "loss": 0.611,
+      "step": 1430
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5055495882563556e-05,
+      "loss": 0.618,
+      "step": 1431
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.505191550304332e-05,
+      "loss": 0.5187,
+      "step": 1432
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5048335123523095e-05,
+      "loss": 0.6462,
+      "step": 1433
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.504475474400286e-05,
+      "loss": 0.5559,
+      "step": 1434
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5041174364482635e-05,
+      "loss": 0.5559,
+      "step": 1435
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.503759398496241e-05,
+      "loss": 0.5214,
+      "step": 1436
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5034013605442174e-05,
+      "loss": 0.5505,
+      "step": 1437
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.503043322592195e-05,
+      "loss": 0.6321,
+      "step": 1438
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.502685284640172e-05,
+      "loss": 0.5473,
+      "step": 1439
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5023272466881494e-05,
+      "loss": 0.6754,
+      "step": 1440
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.501969208736127e-05,
+      "loss": 0.6333,
+      "step": 1441
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.5016111707841034e-05,
+      "loss": 0.6189,
+      "step": 1442
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.501253132832081e-05,
+      "loss": 0.5874,
+      "step": 1443
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.500895094880058e-05,
+      "loss": 0.5685,
+      "step": 1444
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.5005370569280346e-05,
+      "loss": 0.6053,
+      "step": 1445
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.500179018976012e-05,
+      "loss": 0.6454,
+      "step": 1446
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4998209810239886e-05,
+      "loss": 0.5845,
+      "step": 1447
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.499462943071966e-05,
+      "loss": 0.5664,
+      "step": 1448
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.499104905119943e-05,
+      "loss": 0.5718,
+      "step": 1449
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.49874686716792e-05,
+      "loss": 0.5774,
+      "step": 1450
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.498388829215897e-05,
+      "loss": 0.6617,
+      "step": 1451
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.498030791263874e-05,
+      "loss": 0.6046,
+      "step": 1452
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.497672753311851e-05,
+      "loss": 0.5685,
+      "step": 1453
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4973147153598285e-05,
+      "loss": 0.6086,
+      "step": 1454
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.496956677407805e-05,
+      "loss": 0.5459,
+      "step": 1455
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4965986394557824e-05,
+      "loss": 0.6609,
+      "step": 1456
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.49624060150376e-05,
+      "loss": 0.6074,
+      "step": 1457
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4958825635517364e-05,
+      "loss": 0.6126,
+      "step": 1458
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.495524525599714e-05,
+      "loss": 0.5966,
+      "step": 1459
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4951664876476904e-05,
+      "loss": 0.6241,
+      "step": 1460
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.494808449695668e-05,
+      "loss": 0.5558,
+      "step": 1461
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.494450411743645e-05,
+      "loss": 0.6559,
+      "step": 1462
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.494092373791622e-05,
+      "loss": 0.6989,
+      "step": 1463
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4937343358395996e-05,
+      "loss": 0.601,
+      "step": 1464
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.493376297887576e-05,
+      "loss": 0.6591,
+      "step": 1465
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4930182599355536e-05,
+      "loss": 0.6226,
+      "step": 1466
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.492660221983531e-05,
+      "loss": 0.5921,
+      "step": 1467
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4923021840315076e-05,
+      "loss": 0.6414,
+      "step": 1468
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.491944146079485e-05,
+      "loss": 0.6509,
+      "step": 1469
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.4915861081274615e-05,
+      "loss": 0.555,
+      "step": 1470
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.491228070175439e-05,
+      "loss": 0.6069,
+      "step": 1471
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.490870032223416e-05,
+      "loss": 0.6315,
+      "step": 1472
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.490511994271393e-05,
+      "loss": 0.573,
+      "step": 1473
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.49015395631937e-05,
+      "loss": 0.576,
+      "step": 1474
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4897959183673474e-05,
+      "loss": 0.5823,
+      "step": 1475
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.489437880415324e-05,
+      "loss": 0.6114,
+      "step": 1476
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4890798424633014e-05,
+      "loss": 0.5705,
+      "step": 1477
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.488721804511278e-05,
+      "loss": 0.6238,
+      "step": 1478
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4883637665592554e-05,
+      "loss": 0.5922,
+      "step": 1479
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.488005728607233e-05,
+      "loss": 0.5138,
+      "step": 1480
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.487647690655209e-05,
+      "loss": 0.5931,
+      "step": 1481
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4872896527031866e-05,
+      "loss": 0.6637,
+      "step": 1482
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.486931614751164e-05,
+      "loss": 0.5682,
+      "step": 1483
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4865735767991406e-05,
+      "loss": 0.5952,
+      "step": 1484
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.486215538847118e-05,
+      "loss": 0.6338,
+      "step": 1485
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4858575008950945e-05,
+      "loss": 0.5874,
+      "step": 1486
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4854994629430725e-05,
+      "loss": 0.5771,
+      "step": 1487
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.485141424991049e-05,
+      "loss": 0.5712,
+      "step": 1488
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4847833870390265e-05,
+      "loss": 0.5979,
+      "step": 1489
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.484425349087004e-05,
+      "loss": 0.5108,
+      "step": 1490
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4840673111349805e-05,
+      "loss": 0.5922,
+      "step": 1491
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.483709273182958e-05,
+      "loss": 0.5498,
+      "step": 1492
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.483351235230935e-05,
+      "loss": 0.5532,
+      "step": 1493
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.482993197278912e-05,
+      "loss": 0.6701,
+      "step": 1494
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.482635159326889e-05,
+      "loss": 0.6121,
+      "step": 1495
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.482277121374866e-05,
+      "loss": 0.5311,
+      "step": 1496
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.481919083422843e-05,
+      "loss": 0.417,
+      "step": 1497
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.4815610454708203e-05,
+      "loss": 0.5659,
+      "step": 1498
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.481203007518797e-05,
+      "loss": 0.6641,
+      "step": 1499
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.480844969566774e-05,
+      "loss": 0.6595,
+      "step": 1500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4804869316147516e-05,
+      "loss": 0.6554,
+      "step": 1501
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.480128893662728e-05,
+      "loss": 0.6241,
+      "step": 1502
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4797708557107056e-05,
+      "loss": 0.678,
+      "step": 1503
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.479412817758682e-05,
+      "loss": 0.5559,
+      "step": 1504
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4790547798066595e-05,
+      "loss": 0.6223,
+      "step": 1505
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.478696741854637e-05,
+      "loss": 0.4652,
+      "step": 1506
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4783387039026135e-05,
+      "loss": 0.4802,
+      "step": 1507
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.477980665950591e-05,
+      "loss": 0.6548,
+      "step": 1508
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.477622627998568e-05,
+      "loss": 0.5873,
+      "step": 1509
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4772645900465455e-05,
+      "loss": 0.5789,
+      "step": 1510
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.476906552094523e-05,
+      "loss": 0.6162,
+      "step": 1511
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4765485141424994e-05,
+      "loss": 0.6971,
+      "step": 1512
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.476190476190477e-05,
+      "loss": 0.6091,
+      "step": 1513
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4758324382384534e-05,
+      "loss": 0.6469,
+      "step": 1514
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.475474400286431e-05,
+      "loss": 0.5736,
+      "step": 1515
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.475116362334408e-05,
+      "loss": 0.5954,
+      "step": 1516
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4747583243823847e-05,
+      "loss": 0.5116,
+      "step": 1517
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.474400286430362e-05,
+      "loss": 0.5675,
+      "step": 1518
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.474042248478339e-05,
+      "loss": 0.5664,
+      "step": 1519
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.473684210526316e-05,
+      "loss": 0.6149,
+      "step": 1520
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.473326172574293e-05,
+      "loss": 0.6712,
+      "step": 1521
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.47296813462227e-05,
+      "loss": 0.5912,
+      "step": 1522
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.472610096670247e-05,
+      "loss": 0.5311,
+      "step": 1523
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4722520587182245e-05,
+      "loss": 0.5679,
+      "step": 1524
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.471894020766201e-05,
+      "loss": 0.5523,
+      "step": 1525
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.4715359828141785e-05,
+      "loss": 0.5278,
+      "step": 1526
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.471177944862155e-05,
+      "loss": 0.5091,
+      "step": 1527
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4708199069101324e-05,
+      "loss": 0.5762,
+      "step": 1528
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.47046186895811e-05,
+      "loss": 0.5392,
+      "step": 1529
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4701038310060864e-05,
+      "loss": 0.6172,
+      "step": 1530
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.469745793054064e-05,
+      "loss": 0.6087,
+      "step": 1531
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.469387755102041e-05,
+      "loss": 0.5225,
+      "step": 1532
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4690297171500184e-05,
+      "loss": 0.5681,
+      "step": 1533
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.468671679197996e-05,
+      "loss": 0.5435,
+      "step": 1534
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.468313641245972e-05,
+      "loss": 0.5714,
+      "step": 1535
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4679556032939496e-05,
+      "loss": 0.6349,
+      "step": 1536
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.467597565341927e-05,
+      "loss": 0.5765,
+      "step": 1537
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4672395273899036e-05,
+      "loss": 0.5316,
+      "step": 1538
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.466881489437881e-05,
+      "loss": 0.5947,
+      "step": 1539
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4665234514858576e-05,
+      "loss": 0.5568,
+      "step": 1540
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.466165413533835e-05,
+      "loss": 0.58,
+      "step": 1541
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.465807375581812e-05,
+      "loss": 0.6005,
+      "step": 1542
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.465449337629789e-05,
+      "loss": 0.4952,
+      "step": 1543
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.465091299677766e-05,
+      "loss": 0.6461,
+      "step": 1544
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.464733261725743e-05,
+      "loss": 0.5788,
+      "step": 1545
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.46437522377372e-05,
+      "loss": 0.5522,
+      "step": 1546
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4640171858216974e-05,
+      "loss": 0.5527,
+      "step": 1547
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.463659147869674e-05,
+      "loss": 0.6151,
+      "step": 1548
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4633011099176514e-05,
+      "loss": 0.6223,
+      "step": 1549
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.462943071965629e-05,
+      "loss": 0.5724,
+      "step": 1550
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4625850340136054e-05,
+      "loss": 0.6143,
+      "step": 1551
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.462226996061583e-05,
+      "loss": 0.6349,
+      "step": 1552
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.461868958109559e-05,
+      "loss": 0.5159,
+      "step": 1553
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.4615109201575366e-05,
+      "loss": 0.6576,
+      "step": 1554
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.461152882205514e-05,
+      "loss": 0.596,
+      "step": 1555
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4607948442534906e-05,
+      "loss": 0.6156,
+      "step": 1556
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4604368063014686e-05,
+      "loss": 0.5106,
+      "step": 1557
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.460078768349445e-05,
+      "loss": 0.5206,
+      "step": 1558
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4597207303974226e-05,
+      "loss": 0.6039,
+      "step": 1559
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4593626924454e-05,
+      "loss": 0.4772,
+      "step": 1560
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4590046544933765e-05,
+      "loss": 0.4786,
+      "step": 1561
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.458646616541354e-05,
+      "loss": 0.548,
+      "step": 1562
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4582885785893305e-05,
+      "loss": 0.5059,
+      "step": 1563
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.457930540637308e-05,
+      "loss": 0.6421,
+      "step": 1564
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.457572502685285e-05,
+      "loss": 0.6247,
+      "step": 1565
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.457214464733262e-05,
+      "loss": 0.5947,
+      "step": 1566
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.456856426781239e-05,
+      "loss": 0.5823,
+      "step": 1567
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4564983888292164e-05,
+      "loss": 0.6371,
+      "step": 1568
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.456140350877193e-05,
+      "loss": 0.6449,
+      "step": 1569
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4557823129251704e-05,
+      "loss": 0.6754,
+      "step": 1570
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.455424274973147e-05,
+      "loss": 0.4881,
+      "step": 1571
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.455066237021124e-05,
+      "loss": 0.7085,
+      "step": 1572
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4547081990691016e-05,
+      "loss": 0.5704,
+      "step": 1573
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.454350161117078e-05,
+      "loss": 0.6045,
+      "step": 1574
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4539921231650556e-05,
+      "loss": 0.5463,
+      "step": 1575
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.453634085213033e-05,
+      "loss": 0.572,
+      "step": 1576
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4532760472610095e-05,
+      "loss": 0.5698,
+      "step": 1577
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.452918009308987e-05,
+      "loss": 0.5057,
+      "step": 1578
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4525599713569635e-05,
+      "loss": 0.6738,
+      "step": 1579
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4522019334049415e-05,
+      "loss": 0.6001,
+      "step": 1580
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.451843895452918e-05,
+      "loss": 0.5715,
+      "step": 1581
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.4514858575008955e-05,
+      "loss": 0.6163,
+      "step": 1582
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.451127819548873e-05,
+      "loss": 0.6159,
+      "step": 1583
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4507697815968494e-05,
+      "loss": 0.6334,
+      "step": 1584
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.450411743644827e-05,
+      "loss": 0.5583,
+      "step": 1585
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.450053705692804e-05,
+      "loss": 0.6087,
+      "step": 1586
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.449695667740781e-05,
+      "loss": 0.5355,
+      "step": 1587
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.449337629788758e-05,
+      "loss": 0.4843,
+      "step": 1588
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.448979591836735e-05,
+      "loss": 0.5777,
+      "step": 1589
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.448621553884712e-05,
+      "loss": 0.5346,
+      "step": 1590
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.448263515932689e-05,
+      "loss": 0.6187,
+      "step": 1591
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.447905477980666e-05,
+      "loss": 0.6896,
+      "step": 1592
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.447547440028643e-05,
+      "loss": 0.6801,
+      "step": 1593
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4471894020766206e-05,
+      "loss": 0.4759,
+      "step": 1594
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.446831364124597e-05,
+      "loss": 0.551,
+      "step": 1595
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4464733261725745e-05,
+      "loss": 0.5738,
+      "step": 1596
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.446115288220551e-05,
+      "loss": 0.5983,
+      "step": 1597
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4457572502685285e-05,
+      "loss": 0.5627,
+      "step": 1598
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.445399212316506e-05,
+      "loss": 0.4733,
+      "step": 1599
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4450411743644825e-05,
+      "loss": 0.5362,
+      "step": 1600
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.44468313641246e-05,
+      "loss": 0.5614,
+      "step": 1601
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.444325098460437e-05,
+      "loss": 0.5023,
+      "step": 1602
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4439670605084144e-05,
+      "loss": 0.5647,
+      "step": 1603
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.443609022556392e-05,
+      "loss": 0.5889,
+      "step": 1604
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4432509846043684e-05,
+      "loss": 0.5965,
+      "step": 1605
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.442892946652346e-05,
+      "loss": 0.5547,
+      "step": 1606
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.442534908700322e-05,
+      "loss": 0.6271,
+      "step": 1607
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4421768707482997e-05,
+      "loss": 0.5556,
+      "step": 1608
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.441818832796277e-05,
+      "loss": 0.5789,
+      "step": 1609
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.4414607948442536e-05,
+      "loss": 0.517,
+      "step": 1610
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.441102756892231e-05,
+      "loss": 0.6455,
+      "step": 1611
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.440744718940208e-05,
+      "loss": 0.589,
+      "step": 1612
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.440386680988185e-05,
+      "loss": 0.6421,
+      "step": 1613
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.440028643036162e-05,
+      "loss": 0.5981,
+      "step": 1614
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.439670605084139e-05,
+      "loss": 0.6576,
+      "step": 1615
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.439312567132116e-05,
+      "loss": 0.5703,
+      "step": 1616
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4389545291800935e-05,
+      "loss": 0.4365,
+      "step": 1617
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.43859649122807e-05,
+      "loss": 0.5076,
+      "step": 1618
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4382384532760475e-05,
+      "loss": 0.5462,
+      "step": 1619
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.437880415324025e-05,
+      "loss": 0.6459,
+      "step": 1620
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4375223773720014e-05,
+      "loss": 0.5757,
+      "step": 1621
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.437164339419979e-05,
+      "loss": 0.6967,
+      "step": 1622
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4368063014679554e-05,
+      "loss": 0.5278,
+      "step": 1623
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.436448263515933e-05,
+      "loss": 0.5205,
+      "step": 1624
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.43609022556391e-05,
+      "loss": 0.6064,
+      "step": 1625
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4357321876118866e-05,
+      "loss": 0.6615,
+      "step": 1626
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4353741496598646e-05,
+      "loss": 0.6491,
+      "step": 1627
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.435016111707841e-05,
+      "loss": 0.6626,
+      "step": 1628
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4346580737558186e-05,
+      "loss": 0.593,
+      "step": 1629
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.434300035803796e-05,
+      "loss": 0.6242,
+      "step": 1630
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4339419978517726e-05,
+      "loss": 0.6035,
+      "step": 1631
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.43358395989975e-05,
+      "loss": 0.533,
+      "step": 1632
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.4332259219477265e-05,
+      "loss": 0.7427,
+      "step": 1633
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.432867883995704e-05,
+      "loss": 0.6095,
+      "step": 1634
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.432509846043681e-05,
+      "loss": 0.5284,
+      "step": 1635
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.432151808091658e-05,
+      "loss": 0.5954,
+      "step": 1636
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.431793770139635e-05,
+      "loss": 0.66,
+      "step": 1637
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.431435732187612e-05,
+      "loss": 0.513,
+      "step": 1638
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.431077694235589e-05,
+      "loss": 0.5648,
+      "step": 1639
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4307196562835664e-05,
+      "loss": 0.5534,
+      "step": 1640
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.430361618331543e-05,
+      "loss": 0.5335,
+      "step": 1641
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4300035803795204e-05,
+      "loss": 0.5943,
+      "step": 1642
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.429645542427498e-05,
+      "loss": 0.5934,
+      "step": 1643
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.429287504475474e-05,
+      "loss": 0.5636,
+      "step": 1644
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4289294665234516e-05,
+      "loss": 0.5353,
+      "step": 1645
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.428571428571428e-05,
+      "loss": 0.5416,
+      "step": 1646
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4282133906194056e-05,
+      "loss": 0.6131,
+      "step": 1647
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.427855352667383e-05,
+      "loss": 0.4745,
+      "step": 1648
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4274973147153596e-05,
+      "loss": 0.5616,
+      "step": 1649
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4271392767633376e-05,
+      "loss": 0.5155,
+      "step": 1650
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.426781238811314e-05,
+      "loss": 0.5055,
+      "step": 1651
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4264232008592915e-05,
+      "loss": 0.6725,
+      "step": 1652
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.426065162907269e-05,
+      "loss": 0.586,
+      "step": 1653
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4257071249552455e-05,
+      "loss": 0.5439,
+      "step": 1654
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.425349087003223e-05,
+      "loss": 0.558,
+      "step": 1655
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4249910490511994e-05,
+      "loss": 0.5842,
+      "step": 1656
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.424633011099177e-05,
+      "loss": 0.568,
+      "step": 1657
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.424274973147154e-05,
+      "loss": 0.7056,
+      "step": 1658
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.423916935195131e-05,
+      "loss": 0.6327,
+      "step": 1659
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.423558897243108e-05,
+      "loss": 0.5781,
+      "step": 1660
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4232008592910854e-05,
+      "loss": 0.4768,
+      "step": 1661
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.422842821339062e-05,
+      "loss": 0.5787,
+      "step": 1662
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.422484783387039e-05,
+      "loss": 0.6237,
+      "step": 1663
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.422126745435016e-05,
+      "loss": 0.6425,
+      "step": 1664
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.421768707482993e-05,
+      "loss": 0.5489,
+      "step": 1665
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.4214106695309706e-05,
+      "loss": 0.624,
+      "step": 1666
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.421052631578947e-05,
+      "loss": 0.5854,
+      "step": 1667
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4206945936269245e-05,
+      "loss": 0.511,
+      "step": 1668
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.420336555674902e-05,
+      "loss": 0.5266,
+      "step": 1669
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4199785177228785e-05,
+      "loss": 0.4752,
+      "step": 1670
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.419620479770856e-05,
+      "loss": 0.5736,
+      "step": 1671
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4192624418188325e-05,
+      "loss": 0.5388,
+      "step": 1672
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.41890440386681e-05,
+      "loss": 0.5742,
+      "step": 1673
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.418546365914787e-05,
+      "loss": 0.5816,
+      "step": 1674
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4181883279627644e-05,
+      "loss": 0.5309,
+      "step": 1675
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.417830290010742e-05,
+      "loss": 0.7421,
+      "step": 1676
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4174722520587184e-05,
+      "loss": 0.546,
+      "step": 1677
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.417114214106696e-05,
+      "loss": 0.4903,
+      "step": 1678
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.416756176154673e-05,
+      "loss": 0.5564,
+      "step": 1679
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.41639813820265e-05,
+      "loss": 0.5716,
+      "step": 1680
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.416040100250627e-05,
+      "loss": 0.577,
+      "step": 1681
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4156820622986036e-05,
+      "loss": 0.6849,
+      "step": 1682
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.415324024346581e-05,
+      "loss": 0.498,
+      "step": 1683
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.414965986394558e-05,
+      "loss": 0.5434,
+      "step": 1684
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.414607948442535e-05,
+      "loss": 0.4949,
+      "step": 1685
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.414249910490512e-05,
+      "loss": 0.5895,
+      "step": 1686
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4138918725384895e-05,
+      "loss": 0.5622,
+      "step": 1687
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.413533834586466e-05,
+      "loss": 0.5697,
+      "step": 1688
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4131757966344435e-05,
+      "loss": 0.5218,
+      "step": 1689
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.41281775868242e-05,
+      "loss": 0.5226,
+      "step": 1690
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4124597207303975e-05,
+      "loss": 0.475,
+      "step": 1691
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.412101682778375e-05,
+      "loss": 0.7018,
+      "step": 1692
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.4117436448263514e-05,
+      "loss": 0.6124,
+      "step": 1693
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.411385606874329e-05,
+      "loss": 0.6216,
+      "step": 1694
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.411027568922306e-05,
+      "loss": 0.5741,
+      "step": 1695
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.410669530970283e-05,
+      "loss": 0.6843,
+      "step": 1696
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.410311493018261e-05,
+      "loss": 0.5556,
+      "step": 1697
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.409953455066237e-05,
+      "loss": 0.5489,
+      "step": 1698
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4095954171142147e-05,
+      "loss": 0.5112,
+      "step": 1699
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.409237379162191e-05,
+      "loss": 0.5657,
+      "step": 1700
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4088793412101686e-05,
+      "loss": 0.5506,
+      "step": 1701
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.408521303258146e-05,
+      "loss": 0.5791,
+      "step": 1702
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4081632653061226e-05,
+      "loss": 0.6327,
+      "step": 1703
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4078052273541e-05,
+      "loss": 0.5488,
+      "step": 1704
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.407447189402077e-05,
+      "loss": 0.5774,
+      "step": 1705
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.407089151450054e-05,
+      "loss": 0.6631,
+      "step": 1706
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.406731113498031e-05,
+      "loss": 0.5484,
+      "step": 1707
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.406373075546008e-05,
+      "loss": 0.5741,
+      "step": 1708
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.406015037593985e-05,
+      "loss": 0.5808,
+      "step": 1709
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4056569996419625e-05,
+      "loss": 0.5774,
+      "step": 1710
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.405298961689939e-05,
+      "loss": 0.5927,
+      "step": 1711
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4049409237379164e-05,
+      "loss": 0.5192,
+      "step": 1712
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.404582885785894e-05,
+      "loss": 0.5443,
+      "step": 1713
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4042248478338704e-05,
+      "loss": 0.5761,
+      "step": 1714
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.403866809881848e-05,
+      "loss": 0.4661,
+      "step": 1715
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.403508771929824e-05,
+      "loss": 0.6437,
+      "step": 1716
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4031507339778016e-05,
+      "loss": 0.5932,
+      "step": 1717
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.402792696025779e-05,
+      "loss": 0.5955,
+      "step": 1718
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4024346580737556e-05,
+      "loss": 0.5573,
+      "step": 1719
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4020766201217336e-05,
+      "loss": 0.6023,
+      "step": 1720
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.40171858216971e-05,
+      "loss": 0.6293,
+      "step": 1721
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.4013605442176876e-05,
+      "loss": 0.6128,
+      "step": 1722
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.401002506265665e-05,
+      "loss": 0.6772,
+      "step": 1723
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.4006444683136415e-05,
+      "loss": 0.5513,
+      "step": 1724
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.400286430361619e-05,
+      "loss": 0.5122,
+      "step": 1725
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3999283924095955e-05,
+      "loss": 0.6084,
+      "step": 1726
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.399570354457573e-05,
+      "loss": 0.5187,
+      "step": 1727
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.39921231650555e-05,
+      "loss": 0.5665,
+      "step": 1728
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.398854278553527e-05,
+      "loss": 0.5808,
+      "step": 1729
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.398496240601504e-05,
+      "loss": 0.52,
+      "step": 1730
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3981382026494814e-05,
+      "loss": 0.5874,
+      "step": 1731
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.397780164697458e-05,
+      "loss": 0.6608,
+      "step": 1732
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3974221267454354e-05,
+      "loss": 0.6173,
+      "step": 1733
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.397064088793412e-05,
+      "loss": 0.6382,
+      "step": 1734
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.396706050841389e-05,
+      "loss": 0.58,
+      "step": 1735
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3963480128893666e-05,
+      "loss": 0.5153,
+      "step": 1736
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.395989974937343e-05,
+      "loss": 0.4597,
+      "step": 1737
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3956319369853206e-05,
+      "loss": 0.5951,
+      "step": 1738
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.395273899033297e-05,
+      "loss": 0.6288,
+      "step": 1739
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3949158610812746e-05,
+      "loss": 0.5175,
+      "step": 1740
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.394557823129252e-05,
+      "loss": 0.532,
+      "step": 1741
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3941997851772285e-05,
+      "loss": 0.699,
+      "step": 1742
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.393841747225206e-05,
+      "loss": 0.5902,
+      "step": 1743
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.393483709273183e-05,
+      "loss": 0.532,
+      "step": 1744
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3931256713211605e-05,
+      "loss": 0.4617,
+      "step": 1745
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.392767633369138e-05,
+      "loss": 0.4764,
+      "step": 1746
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3924095954171144e-05,
+      "loss": 0.6017,
+      "step": 1747
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.392051557465092e-05,
+      "loss": 0.5482,
+      "step": 1748
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.3916935195130684e-05,
+      "loss": 0.6826,
+      "step": 1749
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.391335481561046e-05,
+      "loss": 0.5642,
+      "step": 1750
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.390977443609023e-05,
+      "loss": 0.5885,
+      "step": 1751
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.390619405657e-05,
+      "loss": 0.5092,
+      "step": 1752
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.390261367704977e-05,
+      "loss": 0.504,
+      "step": 1753
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.389903329752954e-05,
+      "loss": 0.5721,
+      "step": 1754
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.389545291800931e-05,
+      "loss": 0.5995,
+      "step": 1755
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.389187253848908e-05,
+      "loss": 0.5277,
+      "step": 1756
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.388829215896885e-05,
+      "loss": 0.6987,
+      "step": 1757
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.388471177944862e-05,
+      "loss": 0.6451,
+      "step": 1758
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3881131399928395e-05,
+      "loss": 0.5194,
+      "step": 1759
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.387755102040816e-05,
+      "loss": 0.6607,
+      "step": 1760
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3873970640887935e-05,
+      "loss": 0.5656,
+      "step": 1761
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.387039026136771e-05,
+      "loss": 0.5351,
+      "step": 1762
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3866809881847475e-05,
+      "loss": 0.5519,
+      "step": 1763
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.386322950232725e-05,
+      "loss": 0.5667,
+      "step": 1764
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3859649122807014e-05,
+      "loss": 0.7036,
+      "step": 1765
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.385606874328679e-05,
+      "loss": 0.743,
+      "step": 1766
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.385248836376656e-05,
+      "loss": 0.5666,
+      "step": 1767
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3848907984246334e-05,
+      "loss": 0.5805,
+      "step": 1768
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.384532760472611e-05,
+      "loss": 0.6224,
+      "step": 1769
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3841747225205873e-05,
+      "loss": 0.6225,
+      "step": 1770
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.383816684568565e-05,
+      "loss": 0.5741,
+      "step": 1771
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.383458646616542e-05,
+      "loss": 0.6223,
+      "step": 1772
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3831006086645186e-05,
+      "loss": 0.6328,
+      "step": 1773
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.382742570712496e-05,
+      "loss": 0.5481,
+      "step": 1774
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.3823845327604726e-05,
+      "loss": 0.7021,
+      "step": 1775
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.38202649480845e-05,
+      "loss": 0.6657,
+      "step": 1776
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.381668456856427e-05,
+      "loss": 0.557,
+      "step": 1777
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.381310418904404e-05,
+      "loss": 0.4988,
+      "step": 1778
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.380952380952381e-05,
+      "loss": 0.623,
+      "step": 1779
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3805943430003585e-05,
+      "loss": 0.6144,
+      "step": 1780
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.380236305048335e-05,
+      "loss": 0.5874,
+      "step": 1781
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3798782670963125e-05,
+      "loss": 0.5982,
+      "step": 1782
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.379520229144289e-05,
+      "loss": 0.6197,
+      "step": 1783
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3791621911922664e-05,
+      "loss": 0.583,
+      "step": 1784
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.378804153240244e-05,
+      "loss": 0.537,
+      "step": 1785
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3784461152882204e-05,
+      "loss": 0.5788,
+      "step": 1786
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.378088077336198e-05,
+      "loss": 0.5535,
+      "step": 1787
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.377730039384175e-05,
+      "loss": 0.5758,
+      "step": 1788
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3773720014321517e-05,
+      "loss": 0.4873,
+      "step": 1789
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3770139634801297e-05,
+      "loss": 0.6618,
+      "step": 1790
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.376655925528106e-05,
+      "loss": 0.6474,
+      "step": 1791
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3762978875760836e-05,
+      "loss": 0.5596,
+      "step": 1792
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.37593984962406e-05,
+      "loss": 0.5316,
+      "step": 1793
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3755818116720376e-05,
+      "loss": 0.6508,
+      "step": 1794
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.375223773720015e-05,
+      "loss": 0.5333,
+      "step": 1795
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3748657357679915e-05,
+      "loss": 0.5417,
+      "step": 1796
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.374507697815969e-05,
+      "loss": 0.595,
+      "step": 1797
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.374149659863946e-05,
+      "loss": 0.5376,
+      "step": 1798
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.373791621911923e-05,
+      "loss": 0.591,
+      "step": 1799
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3734335839599e-05,
+      "loss": 0.6119,
+      "step": 1800
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.373075546007877e-05,
+      "loss": 0.614,
+      "step": 1801
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.372717508055854e-05,
+      "loss": 0.6321,
+      "step": 1802
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3723594701038314e-05,
+      "loss": 0.6899,
+      "step": 1803
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.372001432151808e-05,
+      "loss": 0.5262,
+      "step": 1804
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.3716433941997854e-05,
+      "loss": 0.5089,
+      "step": 1805
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.371285356247763e-05,
+      "loss": 0.553,
+      "step": 1806
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.370927318295739e-05,
+      "loss": 0.454,
+      "step": 1807
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3705692803437166e-05,
+      "loss": 0.5642,
+      "step": 1808
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.370211242391693e-05,
+      "loss": 0.6454,
+      "step": 1809
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3698532044396706e-05,
+      "loss": 0.6849,
+      "step": 1810
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.369495166487648e-05,
+      "loss": 0.6618,
+      "step": 1811
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3691371285356246e-05,
+      "loss": 0.5551,
+      "step": 1812
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.368779090583602e-05,
+      "loss": 0.615,
+      "step": 1813
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.368421052631579e-05,
+      "loss": 0.5259,
+      "step": 1814
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3680630146795565e-05,
+      "loss": 0.5437,
+      "step": 1815
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.367704976727534e-05,
+      "loss": 0.5036,
+      "step": 1816
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3673469387755105e-05,
+      "loss": 0.6283,
+      "step": 1817
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.366988900823488e-05,
+      "loss": 0.57,
+      "step": 1818
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3666308628714644e-05,
+      "loss": 0.5112,
+      "step": 1819
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.366272824919442e-05,
+      "loss": 0.5382,
+      "step": 1820
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.365914786967419e-05,
+      "loss": 0.4672,
+      "step": 1821
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.365556749015396e-05,
+      "loss": 0.633,
+      "step": 1822
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.365198711063373e-05,
+      "loss": 0.5726,
+      "step": 1823
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3648406731113504e-05,
+      "loss": 0.5635,
+      "step": 1824
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.364482635159327e-05,
+      "loss": 0.5262,
+      "step": 1825
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.364124597207304e-05,
+      "loss": 0.4827,
+      "step": 1826
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.363766559255281e-05,
+      "loss": 0.6661,
+      "step": 1827
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.363408521303258e-05,
+      "loss": 0.5094,
+      "step": 1828
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3630504833512356e-05,
+      "loss": 0.6235,
+      "step": 1829
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.362692445399212e-05,
+      "loss": 0.6308,
+      "step": 1830
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3623344074471896e-05,
+      "loss": 0.6117,
+      "step": 1831
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.361976369495166e-05,
+      "loss": 0.5728,
+      "step": 1832
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3616183315431435e-05,
+      "loss": 0.5946,
+      "step": 1833
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.361260293591121e-05,
+      "loss": 0.5067,
+      "step": 1834
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.3609022556390975e-05,
+      "loss": 0.6609,
+      "step": 1835
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.360544217687075e-05,
+      "loss": 0.6028,
+      "step": 1836
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.360186179735052e-05,
+      "loss": 0.6133,
+      "step": 1837
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3598281417830294e-05,
+      "loss": 0.5761,
+      "step": 1838
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.359470103831007e-05,
+      "loss": 0.555,
+      "step": 1839
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3591120658789834e-05,
+      "loss": 0.5495,
+      "step": 1840
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.358754027926961e-05,
+      "loss": 0.5613,
+      "step": 1841
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.358395989974938e-05,
+      "loss": 0.5829,
+      "step": 1842
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.358037952022915e-05,
+      "loss": 0.5161,
+      "step": 1843
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.357679914070892e-05,
+      "loss": 0.5699,
+      "step": 1844
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3573218761188686e-05,
+      "loss": 0.6122,
+      "step": 1845
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.356963838166846e-05,
+      "loss": 0.5529,
+      "step": 1846
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.356605800214823e-05,
+      "loss": 0.6371,
+      "step": 1847
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3562477622628e-05,
+      "loss": 0.6103,
+      "step": 1848
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.355889724310777e-05,
+      "loss": 0.5255,
+      "step": 1849
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.355531686358754e-05,
+      "loss": 0.5652,
+      "step": 1850
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.355173648406731e-05,
+      "loss": 0.5358,
+      "step": 1851
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3548156104547085e-05,
+      "loss": 0.5909,
+      "step": 1852
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.354457572502685e-05,
+      "loss": 0.512,
+      "step": 1853
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3540995345506625e-05,
+      "loss": 0.4926,
+      "step": 1854
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.35374149659864e-05,
+      "loss": 0.5834,
+      "step": 1855
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3533834586466164e-05,
+      "loss": 0.5648,
+      "step": 1856
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.353025420694594e-05,
+      "loss": 0.5687,
+      "step": 1857
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3526673827425704e-05,
+      "loss": 0.5412,
+      "step": 1858
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.352309344790548e-05,
+      "loss": 0.5704,
+      "step": 1859
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.351951306838525e-05,
+      "loss": 0.7048,
+      "step": 1860
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.3515932688865023e-05,
+      "loss": 0.6689,
+      "step": 1861
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.35123523093448e-05,
+      "loss": 0.4872,
+      "step": 1862
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.350877192982456e-05,
+      "loss": 0.5462,
+      "step": 1863
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3505191550304336e-05,
+      "loss": 0.6642,
+      "step": 1864
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.350161117078411e-05,
+      "loss": 0.5403,
+      "step": 1865
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3498030791263876e-05,
+      "loss": 0.5752,
+      "step": 1866
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.349445041174365e-05,
+      "loss": 0.6575,
+      "step": 1867
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3490870032223415e-05,
+      "loss": 0.5337,
+      "step": 1868
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.348728965270319e-05,
+      "loss": 0.5871,
+      "step": 1869
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.348370927318296e-05,
+      "loss": 0.5808,
+      "step": 1870
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.348012889366273e-05,
+      "loss": 0.5579,
+      "step": 1871
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.34765485141425e-05,
+      "loss": 0.5518,
+      "step": 1872
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3472968134622275e-05,
+      "loss": 0.5148,
+      "step": 1873
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.346938775510204e-05,
+      "loss": 0.6324,
+      "step": 1874
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3465807375581814e-05,
+      "loss": 0.566,
+      "step": 1875
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.346222699606158e-05,
+      "loss": 0.6034,
+      "step": 1876
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3458646616541354e-05,
+      "loss": 0.5314,
+      "step": 1877
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.345506623702113e-05,
+      "loss": 0.5813,
+      "step": 1878
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3451485857500893e-05,
+      "loss": 0.5657,
+      "step": 1879
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3447905477980667e-05,
+      "loss": 0.6374,
+      "step": 1880
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.344432509846044e-05,
+      "loss": 0.485,
+      "step": 1881
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3440744718940206e-05,
+      "loss": 0.7071,
+      "step": 1882
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.343716433941998e-05,
+      "loss": 0.5834,
+      "step": 1883
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.343358395989975e-05,
+      "loss": 0.5357,
+      "step": 1884
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3430003580379526e-05,
+      "loss": 0.5815,
+      "step": 1885
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.342642320085929e-05,
+      "loss": 0.4756,
+      "step": 1886
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3422842821339065e-05,
+      "loss": 0.6267,
+      "step": 1887
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.341926244181884e-05,
+      "loss": 0.5375,
+      "step": 1888
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.3415682062298605e-05,
+      "loss": 0.5487,
+      "step": 1889
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.341210168277838e-05,
+      "loss": 0.6143,
+      "step": 1890
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.340852130325815e-05,
+      "loss": 0.4972,
+      "step": 1891
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.340494092373792e-05,
+      "loss": 0.6352,
+      "step": 1892
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.340136054421769e-05,
+      "loss": 0.4939,
+      "step": 1893
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.339778016469746e-05,
+      "loss": 0.5399,
+      "step": 1894
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.339419978517723e-05,
+      "loss": 0.5985,
+      "step": 1895
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3390619405657004e-05,
+      "loss": 0.4921,
+      "step": 1896
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.338703902613677e-05,
+      "loss": 0.5539,
+      "step": 1897
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.338345864661654e-05,
+      "loss": 0.5238,
+      "step": 1898
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3379878267096316e-05,
+      "loss": 0.5503,
+      "step": 1899
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.337629788757608e-05,
+      "loss": 0.5695,
+      "step": 1900
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3372717508055856e-05,
+      "loss": 0.5713,
+      "step": 1901
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.336913712853562e-05,
+      "loss": 0.4825,
+      "step": 1902
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3365556749015396e-05,
+      "loss": 0.5135,
+      "step": 1903
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.336197636949517e-05,
+      "loss": 0.5813,
+      "step": 1904
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3358395989974935e-05,
+      "loss": 0.5237,
+      "step": 1905
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.335481561045471e-05,
+      "loss": 0.5004,
+      "step": 1906
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.335123523093448e-05,
+      "loss": 0.5308,
+      "step": 1907
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3347654851414255e-05,
+      "loss": 0.5714,
+      "step": 1908
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.334407447189403e-05,
+      "loss": 0.5937,
+      "step": 1909
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3340494092373794e-05,
+      "loss": 0.5475,
+      "step": 1910
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.333691371285357e-05,
+      "loss": 0.5961,
+      "step": 1911
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.3333333333333334e-05,
+      "loss": 0.5257,
+      "step": 1912
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.332975295381311e-05,
+      "loss": 0.6015,
+      "step": 1913
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.332617257429288e-05,
+      "loss": 0.5174,
+      "step": 1914
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.332259219477265e-05,
+      "loss": 0.5269,
+      "step": 1915
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.331901181525242e-05,
+      "loss": 0.5751,
+      "step": 1916
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.331543143573219e-05,
+      "loss": 0.5195,
+      "step": 1917
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.331185105621196e-05,
+      "loss": 0.6208,
+      "step": 1918
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.330827067669173e-05,
+      "loss": 0.6556,
+      "step": 1919
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.33046902971715e-05,
+      "loss": 0.5423,
+      "step": 1920
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.330110991765127e-05,
+      "loss": 0.6356,
+      "step": 1921
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3297529538131046e-05,
+      "loss": 0.6353,
+      "step": 1922
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.329394915861081e-05,
+      "loss": 0.6217,
+      "step": 1923
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3290368779090585e-05,
+      "loss": 0.5675,
+      "step": 1924
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.328678839957035e-05,
+      "loss": 0.5722,
+      "step": 1925
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3283208020050125e-05,
+      "loss": 0.5821,
+      "step": 1926
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.32796276405299e-05,
+      "loss": 0.5778,
+      "step": 1927
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3276047261009664e-05,
+      "loss": 0.5082,
+      "step": 1928
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.327246688148944e-05,
+      "loss": 0.5407,
+      "step": 1929
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.326888650196921e-05,
+      "loss": 0.5227,
+      "step": 1930
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3265306122448984e-05,
+      "loss": 0.5279,
+      "step": 1931
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.326172574292876e-05,
+      "loss": 0.5863,
+      "step": 1932
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3258145363408524e-05,
+      "loss": 0.5001,
+      "step": 1933
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.32545649838883e-05,
+      "loss": 0.5613,
+      "step": 1934
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.325098460436807e-05,
+      "loss": 0.6068,
+      "step": 1935
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3247404224847836e-05,
+      "loss": 0.5651,
+      "step": 1936
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.324382384532761e-05,
+      "loss": 0.5495,
+      "step": 1937
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3240243465807376e-05,
+      "loss": 0.6224,
+      "step": 1938
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.323666308628715e-05,
+      "loss": 0.4967,
+      "step": 1939
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.323308270676692e-05,
+      "loss": 0.5605,
+      "step": 1940
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.322950232724669e-05,
+      "loss": 0.5213,
+      "step": 1941
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.322592194772646e-05,
+      "loss": 0.6543,
+      "step": 1942
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.322234156820623e-05,
+      "loss": 0.5497,
+      "step": 1943
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3218761188686e-05,
+      "loss": 0.5512,
+      "step": 1944
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3215180809165775e-05,
+      "loss": 0.5526,
+      "step": 1945
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.321160042964554e-05,
+      "loss": 0.5893,
+      "step": 1946
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.3208020050125314e-05,
+      "loss": 0.5441,
+      "step": 1947
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.320443967060509e-05,
+      "loss": 0.564,
+      "step": 1948
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3200859291084854e-05,
+      "loss": 0.6024,
+      "step": 1949
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.319727891156463e-05,
+      "loss": 0.5891,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3193698532044394e-05,
+      "loss": 0.5871,
+      "step": 1951
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.319011815252417e-05,
+      "loss": 0.5741,
+      "step": 1952
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.318653777300394e-05,
+      "loss": 0.595,
+      "step": 1953
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.318295739348371e-05,
+      "loss": 0.6189,
+      "step": 1954
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3179377013963486e-05,
+      "loss": 0.4817,
+      "step": 1955
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.317579663444325e-05,
+      "loss": 0.5876,
+      "step": 1956
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3172216254923026e-05,
+      "loss": 0.4917,
+      "step": 1957
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.31686358754028e-05,
+      "loss": 0.6094,
+      "step": 1958
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3165055495882565e-05,
+      "loss": 0.6331,
+      "step": 1959
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.316147511636234e-05,
+      "loss": 0.6518,
+      "step": 1960
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3157894736842105e-05,
+      "loss": 0.5646,
+      "step": 1961
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.315431435732188e-05,
+      "loss": 0.6589,
+      "step": 1962
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.315073397780165e-05,
+      "loss": 0.6563,
+      "step": 1963
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.314715359828142e-05,
+      "loss": 0.6513,
+      "step": 1964
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.314357321876119e-05,
+      "loss": 0.5403,
+      "step": 1965
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3139992839240964e-05,
+      "loss": 0.5432,
+      "step": 1966
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.313641245972073e-05,
+      "loss": 0.5361,
+      "step": 1967
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3132832080200504e-05,
+      "loss": 0.612,
+      "step": 1968
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.312925170068027e-05,
+      "loss": 0.6483,
+      "step": 1969
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3125671321160043e-05,
+      "loss": 0.5553,
+      "step": 1970
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3122090941639817e-05,
+      "loss": 0.614,
+      "step": 1971
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.311851056211958e-05,
+      "loss": 0.703,
+      "step": 1972
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3114930182599356e-05,
+      "loss": 0.6046,
+      "step": 1973
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.311134980307913e-05,
+      "loss": 0.4771,
+      "step": 1974
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3107769423558896e-05,
+      "loss": 0.5291,
+      "step": 1975
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.310418904403867e-05,
+      "loss": 0.6021,
+      "step": 1976
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.310060866451844e-05,
+      "loss": 0.5522,
+      "step": 1977
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3097028284998215e-05,
+      "loss": 0.5809,
+      "step": 1978
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.309344790547798e-05,
+      "loss": 0.5429,
+      "step": 1979
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3089867525957755e-05,
+      "loss": 0.5375,
+      "step": 1980
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.308628714643753e-05,
+      "loss": 0.6042,
+      "step": 1981
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3082706766917295e-05,
+      "loss": 0.4932,
+      "step": 1982
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.307912638739707e-05,
+      "loss": 0.4387,
+      "step": 1983
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.307554600787684e-05,
+      "loss": 0.5163,
+      "step": 1984
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.307196562835661e-05,
+      "loss": 0.6011,
+      "step": 1985
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.306838524883638e-05,
+      "loss": 0.5874,
+      "step": 1986
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.306480486931615e-05,
+      "loss": 0.5966,
+      "step": 1987
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.306122448979592e-05,
+      "loss": 0.6274,
+      "step": 1988
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.305764411027569e-05,
+      "loss": 0.5004,
+      "step": 1989
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.305406373075546e-05,
+      "loss": 0.6205,
+      "step": 1990
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.305048335123523e-05,
+      "loss": 0.5466,
+      "step": 1991
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3046902971715006e-05,
+      "loss": 0.5868,
+      "step": 1992
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.304332259219477e-05,
+      "loss": 0.4791,
+      "step": 1993
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3039742212674546e-05,
+      "loss": 0.5104,
+      "step": 1994
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.303616183315431e-05,
+      "loss": 0.5207,
+      "step": 1995
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3032581453634085e-05,
+      "loss": 0.6129,
+      "step": 1996
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.302900107411386e-05,
+      "loss": 0.6093,
+      "step": 1997
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3025420694593625e-05,
+      "loss": 0.5994,
+      "step": 1998
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.30218403150734e-05,
+      "loss": 0.5276,
+      "step": 1999
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.301825993555317e-05,
+      "loss": 0.5606,
+      "step": 2000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3014679556032944e-05,
+      "loss": 0.6799,
+      "step": 2001
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.301109917651272e-05,
+      "loss": 0.5721,
+      "step": 2002
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.3007518796992484e-05,
+      "loss": 0.5478,
+      "step": 2003
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.300393841747226e-05,
+      "loss": 0.5859,
+      "step": 2004
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.3000358037952024e-05,
+      "loss": 0.5824,
+      "step": 2005
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.29967776584318e-05,
+      "loss": 0.6479,
+      "step": 2006
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.299319727891157e-05,
+      "loss": 0.617,
+      "step": 2007
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2989616899391336e-05,
+      "loss": 0.4881,
+      "step": 2008
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.298603651987111e-05,
+      "loss": 0.5721,
+      "step": 2009
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.298245614035088e-05,
+      "loss": 0.586,
+      "step": 2010
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.297887576083065e-05,
+      "loss": 0.6066,
+      "step": 2011
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.297529538131042e-05,
+      "loss": 0.5018,
+      "step": 2012
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.297171500179019e-05,
+      "loss": 0.5615,
+      "step": 2013
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.296813462226996e-05,
+      "loss": 0.4561,
+      "step": 2014
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2964554242749735e-05,
+      "loss": 0.4996,
+      "step": 2015
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.29609738632295e-05,
+      "loss": 0.6318,
+      "step": 2016
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2957393483709275e-05,
+      "loss": 0.5684,
+      "step": 2017
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.295381310418904e-05,
+      "loss": 0.5273,
+      "step": 2018
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2950232724668814e-05,
+      "loss": 0.544,
+      "step": 2019
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.294665234514859e-05,
+      "loss": 0.5642,
+      "step": 2020
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2943071965628354e-05,
+      "loss": 0.5859,
+      "step": 2021
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.293949158610813e-05,
+      "loss": 0.5463,
+      "step": 2022
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.29359112065879e-05,
+      "loss": 0.6638,
+      "step": 2023
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2932330827067674e-05,
+      "loss": 0.5449,
+      "step": 2024
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.292875044754745e-05,
+      "loss": 0.6859,
+      "step": 2025
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.292517006802721e-05,
+      "loss": 0.6469,
+      "step": 2026
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2921589688506986e-05,
+      "loss": 0.6123,
+      "step": 2027
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.291800930898676e-05,
+      "loss": 0.5294,
+      "step": 2028
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2914428929466526e-05,
+      "loss": 0.5566,
+      "step": 2029
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.29108485499463e-05,
+      "loss": 0.5503,
+      "step": 2030
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.2907268170426066e-05,
+      "loss": 0.57,
+      "step": 2031
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.290368779090584e-05,
+      "loss": 0.5174,
+      "step": 2032
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.290010741138561e-05,
+      "loss": 0.6053,
+      "step": 2033
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.289652703186538e-05,
+      "loss": 0.591,
+      "step": 2034
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.289294665234515e-05,
+      "loss": 0.5899,
+      "step": 2035
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.288936627282492e-05,
+      "loss": 0.5635,
+      "step": 2036
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.288578589330469e-05,
+      "loss": 0.5854,
+      "step": 2037
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2882205513784464e-05,
+      "loss": 0.5211,
+      "step": 2038
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.287862513426423e-05,
+      "loss": 0.5171,
+      "step": 2039
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2875044754744004e-05,
+      "loss": 0.5874,
+      "step": 2040
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.287146437522378e-05,
+      "loss": 0.5348,
+      "step": 2041
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2867883995703544e-05,
+      "loss": 0.5766,
+      "step": 2042
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.286430361618332e-05,
+      "loss": 0.4977,
+      "step": 2043
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.286072323666308e-05,
+      "loss": 0.5898,
+      "step": 2044
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2857142857142856e-05,
+      "loss": 0.5536,
+      "step": 2045
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.285356247762263e-05,
+      "loss": 0.4984,
+      "step": 2046
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.28499820981024e-05,
+      "loss": 0.5804,
+      "step": 2047
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2846401718582176e-05,
+      "loss": 0.5773,
+      "step": 2048
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.284282133906194e-05,
+      "loss": 0.5862,
+      "step": 2049
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2839240959541715e-05,
+      "loss": 0.4808,
+      "step": 2050
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.283566058002149e-05,
+      "loss": 0.6103,
+      "step": 2051
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2832080200501255e-05,
+      "loss": 0.5211,
+      "step": 2052
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.282849982098103e-05,
+      "loss": 0.4812,
+      "step": 2053
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2824919441460795e-05,
+      "loss": 0.4791,
+      "step": 2054
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.282133906194057e-05,
+      "loss": 0.5875,
+      "step": 2055
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.281775868242034e-05,
+      "loss": 0.5793,
+      "step": 2056
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.281417830290011e-05,
+      "loss": 0.5838,
+      "step": 2057
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.281059792337988e-05,
+      "loss": 0.5305,
+      "step": 2058
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.2807017543859654e-05,
+      "loss": 0.5481,
+      "step": 2059
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.280343716433942e-05,
+      "loss": 0.6856,
+      "step": 2060
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2799856784819193e-05,
+      "loss": 0.6025,
+      "step": 2061
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.279627640529896e-05,
+      "loss": 0.4721,
+      "step": 2062
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.279269602577873e-05,
+      "loss": 0.5398,
+      "step": 2063
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2789115646258506e-05,
+      "loss": 0.5973,
+      "step": 2064
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.278553526673827e-05,
+      "loss": 0.465,
+      "step": 2065
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2781954887218046e-05,
+      "loss": 0.5377,
+      "step": 2066
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.277837450769782e-05,
+      "loss": 0.4835,
+      "step": 2067
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2774794128177585e-05,
+      "loss": 0.5671,
+      "step": 2068
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.277121374865736e-05,
+      "loss": 0.6365,
+      "step": 2069
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2767633369137125e-05,
+      "loss": 0.4857,
+      "step": 2070
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2764052989616905e-05,
+      "loss": 0.5312,
+      "step": 2071
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.276047261009667e-05,
+      "loss": 0.5724,
+      "step": 2072
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2756892230576445e-05,
+      "loss": 0.5799,
+      "step": 2073
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.275331185105622e-05,
+      "loss": 0.4915,
+      "step": 2074
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2749731471535984e-05,
+      "loss": 0.5586,
+      "step": 2075
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.274615109201576e-05,
+      "loss": 0.5392,
+      "step": 2076
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.274257071249553e-05,
+      "loss": 0.6364,
+      "step": 2077
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.27389903329753e-05,
+      "loss": 0.5663,
+      "step": 2078
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.273540995345507e-05,
+      "loss": 0.509,
+      "step": 2079
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2731829573934837e-05,
+      "loss": 0.6537,
+      "step": 2080
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.272824919441461e-05,
+      "loss": 0.5423,
+      "step": 2081
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.272466881489438e-05,
+      "loss": 0.6125,
+      "step": 2082
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.272108843537415e-05,
+      "loss": 0.5696,
+      "step": 2083
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.271750805585392e-05,
+      "loss": 0.4752,
+      "step": 2084
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2713927676333696e-05,
+      "loss": 0.4748,
+      "step": 2085
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.271034729681346e-05,
+      "loss": 0.4299,
+      "step": 2086
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2706766917293235e-05,
+      "loss": 0.5412,
+      "step": 2087
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.2703186537773e-05,
+      "loss": 0.5774,
+      "step": 2088
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2699606158252775e-05,
+      "loss": 0.6302,
+      "step": 2089
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.269602577873255e-05,
+      "loss": 0.533,
+      "step": 2090
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2692445399212315e-05,
+      "loss": 0.5917,
+      "step": 2091
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.268886501969209e-05,
+      "loss": 0.5974,
+      "step": 2092
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.268528464017186e-05,
+      "loss": 0.5,
+      "step": 2093
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2681704260651634e-05,
+      "loss": 0.5873,
+      "step": 2094
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.267812388113141e-05,
+      "loss": 0.5619,
+      "step": 2095
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2674543501611174e-05,
+      "loss": 0.5386,
+      "step": 2096
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.267096312209095e-05,
+      "loss": 0.5437,
+      "step": 2097
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.266738274257071e-05,
+      "loss": 0.4374,
+      "step": 2098
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2663802363050486e-05,
+      "loss": 0.5232,
+      "step": 2099
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.266022198353026e-05,
+      "loss": 0.5258,
+      "step": 2100
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2656641604010026e-05,
+      "loss": 0.57,
+      "step": 2101
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.26530612244898e-05,
+      "loss": 0.6091,
+      "step": 2102
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.264948084496957e-05,
+      "loss": 0.5511,
+      "step": 2103
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.264590046544934e-05,
+      "loss": 0.5616,
+      "step": 2104
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.264232008592911e-05,
+      "loss": 0.5765,
+      "step": 2105
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.263873970640888e-05,
+      "loss": 0.5844,
+      "step": 2106
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.263515932688865e-05,
+      "loss": 0.5729,
+      "step": 2107
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2631578947368425e-05,
+      "loss": 0.5336,
+      "step": 2108
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.262799856784819e-05,
+      "loss": 0.5787,
+      "step": 2109
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2624418188327964e-05,
+      "loss": 0.5108,
+      "step": 2110
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.262083780880774e-05,
+      "loss": 0.5289,
+      "step": 2111
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2617257429287504e-05,
+      "loss": 0.5858,
+      "step": 2112
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.261367704976728e-05,
+      "loss": 0.5633,
+      "step": 2113
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.2610096670247044e-05,
+      "loss": 0.4087,
+      "step": 2114
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.260651629072682e-05,
+      "loss": 0.5981,
+      "step": 2115
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.260293591120659e-05,
+      "loss": 0.4648,
+      "step": 2116
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.259935553168636e-05,
+      "loss": 0.587,
+      "step": 2117
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2595775152166136e-05,
+      "loss": 0.5355,
+      "step": 2118
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.25921947726459e-05,
+      "loss": 0.6453,
+      "step": 2119
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2588614393125676e-05,
+      "loss": 0.5995,
+      "step": 2120
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.258503401360545e-05,
+      "loss": 0.6684,
+      "step": 2121
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2581453634085216e-05,
+      "loss": 0.4764,
+      "step": 2122
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.257787325456499e-05,
+      "loss": 0.5348,
+      "step": 2123
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2574292875044755e-05,
+      "loss": 0.5158,
+      "step": 2124
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.257071249552453e-05,
+      "loss": 0.4682,
+      "step": 2125
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.25671321160043e-05,
+      "loss": 0.6355,
+      "step": 2126
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.256355173648407e-05,
+      "loss": 0.4701,
+      "step": 2127
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.255997135696384e-05,
+      "loss": 0.6371,
+      "step": 2128
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.255639097744361e-05,
+      "loss": 0.5596,
+      "step": 2129
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.255281059792338e-05,
+      "loss": 0.6239,
+      "step": 2130
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2549230218403154e-05,
+      "loss": 0.559,
+      "step": 2131
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.254564983888292e-05,
+      "loss": 0.5023,
+      "step": 2132
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2542069459362694e-05,
+      "loss": 0.5952,
+      "step": 2133
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.253848907984247e-05,
+      "loss": 0.5605,
+      "step": 2134
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.253490870032223e-05,
+      "loss": 0.6099,
+      "step": 2135
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2531328320802006e-05,
+      "loss": 0.5917,
+      "step": 2136
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.252774794128177e-05,
+      "loss": 0.6167,
+      "step": 2137
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2524167561761546e-05,
+      "loss": 0.623,
+      "step": 2138
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.252058718224132e-05,
+      "loss": 0.5797,
+      "step": 2139
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2517006802721085e-05,
+      "loss": 0.5233,
+      "step": 2140
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2513426423200865e-05,
+      "loss": 0.5531,
+      "step": 2141
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.250984604368063e-05,
+      "loss": 0.5289,
+      "step": 2142
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.2506265664160405e-05,
+      "loss": 0.5163,
+      "step": 2143
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.250268528464018e-05,
+      "loss": 0.5263,
+      "step": 2144
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2499104905119945e-05,
+      "loss": 0.5504,
+      "step": 2145
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.249552452559972e-05,
+      "loss": 0.4836,
+      "step": 2146
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2491944146079484e-05,
+      "loss": 0.4972,
+      "step": 2147
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.248836376655926e-05,
+      "loss": 0.4438,
+      "step": 2148
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.248478338703903e-05,
+      "loss": 0.5397,
+      "step": 2149
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.24812030075188e-05,
+      "loss": 0.5417,
+      "step": 2150
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.247762262799857e-05,
+      "loss": 0.4724,
+      "step": 2151
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2474042248478343e-05,
+      "loss": 0.5661,
+      "step": 2152
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.247046186895811e-05,
+      "loss": 0.5167,
+      "step": 2153
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.246688148943788e-05,
+      "loss": 0.5365,
+      "step": 2154
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.246330110991765e-05,
+      "loss": 0.6677,
+      "step": 2155
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.245972073039742e-05,
+      "loss": 0.5533,
+      "step": 2156
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2456140350877196e-05,
+      "loss": 0.5394,
+      "step": 2157
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.245255997135696e-05,
+      "loss": 0.5488,
+      "step": 2158
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2448979591836735e-05,
+      "loss": 0.4618,
+      "step": 2159
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.244539921231651e-05,
+      "loss": 0.5091,
+      "step": 2160
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2441818832796275e-05,
+      "loss": 0.5101,
+      "step": 2161
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.243823845327605e-05,
+      "loss": 0.6284,
+      "step": 2162
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2434658073755815e-05,
+      "loss": 0.5554,
+      "step": 2163
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2431077694235595e-05,
+      "loss": 0.7178,
+      "step": 2164
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.242749731471536e-05,
+      "loss": 0.5856,
+      "step": 2165
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2423916935195134e-05,
+      "loss": 0.5083,
+      "step": 2166
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.242033655567491e-05,
+      "loss": 0.5212,
+      "step": 2167
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2416756176154674e-05,
+      "loss": 0.4753,
+      "step": 2168
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.241317579663445e-05,
+      "loss": 0.6104,
+      "step": 2169
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.240959541711422e-05,
+      "loss": 0.6465,
+      "step": 2170
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.2406015037593987e-05,
+      "loss": 0.5176,
+      "step": 2171
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.240243465807376e-05,
+      "loss": 0.5056,
+      "step": 2172
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2398854278553526e-05,
+      "loss": 0.4473,
+      "step": 2173
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.23952738990333e-05,
+      "loss": 0.5411,
+      "step": 2174
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.239169351951307e-05,
+      "loss": 0.4795,
+      "step": 2175
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.238811313999284e-05,
+      "loss": 0.4843,
+      "step": 2176
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.238453276047261e-05,
+      "loss": 0.5097,
+      "step": 2177
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2380952380952385e-05,
+      "loss": 0.559,
+      "step": 2178
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.237737200143215e-05,
+      "loss": 0.5143,
+      "step": 2179
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2373791621911925e-05,
+      "loss": 0.6142,
+      "step": 2180
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.237021124239169e-05,
+      "loss": 0.4832,
+      "step": 2181
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2366630862871465e-05,
+      "loss": 0.5247,
+      "step": 2182
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.236305048335124e-05,
+      "loss": 0.7269,
+      "step": 2183
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2359470103831004e-05,
+      "loss": 0.4907,
+      "step": 2184
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.235588972431078e-05,
+      "loss": 0.4738,
+      "step": 2185
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.235230934479055e-05,
+      "loss": 0.5321,
+      "step": 2186
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2348728965270324e-05,
+      "loss": 0.5884,
+      "step": 2187
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.23451485857501e-05,
+      "loss": 0.5451,
+      "step": 2188
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.234156820622986e-05,
+      "loss": 0.6522,
+      "step": 2189
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2337987826709636e-05,
+      "loss": 0.5042,
+      "step": 2190
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.23344074471894e-05,
+      "loss": 0.5528,
+      "step": 2191
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2330827067669176e-05,
+      "loss": 0.5181,
+      "step": 2192
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.232724668814895e-05,
+      "loss": 0.5564,
+      "step": 2193
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.2323666308628716e-05,
+      "loss": 0.491,
+      "step": 2194
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.232008592910849e-05,
+      "loss": 0.5331,
+      "step": 2195
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.231650554958826e-05,
+      "loss": 0.5436,
+      "step": 2196
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.231292517006803e-05,
+      "loss": 0.6572,
+      "step": 2197
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.23093447905478e-05,
+      "loss": 0.4474,
+      "step": 2198
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.230576441102757e-05,
+      "loss": 0.655,
+      "step": 2199
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.230218403150734e-05,
+      "loss": 0.6144,
+      "step": 2200
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2298603651987114e-05,
+      "loss": 0.6665,
+      "step": 2201
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.229502327246688e-05,
+      "loss": 0.5865,
+      "step": 2202
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2291442892946654e-05,
+      "loss": 0.5429,
+      "step": 2203
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.228786251342643e-05,
+      "loss": 0.5418,
+      "step": 2204
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2284282133906194e-05,
+      "loss": 0.6338,
+      "step": 2205
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.228070175438597e-05,
+      "loss": 0.6404,
+      "step": 2206
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.227712137486573e-05,
+      "loss": 0.6367,
+      "step": 2207
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2273540995345506e-05,
+      "loss": 0.496,
+      "step": 2208
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.226996061582528e-05,
+      "loss": 0.5408,
+      "step": 2209
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2266380236305046e-05,
+      "loss": 0.4896,
+      "step": 2210
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2262799856784826e-05,
+      "loss": 0.6205,
+      "step": 2211
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.225921947726459e-05,
+      "loss": 0.4627,
+      "step": 2212
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2255639097744366e-05,
+      "loss": 0.5881,
+      "step": 2213
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.225205871822414e-05,
+      "loss": 0.584,
+      "step": 2214
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2248478338703905e-05,
+      "loss": 0.5998,
+      "step": 2215
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.224489795918368e-05,
+      "loss": 0.515,
+      "step": 2216
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2241317579663445e-05,
+      "loss": 0.5199,
+      "step": 2217
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.223773720014322e-05,
+      "loss": 0.5396,
+      "step": 2218
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.223415682062299e-05,
+      "loss": 0.4506,
+      "step": 2219
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.223057644110276e-05,
+      "loss": 0.5077,
+      "step": 2220
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.222699606158253e-05,
+      "loss": 0.558,
+      "step": 2221
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2223415682062304e-05,
+      "loss": 0.5803,
+      "step": 2222
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.221983530254207e-05,
+      "loss": 0.5788,
+      "step": 2223
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2216254923021844e-05,
+      "loss": 0.5725,
+      "step": 2224
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.221267454350161e-05,
+      "loss": 0.6469,
+      "step": 2225
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.220909416398138e-05,
+      "loss": 0.47,
+      "step": 2226
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.2205513784461156e-05,
+      "loss": 0.5209,
+      "step": 2227
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.220193340494092e-05,
+      "loss": 0.5511,
+      "step": 2228
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2198353025420696e-05,
+      "loss": 0.6221,
+      "step": 2229
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.219477264590046e-05,
+      "loss": 0.5558,
+      "step": 2230
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2191192266380235e-05,
+      "loss": 0.518,
+      "step": 2231
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.218761188686001e-05,
+      "loss": 0.5445,
+      "step": 2232
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2184031507339775e-05,
+      "loss": 0.4157,
+      "step": 2233
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2180451127819555e-05,
+      "loss": 0.546,
+      "step": 2234
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.217687074829932e-05,
+      "loss": 0.4725,
+      "step": 2235
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2173290368779095e-05,
+      "loss": 0.5547,
+      "step": 2236
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.216970998925887e-05,
+      "loss": 0.591,
+      "step": 2237
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2166129609738634e-05,
+      "loss": 0.5509,
+      "step": 2238
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.216254923021841e-05,
+      "loss": 0.5257,
+      "step": 2239
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.215896885069818e-05,
+      "loss": 0.5227,
+      "step": 2240
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.215538847117795e-05,
+      "loss": 0.4425,
+      "step": 2241
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.215180809165772e-05,
+      "loss": 0.544,
+      "step": 2242
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.214822771213749e-05,
+      "loss": 0.4618,
+      "step": 2243
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.214464733261726e-05,
+      "loss": 0.6331,
+      "step": 2244
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.214106695309703e-05,
+      "loss": 0.5991,
+      "step": 2245
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.21374865735768e-05,
+      "loss": 0.5319,
+      "step": 2246
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.213390619405657e-05,
+      "loss": 0.4892,
+      "step": 2247
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.213032581453634e-05,
+      "loss": 0.5594,
+      "step": 2248
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.212674543501611e-05,
+      "loss": 0.5337,
+      "step": 2249
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2123165055495885e-05,
+      "loss": 0.5779,
+      "step": 2250
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.211958467597565e-05,
+      "loss": 0.5536,
+      "step": 2251
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2116004296455425e-05,
+      "loss": 0.5581,
+      "step": 2252
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.21124239169352e-05,
+      "loss": 0.5361,
+      "step": 2253
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2108843537414965e-05,
+      "loss": 0.4488,
+      "step": 2254
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.210526315789474e-05,
+      "loss": 0.5706,
+      "step": 2255
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2101682778374504e-05,
+      "loss": 0.5284,
+      "step": 2256
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2098102398854284e-05,
+      "loss": 0.5194,
+      "step": 2257
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.209452201933405e-05,
+      "loss": 0.7088,
+      "step": 2258
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2090941639813824e-05,
+      "loss": 0.5639,
+      "step": 2259
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.20873612602936e-05,
+      "loss": 0.544,
+      "step": 2260
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.208378088077336e-05,
+      "loss": 0.5429,
+      "step": 2261
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2080200501253137e-05,
+      "loss": 0.5084,
+      "step": 2262
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.207662012173291e-05,
+      "loss": 0.6323,
+      "step": 2263
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2073039742212676e-05,
+      "loss": 0.5383,
+      "step": 2264
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.206945936269245e-05,
+      "loss": 0.647,
+      "step": 2265
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2065878983172216e-05,
+      "loss": 0.5436,
+      "step": 2266
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.206229860365199e-05,
+      "loss": 0.533,
+      "step": 2267
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.205871822413176e-05,
+      "loss": 0.5984,
+      "step": 2268
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.205513784461153e-05,
+      "loss": 0.6679,
+      "step": 2269
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.20515574650913e-05,
+      "loss": 0.5189,
+      "step": 2270
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2047977085571075e-05,
+      "loss": 0.5295,
+      "step": 2271
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.204439670605084e-05,
+      "loss": 0.4657,
+      "step": 2272
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2040816326530615e-05,
+      "loss": 0.4519,
+      "step": 2273
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.203723594701038e-05,
+      "loss": 0.5492,
+      "step": 2274
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2033655567490154e-05,
+      "loss": 0.6169,
+      "step": 2275
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.203007518796993e-05,
+      "loss": 0.5324,
+      "step": 2276
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2026494808449694e-05,
+      "loss": 0.6099,
+      "step": 2277
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.202291442892947e-05,
+      "loss": 0.591,
+      "step": 2278
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.201933404940924e-05,
+      "loss": 0.5055,
+      "step": 2279
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2015753669889006e-05,
+      "loss": 0.5024,
+      "step": 2280
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2012173290368786e-05,
+      "loss": 0.5452,
+      "step": 2281
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.200859291084855e-05,
+      "loss": 0.6194,
+      "step": 2282
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.2005012531328326e-05,
+      "loss": 0.5482,
+      "step": 2283
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.200143215180809e-05,
+      "loss": 0.5596,
+      "step": 2284
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1997851772287866e-05,
+      "loss": 0.5726,
+      "step": 2285
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.199427139276764e-05,
+      "loss": 0.4555,
+      "step": 2286
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1990691013247405e-05,
+      "loss": 0.5135,
+      "step": 2287
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.198711063372718e-05,
+      "loss": 0.4957,
+      "step": 2288
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.198353025420695e-05,
+      "loss": 0.5383,
+      "step": 2289
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.197994987468672e-05,
+      "loss": 0.5974,
+      "step": 2290
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.197636949516649e-05,
+      "loss": 0.583,
+      "step": 2291
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.197278911564626e-05,
+      "loss": 0.4856,
+      "step": 2292
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.196920873612603e-05,
+      "loss": 0.485,
+      "step": 2293
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1965628356605804e-05,
+      "loss": 0.4997,
+      "step": 2294
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.196204797708557e-05,
+      "loss": 0.5476,
+      "step": 2295
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1958467597565344e-05,
+      "loss": 0.5287,
+      "step": 2296
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.195488721804512e-05,
+      "loss": 0.4709,
+      "step": 2297
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.195130683852488e-05,
+      "loss": 0.5736,
+      "step": 2298
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1947726459004656e-05,
+      "loss": 0.4615,
+      "step": 2299
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.194414607948442e-05,
+      "loss": 0.5306,
+      "step": 2300
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1940565699964196e-05,
+      "loss": 0.515,
+      "step": 2301
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.193698532044397e-05,
+      "loss": 0.4308,
+      "step": 2302
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1933404940923736e-05,
+      "loss": 0.554,
+      "step": 2303
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1929824561403516e-05,
+      "loss": 0.571,
+      "step": 2304
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.192624418188328e-05,
+      "loss": 0.4434,
+      "step": 2305
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1922663802363055e-05,
+      "loss": 0.5615,
+      "step": 2306
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.191908342284283e-05,
+      "loss": 0.6195,
+      "step": 2307
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1915503043322595e-05,
+      "loss": 0.5385,
+      "step": 2308
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.191192266380237e-05,
+      "loss": 0.5275,
+      "step": 2309
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.1908342284282134e-05,
+      "loss": 0.4919,
+      "step": 2310
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.190476190476191e-05,
+      "loss": 0.5725,
+      "step": 2311
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.190118152524168e-05,
+      "loss": 0.6384,
+      "step": 2312
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.189760114572145e-05,
+      "loss": 0.4696,
+      "step": 2313
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.189402076620122e-05,
+      "loss": 0.6236,
+      "step": 2314
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1890440386680994e-05,
+      "loss": 0.5283,
+      "step": 2315
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.188686000716076e-05,
+      "loss": 0.4128,
+      "step": 2316
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.188327962764053e-05,
+      "loss": 0.5784,
+      "step": 2317
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.18796992481203e-05,
+      "loss": 0.4659,
+      "step": 2318
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.187611886860007e-05,
+      "loss": 0.4935,
+      "step": 2319
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1872538489079846e-05,
+      "loss": 0.6415,
+      "step": 2320
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.186895810955961e-05,
+      "loss": 0.521,
+      "step": 2321
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1865377730039386e-05,
+      "loss": 0.4499,
+      "step": 2322
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.186179735051915e-05,
+      "loss": 0.5992,
+      "step": 2323
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1858216970998925e-05,
+      "loss": 0.65,
+      "step": 2324
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.18546365914787e-05,
+      "loss": 0.504,
+      "step": 2325
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1851056211958465e-05,
+      "loss": 0.5237,
+      "step": 2326
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1847475832438245e-05,
+      "loss": 0.5104,
+      "step": 2327
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.184389545291801e-05,
+      "loss": 0.5229,
+      "step": 2328
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1840315073397784e-05,
+      "loss": 0.5868,
+      "step": 2329
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.183673469387756e-05,
+      "loss": 0.5524,
+      "step": 2330
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1833154314357324e-05,
+      "loss": 0.5459,
+      "step": 2331
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.18295739348371e-05,
+      "loss": 0.505,
+      "step": 2332
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.182599355531687e-05,
+      "loss": 0.608,
+      "step": 2333
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.182241317579664e-05,
+      "loss": 0.4771,
+      "step": 2334
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.181883279627641e-05,
+      "loss": 0.468,
+      "step": 2335
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.1815252416756176e-05,
+      "loss": 0.606,
+      "step": 2336
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.181167203723595e-05,
+      "loss": 0.537,
+      "step": 2337
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.180809165771572e-05,
+      "loss": 0.5457,
+      "step": 2338
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.180451127819549e-05,
+      "loss": 0.564,
+      "step": 2339
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.180093089867526e-05,
+      "loss": 0.5815,
+      "step": 2340
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.179735051915503e-05,
+      "loss": 0.5523,
+      "step": 2341
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.17937701396348e-05,
+      "loss": 0.5493,
+      "step": 2342
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1790189760114575e-05,
+      "loss": 0.4832,
+      "step": 2343
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.178660938059434e-05,
+      "loss": 0.5607,
+      "step": 2344
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1783029001074115e-05,
+      "loss": 0.5883,
+      "step": 2345
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.177944862155389e-05,
+      "loss": 0.7145,
+      "step": 2346
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1775868242033654e-05,
+      "loss": 0.5397,
+      "step": 2347
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.177228786251343e-05,
+      "loss": 0.5006,
+      "step": 2348
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1768707482993194e-05,
+      "loss": 0.5218,
+      "step": 2349
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.176512710347297e-05,
+      "loss": 0.6663,
+      "step": 2350
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.176154672395275e-05,
+      "loss": 0.5234,
+      "step": 2351
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.175796634443251e-05,
+      "loss": 0.6004,
+      "step": 2352
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1754385964912287e-05,
+      "loss": 0.5742,
+      "step": 2353
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.175080558539205e-05,
+      "loss": 0.6768,
+      "step": 2354
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1747225205871826e-05,
+      "loss": 0.5252,
+      "step": 2355
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.17436448263516e-05,
+      "loss": 0.5892,
+      "step": 2356
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1740064446831366e-05,
+      "loss": 0.5906,
+      "step": 2357
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.173648406731114e-05,
+      "loss": 0.5081,
+      "step": 2358
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1732903687790905e-05,
+      "loss": 0.5539,
+      "step": 2359
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.172932330827068e-05,
+      "loss": 0.5165,
+      "step": 2360
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.172574292875045e-05,
+      "loss": 0.512,
+      "step": 2361
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.172216254923022e-05,
+      "loss": 0.5394,
+      "step": 2362
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.171858216970999e-05,
+      "loss": 0.6091,
+      "step": 2363
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1715001790189765e-05,
+      "loss": 0.6027,
+      "step": 2364
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.171142141066953e-05,
+      "loss": 0.5369,
+      "step": 2365
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1707841031149304e-05,
+      "loss": 0.5631,
+      "step": 2366
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.170426065162907e-05,
+      "loss": 0.5243,
+      "step": 2367
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.1700680272108844e-05,
+      "loss": 0.5887,
+      "step": 2368
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.169709989258862e-05,
+      "loss": 0.5556,
+      "step": 2369
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.169351951306838e-05,
+      "loss": 0.5364,
+      "step": 2370
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1689939133548156e-05,
+      "loss": 0.6018,
+      "step": 2371
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.168635875402793e-05,
+      "loss": 0.5348,
+      "step": 2372
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1682778374507696e-05,
+      "loss": 0.5186,
+      "step": 2373
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1679197994987476e-05,
+      "loss": 0.4728,
+      "step": 2374
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.167561761546724e-05,
+      "loss": 0.4528,
+      "step": 2375
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1672037235947016e-05,
+      "loss": 0.5571,
+      "step": 2376
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.166845685642678e-05,
+      "loss": 0.4043,
+      "step": 2377
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1664876476906555e-05,
+      "loss": 0.4134,
+      "step": 2378
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.166129609738633e-05,
+      "loss": 0.4629,
+      "step": 2379
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1657715717866095e-05,
+      "loss": 0.51,
+      "step": 2380
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.165413533834587e-05,
+      "loss": 0.5561,
+      "step": 2381
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.165055495882564e-05,
+      "loss": 0.588,
+      "step": 2382
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.164697457930541e-05,
+      "loss": 0.4873,
+      "step": 2383
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.164339419978518e-05,
+      "loss": 0.6017,
+      "step": 2384
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.163981382026495e-05,
+      "loss": 0.594,
+      "step": 2385
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.163623344074472e-05,
+      "loss": 0.5883,
+      "step": 2386
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1632653061224494e-05,
+      "loss": 0.4341,
+      "step": 2387
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.162907268170426e-05,
+      "loss": 0.5117,
+      "step": 2388
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.162549230218403e-05,
+      "loss": 0.5621,
+      "step": 2389
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1621911922663806e-05,
+      "loss": 0.5342,
+      "step": 2390
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.161833154314357e-05,
+      "loss": 0.5463,
+      "step": 2391
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1614751163623346e-05,
+      "loss": 0.6766,
+      "step": 2392
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.161117078410311e-05,
+      "loss": 0.7212,
+      "step": 2393
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1607590404582886e-05,
+      "loss": 0.5314,
+      "step": 2394
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.160401002506266e-05,
+      "loss": 0.5283,
+      "step": 2395
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.1600429645542425e-05,
+      "loss": 0.4255,
+      "step": 2396
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1596849266022205e-05,
+      "loss": 0.6521,
+      "step": 2397
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.159326888650197e-05,
+      "loss": 0.5211,
+      "step": 2398
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1589688506981745e-05,
+      "loss": 0.4782,
+      "step": 2399
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.158610812746152e-05,
+      "loss": 0.5121,
+      "step": 2400
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1582527747941284e-05,
+      "loss": 0.5334,
+      "step": 2401
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.157894736842106e-05,
+      "loss": 0.5247,
+      "step": 2402
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1575366988900824e-05,
+      "loss": 0.5578,
+      "step": 2403
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.15717866093806e-05,
+      "loss": 0.6368,
+      "step": 2404
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.156820622986037e-05,
+      "loss": 0.4942,
+      "step": 2405
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.156462585034014e-05,
+      "loss": 0.4956,
+      "step": 2406
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.156104547081991e-05,
+      "loss": 0.6905,
+      "step": 2407
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.155746509129968e-05,
+      "loss": 0.5103,
+      "step": 2408
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.155388471177945e-05,
+      "loss": 0.5185,
+      "step": 2409
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.155030433225922e-05,
+      "loss": 0.5853,
+      "step": 2410
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.154672395273899e-05,
+      "loss": 0.5051,
+      "step": 2411
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.154314357321876e-05,
+      "loss": 0.5335,
+      "step": 2412
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1539563193698536e-05,
+      "loss": 0.6019,
+      "step": 2413
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.15359828141783e-05,
+      "loss": 0.6518,
+      "step": 2414
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1532402434658075e-05,
+      "loss": 0.4623,
+      "step": 2415
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.152882205513784e-05,
+      "loss": 0.6096,
+      "step": 2416
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1525241675617615e-05,
+      "loss": 0.5918,
+      "step": 2417
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.152166129609739e-05,
+      "loss": 0.5264,
+      "step": 2418
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1518080916577154e-05,
+      "loss": 0.5305,
+      "step": 2419
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.151450053705693e-05,
+      "loss": 0.6035,
+      "step": 2420
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.15109201575367e-05,
+      "loss": 0.4898,
+      "step": 2421
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1507339778016474e-05,
+      "loss": 0.6575,
+      "step": 2422
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.150375939849625e-05,
+      "loss": 0.4837,
+      "step": 2423
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1500179018976013e-05,
+      "loss": 0.6948,
+      "step": 2424
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.149659863945579e-05,
+      "loss": 0.4972,
+      "step": 2425
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.149301825993556e-05,
+      "loss": 0.6387,
+      "step": 2426
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1489437880415326e-05,
+      "loss": 0.5223,
+      "step": 2427
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.14858575008951e-05,
+      "loss": 0.5148,
+      "step": 2428
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1482277121374866e-05,
+      "loss": 0.6471,
+      "step": 2429
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.147869674185464e-05,
+      "loss": 0.5352,
+      "step": 2430
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.147511636233441e-05,
+      "loss": 0.5651,
+      "step": 2431
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.147153598281418e-05,
+      "loss": 0.52,
+      "step": 2432
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.146795560329395e-05,
+      "loss": 0.4967,
+      "step": 2433
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.146437522377372e-05,
+      "loss": 0.5273,
+      "step": 2434
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.146079484425349e-05,
+      "loss": 0.4397,
+      "step": 2435
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1457214464733265e-05,
+      "loss": 0.5105,
+      "step": 2436
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.145363408521303e-05,
+      "loss": 0.5218,
+      "step": 2437
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1450053705692804e-05,
+      "loss": 0.4524,
+      "step": 2438
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.144647332617258e-05,
+      "loss": 0.4178,
+      "step": 2439
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1442892946652344e-05,
+      "loss": 0.5054,
+      "step": 2440
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.143931256713212e-05,
+      "loss": 0.5648,
+      "step": 2441
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1435732187611883e-05,
+      "loss": 0.4437,
+      "step": 2442
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1432151808091657e-05,
+      "loss": 0.5713,
+      "step": 2443
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1428571428571437e-05,
+      "loss": 0.474,
+      "step": 2444
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.14249910490512e-05,
+      "loss": 0.5267,
+      "step": 2445
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1421410669530976e-05,
+      "loss": 0.5331,
+      "step": 2446
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.141783029001074e-05,
+      "loss": 0.5616,
+      "step": 2447
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1414249910490516e-05,
+      "loss": 0.493,
+      "step": 2448
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.141066953097029e-05,
+      "loss": 0.8337,
+      "step": 2449
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1407089151450055e-05,
+      "loss": 0.4678,
+      "step": 2450
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.140350877192983e-05,
+      "loss": 0.5144,
+      "step": 2451
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1399928392409595e-05,
+      "loss": 0.5041,
+      "step": 2452
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.139634801288937e-05,
+      "loss": 0.5506,
+      "step": 2453
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.139276763336914e-05,
+      "loss": 0.6357,
+      "step": 2454
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.138918725384891e-05,
+      "loss": 0.6384,
+      "step": 2455
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.138560687432868e-05,
+      "loss": 0.5246,
+      "step": 2456
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1382026494808454e-05,
+      "loss": 0.5569,
+      "step": 2457
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.137844611528822e-05,
+      "loss": 0.4647,
+      "step": 2458
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1374865735767994e-05,
+      "loss": 0.5038,
+      "step": 2459
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.137128535624776e-05,
+      "loss": 0.6412,
+      "step": 2460
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.136770497672753e-05,
+      "loss": 0.5624,
+      "step": 2461
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1364124597207307e-05,
+      "loss": 0.5715,
+      "step": 2462
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.136054421768707e-05,
+      "loss": 0.4044,
+      "step": 2463
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1356963838166846e-05,
+      "loss": 0.5386,
+      "step": 2464
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.135338345864662e-05,
+      "loss": 0.4954,
+      "step": 2465
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1349803079126386e-05,
+      "loss": 0.5062,
+      "step": 2466
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.134622269960616e-05,
+      "loss": 0.4768,
+      "step": 2467
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.134264232008593e-05,
+      "loss": 0.519,
+      "step": 2468
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1339061940565705e-05,
+      "loss": 0.5731,
+      "step": 2469
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.133548156104547e-05,
+      "loss": 0.5175,
+      "step": 2470
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1331901181525245e-05,
+      "loss": 0.5512,
+      "step": 2471
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.132832080200502e-05,
+      "loss": 0.465,
+      "step": 2472
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1324740422484784e-05,
+      "loss": 0.6735,
+      "step": 2473
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.132116004296456e-05,
+      "loss": 0.5998,
+      "step": 2474
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.131757966344433e-05,
+      "loss": 0.5154,
+      "step": 2475
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.13139992839241e-05,
+      "loss": 0.5417,
+      "step": 2476
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.131041890440387e-05,
+      "loss": 0.5574,
+      "step": 2477
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.130683852488364e-05,
+      "loss": 0.5108,
+      "step": 2478
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.130325814536341e-05,
+      "loss": 0.6596,
+      "step": 2479
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.129967776584318e-05,
+      "loss": 0.5379,
+      "step": 2480
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.129609738632295e-05,
+      "loss": 0.5656,
+      "step": 2481
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.129251700680272e-05,
+      "loss": 0.4785,
+      "step": 2482
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1288936627282496e-05,
+      "loss": 0.5957,
+      "step": 2483
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.128535624776226e-05,
+      "loss": 0.6194,
+      "step": 2484
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1281775868242036e-05,
+      "loss": 0.5529,
+      "step": 2485
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.12781954887218e-05,
+      "loss": 0.4594,
+      "step": 2486
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1274615109201575e-05,
+      "loss": 0.4822,
+      "step": 2487
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.127103472968135e-05,
+      "loss": 0.605,
+      "step": 2488
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1267454350161115e-05,
+      "loss": 0.5225,
+      "step": 2489
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.126387397064089e-05,
+      "loss": 0.5966,
+      "step": 2490
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.126029359112066e-05,
+      "loss": 0.5817,
+      "step": 2491
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1256713211600434e-05,
+      "loss": 0.5759,
+      "step": 2492
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.125313283208021e-05,
+      "loss": 0.5097,
+      "step": 2493
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1249552452559974e-05,
+      "loss": 0.5724,
+      "step": 2494
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.124597207303975e-05,
+      "loss": 0.4335,
+      "step": 2495
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1242391693519514e-05,
+      "loss": 0.543,
+      "step": 2496
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.123881131399929e-05,
+      "loss": 0.5588,
+      "step": 2497
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.123523093447906e-05,
+      "loss": 0.6751,
+      "step": 2498
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1231650554958826e-05,
+      "loss": 0.5191,
+      "step": 2499
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.12280701754386e-05,
+      "loss": 0.5407,
+      "step": 2500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.122448979591837e-05,
+      "loss": 0.6184,
+      "step": 2501
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.122090941639814e-05,
+      "loss": 0.5132,
+      "step": 2502
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.121732903687791e-05,
+      "loss": 0.5673,
+      "step": 2503
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.121374865735768e-05,
+      "loss": 0.4531,
+      "step": 2504
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.121016827783745e-05,
+      "loss": 0.5522,
+      "step": 2505
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1206587898317225e-05,
+      "loss": 0.5427,
+      "step": 2506
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.120300751879699e-05,
+      "loss": 0.4971,
+      "step": 2507
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1199427139276765e-05,
+      "loss": 0.555,
+      "step": 2508
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.119584675975654e-05,
+      "loss": 0.5698,
+      "step": 2509
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1192266380236304e-05,
+      "loss": 0.4637,
+      "step": 2510
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.118868600071608e-05,
+      "loss": 0.5228,
+      "step": 2511
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1185105621195844e-05,
+      "loss": 0.6098,
+      "step": 2512
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.118152524167562e-05,
+      "loss": 0.6346,
+      "step": 2513
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.117794486215539e-05,
+      "loss": 0.4238,
+      "step": 2514
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1174364482635163e-05,
+      "loss": 0.5675,
+      "step": 2515
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.117078410311494e-05,
+      "loss": 0.649,
+      "step": 2516
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.11672037235947e-05,
+      "loss": 0.4846,
+      "step": 2517
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1163623344074476e-05,
+      "loss": 0.5403,
+      "step": 2518
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.116004296455425e-05,
+      "loss": 0.5296,
+      "step": 2519
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1156462585034016e-05,
+      "loss": 0.5902,
+      "step": 2520
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.115288220551379e-05,
+      "loss": 0.4368,
+      "step": 2521
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1149301825993555e-05,
+      "loss": 0.5584,
+      "step": 2522
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.114572144647333e-05,
+      "loss": 0.4757,
+      "step": 2523
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.11421410669531e-05,
+      "loss": 0.5877,
+      "step": 2524
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.113856068743287e-05,
+      "loss": 0.4914,
+      "step": 2525
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.113498030791264e-05,
+      "loss": 0.5404,
+      "step": 2526
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.113139992839241e-05,
+      "loss": 0.5013,
+      "step": 2527
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.112781954887218e-05,
+      "loss": 0.439,
+      "step": 2528
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1124239169351954e-05,
+      "loss": 0.6282,
+      "step": 2529
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.112065878983172e-05,
+      "loss": 0.4511,
+      "step": 2530
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1117078410311494e-05,
+      "loss": 0.485,
+      "step": 2531
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.111349803079127e-05,
+      "loss": 0.5204,
+      "step": 2532
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1109917651271033e-05,
+      "loss": 0.3987,
+      "step": 2533
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1106337271750807e-05,
+      "loss": 0.4939,
+      "step": 2534
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.110275689223057e-05,
+      "loss": 0.5289,
+      "step": 2535
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1099176512710346e-05,
+      "loss": 0.5365,
+      "step": 2536
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.109559613319012e-05,
+      "loss": 0.5058,
+      "step": 2537
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.109201575366989e-05,
+      "loss": 0.6194,
+      "step": 2538
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1088435374149666e-05,
+      "loss": 0.6293,
+      "step": 2539
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.108485499462943e-05,
+      "loss": 0.5681,
+      "step": 2540
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1081274615109205e-05,
+      "loss": 0.5682,
+      "step": 2541
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.107769423558898e-05,
+      "loss": 0.4808,
+      "step": 2542
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1074113856068745e-05,
+      "loss": 0.6183,
+      "step": 2543
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.107053347654852e-05,
+      "loss": 0.5851,
+      "step": 2544
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1066953097028285e-05,
+      "loss": 0.4849,
+      "step": 2545
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.106337271750806e-05,
+      "loss": 0.5878,
+      "step": 2546
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.105979233798783e-05,
+      "loss": 0.5088,
+      "step": 2547
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.10562119584676e-05,
+      "loss": 0.5229,
+      "step": 2548
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.105263157894737e-05,
+      "loss": 0.5058,
+      "step": 2549
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1049051199427144e-05,
+      "loss": 0.5277,
+      "step": 2550
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.104547081990691e-05,
+      "loss": 0.625,
+      "step": 2551
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.104189044038668e-05,
+      "loss": 0.6354,
+      "step": 2552
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.103831006086645e-05,
+      "loss": 0.475,
+      "step": 2553
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.103472968134622e-05,
+      "loss": 0.6019,
+      "step": 2554
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1031149301825996e-05,
+      "loss": 0.4884,
+      "step": 2555
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.102756892230576e-05,
+      "loss": 0.5939,
+      "step": 2556
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1023988542785536e-05,
+      "loss": 0.5665,
+      "step": 2557
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.102040816326531e-05,
+      "loss": 0.5886,
+      "step": 2558
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1016827783745075e-05,
+      "loss": 0.4418,
+      "step": 2559
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.101324740422485e-05,
+      "loss": 0.5431,
+      "step": 2560
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.100966702470462e-05,
+      "loss": 0.5075,
+      "step": 2561
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1006086645184395e-05,
+      "loss": 0.4522,
+      "step": 2562
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.100250626566416e-05,
+      "loss": 0.3914,
+      "step": 2563
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0998925886143934e-05,
+      "loss": 0.5087,
+      "step": 2564
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.099534550662371e-05,
+      "loss": 0.4733,
+      "step": 2565
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0991765127103474e-05,
+      "loss": 0.6528,
+      "step": 2566
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.098818474758325e-05,
+      "loss": 0.4891,
+      "step": 2567
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.098460436806302e-05,
+      "loss": 0.5211,
+      "step": 2568
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.098102398854279e-05,
+      "loss": 0.5317,
+      "step": 2569
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.097744360902256e-05,
+      "loss": 0.5221,
+      "step": 2570
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0973863229502326e-05,
+      "loss": 0.6378,
+      "step": 2571
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.09702828499821e-05,
+      "loss": 0.5128,
+      "step": 2572
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.096670247046187e-05,
+      "loss": 0.5679,
+      "step": 2573
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.096312209094164e-05,
+      "loss": 0.5375,
+      "step": 2574
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.095954171142141e-05,
+      "loss": 0.6255,
+      "step": 2575
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0955961331901186e-05,
+      "loss": 0.5159,
+      "step": 2576
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.095238095238095e-05,
+      "loss": 0.4957,
+      "step": 2577
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0948800572860725e-05,
+      "loss": 0.5877,
+      "step": 2578
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.094522019334049e-05,
+      "loss": 0.5147,
+      "step": 2579
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0941639813820265e-05,
+      "loss": 0.5506,
+      "step": 2580
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.093805943430004e-05,
+      "loss": 0.5598,
+      "step": 2581
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0934479054779804e-05,
+      "loss": 0.4638,
+      "step": 2582
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.093089867525958e-05,
+      "loss": 0.4809,
+      "step": 2583
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.092731829573935e-05,
+      "loss": 0.589,
+      "step": 2584
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0923737916219124e-05,
+      "loss": 0.4937,
+      "step": 2585
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.09201575366989e-05,
+      "loss": 0.4035,
+      "step": 2586
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0916577157178664e-05,
+      "loss": 0.5323,
+      "step": 2587
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.091299677765844e-05,
+      "loss": 0.5327,
+      "step": 2588
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.09094163981382e-05,
+      "loss": 0.5307,
+      "step": 2589
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0905836018617976e-05,
+      "loss": 0.4069,
+      "step": 2590
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.090225563909775e-05,
+      "loss": 0.573,
+      "step": 2591
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.0898675259577516e-05,
+      "loss": 0.5012,
+      "step": 2592
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.089509488005729e-05,
+      "loss": 0.5479,
+      "step": 2593
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.089151450053706e-05,
+      "loss": 0.4372,
+      "step": 2594
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.088793412101683e-05,
+      "loss": 0.6049,
+      "step": 2595
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.08843537414966e-05,
+      "loss": 0.4684,
+      "step": 2596
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.088077336197637e-05,
+      "loss": 0.5218,
+      "step": 2597
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.087719298245614e-05,
+      "loss": 0.6338,
+      "step": 2598
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0873612602935915e-05,
+      "loss": 0.4834,
+      "step": 2599
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.087003222341568e-05,
+      "loss": 0.7416,
+      "step": 2600
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0866451843895454e-05,
+      "loss": 0.6249,
+      "step": 2601
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.086287146437523e-05,
+      "loss": 0.5054,
+      "step": 2602
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0859291084854994e-05,
+      "loss": 0.4821,
+      "step": 2603
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.085571070533477e-05,
+      "loss": 0.4804,
+      "step": 2604
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0852130325814534e-05,
+      "loss": 0.5939,
+      "step": 2605
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.084854994629431e-05,
+      "loss": 0.514,
+      "step": 2606
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.084496956677408e-05,
+      "loss": 0.5561,
+      "step": 2607
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.084138918725385e-05,
+      "loss": 0.4916,
+      "step": 2608
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0837808807733626e-05,
+      "loss": 0.4342,
+      "step": 2609
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.083422842821339e-05,
+      "loss": 0.4551,
+      "step": 2610
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0830648048693166e-05,
+      "loss": 0.5356,
+      "step": 2611
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.082706766917294e-05,
+      "loss": 0.4579,
+      "step": 2612
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0823487289652705e-05,
+      "loss": 0.515,
+      "step": 2613
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.081990691013248e-05,
+      "loss": 0.5724,
+      "step": 2614
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0816326530612245e-05,
+      "loss": 0.5187,
+      "step": 2615
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.081274615109202e-05,
+      "loss": 0.5579,
+      "step": 2616
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.080916577157179e-05,
+      "loss": 0.5531,
+      "step": 2617
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.080558539205156e-05,
+      "loss": 0.5583,
+      "step": 2618
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.080200501253133e-05,
+      "loss": 0.5475,
+      "step": 2619
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0798424633011104e-05,
+      "loss": 0.5744,
+      "step": 2620
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.079484425349087e-05,
+      "loss": 0.4234,
+      "step": 2621
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0791263873970644e-05,
+      "loss": 0.4558,
+      "step": 2622
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.078768349445041e-05,
+      "loss": 0.603,
+      "step": 2623
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0784103114930183e-05,
+      "loss": 0.5247,
+      "step": 2624
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0780522735409957e-05,
+      "loss": 0.5264,
+      "step": 2625
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.077694235588972e-05,
+      "loss": 0.6069,
+      "step": 2626
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0773361976369496e-05,
+      "loss": 0.5056,
+      "step": 2627
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.076978159684926e-05,
+      "loss": 0.5756,
+      "step": 2628
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0766201217329036e-05,
+      "loss": 0.4936,
+      "step": 2629
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.076262083780881e-05,
+      "loss": 0.6088,
+      "step": 2630
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.075904045828858e-05,
+      "loss": 0.4868,
+      "step": 2631
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0755460078768355e-05,
+      "loss": 0.5276,
+      "step": 2632
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.075187969924812e-05,
+      "loss": 0.5229,
+      "step": 2633
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0748299319727895e-05,
+      "loss": 0.5948,
+      "step": 2634
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.074471894020767e-05,
+      "loss": 0.5507,
+      "step": 2635
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0741138560687435e-05,
+      "loss": 0.5184,
+      "step": 2636
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.073755818116721e-05,
+      "loss": 0.5962,
+      "step": 2637
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0733977801646974e-05,
+      "loss": 0.592,
+      "step": 2638
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.073039742212675e-05,
+      "loss": 0.504,
+      "step": 2639
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.072681704260652e-05,
+      "loss": 0.5334,
+      "step": 2640
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.072323666308629e-05,
+      "loss": 0.5653,
+      "step": 2641
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.071965628356606e-05,
+      "loss": 0.5319,
+      "step": 2642
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.071607590404583e-05,
+      "loss": 0.6091,
+      "step": 2643
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.07124955245256e-05,
+      "loss": 0.554,
+      "step": 2644
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.070891514500537e-05,
+      "loss": 0.545,
+      "step": 2645
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.070533476548514e-05,
+      "loss": 0.5713,
+      "step": 2646
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.070175438596491e-05,
+      "loss": 0.6839,
+      "step": 2647
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0698174006444686e-05,
+      "loss": 0.5031,
+      "step": 2648
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.069459362692445e-05,
+      "loss": 0.4836,
+      "step": 2649
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0691013247404225e-05,
+      "loss": 0.5058,
+      "step": 2650
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0687432867884e-05,
+      "loss": 0.56,
+      "step": 2651
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0683852488363765e-05,
+      "loss": 0.4621,
+      "step": 2652
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.068027210884354e-05,
+      "loss": 0.5036,
+      "step": 2653
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.067669172932331e-05,
+      "loss": 0.4893,
+      "step": 2654
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0673111349803084e-05,
+      "loss": 0.5516,
+      "step": 2655
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.066953097028285e-05,
+      "loss": 0.575,
+      "step": 2656
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0665950590762624e-05,
+      "loss": 0.417,
+      "step": 2657
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.06623702112424e-05,
+      "loss": 0.3888,
+      "step": 2658
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0658789831722164e-05,
+      "loss": 0.5372,
+      "step": 2659
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.065520945220194e-05,
+      "loss": 0.5202,
+      "step": 2660
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.065162907268171e-05,
+      "loss": 0.5037,
+      "step": 2661
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0648048693161476e-05,
+      "loss": 0.5272,
+      "step": 2662
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.064446831364125e-05,
+      "loss": 0.5046,
+      "step": 2663
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0640887934121016e-05,
+      "loss": 0.4577,
+      "step": 2664
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.063730755460079e-05,
+      "loss": 0.5984,
+      "step": 2665
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.063372717508056e-05,
+      "loss": 0.4086,
+      "step": 2666
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.063014679556033e-05,
+      "loss": 0.4369,
+      "step": 2667
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.06265664160401e-05,
+      "loss": 0.5723,
+      "step": 2668
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0622986036519875e-05,
+      "loss": 0.5225,
+      "step": 2669
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.061940565699964e-05,
+      "loss": 0.5087,
+      "step": 2670
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0615825277479415e-05,
+      "loss": 0.4986,
+      "step": 2671
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.061224489795918e-05,
+      "loss": 0.5246,
+      "step": 2672
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0608664518438954e-05,
+      "loss": 0.7454,
+      "step": 2673
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.060508413891873e-05,
+      "loss": 0.5162,
+      "step": 2674
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.0601503759398494e-05,
+      "loss": 0.4386,
+      "step": 2675
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.059792337987827e-05,
+      "loss": 0.5048,
+      "step": 2676
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.059434300035804e-05,
+      "loss": 0.5507,
+      "step": 2677
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0590762620837814e-05,
+      "loss": 0.517,
+      "step": 2678
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.058718224131759e-05,
+      "loss": 0.6341,
+      "step": 2679
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.058360186179735e-05,
+      "loss": 0.4504,
+      "step": 2680
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0580021482277126e-05,
+      "loss": 0.6321,
+      "step": 2681
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.057644110275689e-05,
+      "loss": 0.5665,
+      "step": 2682
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0572860723236666e-05,
+      "loss": 0.5031,
+      "step": 2683
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.056928034371644e-05,
+      "loss": 0.6435,
+      "step": 2684
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0565699964196206e-05,
+      "loss": 0.451,
+      "step": 2685
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.056211958467598e-05,
+      "loss": 0.5544,
+      "step": 2686
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.055853920515575e-05,
+      "loss": 0.5584,
+      "step": 2687
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.055495882563552e-05,
+      "loss": 0.5872,
+      "step": 2688
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.055137844611529e-05,
+      "loss": 0.5198,
+      "step": 2689
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.054779806659506e-05,
+      "loss": 0.5689,
+      "step": 2690
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.054421768707483e-05,
+      "loss": 0.5051,
+      "step": 2691
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0540637307554604e-05,
+      "loss": 0.6671,
+      "step": 2692
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.053705692803437e-05,
+      "loss": 0.5824,
+      "step": 2693
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0533476548514144e-05,
+      "loss": 0.5052,
+      "step": 2694
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.052989616899392e-05,
+      "loss": 0.4918,
+      "step": 2695
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0526315789473684e-05,
+      "loss": 0.5857,
+      "step": 2696
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.052273540995346e-05,
+      "loss": 0.5447,
+      "step": 2697
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.051915503043322e-05,
+      "loss": 0.6089,
+      "step": 2698
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0515574650912996e-05,
+      "loss": 0.5031,
+      "step": 2699
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.051199427139277e-05,
+      "loss": 0.5294,
+      "step": 2700
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.050841389187254e-05,
+      "loss": 0.5118,
+      "step": 2701
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0504833512352316e-05,
+      "loss": 0.4631,
+      "step": 2702
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.050125313283208e-05,
+      "loss": 0.5845,
+      "step": 2703
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.0497672753311855e-05,
+      "loss": 0.6666,
+      "step": 2704
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.049409237379163e-05,
+      "loss": 0.5143,
+      "step": 2705
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0490511994271395e-05,
+      "loss": 0.5966,
+      "step": 2706
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.048693161475117e-05,
+      "loss": 0.5018,
+      "step": 2707
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0483351235230935e-05,
+      "loss": 0.4892,
+      "step": 2708
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.047977085571071e-05,
+      "loss": 0.5877,
+      "step": 2709
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.047619047619048e-05,
+      "loss": 0.5866,
+      "step": 2710
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.047261009667025e-05,
+      "loss": 0.5306,
+      "step": 2711
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.046902971715002e-05,
+      "loss": 0.4555,
+      "step": 2712
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0465449337629794e-05,
+      "loss": 0.482,
+      "step": 2713
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.046186895810956e-05,
+      "loss": 0.5948,
+      "step": 2714
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0458288578589333e-05,
+      "loss": 0.5513,
+      "step": 2715
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.04547081990691e-05,
+      "loss": 0.5044,
+      "step": 2716
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.045112781954887e-05,
+      "loss": 0.4561,
+      "step": 2717
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0447547440028646e-05,
+      "loss": 0.5431,
+      "step": 2718
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.044396706050841e-05,
+      "loss": 0.5417,
+      "step": 2719
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0440386680988186e-05,
+      "loss": 0.4798,
+      "step": 2720
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.043680630146795e-05,
+      "loss": 0.4288,
+      "step": 2721
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0433225921947725e-05,
+      "loss": 0.4856,
+      "step": 2722
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.04296455424275e-05,
+      "loss": 0.546,
+      "step": 2723
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.042606516290727e-05,
+      "loss": 0.6363,
+      "step": 2724
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0422484783387045e-05,
+      "loss": 0.5515,
+      "step": 2725
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.041890440386681e-05,
+      "loss": 0.4827,
+      "step": 2726
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0415324024346585e-05,
+      "loss": 0.6097,
+      "step": 2727
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.041174364482636e-05,
+      "loss": 0.566,
+      "step": 2728
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0408163265306124e-05,
+      "loss": 0.5457,
+      "step": 2729
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.04045828857859e-05,
+      "loss": 0.5566,
+      "step": 2730
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.040100250626567e-05,
+      "loss": 0.5033,
+      "step": 2731
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.039742212674544e-05,
+      "loss": 0.6272,
+      "step": 2732
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.039384174722521e-05,
+      "loss": 0.4215,
+      "step": 2733
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0390261367704977e-05,
+      "loss": 0.4577,
+      "step": 2734
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.038668098818475e-05,
+      "loss": 0.5127,
+      "step": 2735
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.038310060866452e-05,
+      "loss": 0.6176,
+      "step": 2736
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.037952022914429e-05,
+      "loss": 0.488,
+      "step": 2737
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.037593984962406e-05,
+      "loss": 0.5688,
+      "step": 2738
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.037235947010383e-05,
+      "loss": 0.4817,
+      "step": 2739
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.03687790905836e-05,
+      "loss": 0.4249,
+      "step": 2740
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0365198711063375e-05,
+      "loss": 0.5686,
+      "step": 2741
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.036161833154314e-05,
+      "loss": 0.5857,
+      "step": 2742
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0358037952022915e-05,
+      "loss": 0.5107,
+      "step": 2743
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.035445757250269e-05,
+      "loss": 0.6573,
+      "step": 2744
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0350877192982455e-05,
+      "loss": 0.5547,
+      "step": 2745
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.034729681346223e-05,
+      "loss": 0.5256,
+      "step": 2746
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0343716433941994e-05,
+      "loss": 0.4825,
+      "step": 2747
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0340136054421774e-05,
+      "loss": 0.6603,
+      "step": 2748
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.033655567490154e-05,
+      "loss": 0.5491,
+      "step": 2749
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0332975295381314e-05,
+      "loss": 0.5546,
+      "step": 2750
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.032939491586109e-05,
+      "loss": 0.4856,
+      "step": 2751
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.032581453634085e-05,
+      "loss": 0.5333,
+      "step": 2752
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0322234156820626e-05,
+      "loss": 0.5669,
+      "step": 2753
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.03186537773004e-05,
+      "loss": 0.5927,
+      "step": 2754
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0315073397780166e-05,
+      "loss": 0.6163,
+      "step": 2755
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.031149301825994e-05,
+      "loss": 0.451,
+      "step": 2756
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.0307912638739706e-05,
+      "loss": 0.5079,
+      "step": 2757
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.030433225921948e-05,
+      "loss": 0.4808,
+      "step": 2758
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.030075187969925e-05,
+      "loss": 0.5342,
+      "step": 2759
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.029717150017902e-05,
+      "loss": 0.4905,
+      "step": 2760
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.029359112065879e-05,
+      "loss": 0.4659,
+      "step": 2761
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0290010741138565e-05,
+      "loss": 0.4424,
+      "step": 2762
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.028643036161833e-05,
+      "loss": 0.5626,
+      "step": 2763
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0282849982098104e-05,
+      "loss": 0.562,
+      "step": 2764
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.027926960257787e-05,
+      "loss": 0.4962,
+      "step": 2765
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0275689223057644e-05,
+      "loss": 0.5267,
+      "step": 2766
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.027210884353742e-05,
+      "loss": 0.4529,
+      "step": 2767
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0268528464017184e-05,
+      "loss": 0.5205,
+      "step": 2768
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.026494808449696e-05,
+      "loss": 0.3947,
+      "step": 2769
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.026136770497673e-05,
+      "loss": 0.5006,
+      "step": 2770
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.02577873254565e-05,
+      "loss": 0.4882,
+      "step": 2771
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0254206945936276e-05,
+      "loss": 0.3546,
+      "step": 2772
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.025062656641604e-05,
+      "loss": 0.5532,
+      "step": 2773
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0247046186895816e-05,
+      "loss": 0.4751,
+      "step": 2774
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.024346580737558e-05,
+      "loss": 0.3688,
+      "step": 2775
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0239885427855356e-05,
+      "loss": 0.432,
+      "step": 2776
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.023630504833513e-05,
+      "loss": 0.5745,
+      "step": 2777
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0232724668814895e-05,
+      "loss": 0.4726,
+      "step": 2778
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.022914428929467e-05,
+      "loss": 0.5436,
+      "step": 2779
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.022556390977444e-05,
+      "loss": 0.5466,
+      "step": 2780
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.022198353025421e-05,
+      "loss": 0.4763,
+      "step": 2781
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.021840315073398e-05,
+      "loss": 0.5238,
+      "step": 2782
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.021482277121375e-05,
+      "loss": 0.5015,
+      "step": 2783
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.021124239169352e-05,
+      "loss": 0.6661,
+      "step": 2784
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0207662012173294e-05,
+      "loss": 0.5508,
+      "step": 2785
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.020408163265306e-05,
+      "loss": 0.4218,
+      "step": 2786
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.0200501253132834e-05,
+      "loss": 0.6453,
+      "step": 2787
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.019692087361261e-05,
+      "loss": 0.5053,
+      "step": 2788
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.019334049409237e-05,
+      "loss": 0.6231,
+      "step": 2789
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0189760114572146e-05,
+      "loss": 0.4891,
+      "step": 2790
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.018617973505191e-05,
+      "loss": 0.4705,
+      "step": 2791
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0182599355531686e-05,
+      "loss": 0.4452,
+      "step": 2792
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.017901897601146e-05,
+      "loss": 0.5777,
+      "step": 2793
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.017543859649123e-05,
+      "loss": 0.6075,
+      "step": 2794
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0171858216971005e-05,
+      "loss": 0.5102,
+      "step": 2795
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.016827783745077e-05,
+      "loss": 0.5495,
+      "step": 2796
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0164697457930545e-05,
+      "loss": 0.5328,
+      "step": 2797
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.016111707841032e-05,
+      "loss": 0.5199,
+      "step": 2798
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0157536698890085e-05,
+      "loss": 0.4757,
+      "step": 2799
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.015395631936986e-05,
+      "loss": 0.6732,
+      "step": 2800
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.0150375939849624e-05,
+      "loss": 0.6048,
+      "step": 2801
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.01467955603294e-05,
+      "loss": 0.4561,
+      "step": 2802
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.014321518080917e-05,
+      "loss": 0.6172,
+      "step": 2803
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.7440963318095662,
+      "eval_loss": 0.5148146748542786,
+      "eval_runtime": 25.8168,
+      "eval_samples_per_second": 1737.045,
+      "eval_steps_per_second": 27.153,
+      "step": 2803
+    }
+  ],
+  "max_steps": 14015,
+  "num_train_epochs": 5,
+  "total_flos": 1488494762635392.0,
+  "trial_name": null,
+  "trial_params": null
+}