diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,6451 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "global_step": 1071,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.3255813953488374e-07,
+      "loss": 0.6927,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.651162790697675e-07,
+      "loss": 0.7448,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.976744186046513e-07,
+      "loss": 0.601,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.30232558139535e-07,
+      "loss": 0.608,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 1.1627906976744188e-06,
+      "loss": 0.6313,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3953488372093025e-06,
+      "loss": 0.6369,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.6279069767441862e-06,
+      "loss": 0.5791,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.86046511627907e-06,
+      "loss": 0.6322,
+      "step": 8
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.0930232558139536e-06,
+      "loss": 0.6634,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.3255813953488376e-06,
+      "loss": 0.5336,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.558139534883721e-06,
+      "loss": 0.6026,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.790697674418605e-06,
+      "loss": 0.5601,
+      "step": 12
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.0232558139534885e-06,
+      "loss": 0.5419,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.2558139534883724e-06,
+      "loss": 0.5552,
+      "step": 14
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.4883720930232564e-06,
+      "loss": 0.6033,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 3.72093023255814e-06,
+      "loss": 0.5317,
+      "step": 16
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.953488372093024e-06,
+      "loss": 0.6444,
+      "step": 17
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.186046511627907e-06,
+      "loss": 0.5289,
+      "step": 18
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.418604651162791e-06,
+      "loss": 0.5654,
+      "step": 19
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.651162790697675e-06,
+      "loss": 0.4864,
+      "step": 20
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.883720930232559e-06,
+      "loss": 0.512,
+      "step": 21
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.116279069767442e-06,
+      "loss": 0.5533,
+      "step": 22
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 5.348837209302326e-06,
+      "loss": 0.4778,
+      "step": 23
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.58139534883721e-06,
+      "loss": 0.5771,
+      "step": 24
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 5.8139534883720935e-06,
+      "loss": 0.5388,
+      "step": 25
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 6.046511627906977e-06,
+      "loss": 0.4788,
+      "step": 26
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 6.279069767441861e-06,
+      "loss": 0.5733,
+      "step": 27
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 6.511627906976745e-06,
+      "loss": 0.499,
+      "step": 28
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 6.744186046511628e-06,
+      "loss": 0.5316,
+      "step": 29
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 6.976744186046513e-06,
+      "loss": 0.5203,
+      "step": 30
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 7.209302325581395e-06,
+      "loss": 0.5197,
+      "step": 31
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 7.44186046511628e-06,
+      "loss": 0.5622,
+      "step": 32
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 7.674418604651164e-06,
+      "loss": 0.4875,
+      "step": 33
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 7.906976744186048e-06,
+      "loss": 0.5775,
+      "step": 34
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 8.139534883720931e-06,
+      "loss": 0.5145,
+      "step": 35
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 8.372093023255815e-06,
+      "loss": 0.4491,
+      "step": 36
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 8.604651162790698e-06,
+      "loss": 0.3912,
+      "step": 37
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 8.837209302325582e-06,
+      "loss": 0.4365,
+      "step": 38
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.069767441860465e-06,
+      "loss": 0.4763,
+      "step": 39
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.30232558139535e-06,
+      "loss": 0.5241,
+      "step": 40
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 9.534883720930234e-06,
+      "loss": 0.5922,
+      "step": 41
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.767441860465117e-06,
+      "loss": 0.5252,
+      "step": 42
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1e-05,
+      "loss": 0.5626,
+      "step": 43
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 9.999976651811724e-06,
+      "loss": 0.5012,
+      "step": 44
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.99990660746495e-06,
+      "loss": 0.5336,
+      "step": 45
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.99978986761384e-06,
+      "loss": 0.5074,
+      "step": 46
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.999626433348664e-06,
+      "loss": 0.4718,
+      "step": 47
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 9.999416306195775e-06,
+      "loss": 0.4969,
+      "step": 48
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.99915948811761e-06,
+      "loss": 0.4878,
+      "step": 49
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.998855981512665e-06,
+      "loss": 0.552,
+      "step": 50
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.998505789215469e-06,
+      "loss": 0.4912,
+      "step": 51
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.998108914496567e-06,
+      "loss": 0.4372,
+      "step": 52
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.99766536106248e-06,
+      "loss": 0.4957,
+      "step": 53
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.997175133055676e-06,
+      "loss": 0.5149,
+      "step": 54
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.996638235054527e-06,
+      "loss": 0.4562,
+      "step": 55
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.996054672073276e-06,
+      "loss": 0.5329,
+      "step": 56
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.995424449561974e-06,
+      "loss": 0.4609,
+      "step": 57
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 9.994747573406444e-06,
+      "loss": 0.5044,
+      "step": 58
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.994024049928222e-06,
+      "loss": 0.5051,
+      "step": 59
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.993253885884488e-06,
+      "loss": 0.4964,
+      "step": 60
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.992437088468016e-06,
+      "loss": 0.4759,
+      "step": 61
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 9.991573665307108e-06,
+      "loss": 0.4139,
+      "step": 62
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.990663624465504e-06,
+      "loss": 0.4862,
+      "step": 63
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.989706974442329e-06,
+      "loss": 0.4812,
+      "step": 64
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.988703724172e-06,
+      "loss": 0.4264,
+      "step": 65
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 9.987653883024147e-06,
+      "loss": 0.4633,
+      "step": 66
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.986557460803527e-06,
+      "loss": 0.5227,
+      "step": 67
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.98541446774993e-06,
+      "loss": 0.5213,
+      "step": 68
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 9.98422491453808e-06,
+      "loss": 0.5079,
+      "step": 69
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.982988812277544e-06,
+      "loss": 0.4164,
+      "step": 70
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.98170617251262e-06,
+      "loss": 0.5647,
+      "step": 71
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.980377007222236e-06,
+      "loss": 0.4137,
+      "step": 72
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 9.979001328819828e-06,
+      "loss": 0.5746,
+      "step": 73
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.97757915015324e-06,
+      "loss": 0.4513,
+      "step": 74
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.976110484504587e-06,
+      "loss": 0.4904,
+      "step": 75
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 9.974595345590146e-06,
+      "loss": 0.428,
+      "step": 76
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.97303374756021e-06,
+      "loss": 0.4855,
+      "step": 77
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.971425704998979e-06,
+      "loss": 0.5235,
+      "step": 78
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.969771232924404e-06,
+      "loss": 0.4995,
+      "step": 79
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 9.968070346788052e-06,
+      "loss": 0.4523,
+      "step": 80
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.96632306247497e-06,
+      "loss": 0.498,
+      "step": 81
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.964529396303524e-06,
+      "loss": 0.4923,
+      "step": 82
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 9.962689365025259e-06,
+      "loss": 0.4472,
+      "step": 83
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.960802985824734e-06,
+      "loss": 0.4925,
+      "step": 84
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.958870276319364e-06,
+      "loss": 0.4267,
+      "step": 85
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.95689125455925e-06,
+      "loss": 0.412,
+      "step": 86
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 9.954865939027028e-06,
+      "loss": 0.5546,
+      "step": 87
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.952794348637674e-06,
+      "loss": 0.4652,
+      "step": 88
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.95067650273834e-06,
+      "loss": 0.4338,
+      "step": 89
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.948512421108175e-06,
+      "loss": 0.4527,
+      "step": 90
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 9.94630212395813e-06,
+      "loss": 0.435,
+      "step": 91
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.944045631930782e-06,
+      "loss": 0.4558,
+      "step": 92
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.941742966100128e-06,
+      "loss": 0.4389,
+      "step": 93
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 9.939394147971398e-06,
+      "loss": 0.4451,
+      "step": 94
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.936999199480854e-06,
+      "loss": 0.4665,
+      "step": 95
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.934558142995577e-06,
+      "loss": 0.4368,
+      "step": 96
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.932071001313265e-06,
+      "loss": 0.4095,
+      "step": 97
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 9.929537797662022e-06,
+      "loss": 0.484,
+      "step": 98
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.926958555700134e-06,
+      "loss": 0.4774,
+      "step": 99
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.924333299515849e-06,
+      "loss": 0.3817,
+      "step": 100
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 9.92166205362716e-06,
+      "loss": 0.5114,
+      "step": 101
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.918944842981564e-06,
+      "loss": 0.4381,
+      "step": 102
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.916181692955841e-06,
+      "loss": 0.4958,
+      "step": 103
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.913372629355814e-06,
+      "loss": 0.4138,
+      "step": 104
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 9.910517678416097e-06,
+      "loss": 0.3692,
+      "step": 105
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.907616866799862e-06,
+      "loss": 0.4465,
+      "step": 106
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.90467022159859e-06,
+      "loss": 0.4548,
+      "step": 107
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 9.90167777033181e-06,
+      "loss": 0.5527,
+      "step": 108
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.89863954094685e-06,
+      "loss": 0.4042,
+      "step": 109
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.89555556181857e-06,
+      "loss": 0.4588,
+      "step": 110
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.8924258617491e-06,
+      "loss": 0.4638,
+      "step": 111
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 9.88925046996757e-06,
+      "loss": 0.4651,
+      "step": 112
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.886029416129837e-06,
+      "loss": 0.5044,
+      "step": 113
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.882762730318211e-06,
+      "loss": 0.5116,
+      "step": 114
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.879450443041172e-06,
+      "loss": 0.4827,
+      "step": 115
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 9.876092585233081e-06,
+      "loss": 0.4601,
+      "step": 116
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.872689188253895e-06,
+      "loss": 0.4559,
+      "step": 117
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.869240283888879e-06,
+      "loss": 0.4334,
+      "step": 118
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 9.865745904348296e-06,
+      "loss": 0.4826,
+      "step": 119
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.862206082267123e-06,
+      "loss": 0.4246,
+      "step": 120
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.85862085070473e-06,
+      "loss": 0.4679,
+      "step": 121
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.854990243144583e-06,
+      "loss": 0.4472,
+      "step": 122
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 9.851314293493923e-06,
+      "loss": 0.486,
+      "step": 123
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.847593036083457e-06,
+      "loss": 0.4428,
+      "step": 124
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.843826505667038e-06,
+      "loss": 0.4553,
+      "step": 125
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 9.840014737421321e-06,
+      "loss": 0.4421,
+      "step": 126
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.836157766945467e-06,
+      "loss": 0.4958,
+      "step": 127
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.832255630260781e-06,
+      "loss": 0.386,
+      "step": 128
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.828308363810392e-06,
+      "loss": 0.4448,
+      "step": 129
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 9.82431600445891e-06,
+      "loss": 0.4408,
+      "step": 130
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.820278589492076e-06,
+      "loss": 0.4476,
+      "step": 131
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.816196156616422e-06,
+      "loss": 0.4903,
+      "step": 132
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 9.812068743958912e-06,
+      "loss": 0.4279,
+      "step": 133
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.807896390066587e-06,
+      "loss": 0.4443,
+      "step": 134
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.80367913390621e-06,
+      "loss": 0.4137,
+      "step": 135
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.7994170148639e-06,
+      "loss": 0.4149,
+      "step": 136
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 9.795110072744756e-06,
+      "loss": 0.408,
+      "step": 137
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.790758347772498e-06,
+      "loss": 0.5097,
+      "step": 138
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.786361880589084e-06,
+      "loss": 0.3729,
+      "step": 139
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.781920712254332e-06,
+      "loss": 0.4598,
+      "step": 140
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 9.777434884245533e-06,
+      "loss": 0.4264,
+      "step": 141
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.772904438457071e-06,
+      "loss": 0.478,
+      "step": 142
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.768329417200029e-06,
+      "loss": 0.485,
+      "step": 143
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 9.763709863201789e-06,
+      "loss": 0.388,
+      "step": 144
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.759045819605635e-06,
+      "loss": 0.5051,
+      "step": 145
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.754337329970355e-06,
+      "loss": 0.4529,
+      "step": 146
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.749584438269833e-06,
+      "loss": 0.4182,
+      "step": 147
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 9.74478718889263e-06,
+      "loss": 0.4498,
+      "step": 148
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.739945626641579e-06,
+      "loss": 0.4771,
+      "step": 149
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.735059796733363e-06,
+      "loss": 0.4334,
+      "step": 150
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 9.730129744798096e-06,
+      "loss": 0.4073,
+      "step": 151
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.725155516878886e-06,
+      "loss": 0.4624,
+      "step": 152
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.720137159431418e-06,
+      "loss": 0.4276,
+      "step": 153
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.715074719323515e-06,
+      "loss": 0.4601,
+      "step": 154
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 9.709968243834698e-06,
+      "loss": 0.3997,
+      "step": 155
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.704817780655746e-06,
+      "loss": 0.4088,
+      "step": 156
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.699623377888256e-06,
+      "loss": 0.4274,
+      "step": 157
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 9.694385084044185e-06,
+      "loss": 0.438,
+      "step": 158
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.689102948045398e-06,
+      "loss": 0.467,
+      "step": 159
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.683777019223219e-06,
+      "loss": 0.483,
+      "step": 160
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.678407347317967e-06,
+      "loss": 0.3738,
+      "step": 161
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 9.67299398247848e-06,
+      "loss": 0.4568,
+      "step": 162
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.667536975261667e-06,
+      "loss": 0.4092,
+      "step": 163
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.66203637663202e-06,
+      "loss": 0.3938,
+      "step": 164
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.656492237961143e-06,
+      "loss": 0.3977,
+      "step": 165
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 9.650904611027275e-06,
+      "loss": 0.4375,
+      "step": 166
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.6452735480148e-06,
+      "loss": 0.4416,
+      "step": 167
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.639599101513769e-06,
+      "loss": 0.3981,
+      "step": 168
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 9.633881324519397e-06,
+      "loss": 0.4188,
+      "step": 169
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.628120270431579e-06,
+      "loss": 0.3862,
+      "step": 170
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.622315993054384e-06,
+      "loss": 0.4423,
+      "step": 171
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.616468546595556e-06,
+      "loss": 0.4882,
+      "step": 172
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 9.61057798566601e-06,
+      "loss": 0.5166,
+      "step": 173
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.604644365279316e-06,
+      "loss": 0.3763,
+      "step": 174
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.598667740851187e-06,
+      "loss": 0.4491,
+      "step": 175
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 9.592648168198961e-06,
+      "loss": 0.454,
+      "step": 176
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.586585703541092e-06,
+      "loss": 0.47,
+      "step": 177
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.580480403496599e-06,
+      "loss": 0.4112,
+      "step": 178
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.574332325084564e-06,
+      "loss": 0.4314,
+      "step": 179
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 9.568141525723582e-06,
+      "loss": 0.421,
+      "step": 180
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.561908063231234e-06,
+      "loss": 0.4147,
+      "step": 181
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.555631995823543e-06,
+      "loss": 0.4774,
+      "step": 182
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 9.549313382114427e-06,
+      "loss": 0.4661,
+      "step": 183
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.542952281115163e-06,
+      "loss": 0.4121,
+      "step": 184
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.536548752233822e-06,
+      "loss": 0.3729,
+      "step": 185
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.530102855274723e-06,
+      "loss": 0.4313,
+      "step": 186
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.523614650437876e-06,
+      "loss": 0.4061,
+      "step": 187
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.517084198318408e-06,
+      "loss": 0.4221,
+      "step": 188
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.51051155990601e-06,
+      "loss": 0.3687,
+      "step": 189
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.503896796584363e-06,
+      "loss": 0.3516,
+      "step": 190
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.497239970130561e-06,
+      "loss": 0.4554,
+      "step": 191
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.490541142714542e-06,
+      "loss": 0.4348,
+      "step": 192
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.483800376898496e-06,
+      "loss": 0.4288,
+      "step": 193
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.477017735636294e-06,
+      "loss": 0.4179,
+      "step": 194
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.470193282272886e-06,
+      "loss": 0.4147,
+      "step": 195
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.463327080543726e-06,
+      "loss": 0.4064,
+      "step": 196
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.456419194574158e-06,
+      "loss": 0.4676,
+      "step": 197
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.449469688878833e-06,
+      "loss": 0.4436,
+      "step": 198
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.442478628361098e-06,
+      "loss": 0.4625,
+      "step": 199
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.435446078312389e-06,
+      "loss": 0.4579,
+      "step": 200
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 9.428372104411632e-06,
+      "loss": 0.4074,
+      "step": 201
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.421256772724612e-06,
+      "loss": 0.4269,
+      "step": 202
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.414100149703373e-06,
+      "loss": 0.3968,
+      "step": 203
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.406902302185587e-06,
+      "loss": 0.4562,
+      "step": 204
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 9.399663297393937e-06,
+      "loss": 0.4622,
+      "step": 205
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.392383202935476e-06,
+      "loss": 0.4091,
+      "step": 206
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.385062086801013e-06,
+      "loss": 0.4276,
+      "step": 207
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 9.377700017364469e-06,
+      "loss": 0.3799,
+      "step": 208
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.370297063382235e-06,
+      "loss": 0.3638,
+      "step": 209
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.362853293992535e-06,
+      "loss": 0.4232,
+      "step": 210
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.355368778714784e-06,
+      "loss": 0.4238,
+      "step": 211
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 9.347843587448931e-06,
+      "loss": 0.458,
+      "step": 212
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.340277790474804e-06,
+      "loss": 0.4491,
+      "step": 213
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.33267145845147e-06,
+      "loss": 0.3917,
+      "step": 214
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 9.325024662416553e-06,
+      "loss": 0.4247,
+      "step": 215
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.31733747378559e-06,
+      "loss": 0.3902,
+      "step": 216
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.30960996435135e-06,
+      "loss": 0.4447,
+      "step": 217
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.301842206283173e-06,
+      "loss": 0.3886,
+      "step": 218
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 9.294034272126286e-06,
+      "loss": 0.4156,
+      "step": 219
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.286186234801142e-06,
+      "loss": 0.4028,
+      "step": 220
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.278298167602716e-06,
+      "loss": 0.4145,
+      "step": 221
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.270370144199843e-06,
+      "loss": 0.3846,
+      "step": 222
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 9.262402238634514e-06,
+      "loss": 0.5301,
+      "step": 223
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.254394525321195e-06,
+      "loss": 0.4042,
+      "step": 224
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.246347079046124e-06,
+      "loss": 0.366,
+      "step": 225
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 9.238259974966617e-06,
+      "loss": 0.4007,
+      "step": 226
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.230133288610366e-06,
+      "loss": 0.3255,
+      "step": 227
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.221967095874733e-06,
+      "loss": 0.389,
+      "step": 228
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.213761473026039e-06,
+      "loss": 0.3687,
+      "step": 229
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 9.205516496698854e-06,
+      "loss": 0.4824,
+      "step": 230
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 9.197232243895285e-06,
+      "loss": 0.4489,
+      "step": 231
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 9.188908791984245e-06,
+      "loss": 0.3804,
+      "step": 232
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 9.180546218700748e-06,
+      "loss": 0.3578,
+      "step": 233
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.172144602145165e-06,
+      "loss": 0.3939,
+      "step": 234
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.163704020782507e-06,
+      "loss": 0.392,
+      "step": 235
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.155224553441686e-06,
+      "loss": 0.4111,
+      "step": 236
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 9.146706279314786e-06,
+      "loss": 0.4145,
+      "step": 237
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.138149277956307e-06,
+      "loss": 0.4002,
+      "step": 238
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.129553629282448e-06,
+      "loss": 0.4292,
+      "step": 239
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 9.120919413570335e-06,
+      "loss": 0.3822,
+      "step": 240
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.112246711457284e-06,
+      "loss": 0.4147,
+      "step": 241
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.10353560394005e-06,
+      "loss": 0.3618,
+      "step": 242
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.094786172374066e-06,
+      "loss": 0.42,
+      "step": 243
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 9.08599849847268e-06,
+      "loss": 0.4743,
+      "step": 244
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.077172664306396e-06,
+      "loss": 0.4833,
+      "step": 245
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.068308752302113e-06,
+      "loss": 0.4025,
+      "step": 246
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.059406845242343e-06,
+      "loss": 0.4791,
+      "step": 247
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 9.050467026264448e-06,
+      "loss": 0.4281,
+      "step": 248
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.041489378859856e-06,
+      "loss": 0.4768,
+      "step": 249
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.03247398687329e-06,
+      "loss": 0.4043,
+      "step": 250
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 9.023420934501981e-06,
+      "loss": 0.415,
+      "step": 251
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 9.014330306294872e-06,
+      "loss": 0.426,
+      "step": 252
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 9.005202187151845e-06,
+      "loss": 0.3954,
+      "step": 253
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.996036662322917e-06,
+      "loss": 0.3713,
+      "step": 254
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 8.98683381740745e-06,
+      "loss": 0.3486,
+      "step": 255
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.977593738353346e-06,
+      "loss": 0.4363,
+      "step": 256
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.968316511456241e-06,
+      "loss": 0.4373,
+      "step": 257
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 8.95900222335872e-06,
+      "loss": 0.4468,
+      "step": 258
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.949650961049479e-06,
+      "loss": 0.431,
+      "step": 259
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.94026281186253e-06,
+      "loss": 0.4037,
+      "step": 260
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.930837863476386e-06,
+      "loss": 0.374,
+      "step": 261
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 8.921376203913235e-06,
+      "loss": 0.3683,
+      "step": 262
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.911877921538117e-06,
+      "loss": 0.3875,
+      "step": 263
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.902343105058108e-06,
+      "loss": 0.4048,
+      "step": 264
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 8.892771843521487e-06,
+      "loss": 0.3917,
+      "step": 265
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.883164226316897e-06,
+      "loss": 0.4277,
+      "step": 266
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.87352034317252e-06,
+      "loss": 0.3859,
+      "step": 267
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.863840284155238e-06,
+      "loss": 0.3653,
+      "step": 268
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 8.854124139669786e-06,
+      "loss": 0.3957,
+      "step": 269
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.844372000457912e-06,
+      "loss": 0.3671,
+      "step": 270
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.83458395759753e-06,
+      "loss": 0.457,
+      "step": 271
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.824760102501865e-06,
+      "loss": 0.3626,
+      "step": 272
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 8.814900526918608e-06,
+      "loss": 0.4397,
+      "step": 273
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.805005322929046e-06,
+      "loss": 0.431,
+      "step": 274
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.795074582947214e-06,
+      "loss": 0.4255,
+      "step": 275
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 8.785108399719029e-06,
+      "loss": 0.4257,
+      "step": 276
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.775106866321419e-06,
+      "loss": 0.4257,
+      "step": 277
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.765070076161457e-06,
+      "loss": 0.4857,
+      "step": 278
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.754998122975489e-06,
+      "loss": 0.4125,
+      "step": 279
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 8.744891100828261e-06,
+      "loss": 0.3956,
+      "step": 280
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.734749104112032e-06,
+      "loss": 0.3936,
+      "step": 281
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.724572227545707e-06,
+      "loss": 0.4113,
+      "step": 282
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 8.714360566173932e-06,
+      "loss": 0.4365,
+      "step": 283
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.704114215366228e-06,
+      "loss": 0.4065,
+      "step": 284
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.693833270816083e-06,
+      "loss": 0.3675,
+      "step": 285
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.683517828540074e-06,
+      "loss": 0.4144,
+      "step": 286
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 8.67316798487695e-06,
+      "loss": 0.3411,
+      "step": 287
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 8.662783836486751e-06,
+      "loss": 0.4594,
+      "step": 288
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 8.652365480349904e-06,
+      "loss": 0.3681,
+      "step": 289
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 8.641913013766301e-06,
+      "loss": 0.409,
+      "step": 290
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 8.631426534354404e-06,
+      "loss": 0.4461,
+      "step": 291
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 8.620906140050332e-06,
+      "loss": 0.4531,
+      "step": 292
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 8.610351929106944e-06,
+      "loss": 0.3565,
+      "step": 293
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 8.599764000092921e-06,
+      "loss": 0.3538,
+      "step": 294
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.589142451891849e-06,
+      "loss": 0.4282,
+      "step": 295
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.57848738370129e-06,
+      "loss": 0.3847,
+      "step": 296
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.56779889503186e-06,
+      "loss": 0.4043,
+      "step": 297
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 8.557077085706295e-06,
+      "loss": 0.3414,
+      "step": 298
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.546322055858526e-06,
+      "loss": 0.4686,
+      "step": 299
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.535533905932739e-06,
+      "loss": 0.3921,
+      "step": 300
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 8.524712736682433e-06,
+      "loss": 0.3492,
+      "step": 301
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.513858649169488e-06,
+      "loss": 0.366,
+      "step": 302
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.502971744763216e-06,
+      "loss": 0.4565,
+      "step": 303
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.492052125139414e-06,
+      "loss": 0.3662,
+      "step": 304
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 8.481099892279418e-06,
+      "loss": 0.3683,
+      "step": 305
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 8.470115148469143e-06,
+      "loss": 0.3907,
+      "step": 306
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 8.459097996298137e-06,
+      "loss": 0.3461,
+      "step": 307
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 8.448048538658618e-06,
+      "loss": 0.3795,
+      "step": 308
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.43696687874451e-06,
+      "loss": 0.3642,
+      "step": 309
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.42585312005049e-06,
+      "loss": 0.3721,
+      "step": 310
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.414707366371006e-06,
+      "loss": 0.3658,
+      "step": 311
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.403529721799325e-06,
+      "loss": 0.3317,
+      "step": 312
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.392320290726543e-06,
+      "loss": 0.4261,
+      "step": 313
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.381079177840625e-06,
+      "loss": 0.3434,
+      "step": 314
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.369806488125418e-06,
+      "loss": 0.3783,
+      "step": 315
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.358502326859674e-06,
+      "loss": 0.3742,
+      "step": 316
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.347166799616069e-06,
+      "loss": 0.4285,
+      "step": 317
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.335800012260211e-06,
+      "loss": 0.4039,
+      "step": 318
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 8.324402070949658e-06,
+      "loss": 0.4683,
+      "step": 319
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.312973082132922e-06,
+      "loss": 0.3564,
+      "step": 320
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.301513152548474e-06,
+      "loss": 0.3781,
+      "step": 321
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.290022389223754e-06,
+      "loss": 0.3251,
+      "step": 322
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 8.278500899474162e-06,
+      "loss": 0.4771,
+      "step": 323
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 8.266948790902064e-06,
+      "loss": 0.3764,
+      "step": 324
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 8.255366171395783e-06,
+      "loss": 0.3995,
+      "step": 325
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 8.243753149128589e-06,
+      "loss": 0.3285,
+      "step": 326
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 8.232109832557696e-06,
+      "loss": 0.3995,
+      "step": 327
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 8.220436330423243e-06,
+      "loss": 0.3563,
+      "step": 328
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 8.208732751747281e-06,
+      "loss": 0.4447,
+      "step": 329
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 8.196999205832752e-06,
+      "loss": 0.429,
+      "step": 330
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.18523580226247e-06,
+      "loss": 0.3909,
+      "step": 331
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.173442650898103e-06,
+      "loss": 0.3766,
+      "step": 332
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 8.161619861879136e-06,
+      "loss": 0.4008,
+      "step": 333
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 8.149767545621852e-06,
+      "loss": 0.4851,
+      "step": 334
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 8.137885812818296e-06,
+      "loss": 0.3957,
+      "step": 335
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 8.12597477443524e-06,
+      "loss": 0.4399,
+      "step": 336
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 8.114034541713152e-06,
+      "loss": 0.4273,
+      "step": 337
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 8.10206522616515e-06,
+      "loss": 0.38,
+      "step": 338
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 8.090066939575972e-06,
+      "loss": 0.395,
+      "step": 339
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 8.078039794000915e-06,
+      "loss": 0.4267,
+      "step": 340
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.065983901764807e-06,
+      "loss": 0.4473,
+      "step": 341
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.05389937546094e-06,
+      "loss": 0.3366,
+      "step": 342
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.041786327950037e-06,
+      "loss": 0.3739,
+      "step": 343
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.029644872359182e-06,
+      "loss": 0.4598,
+      "step": 344
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 8.017475122080767e-06,
+      "loss": 0.3657,
+      "step": 345
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 8.005277190771447e-06,
+      "loss": 0.4515,
+      "step": 346
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.993051192351056e-06,
+      "loss": 0.3794,
+      "step": 347
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 7.980797241001563e-06,
+      "loss": 0.3958,
+      "step": 348
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.96851545116599e-06,
+      "loss": 0.3708,
+      "step": 349
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.956205937547354e-06,
+      "loss": 0.3679,
+      "step": 350
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 7.943868815107594e-06,
+      "loss": 0.3969,
+      "step": 351
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.931504199066491e-06,
+      "loss": 0.3543,
+      "step": 352
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.919112204900597e-06,
+      "loss": 0.3819,
+      "step": 353
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.90669294834216e-06,
+      "loss": 0.3106,
+      "step": 354
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.894246545378037e-06,
+      "loss": 0.4005,
+      "step": 355
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.881773112248607e-06,
+      "loss": 0.3905,
+      "step": 356
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.869272765446701e-06,
+      "loss": 0.3117,
+      "step": 357
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.856745621716495e-06,
+      "loss": 0.2894,
+      "step": 358
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 7.844191798052438e-06,
+      "loss": 0.3168,
+      "step": 359
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 7.831611411698141e-06,
+      "loss": 0.294,
+      "step": 360
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 7.819004580145298e-06,
+      "loss": 0.2639,
+      "step": 361
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 7.80637142113258e-06,
+      "loss": 0.2518,
+      "step": 362
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 7.793712052644535e-06,
+      "loss": 0.3016,
+      "step": 363
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 7.781026592910493e-06,
+      "loss": 0.2549,
+      "step": 364
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 7.768315160403453e-06,
+      "loss": 0.2546,
+      "step": 365
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 7.755577873838985e-06,
+      "loss": 0.2394,
+      "step": 366
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 7.742814852174112e-06,
+      "loss": 0.2611,
+      "step": 367
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 7.730026214606207e-06,
+      "loss": 0.2451,
+      "step": 368
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 7.71721208057188e-06,
+      "loss": 0.2605,
+      "step": 369
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 7.704372569745857e-06,
+      "loss": 0.2518,
+      "step": 370
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 7.691507802039861e-06,
+      "loss": 0.2908,
+      "step": 371
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 7.678617897601501e-06,
+      "loss": 0.2571,
+      "step": 372
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 7.665702976813142e-06,
+      "loss": 0.2615,
+      "step": 373
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 7.65276316029079e-06,
+      "loss": 0.2451,
+      "step": 374
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 7.639798568882947e-06,
+      "loss": 0.3141,
+      "step": 375
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 7.626809323669506e-06,
+      "loss": 0.2424,
+      "step": 376
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 7.613795545960602e-06,
+      "loss": 0.2282,
+      "step": 377
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 7.60075735729549e-06,
+      "loss": 0.3148,
+      "step": 378
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 7.5876948794414015e-06,
+      "loss": 0.2744,
+      "step": 379
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 7.5746082343924146e-06,
+      "loss": 0.2612,
+      "step": 380
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 7.561497544368309e-06,
+      "loss": 0.2645,
+      "step": 381
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 7.5483629318134285e-06,
+      "loss": 0.265,
+      "step": 382
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 7.535204519395538e-06,
+      "loss": 0.2622,
+      "step": 383
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 7.522022430004672e-06,
+      "loss": 0.2753,
+      "step": 384
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 7.508816786751991e-06,
+      "loss": 0.2604,
+      "step": 385
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 7.495587712968637e-06,
+      "loss": 0.2381,
+      "step": 386
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 7.482335332204568e-06,
+      "loss": 0.2531,
+      "step": 387
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 7.469059768227419e-06,
+      "loss": 0.2846,
+      "step": 388
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 7.455761145021335e-06,
+      "loss": 0.2687,
+      "step": 389
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 7.4424395867858224e-06,
+      "loss": 0.2542,
+      "step": 390
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 7.429095217934578e-06,
+      "loss": 0.2327,
+      "step": 391
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 7.415728163094338e-06,
+      "loss": 0.2615,
+      "step": 392
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 7.402338547103708e-06,
+      "loss": 0.2715,
+      "step": 393
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 7.388926495011996e-06,
+      "loss": 0.2489,
+      "step": 394
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 7.375492132078051e-06,
+      "loss": 0.2909,
+      "step": 395
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 7.362035583769087e-06,
+      "loss": 0.2511,
+      "step": 396
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 7.348556975759512e-06,
+      "loss": 0.2984,
+      "step": 397
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 7.335056433929758e-06,
+      "loss": 0.2572,
+      "step": 398
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 7.321534084365101e-06,
+      "loss": 0.2762,
+      "step": 399
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 7.307990053354489e-06,
+      "loss": 0.2408,
+      "step": 400
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 7.294424467389354e-06,
+      "loss": 0.2531,
+      "step": 401
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 7.280837453162437e-06,
+      "loss": 0.2636,
+      "step": 402
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 7.267229137566607e-06,
+      "loss": 0.2067,
+      "step": 403
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 7.2535996476936696e-06,
+      "loss": 0.2423,
+      "step": 404
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 7.239949110833182e-06,
+      "loss": 0.2319,
+      "step": 405
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 7.2262776544712665e-06,
+      "loss": 0.234,
+      "step": 406
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 7.2125854062894184e-06,
+      "loss": 0.244,
+      "step": 407
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 7.198872494163312e-06,
+      "loss": 0.2745,
+      "step": 408
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.185139046161611e-06,
+      "loss": 0.2872,
+      "step": 409
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.171385190544766e-06,
+      "loss": 0.2451,
+      "step": 410
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.15761105576382e-06,
+      "loss": 0.2475,
+      "step": 411
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 7.143816770459211e-06,
+      "loss": 0.2838,
+      "step": 412
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.130002463459569e-06,
+      "loss": 0.3024,
+      "step": 413
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.1161682637805065e-06,
+      "loss": 0.2334,
+      "step": 414
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 7.102314300623425e-06,
+      "loss": 0.2765,
+      "step": 415
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.088440703374302e-06,
+      "loss": 0.2582,
+      "step": 416
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.074547601602479e-06,
+      "loss": 0.2281,
+      "step": 417
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.060635125059461e-06,
+      "loss": 0.2682,
+      "step": 418
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 7.0467034036776945e-06,
+      "loss": 0.29,
+      "step": 419
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.032752567569362e-06,
+      "loss": 0.3024,
+      "step": 420
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.018782747025161e-06,
+      "loss": 0.2621,
+      "step": 421
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 7.004794072513096e-06,
+      "loss": 0.2229,
+      "step": 422
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 6.990786674677246e-06,
+      "loss": 0.2826,
+      "step": 423
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 6.976760684336556e-06,
+      "loss": 0.3023,
+      "step": 424
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 6.962716232483612e-06,
+      "loss": 0.2715,
+      "step": 425
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 6.948653450283416e-06,
+      "loss": 0.2765,
+      "step": 426
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 6.934572469072163e-06,
+      "loss": 0.2701,
+      "step": 427
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 6.920473420356013e-06,
+      "loss": 0.2763,
+      "step": 428
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 6.9063564358098636e-06,
+      "loss": 0.2193,
+      "step": 429
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 6.89222164727612e-06,
+      "loss": 0.2581,
+      "step": 430
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 6.878069186763466e-06,
+      "loss": 0.2432,
+      "step": 431
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 6.8638991864456205e-06,
+      "loss": 0.2703,
+      "step": 432
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 6.849711778660124e-06,
+      "loss": 0.2911,
+      "step": 433
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.835507095907082e-06,
+      "loss": 0.2592,
+      "step": 434
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.821285270847934e-06,
+      "loss": 0.2342,
+      "step": 435
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.807046436304224e-06,
+      "loss": 0.2737,
+      "step": 436
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 6.792790725256347e-06,
+      "loss": 0.2817,
+      "step": 437
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.77851827084231e-06,
+      "loss": 0.2721,
+      "step": 438
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.764229206356498e-06,
+      "loss": 0.2629,
+      "step": 439
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 6.749923665248419e-06,
+      "loss": 0.2522,
+      "step": 440
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.735601781121454e-06,
+      "loss": 0.2518,
+      "step": 441
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.7212636877316285e-06,
+      "loss": 0.2774,
+      "step": 442
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.706909518986341e-06,
+      "loss": 0.3236,
+      "step": 443
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 6.692539408943124e-06,
+      "loss": 0.2238,
+      "step": 444
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.678153491808394e-06,
+      "loss": 0.3037,
+      "step": 445
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.6637519019361895e-06,
+      "loss": 0.3043,
+      "step": 446
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.649334773826924e-06,
+      "loss": 0.2897,
+      "step": 447
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 6.6349022421261275e-06,
+      "loss": 0.2578,
+      "step": 448
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.6204544416231865e-06,
+      "loss": 0.2662,
+      "step": 449
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.6059915072500845e-06,
+      "loss": 0.2668,
+      "step": 450
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 6.591513574080152e-06,
+      "loss": 0.2883,
+      "step": 451
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.577020777326789e-06,
+      "loss": 0.2021,
+      "step": 452
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.562513252342216e-06,
+      "loss": 0.2378,
+      "step": 453
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.547991134616204e-06,
+      "loss": 0.2307,
+      "step": 454
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 6.5334545597748075e-06,
+      "loss": 0.2687,
+      "step": 455
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.5189036635791e-06,
+      "loss": 0.2618,
+      "step": 456
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.5043385819239095e-06,
+      "loss": 0.2482,
+      "step": 457
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 6.489759450836541e-06,
+      "loss": 0.2757,
+      "step": 458
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 6.475166406475515e-06,
+      "loss": 0.2672,
+      "step": 459
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 6.460559585129289e-06,
+      "loss": 0.2782,
+      "step": 460
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 6.445939123214991e-06,
+      "loss": 0.2237,
+      "step": 461
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 6.431305157277139e-06,
+      "loss": 0.2645,
+      "step": 462
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 6.41665782398637e-06,
+      "loss": 0.2601,
+      "step": 463
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 6.40199726013816e-06,
+      "loss": 0.2728,
+      "step": 464
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 6.387323602651554e-06,
+      "loss": 0.2214,
+      "step": 465
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 6.3726369885678785e-06,
+      "loss": 0.2588,
+      "step": 466
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 6.357937555049465e-06,
+      "loss": 0.3038,
+      "step": 467
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 6.34322543937837e-06,
+      "loss": 0.2508,
+      "step": 468
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 6.328500778955091e-06,
+      "loss": 0.2704,
+      "step": 469
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 6.31376371129729e-06,
+      "loss": 0.2302,
+      "step": 470
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 6.299014374038493e-06,
+      "loss": 0.2266,
+      "step": 471
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 6.284252904926826e-06,
+      "loss": 0.3359,
+      "step": 472
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 6.269479441823712e-06,
+      "loss": 0.2453,
+      "step": 473
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 6.254694122702589e-06,
+      "loss": 0.2817,
+      "step": 474
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 6.239897085647624e-06,
+      "loss": 0.3185,
+      "step": 475
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 6.225088468852418e-06,
+      "loss": 0.2302,
+      "step": 476
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 6.210268410618723e-06,
+      "loss": 0.2464,
+      "step": 477
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 6.1954370493551415e-06,
+      "loss": 0.257,
+      "step": 478
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 6.180594523575838e-06,
+      "loss": 0.2808,
+      "step": 479
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 6.16574097189925e-06,
+      "loss": 0.2954,
+      "step": 480
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 6.150876533046784e-06,
+      "loss": 0.2723,
+      "step": 481
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 6.1360013458415276e-06,
+      "loss": 0.2782,
+      "step": 482
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 6.12111554920695e-06,
+      "loss": 0.2818,
+      "step": 483
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 6.106219282165603e-06,
+      "loss": 0.3094,
+      "step": 484
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 6.091312683837823e-06,
+      "loss": 0.2577,
+      "step": 485
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 6.076395893440442e-06,
+      "loss": 0.2759,
+      "step": 486
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 6.061469050285469e-06,
+      "loss": 0.2581,
+      "step": 487
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 6.0465322937788e-06,
+      "loss": 0.2663,
+      "step": 488
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 6.031585763418919e-06,
+      "loss": 0.2408,
+      "step": 489
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 6.016629598795587e-06,
+      "loss": 0.263,
+      "step": 490
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 6.0016639395885424e-06,
+      "loss": 0.3163,
+      "step": 491
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 5.986688925566198e-06,
+      "loss": 0.2749,
+      "step": 492
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 5.971704696584332e-06,
+      "loss": 0.2943,
+      "step": 493
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 5.956711392584782e-06,
+      "loss": 0.2533,
+      "step": 494
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 5.941709153594146e-06,
+      "loss": 0.2503,
+      "step": 495
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 5.9266981197224615e-06,
+      "loss": 0.2416,
+      "step": 496
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 5.911678431161907e-06,
+      "loss": 0.2947,
+      "step": 497
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 5.8966502281854885e-06,
+      "loss": 0.2824,
+      "step": 498
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 5.881613651145732e-06,
+      "loss": 0.2362,
+      "step": 499
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 5.86656884047337e-06,
+      "loss": 0.284,
+      "step": 500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 5.851515936676031e-06,
+      "loss": 0.2404,
+      "step": 501
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 5.836455080336929e-06,
+      "loss": 0.2494,
+      "step": 502
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 5.821386412113546e-06,
+      "loss": 0.244,
+      "step": 503
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 5.806310072736323e-06,
+      "loss": 0.2121,
+      "step": 504
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 5.791226203007346e-06,
+      "loss": 0.2356,
+      "step": 505
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 5.7761349437990255e-06,
+      "loss": 0.2525,
+      "step": 506
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 5.761036436052788e-06,
+      "loss": 0.2631,
+      "step": 507
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 5.745930820777753e-06,
+      "loss": 0.2787,
+      "step": 508
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 5.7308182390494185e-06,
+      "loss": 0.2859,
+      "step": 509
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 5.7156988320083485e-06,
+      "loss": 0.302,
+      "step": 510
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 5.700572740858847e-06,
+      "loss": 0.2325,
+      "step": 511
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 5.685440106867642e-06,
+      "loss": 0.2593,
+      "step": 512
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 5.6703010713625715e-06,
+      "loss": 0.284,
+      "step": 513
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 5.6551557757312536e-06,
+      "loss": 0.3164,
+      "step": 514
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 5.640004361419776e-06,
+      "loss": 0.2627,
+      "step": 515
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.6248469699313664e-06,
+      "loss": 0.2453,
+      "step": 516
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.609683742825078e-06,
+      "loss": 0.2666,
+      "step": 517
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.594514821714462e-06,
+      "loss": 0.2285,
+      "step": 518
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 5.579340348266251e-06,
+      "loss": 0.232,
+      "step": 519
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.564160464199029e-06,
+      "loss": 0.2661,
+      "step": 520
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.548975311281911e-06,
+      "loss": 0.2903,
+      "step": 521
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.533785031333224e-06,
+      "loss": 0.212,
+      "step": 522
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 5.518589766219173e-06,
+      "loss": 0.2831,
+      "step": 523
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.503389657852519e-06,
+      "loss": 0.2586,
+      "step": 524
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.488184848191265e-06,
+      "loss": 0.2623,
+      "step": 525
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 5.4729754792373094e-06,
+      "loss": 0.2781,
+      "step": 526
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.457761693035139e-06,
+      "loss": 0.2455,
+      "step": 527
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.4425436316704905e-06,
+      "loss": 0.233,
+      "step": 528
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.427321437269027e-06,
+      "loss": 0.2874,
+      "step": 529
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 5.412095251995017e-06,
+      "loss": 0.2762,
+      "step": 530
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 5.396865218049995e-06,
+      "loss": 0.2117,
+      "step": 531
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 5.381631477671439e-06,
+      "loss": 0.2019,
+      "step": 532
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 5.366394173131445e-06,
+      "loss": 0.2418,
+      "step": 533
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 5.351153446735398e-06,
+      "loss": 0.292,
+      "step": 534
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 5.335909440820635e-06,
+      "loss": 0.2047,
+      "step": 535
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 5.320662297755123e-06,
+      "loss": 0.2762,
+      "step": 536
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 5.305412159936133e-06,
+      "loss": 0.2257,
+      "step": 537
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 5.290159169788895e-06,
+      "loss": 0.2712,
+      "step": 538
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 5.27490346976529e-06,
+      "loss": 0.2751,
+      "step": 539
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 5.259645202342496e-06,
+      "loss": 0.2121,
+      "step": 540
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 5.244384510021673e-06,
+      "loss": 0.2269,
+      "step": 541
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 5.2291215353266315e-06,
+      "loss": 0.2338,
+      "step": 542
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 5.21385642080249e-06,
+      "loss": 0.2839,
+      "step": 543
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 5.198589309014358e-06,
+      "loss": 0.2608,
+      "step": 544
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 5.183320342545995e-06,
+      "loss": 0.2805,
+      "step": 545
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 5.168049663998485e-06,
+      "loss": 0.2716,
+      "step": 546
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 5.152777415988894e-06,
+      "loss": 0.2469,
+      "step": 547
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 5.137503741148957e-06,
+      "loss": 0.2477,
+      "step": 548
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 5.122228782123723e-06,
+      "loss": 0.255,
+      "step": 549
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 5.106952681570242e-06,
+      "loss": 0.2523,
+      "step": 550
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 5.091675582156224e-06,
+      "loss": 0.2484,
+      "step": 551
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 5.076397626558704e-06,
+      "loss": 0.2129,
+      "step": 552
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 5.061118957462716e-06,
+      "loss": 0.278,
+      "step": 553
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 5.045839717559958e-06,
+      "loss": 0.2238,
+      "step": 554
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 5.0305600495474586e-06,
+      "loss": 0.2572,
+      "step": 555
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 5.015280096126242e-06,
+      "loss": 0.2564,
+      "step": 556
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 5e-06,
+      "loss": 0.2454,
+      "step": 557
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.98471990387376e-06,
+      "loss": 0.2637,
+      "step": 558
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.969439950452543e-06,
+      "loss": 0.2383,
+      "step": 559
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.954160282440043e-06,
+      "loss": 0.3294,
+      "step": 560
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.938881042537286e-06,
+      "loss": 0.2517,
+      "step": 561
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.923602373441297e-06,
+      "loss": 0.1915,
+      "step": 562
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.908324417843779e-06,
+      "loss": 0.2495,
+      "step": 563
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.89304731842976e-06,
+      "loss": 0.2759,
+      "step": 564
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.877771217876279e-06,
+      "loss": 0.224,
+      "step": 565
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.8624962588510456e-06,
+      "loss": 0.2644,
+      "step": 566
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.847222584011107e-06,
+      "loss": 0.2451,
+      "step": 567
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.831950336001518e-06,
+      "loss": 0.2647,
+      "step": 568
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.8166796574540065e-06,
+      "loss": 0.251,
+      "step": 569
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.801410690985643e-06,
+      "loss": 0.3039,
+      "step": 570
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.7861435791975124e-06,
+      "loss": 0.2954,
+      "step": 571
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.770878464673372e-06,
+      "loss": 0.2354,
+      "step": 572
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.755615489978328e-06,
+      "loss": 0.2619,
+      "step": 573
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.740354797657504e-06,
+      "loss": 0.1938,
+      "step": 574
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.72509653023471e-06,
+      "loss": 0.2161,
+      "step": 575
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.709840830211105e-06,
+      "loss": 0.2381,
+      "step": 576
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.694587840063868e-06,
+      "loss": 0.2734,
+      "step": 577
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.679337702244877e-06,
+      "loss": 0.2595,
+      "step": 578
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.664090559179367e-06,
+      "loss": 0.2435,
+      "step": 579
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.648846553264603e-06,
+      "loss": 0.281,
+      "step": 580
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.633605826868556e-06,
+      "loss": 0.2366,
+      "step": 581
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.618368522328563e-06,
+      "loss": 0.2656,
+      "step": 582
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.603134781950007e-06,
+      "loss": 0.2768,
+      "step": 583
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.587904748004984e-06,
+      "loss": 0.2246,
+      "step": 584
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.5726785627309736e-06,
+      "loss": 0.2524,
+      "step": 585
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.557456368329511e-06,
+      "loss": 0.274,
+      "step": 586
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.542238306964863e-06,
+      "loss": 0.2062,
+      "step": 587
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.527024520762693e-06,
+      "loss": 0.2812,
+      "step": 588
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.511815151808737e-06,
+      "loss": 0.1954,
+      "step": 589
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.496610342147482e-06,
+      "loss": 0.2743,
+      "step": 590
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.48141023378083e-06,
+      "loss": 0.2815,
+      "step": 591
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.466214968666777e-06,
+      "loss": 0.2554,
+      "step": 592
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.45102468871809e-06,
+      "loss": 0.2094,
+      "step": 593
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.435839535800973e-06,
+      "loss": 0.3104,
+      "step": 594
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.420659651733751e-06,
+      "loss": 0.2264,
+      "step": 595
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.40548517828554e-06,
+      "loss": 0.2753,
+      "step": 596
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.3903162571749234e-06,
+      "loss": 0.2285,
+      "step": 597
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.375153030068635e-06,
+      "loss": 0.2325,
+      "step": 598
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.359995638580226e-06,
+      "loss": 0.2275,
+      "step": 599
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.344844224268747e-06,
+      "loss": 0.2193,
+      "step": 600
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.32969892863743e-06,
+      "loss": 0.2567,
+      "step": 601
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.31455989313236e-06,
+      "loss": 0.2786,
+      "step": 602
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.299427259141155e-06,
+      "loss": 0.2325,
+      "step": 603
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.284301167991654e-06,
+      "loss": 0.3322,
+      "step": 604
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.269181760950584e-06,
+      "loss": 0.2769,
+      "step": 605
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.2540691792222485e-06,
+      "loss": 0.2523,
+      "step": 606
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.238963563947212e-06,
+      "loss": 0.2584,
+      "step": 607
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.2238650562009744e-06,
+      "loss": 0.3678,
+      "step": 608
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.2087737969926545e-06,
+      "loss": 0.2174,
+      "step": 609
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.193689927263677e-06,
+      "loss": 0.226,
+      "step": 610
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.178613587886455e-06,
+      "loss": 0.2764,
+      "step": 611
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.163544919663073e-06,
+      "loss": 0.2585,
+      "step": 612
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.148484063323969e-06,
+      "loss": 0.2126,
+      "step": 613
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.133431159526631e-06,
+      "loss": 0.2236,
+      "step": 614
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.1183863488542686e-06,
+      "loss": 0.2242,
+      "step": 615
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.103349771814512e-06,
+      "loss": 0.1996,
+      "step": 616
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.088321568838095e-06,
+      "loss": 0.2272,
+      "step": 617
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.07330188027754e-06,
+      "loss": 0.223,
+      "step": 618
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.058290846405856e-06,
+      "loss": 0.2371,
+      "step": 619
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.043288607415219e-06,
+      "loss": 0.2571,
+      "step": 620
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.02829530341567e-06,
+      "loss": 0.2786,
+      "step": 621
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.013311074433804e-06,
+      "loss": 0.2341,
+      "step": 622
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.998336060411459e-06,
+      "loss": 0.2189,
+      "step": 623
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.983370401204415e-06,
+      "loss": 0.2961,
+      "step": 624
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.968414236581083e-06,
+      "loss": 0.2283,
+      "step": 625
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 3.953467706221202e-06,
+      "loss": 0.2592,
+      "step": 626
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.938530949714533e-06,
+      "loss": 0.2874,
+      "step": 627
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.9236041065595596e-06,
+      "loss": 0.2264,
+      "step": 628
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.908687316162178e-06,
+      "loss": 0.2199,
+      "step": 629
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 3.8937807178344004e-06,
+      "loss": 0.2441,
+      "step": 630
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.878884450793053e-06,
+      "loss": 0.2059,
+      "step": 631
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.863998654158473e-06,
+      "loss": 0.2372,
+      "step": 632
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 3.849123466953217e-06,
+      "loss": 0.3118,
+      "step": 633
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.834259028100753e-06,
+      "loss": 0.2343,
+      "step": 634
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.819405476424164e-06,
+      "loss": 0.2908,
+      "step": 635
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.804562950644861e-06,
+      "loss": 0.3156,
+      "step": 636
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 3.7897315893812796e-06,
+      "loss": 0.2261,
+      "step": 637
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.774911531147582e-06,
+      "loss": 0.264,
+      "step": 638
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.7601029143523767e-06,
+      "loss": 0.2392,
+      "step": 639
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 3.7453058772974115e-06,
+      "loss": 0.2633,
+      "step": 640
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.7305205581762895e-06,
+      "loss": 0.2444,
+      "step": 641
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.715747095073173e-06,
+      "loss": 0.2699,
+      "step": 642
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.7009856259615074e-06,
+      "loss": 0.2313,
+      "step": 643
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 3.686236288702712e-06,
+      "loss": 0.2365,
+      "step": 644
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.6714992210449084e-06,
+      "loss": 0.2798,
+      "step": 645
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.656774560621632e-06,
+      "loss": 0.2253,
+      "step": 646
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 3.642062444950537e-06,
+      "loss": 0.2886,
+      "step": 647
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.6273630114321223e-06,
+      "loss": 0.2505,
+      "step": 648
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.612676397348447e-06,
+      "loss": 0.2618,
+      "step": 649
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.598002739861841e-06,
+      "loss": 0.2485,
+      "step": 650
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 3.5833421760136323e-06,
+      "loss": 0.2087,
+      "step": 651
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.568694842722863e-06,
+      "loss": 0.2245,
+      "step": 652
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.5540608767850106e-06,
+      "loss": 0.2801,
+      "step": 653
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.539440414870712e-06,
+      "loss": 0.2802,
+      "step": 654
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 3.524833593524487e-06,
+      "loss": 0.2236,
+      "step": 655
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.51024054916346e-06,
+      "loss": 0.2599,
+      "step": 656
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.4956614180760918e-06,
+      "loss": 0.2399,
+      "step": 657
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 3.481096336420901e-06,
+      "loss": 0.2618,
+      "step": 658
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.4665454402251937e-06,
+      "loss": 0.2919,
+      "step": 659
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.452008865383797e-06,
+      "loss": 0.2191,
+      "step": 660
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.437486747657785e-06,
+      "loss": 0.2416,
+      "step": 661
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 3.4229792226732124e-06,
+      "loss": 0.2637,
+      "step": 662
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.40848642591985e-06,
+      "loss": 0.2641,
+      "step": 663
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.394008492749917e-06,
+      "loss": 0.2298,
+      "step": 664
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 3.379545558376816e-06,
+      "loss": 0.2111,
+      "step": 665
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.365097757873874e-06,
+      "loss": 0.3268,
+      "step": 666
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.350665226173078e-06,
+      "loss": 0.2197,
+      "step": 667
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.336248098063812e-06,
+      "loss": 0.2612,
+      "step": 668
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 3.321846508191609e-06,
+      "loss": 0.2248,
+      "step": 669
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.307460591056877e-06,
+      "loss": 0.2619,
+      "step": 670
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.293090481013661e-06,
+      "loss": 0.2122,
+      "step": 671
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 3.2787363122683714e-06,
+      "loss": 0.2391,
+      "step": 672
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.2643982188785457e-06,
+      "loss": 0.2486,
+      "step": 673
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.250076334751583e-06,
+      "loss": 0.2526,
+      "step": 674
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.2357707936435013e-06,
+      "loss": 0.186,
+      "step": 675
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 3.2214817291576905e-06,
+      "loss": 0.1957,
+      "step": 676
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.2072092747436546e-06,
+      "loss": 0.1839,
+      "step": 677
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.1929535636957774e-06,
+      "loss": 0.1933,
+      "step": 678
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.1787147291520675e-06,
+      "loss": 0.2199,
+      "step": 679
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 3.164492904092921e-06,
+      "loss": 0.2066,
+      "step": 680
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1502882213398776e-06,
+      "loss": 0.2042,
+      "step": 681
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.136100813554381e-06,
+      "loss": 0.2614,
+      "step": 682
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 3.1219308132365365e-06,
+      "loss": 0.2325,
+      "step": 683
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.1077783527238807e-06,
+      "loss": 0.2499,
+      "step": 684
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.093643564190138e-06,
+      "loss": 0.2326,
+      "step": 685
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.0795265796439876e-06,
+      "loss": 0.2546,
+      "step": 686
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 3.0654275309278382e-06,
+      "loss": 0.2792,
+      "step": 687
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.051346549716586e-06,
+      "loss": 0.2725,
+      "step": 688
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.037283767516389e-06,
+      "loss": 0.2501,
+      "step": 689
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 3.0232393156634455e-06,
+      "loss": 0.225,
+      "step": 690
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 3.0092133253227563e-06,
+      "loss": 0.2383,
+      "step": 691
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.995205927486905e-06,
+      "loss": 0.2467,
+      "step": 692
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.9812172529748395e-06,
+      "loss": 0.2534,
+      "step": 693
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 2.9672474324306407e-06,
+      "loss": 0.2734,
+      "step": 694
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.9532965963223076e-06,
+      "loss": 0.2553,
+      "step": 695
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.939364874940541e-06,
+      "loss": 0.2368,
+      "step": 696
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 2.9254523983975224e-06,
+      "loss": 0.2672,
+      "step": 697
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.911559296625701e-06,
+      "loss": 0.2391,
+      "step": 698
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.8976856993765766e-06,
+      "loss": 0.2363,
+      "step": 699
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.883831736219495e-06,
+      "loss": 0.2076,
+      "step": 700
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 2.869997536540435e-06,
+      "loss": 0.2395,
+      "step": 701
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.856183229540789e-06,
+      "loss": 0.2397,
+      "step": 702
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.8423889442361797e-06,
+      "loss": 0.2057,
+      "step": 703
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.828614809455236e-06,
+      "loss": 0.1873,
+      "step": 704
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 2.814860953838389e-06,
+      "loss": 0.2376,
+      "step": 705
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.801127505836687e-06,
+      "loss": 0.234,
+      "step": 706
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.787414593710583e-06,
+      "loss": 0.2346,
+      "step": 707
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 2.7737223455287343e-06,
+      "loss": 0.2419,
+      "step": 708
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.760050889166818e-06,
+      "loss": 0.228,
+      "step": 709
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.746400352306332e-06,
+      "loss": 0.2374,
+      "step": 710
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7327708624333936e-06,
+      "loss": 0.2586,
+      "step": 711
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 2.7191625468375636e-06,
+      "loss": 0.2267,
+      "step": 712
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.705575532610649e-06,
+      "loss": 0.2739,
+      "step": 713
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.692009946645513e-06,
+      "loss": 0.1752,
+      "step": 714
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 2.678465915634899e-06,
+      "loss": 0.1387,
+      "step": 715
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.664943566070244e-06,
+      "loss": 0.1797,
+      "step": 716
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.651443024240489e-06,
+      "loss": 0.151,
+      "step": 717
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.6379644162309135e-06,
+      "loss": 0.1917,
+      "step": 718
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 2.6245078679219503e-06,
+      "loss": 0.1413,
+      "step": 719
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.6110735049880054e-06,
+      "loss": 0.1693,
+      "step": 720
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.597661452896293e-06,
+      "loss": 0.107,
+      "step": 721
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 2.584271836905664e-06,
+      "loss": 0.1442,
+      "step": 722
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.5709047820654236e-06,
+      "loss": 0.1225,
+      "step": 723
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.5575604132141792e-06,
+      "loss": 0.1552,
+      "step": 724
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.5442388549786668e-06,
+      "loss": 0.1425,
+      "step": 725
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 2.530940231772584e-06,
+      "loss": 0.1351,
+      "step": 726
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.517664667795434e-06,
+      "loss": 0.1431,
+      "step": 727
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.5044122870313647e-06,
+      "loss": 0.1248,
+      "step": 728
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.4911832132480108e-06,
+      "loss": 0.1155,
+      "step": 729
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 2.4779775699953303e-06,
+      "loss": 0.1698,
+      "step": 730
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.4647954806044633e-06,
+      "loss": 0.1427,
+      "step": 731
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.451637068186573e-06,
+      "loss": 0.1656,
+      "step": 732
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 2.4385024556316928e-06,
+      "loss": 0.1356,
+      "step": 733
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.425391765607587e-06,
+      "loss": 0.1408,
+      "step": 734
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.412305120558599e-06,
+      "loss": 0.1315,
+      "step": 735
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.3992426427045096e-06,
+      "loss": 0.1521,
+      "step": 736
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 2.386204454039398e-06,
+      "loss": 0.1641,
+      "step": 737
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.373190676330494e-06,
+      "loss": 0.1409,
+      "step": 738
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.3602014311170524e-06,
+      "loss": 0.1427,
+      "step": 739
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 2.3472368397092123e-06,
+      "loss": 0.1567,
+      "step": 740
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.3342970231868573e-06,
+      "loss": 0.1283,
+      "step": 741
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.3213821023984994e-06,
+      "loss": 0.1673,
+      "step": 742
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.308492197960141e-06,
+      "loss": 0.1459,
+      "step": 743
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 2.295627430254145e-06,
+      "loss": 0.1657,
+      "step": 744
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.2827879194281196e-06,
+      "loss": 0.1323,
+      "step": 745
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.269973785393794e-06,
+      "loss": 0.1235,
+      "step": 746
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 2.2571851478258903e-06,
+      "loss": 0.1235,
+      "step": 747
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.244422126161017e-06,
+      "loss": 0.1325,
+      "step": 748
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.2316848395965483e-06,
+      "loss": 0.1384,
+      "step": 749
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.2189734070895086e-06,
+      "loss": 0.1498,
+      "step": 750
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 2.2062879473554654e-06,
+      "loss": 0.1254,
+      "step": 751
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.1936285788674204e-06,
+      "loss": 0.1221,
+      "step": 752
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.180995419854703e-06,
+      "loss": 0.1189,
+      "step": 753
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.1683885883018596e-06,
+      "loss": 0.1279,
+      "step": 754
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 2.155808201947563e-06,
+      "loss": 0.1566,
+      "step": 755
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.1432543782835064e-06,
+      "loss": 0.1535,
+      "step": 756
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.130727234553301e-06,
+      "loss": 0.1288,
+      "step": 757
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 2.118226887751394e-06,
+      "loss": 0.1603,
+      "step": 758
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.105753454621966e-06,
+      "loss": 0.133,
+      "step": 759
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.0933070516578407e-06,
+      "loss": 0.1515,
+      "step": 760
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.0808877950994037e-06,
+      "loss": 0.1649,
+      "step": 761
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 2.0684958009335122e-06,
+      "loss": 0.161,
+      "step": 762
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.0561311848924082e-06,
+      "loss": 0.1665,
+      "step": 763
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.043794062452647e-06,
+      "loss": 0.1321,
+      "step": 764
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 2.031484548834013e-06,
+      "loss": 0.1564,
+      "step": 765
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.0192027589984377e-06,
+      "loss": 0.1422,
+      "step": 766
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 2.0069488076489445e-06,
+      "loss": 0.1456,
+      "step": 767
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.994722809228554e-06,
+      "loss": 0.1304,
+      "step": 768
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.9825248779192323e-06,
+      "loss": 0.107,
+      "step": 769
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.9703551276408204e-06,
+      "loss": 0.1288,
+      "step": 770
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.958213672049964e-06,
+      "loss": 0.1495,
+      "step": 771
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.9461006245390594e-06,
+      "loss": 0.1407,
+      "step": 772
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.9340160982351937e-06,
+      "loss": 0.1619,
+      "step": 773
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.9219602059990855e-06,
+      "loss": 0.1826,
+      "step": 774
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.909933060424029e-06,
+      "loss": 0.1298,
+      "step": 775
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.8979347738348498e-06,
+      "loss": 0.1767,
+      "step": 776
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.8859654582868508e-06,
+      "loss": 0.1339,
+      "step": 777
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.8740252255647616e-06,
+      "loss": 0.1297,
+      "step": 778
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.862114187181705e-06,
+      "loss": 0.1571,
+      "step": 779
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.850232454378149e-06,
+      "loss": 0.1231,
+      "step": 780
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8383801381208644e-06,
+      "loss": 0.1522,
+      "step": 781
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8265573491018978e-06,
+      "loss": 0.1482,
+      "step": 782
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8147641977375313e-06,
+      "loss": 0.1037,
+      "step": 783
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.80300079416725e-06,
+      "loss": 0.1613,
+      "step": 784
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.79126724825272e-06,
+      "loss": 0.1474,
+      "step": 785
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.7795636695767582e-06,
+      "loss": 0.143,
+      "step": 786
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.7678901674423044e-06,
+      "loss": 0.1987,
+      "step": 787
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.7562468508714115e-06,
+      "loss": 0.1546,
+      "step": 788
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.7446338286042196e-06,
+      "loss": 0.1317,
+      "step": 789
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.7330512090979372e-06,
+      "loss": 0.1556,
+      "step": 790
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.7214991005258386e-06,
+      "loss": 0.187,
+      "step": 791
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.7099776107762483e-06,
+      "loss": 0.1368,
+      "step": 792
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.698486847451527e-06,
+      "loss": 0.1936,
+      "step": 793
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.6870269178670795e-06,
+      "loss": 0.1298,
+      "step": 794
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.6755979290503437e-06,
+      "loss": 0.1229,
+      "step": 795
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.6641999877397903e-06,
+      "loss": 0.1778,
+      "step": 796
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.6528332003839325e-06,
+      "loss": 0.1322,
+      "step": 797
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.6414976731403265e-06,
+      "loss": 0.1188,
+      "step": 798
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.6301935118745826e-06,
+      "loss": 0.157,
+      "step": 799
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.618920822159375e-06,
+      "loss": 0.1468,
+      "step": 800
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.6076797092734575e-06,
+      "loss": 0.1864,
+      "step": 801
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.5964702782006753e-06,
+      "loss": 0.1524,
+      "step": 802
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.5852926336289926e-06,
+      "loss": 0.1753,
+      "step": 803
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.5741468799495112e-06,
+      "loss": 0.1539,
+      "step": 804
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.5630331212554906e-06,
+      "loss": 0.1635,
+      "step": 805
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.5519514613413832e-06,
+      "loss": 0.1504,
+      "step": 806
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.5409020037018652e-06,
+      "loss": 0.1828,
+      "step": 807
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.5298848515308584e-06,
+      "loss": 0.1371,
+      "step": 808
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.5189001077205835e-06,
+      "loss": 0.135,
+      "step": 809
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.5079478748605874e-06,
+      "loss": 0.1296,
+      "step": 810
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.4970282552367854e-06,
+      "loss": 0.1351,
+      "step": 811
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.4861413508305128e-06,
+      "loss": 0.1377,
+      "step": 812
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.4752872633175691e-06,
+      "loss": 0.1382,
+      "step": 813
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.4644660940672628e-06,
+      "loss": 0.1753,
+      "step": 814
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.453677944141474e-06,
+      "loss": 0.1477,
+      "step": 815
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.4429229142937062e-06,
+      "loss": 0.1542,
+      "step": 816
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.432201104968141e-06,
+      "loss": 0.1363,
+      "step": 817
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.4215126162987097e-06,
+      "loss": 0.1192,
+      "step": 818
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.4108575481081522e-06,
+      "loss": 0.1431,
+      "step": 819
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.4002359999070797e-06,
+      "loss": 0.115,
+      "step": 820
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.3896480708930576e-06,
+      "loss": 0.1211,
+      "step": 821
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.3790938599496712e-06,
+      "loss": 0.158,
+      "step": 822
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.368573465645599e-06,
+      "loss": 0.1845,
+      "step": 823
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.358086986233701e-06,
+      "loss": 0.1263,
+      "step": 824
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3476345196500978e-06,
+      "loss": 0.1287,
+      "step": 825
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.3372161635132486e-06,
+      "loss": 0.1281,
+      "step": 826
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.3268320151230518e-06,
+      "loss": 0.1348,
+      "step": 827
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.3164821714599296e-06,
+      "loss": 0.1306,
+      "step": 828
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.3061667291839182e-06,
+      "loss": 0.1454,
+      "step": 829
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.295885784633774e-06,
+      "loss": 0.1158,
+      "step": 830
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2856394338260691e-06,
+      "loss": 0.1444,
+      "step": 831
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2754277724542945e-06,
+      "loss": 0.1608,
+      "step": 832
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.2652508958879671e-06,
+      "loss": 0.1605,
+      "step": 833
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.2551088991717409e-06,
+      "loss": 0.1747,
+      "step": 834
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.245001877024512e-06,
+      "loss": 0.1492,
+      "step": 835
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.2349299238385442e-06,
+      "loss": 0.1324,
+      "step": 836
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.224893133678583e-06,
+      "loss": 0.1501,
+      "step": 837
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.2148916002809719e-06,
+      "loss": 0.1993,
+      "step": 838
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.2049254170527857e-06,
+      "loss": 0.1717,
+      "step": 839
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.1949946770709558e-06,
+      "loss": 0.1232,
+      "step": 840
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1850994730813937e-06,
+      "loss": 0.1519,
+      "step": 841
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.175239897498135e-06,
+      "loss": 0.1194,
+      "step": 842
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1654160424024718e-06,
+      "loss": 0.1348,
+      "step": 843
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.1556279995420888e-06,
+      "loss": 0.1219,
+      "step": 844
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1458758603302145e-06,
+      "loss": 0.1477,
+      "step": 845
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1361597158447634e-06,
+      "loss": 0.1305,
+      "step": 846
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.1264796568274811e-06,
+      "loss": 0.1477,
+      "step": 847
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1168357736831042e-06,
+      "loss": 0.1402,
+      "step": 848
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.1072281564785148e-06,
+      "loss": 0.1104,
+      "step": 849
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.097656894941892e-06,
+      "loss": 0.1324,
+      "step": 850
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.088122078461884e-06,
+      "loss": 0.1606,
+      "step": 851
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.078623796086768e-06,
+      "loss": 0.1181,
+      "step": 852
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.0691621365236154e-06,
+      "loss": 0.1688,
+      "step": 853
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.0597371881374713e-06,
+      "loss": 0.1092,
+      "step": 854
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.0503490389505244e-06,
+      "loss": 0.1078,
+      "step": 855
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.0409977766412821e-06,
+      "loss": 0.1275,
+      "step": 856
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.0316834885437594e-06,
+      "loss": 0.1484,
+      "step": 857
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.0224062616466562e-06,
+      "loss": 0.1475,
+      "step": 858
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.013166182592551e-06,
+      "loss": 0.127,
+      "step": 859
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.003963337677083e-06,
+      "loss": 0.1343,
+      "step": 860
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 9.947978128481556e-07,
+      "loss": 0.1652,
+      "step": 861
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 9.856696937051297e-07,
+      "loss": 0.135,
+      "step": 862
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 9.765790654980195e-07,
+      "loss": 0.1257,
+      "step": 863
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 9.6752601312671e-07,
+      "loss": 0.1523,
+      "step": 864
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 9.58510621140145e-07,
+      "loss": 0.1439,
+      "step": 865
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.495329737355541e-07,
+      "loss": 0.1699,
+      "step": 866
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.405931547576591e-07,
+      "loss": 0.125,
+      "step": 867
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.31691247697889e-07,
+      "loss": 0.1483,
+      "step": 868
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 9.228273356936046e-07,
+      "loss": 0.1679,
+      "step": 869
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 9.140015015273224e-07,
+      "loss": 0.1155,
+      "step": 870
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 9.052138276259348e-07,
+      "loss": 0.1459,
+      "step": 871
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 8.96464396059949e-07,
+      "loss": 0.1915,
+      "step": 872
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 8.87753288542717e-07,
+      "loss": 0.1184,
+      "step": 873
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 8.79080586429667e-07,
+      "loss": 0.1365,
+      "step": 874
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 8.704463707175526e-07,
+      "loss": 0.1426,
+      "step": 875
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 8.618507220436939e-07,
+      "loss": 0.1799,
+      "step": 876
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 8.532937206852165e-07,
+      "loss": 0.1651,
+      "step": 877
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 8.447754465583142e-07,
+      "loss": 0.1563,
+      "step": 878
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 8.362959792174941e-07,
+      "loss": 0.1092,
+      "step": 879
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 8.278553978548365e-07,
+      "loss": 0.1384,
+      "step": 880
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 8.194537812992531e-07,
+      "loss": 0.1646,
+      "step": 881
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 8.110912080157552e-07,
+      "loss": 0.1291,
+      "step": 882
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 8.027677561047176e-07,
+      "loss": 0.1518,
+      "step": 883
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 7.944835033011472e-07,
+      "loss": 0.1374,
+      "step": 884
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 7.862385269739625e-07,
+      "loss": 0.1715,
+      "step": 885
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 7.780329041252688e-07,
+      "loss": 0.1667,
+      "step": 886
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 7.698667113896346e-07,
+      "loss": 0.1675,
+      "step": 887
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 7.617400250333834e-07,
+      "loss": 0.1278,
+      "step": 888
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 7.536529209538773e-07,
+      "loss": 0.1403,
+      "step": 889
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 7.45605474678806e-07,
+      "loss": 0.1364,
+      "step": 890
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 7.375977613654861e-07,
+      "loss": 0.1622,
+      "step": 891
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 7.296298558001592e-07,
+      "loss": 0.1282,
+      "step": 892
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 7.217018323972852e-07,
+      "loss": 0.1124,
+      "step": 893
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 7.138137651988597e-07,
+      "loss": 0.1392,
+      "step": 894
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 7.059657278737136e-07,
+      "loss": 0.1285,
+      "step": 895
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 6.981577937168277e-07,
+      "loss": 0.1494,
+      "step": 896
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 6.903900356486504e-07,
+      "loss": 0.1396,
+      "step": 897
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 6.826625262144105e-07,
+      "loss": 0.1341,
+      "step": 898
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 6.749753375834467e-07,
+      "loss": 0.1281,
+      "step": 899
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 6.67328541548532e-07,
+      "loss": 0.1227,
+      "step": 900
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 6.597222095251965e-07,
+      "loss": 0.1357,
+      "step": 901
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 6.52156412551071e-07,
+      "loss": 0.1292,
+      "step": 902
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 6.446312212852162e-07,
+      "loss": 0.1127,
+      "step": 903
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 6.37146706007466e-07,
+      "loss": 0.1273,
+      "step": 904
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 6.29702936617767e-07,
+      "loss": 0.1346,
+      "step": 905
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 6.222999826355325e-07,
+      "loss": 0.1311,
+      "step": 906
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 6.14937913198988e-07,
+      "loss": 0.1429,
+      "step": 907
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 6.076167970645252e-07,
+      "loss": 0.1312,
+      "step": 908
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 6.003367026060647e-07,
+      "loss": 0.1451,
+      "step": 909
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 5.930976978144132e-07,
+      "loss": 0.1429,
+      "step": 910
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 5.858998502966273e-07,
+      "loss": 0.1451,
+      "step": 911
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 5.787432272753885e-07,
+      "loss": 0.1518,
+      "step": 912
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 5.716278955883703e-07,
+      "loss": 0.1287,
+      "step": 913
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 5.645539216876117e-07,
+      "loss": 0.1295,
+      "step": 914
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 5.575213716389039e-07,
+      "loss": 0.1546,
+      "step": 915
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 5.505303111211685e-07,
+      "loss": 0.1399,
+      "step": 916
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 5.435808054258429e-07,
+      "loss": 0.1417,
+      "step": 917
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 5.366729194562747e-07,
+      "loss": 0.1583,
+      "step": 918
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 5.298067177271144e-07,
+      "loss": 0.1375,
+      "step": 919
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 5.229822643637078e-07,
+      "loss": 0.1376,
+      "step": 920
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 5.161996231015049e-07,
+      "loss": 0.1548,
+      "step": 921
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 5.094588572854603e-07,
+      "loss": 0.1573,
+      "step": 922
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 5.027600298694397e-07,
+      "loss": 0.1508,
+      "step": 923
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.961032034156388e-07,
+      "loss": 0.1505,
+      "step": 924
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.89488440093992e-07,
+      "loss": 0.1489,
+      "step": 925
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 4.829158016815944e-07,
+      "loss": 0.171,
+      "step": 926
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.763853495621251e-07,
+      "loss": 0.1372,
+      "step": 927
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.6989714472527705e-07,
+      "loss": 0.1586,
+      "step": 928
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 4.6345124776617847e-07,
+      "loss": 0.1883,
+      "step": 929
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.570477188848377e-07,
+      "loss": 0.1323,
+      "step": 930
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.5068661788557345e-07,
+      "loss": 0.1362,
+      "step": 931
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.4436800417645863e-07,
+      "loss": 0.1486,
+      "step": 932
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 4.3809193676876584e-07,
+      "loss": 0.145,
+      "step": 933
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.318584742764187e-07,
+      "loss": 0.1096,
+      "step": 934
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.2566767491543706e-07,
+      "loss": 0.147,
+      "step": 935
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.195195965034016e-07,
+      "loss": 0.1465,
+      "step": 936
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 4.134142964589105e-07,
+      "loss": 0.158,
+      "step": 937
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.073518318010389e-07,
+      "loss": 0.1444,
+      "step": 938
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.01332259148815e-07,
+      "loss": 0.1462,
+      "step": 939
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 3.953556347206861e-07,
+      "loss": 0.1672,
+      "step": 940
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.894220143339905e-07,
+      "loss": 0.1422,
+      "step": 941
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.8353145340444486e-07,
+      "loss": 0.121,
+      "step": 942
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.776840069456189e-07,
+      "loss": 0.1167,
+      "step": 943
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.71879729568424e-07,
+      "loss": 0.1285,
+      "step": 944
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.661186754806051e-07,
+      "loss": 0.1506,
+      "step": 945
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.604008984862334e-07,
+      "loss": 0.1516,
+      "step": 946
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.5472645198520064e-07,
+      "loss": 0.1494,
+      "step": 947
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.490953889727261e-07,
+      "loss": 0.1407,
+      "step": 948
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.435077620388577e-07,
+      "loss": 0.1343,
+      "step": 949
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.379636233679812e-07,
+      "loss": 0.1606,
+      "step": 950
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.324630247383337e-07,
+      "loss": 0.1006,
+      "step": 951
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.2700601752152117e-07,
+      "loss": 0.1323,
+      "step": 952
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.215926526820351e-07,
+      "loss": 0.1246,
+      "step": 953
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.162229807767808e-07,
+      "loss": 0.153,
+      "step": 954
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.108970519546034e-07,
+      "loss": 0.1462,
+      "step": 955
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.0561491595581695e-07,
+      "loss": 0.1525,
+      "step": 956
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.0037662211174437e-07,
+      "loss": 0.1672,
+      "step": 957
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 2.951822193442544e-07,
+      "loss": 0.1548,
+      "step": 958
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.9003175616530264e-07,
+      "loss": 0.1644,
+      "step": 959
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.8492528067648575e-07,
+      "loss": 0.1405,
+      "step": 960
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.798628405685827e-07,
+      "loss": 0.1523,
+      "step": 961
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 2.74844483121115e-07,
+      "loss": 0.1577,
+      "step": 962
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.698702552019045e-07,
+      "loss": 0.143,
+      "step": 963
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.649402032666365e-07,
+      "loss": 0.1279,
+      "step": 964
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 2.6005437335842155e-07,
+      "loss": 0.1171,
+      "step": 965
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.552128111073715e-07,
+      "loss": 0.1433,
+      "step": 966
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.50415561730169e-07,
+      "loss": 0.1397,
+      "step": 967
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.456626700296455e-07,
+      "loss": 0.1413,
+      "step": 968
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 2.409541803943666e-07,
+      "loss": 0.1546,
+      "step": 969
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.3629013679821343e-07,
+      "loss": 0.1627,
+      "step": 970
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.3167058279997156e-07,
+      "loss": 0.1325,
+      "step": 971
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 2.2709556154292878e-07,
+      "loss": 0.1509,
+      "step": 972
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.2256511575446837e-07,
+      "loss": 0.2011,
+      "step": 973
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.1807928774567e-07,
+      "loss": 0.1583,
+      "step": 974
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.136381194109166e-07,
+      "loss": 0.1532,
+      "step": 975
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 2.0924165222750315e-07,
+      "loss": 0.1786,
+      "step": 976
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.048899272552457e-07,
+      "loss": 0.1352,
+      "step": 977
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 2.0058298513610185e-07,
+      "loss": 0.1452,
+      "step": 978
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.9632086609379041e-07,
+      "loss": 0.1485,
+      "step": 979
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.9210360993341447e-07,
+      "loss": 0.1586,
+      "step": 980
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.879312560410901e-07,
+      "loss": 0.1549,
+      "step": 981
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.8380384338357925e-07,
+      "loss": 0.1337,
+      "step": 982
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.797214105079248e-07,
+      "loss": 0.1719,
+      "step": 983
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7568399554109106e-07,
+      "loss": 0.1122,
+      "step": 984
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7169163618960817e-07,
+      "loss": 0.1197,
+      "step": 985
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.677443697392206e-07,
+      "loss": 0.1947,
+      "step": 986
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6384223305453417e-07,
+      "loss": 0.1433,
+      "step": 987
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.5998526257867852e-07,
+      "loss": 0.1414,
+      "step": 988
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.5617349433296402e-07,
+      "loss": 0.1463,
+      "step": 989
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.5240696391654208e-07,
+      "loss": 0.1608,
+      "step": 990
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.4868570650607816e-07,
+      "loss": 0.1074,
+      "step": 991
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.450097568554193e-07,
+      "loss": 0.1312,
+      "step": 992
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.4137914929527097e-07,
+      "loss": 0.1402,
+      "step": 993
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.3779391773287854e-07,
+      "loss": 0.1508,
+      "step": 994
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.342540956517041e-07,
+      "loss": 0.1663,
+      "step": 995
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.3075971611112237e-07,
+      "loss": 0.1596,
+      "step": 996
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.2731081174610526e-07,
+      "loss": 0.1536,
+      "step": 997
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.2390741476692003e-07,
+      "loss": 0.1732,
+      "step": 998
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.205495569588283e-07,
+      "loss": 0.1408,
+      "step": 999
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.1723726968178917e-07,
+      "loss": 0.1465,
+      "step": 1000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.139705838701638e-07,
+      "loss": 0.1954,
+      "step": 1001
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.1074953003243183e-07,
+      "loss": 0.1808,
+      "step": 1002
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.0757413825090212e-07,
+      "loss": 0.1532,
+      "step": 1003
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.0444443818143135e-07,
+      "loss": 0.1384,
+      "step": 1004
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.0136045905315028e-07,
+      "loss": 0.1556,
+      "step": 1005
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.832222966819015e-08,
+      "loss": 0.1433,
+      "step": 1006
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.532977840141123e-08,
+      "loss": 0.1841,
+      "step": 1007
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 9.238313320013903e-08,
+      "loss": 0.1225,
+      "step": 1008
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.948232158390468e-08,
+      "loss": 0.1418,
+      "step": 1009
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.662737064418725e-08,
+      "loss": 0.1646,
+      "step": 1010
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.381830704415839e-08,
+      "loss": 0.1297,
+      "step": 1011
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 8.105515701843703e-08,
+      "loss": 0.1495,
+      "step": 1012
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 7.833794637284232e-08,
+      "loss": 0.1532,
+      "step": 1013
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 7.566670048415214e-08,
+      "loss": 0.1413,
+      "step": 1014
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 7.30414442998667e-08,
+      "loss": 0.1402,
+      "step": 1015
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 7.046220233797752e-08,
+      "loss": 0.1181,
+      "step": 1016
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 6.792899868673487e-08,
+      "loss": 0.1638,
+      "step": 1017
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 6.544185700442407e-08,
+      "loss": 0.1541,
+      "step": 1018
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 6.300080051914792e-08,
+      "loss": 0.1542,
+      "step": 1019
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 6.060585202860291e-08,
+      "loss": 0.1294,
+      "step": 1020
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 5.825703389987392e-08,
+      "loss": 0.1565,
+      "step": 1021
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 5.5954368069219834e-08,
+      "loss": 0.1174,
+      "step": 1022
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 5.369787604186993e-08,
+      "loss": 0.127,
+      "step": 1023
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 5.148757889182565e-08,
+      "loss": 0.1463,
+      "step": 1024
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.9323497261659635e-08,
+      "loss": 0.1221,
+      "step": 1025
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 4.7205651362326467e-08,
+      "loss": 0.1911,
+      "step": 1026
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.513406097297224e-08,
+      "loss": 0.1578,
+      "step": 1027
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.3108745440749723e-08,
+      "loss": 0.1612,
+      "step": 1028
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 4.112972368063794e-08,
+      "loss": 0.1484,
+      "step": 1029
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.9197014175266226e-08,
+      "loss": 0.1286,
+      "step": 1030
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.731063497474152e-08,
+      "loss": 0.1789,
+      "step": 1031
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.547060369647693e-08,
+      "loss": 0.1222,
+      "step": 1032
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.3676937525032314e-08,
+      "loss": 0.1505,
+      "step": 1033
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.192965321195007e-08,
+      "loss": 0.1181,
+      "step": 1034
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.022876707559796e-08,
+      "loss": 0.1889,
+      "step": 1035
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.8574295001021492e-08,
+      "loss": 0.1401,
+      "step": 1036
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 2.696625243979012e-08,
+      "loss": 0.1614,
+      "step": 1037
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.5404654409856288e-08,
+      "loss": 0.1647,
+      "step": 1038
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.3889515495413297e-08,
+      "loss": 0.122,
+      "step": 1039
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 2.2420849846761517e-08,
+      "loss": 0.1237,
+      "step": 1040
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 2.0998671180172957e-08,
+      "loss": 0.1388,
+      "step": 1041
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.962299277776636e-08,
+      "loss": 0.161,
+      "step": 1042
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.8293827487380623e-08,
+      "loss": 0.1061,
+      "step": 1043
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.701118772245658e-08,
+      "loss": 0.1335,
+      "step": 1044
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.577508546191986e-08,
+      "loss": 0.1294,
+      "step": 1045
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.4585532250070423e-08,
+      "loss": 0.163,
+      "step": 1046
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.3442539196472647e-08,
+      "loss": 0.1408,
+      "step": 1047
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.2346116975853194e-08,
+      "loss": 0.151,
+      "step": 1048
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.1296275828001635e-08,
+      "loss": 0.1396,
+      "step": 1049
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.0293025557672753e-08,
+      "loss": 0.1191,
+      "step": 1050
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 9.336375534497732e-09,
+      "loss": 0.1222,
+      "step": 1051
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 8.426334692893668e-09,
+      "loss": 0.1473,
+      "step": 1052
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 7.56291153198363e-09,
+      "loss": 0.145,
+      "step": 1053
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 6.746114115513402e-09,
+      "loss": 0.1251,
+      "step": 1054
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 5.975950071779313e-09,
+      "loss": 0.1726,
+      "step": 1055
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 5.252426593555515e-09,
+      "loss": 0.1019,
+      "step": 1056
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 4.575550438026266e-09,
+      "loss": 0.1459,
+      "step": 1057
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 3.9453279267248625e-09,
+      "loss": 0.1274,
+      "step": 1058
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.361764945473134e-09,
+      "loss": 0.1764,
+      "step": 1059
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.8248669443253775e-09,
+      "loss": 0.1337,
+      "step": 1060
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 2.334638937521172e-09,
+      "loss": 0.1387,
+      "step": 1061
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.891085503433754e-09,
+      "loss": 0.1169,
+      "step": 1062
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.4942107845317132e-09,
+      "loss": 0.1494,
+      "step": 1063
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.1440184873362514e-09,
+      "loss": 0.1397,
+      "step": 1064
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 8.405118823906489e-10,
+      "loss": 0.1377,
+      "step": 1065
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 5.836938042258489e-10,
+      "loss": 0.1706,
+      "step": 1066
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.735666513371428e-10,
+      "loss": 0.1163,
+      "step": 1067
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.1013238615974486e-10,
+      "loss": 0.167,
+      "step": 1068
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 9.339253505102896e-11,
+      "loss": 0.1865,
+      "step": 1069
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 2.334818827665064e-11,
+      "loss": 0.157,
+      "step": 1070
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0,
+      "loss": 0.1459,
+      "step": 1071
+    },
+    {
+      "epoch": 3.0,
+      "step": 1071,
+      "total_flos": 4.1033935718543524e+18,
+      "train_loss": 0.28119602048096537,
+      "train_runtime": 33404.3503,
+      "train_samples_per_second": 3.072,
+      "train_steps_per_second": 0.032
+    }
+  ],
+  "max_steps": 1071,
+  "num_train_epochs": 3,
+  "total_flos": 4.1033935718543524e+18,
+  "trial_name": null,
+  "trial_params": null
+}