diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,11472 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1907,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.4482758620689656e-07,
+      "loss": 2.5332,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 6.896551724137931e-07,
+      "loss": 2.5918,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.0344827586206898e-06,
+      "loss": 2.6133,
+      "step": 3
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.3793103448275862e-06,
+      "loss": 2.6133,
+      "step": 4
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.724137931034483e-06,
+      "loss": 2.5586,
+      "step": 5
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.0689655172413796e-06,
+      "loss": 2.5215,
+      "step": 6
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.4137931034482762e-06,
+      "loss": 2.2949,
+      "step": 7
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.7586206896551725e-06,
+      "loss": 2.2461,
+      "step": 8
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 3.103448275862069e-06,
+      "loss": 1.7588,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.448275862068966e-06,
+      "loss": 1.4385,
+      "step": 10
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.793103448275862e-06,
+      "loss": 1.5371,
+      "step": 11
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.137931034482759e-06,
+      "loss": 1.1934,
+      "step": 12
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.482758620689656e-06,
+      "loss": 1.1309,
+      "step": 13
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.8275862068965525e-06,
+      "loss": 1.2466,
+      "step": 14
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.172413793103449e-06,
+      "loss": 1.1221,
+      "step": 15
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.517241379310345e-06,
+      "loss": 0.8535,
+      "step": 16
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.862068965517242e-06,
+      "loss": 1.0122,
+      "step": 17
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.206896551724138e-06,
+      "loss": 1.0264,
+      "step": 18
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.551724137931035e-06,
+      "loss": 0.9458,
+      "step": 19
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.896551724137932e-06,
+      "loss": 1.0029,
+      "step": 20
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.241379310344828e-06,
+      "loss": 0.9878,
+      "step": 21
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.586206896551724e-06,
+      "loss": 1.0273,
+      "step": 22
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.93103448275862e-06,
+      "loss": 0.9409,
+      "step": 23
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.275862068965518e-06,
+      "loss": 0.7969,
+      "step": 24
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.620689655172414e-06,
+      "loss": 0.9946,
+      "step": 25
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.965517241379312e-06,
+      "loss": 0.9131,
+      "step": 26
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.310344827586207e-06,
+      "loss": 1.0547,
+      "step": 27
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 9.655172413793105e-06,
+      "loss": 0.9688,
+      "step": 28
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1e-05,
+      "loss": 0.7617,
+      "step": 29
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0344827586206898e-05,
+      "loss": 0.7607,
+      "step": 30
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0689655172413792e-05,
+      "loss": 0.7676,
+      "step": 31
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.103448275862069e-05,
+      "loss": 0.8247,
+      "step": 32
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1379310344827587e-05,
+      "loss": 0.8354,
+      "step": 33
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1724137931034483e-05,
+      "loss": 0.7949,
+      "step": 34
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.206896551724138e-05,
+      "loss": 0.6887,
+      "step": 35
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2413793103448277e-05,
+      "loss": 0.6836,
+      "step": 36
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2758620689655174e-05,
+      "loss": 0.7354,
+      "step": 37
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.310344827586207e-05,
+      "loss": 0.7163,
+      "step": 38
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3448275862068967e-05,
+      "loss": 0.709,
+      "step": 39
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3793103448275863e-05,
+      "loss": 0.7788,
+      "step": 40
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4137931034482759e-05,
+      "loss": 0.668,
+      "step": 41
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4482758620689657e-05,
+      "loss": 0.8237,
+      "step": 42
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.4827586206896554e-05,
+      "loss": 0.6567,
+      "step": 43
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5172413793103448e-05,
+      "loss": 0.6929,
+      "step": 44
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5517241379310346e-05,
+      "loss": 0.6348,
+      "step": 45
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.586206896551724e-05,
+      "loss": 0.8032,
+      "step": 46
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.6206896551724137e-05,
+      "loss": 0.6797,
+      "step": 47
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6551724137931037e-05,
+      "loss": 0.7144,
+      "step": 48
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6896551724137932e-05,
+      "loss": 0.666,
+      "step": 49
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7241379310344828e-05,
+      "loss": 0.6177,
+      "step": 50
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7586206896551724e-05,
+      "loss": 0.6333,
+      "step": 51
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7931034482758623e-05,
+      "loss": 0.6323,
+      "step": 52
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.827586206896552e-05,
+      "loss": 0.6335,
+      "step": 53
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8620689655172415e-05,
+      "loss": 0.5559,
+      "step": 54
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.896551724137931e-05,
+      "loss": 0.668,
+      "step": 55
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.931034482758621e-05,
+      "loss": 0.6821,
+      "step": 56
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9655172413793106e-05,
+      "loss": 0.5564,
+      "step": 57
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2e-05,
+      "loss": 0.564,
+      "step": 58
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999985565696826e-05,
+      "loss": 0.7134,
+      "step": 59
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999942262828967e-05,
+      "loss": 0.5339,
+      "step": 60
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999987009152143e-05,
+      "loss": 0.625,
+      "step": 61
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999769051982578e-05,
+      "loss": 0.5859,
+      "step": 62
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.999963914450408e-05,
+      "loss": 0.5471,
+      "step": 63
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999480369460968e-05,
+      "loss": 0.5779,
+      "step": 64
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999292727311608e-05,
+      "loss": 0.6016,
+      "step": 65
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.9999076218597693e-05,
+      "loss": 0.5852,
+      "step": 66
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9998830843944253e-05,
+      "loss": 0.6384,
+      "step": 67
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999855660405965e-05,
+      "loss": 0.7061,
+      "step": 68
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9998253499735577e-05,
+      "loss": 0.6357,
+      "step": 69
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9997921531847052e-05,
+      "loss": 0.6685,
+      "step": 70
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999756070135243e-05,
+      "loss": 0.665,
+      "step": 71
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999717100929336e-05,
+      "loss": 0.4158,
+      "step": 72
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9996752456794846e-05,
+      "loss": 0.6606,
+      "step": 73
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9996305045065182e-05,
+      "loss": 0.6396,
+      "step": 74
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9995828775395985e-05,
+      "loss": 0.7012,
+      "step": 75
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999532364916218e-05,
+      "loss": 0.5227,
+      "step": 76
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9994789667821996e-05,
+      "loss": 0.6145,
+      "step": 77
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9994226832916957e-05,
+      "loss": 0.6343,
+      "step": 78
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9993635146071903e-05,
+      "loss": 0.7168,
+      "step": 79
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9993014608994935e-05,
+      "loss": 0.6174,
+      "step": 80
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999236522347747e-05,
+      "loss": 0.5701,
+      "step": 81
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999168699139418e-05,
+      "loss": 0.5334,
+      "step": 82
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.999097991470304e-05,
+      "loss": 0.6294,
+      "step": 83
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9990243995445278e-05,
+      "loss": 0.6772,
+      "step": 84
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9989479235745385e-05,
+      "loss": 0.5898,
+      "step": 85
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9988685637811126e-05,
+      "loss": 0.5964,
+      "step": 86
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9987863203933493e-05,
+      "loss": 0.5989,
+      "step": 87
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9987011936486752e-05,
+      "loss": 0.6606,
+      "step": 88
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9986131837928383e-05,
+      "loss": 0.5896,
+      "step": 89
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9985222910799112e-05,
+      "loss": 0.5774,
+      "step": 90
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.998428515772289e-05,
+      "loss": 0.5303,
+      "step": 91
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9983318581406876e-05,
+      "loss": 0.6179,
+      "step": 92
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9982323184641436e-05,
+      "loss": 0.6641,
+      "step": 93
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9981298970300143e-05,
+      "loss": 0.5364,
+      "step": 94
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9980245941339768e-05,
+      "loss": 0.6772,
+      "step": 95
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9979164100800253e-05,
+      "loss": 0.6123,
+      "step": 96
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9978053451804724e-05,
+      "loss": 0.5669,
+      "step": 97
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9976913997559468e-05,
+      "loss": 0.54,
+      "step": 98
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9975745741353927e-05,
+      "loss": 0.5181,
+      "step": 99
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9974548686560703e-05,
+      "loss": 0.6316,
+      "step": 100
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9973322836635517e-05,
+      "loss": 0.6841,
+      "step": 101
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9972068195117235e-05,
+      "loss": 0.6099,
+      "step": 102
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9970784765627824e-05,
+      "loss": 0.5576,
+      "step": 103
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 1.9969472551872373e-05,
+      "loss": 0.5806,
+      "step": 104
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996813155763906e-05,
+      "loss": 0.5171,
+      "step": 105
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996676178679914e-05,
+      "loss": 0.455,
+      "step": 106
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.996536324330696e-05,
+      "loss": 0.5732,
+      "step": 107
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9963935931199917e-05,
+      "loss": 0.6052,
+      "step": 108
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9962479854598463e-05,
+      "loss": 0.5044,
+      "step": 109
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9960995017706082e-05,
+      "loss": 0.5806,
+      "step": 110
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9959481424809302e-05,
+      "loss": 0.6509,
+      "step": 111
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9957939080277647e-05,
+      "loss": 0.6553,
+      "step": 112
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995636798856365e-05,
+      "loss": 0.6082,
+      "step": 113
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.995476815420284e-05,
+      "loss": 0.5627,
+      "step": 114
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9953139581813708e-05,
+      "loss": 0.6858,
+      "step": 115
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9951482276097726e-05,
+      "loss": 0.6074,
+      "step": 116
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9949796241839293e-05,
+      "loss": 0.6179,
+      "step": 117
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9948081483905763e-05,
+      "loss": 0.5569,
+      "step": 118
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.99463380072474e-05,
+      "loss": 0.5391,
+      "step": 119
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9944565816897378e-05,
+      "loss": 0.5583,
+      "step": 120
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9942764917971767e-05,
+      "loss": 0.5793,
+      "step": 121
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9940935315669504e-05,
+      "loss": 0.5811,
+      "step": 122
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 1.9939077015272404e-05,
+      "loss": 0.6172,
+      "step": 123
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.993719002214512e-05,
+      "loss": 0.489,
+      "step": 124
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9935274341735132e-05,
+      "loss": 0.5247,
+      "step": 125
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9933329979572743e-05,
+      "loss": 0.5896,
+      "step": 126
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.993135694127106e-05,
+      "loss": 0.6331,
+      "step": 127
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9929355232525976e-05,
+      "loss": 0.6311,
+      "step": 128
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.992732485911613e-05,
+      "loss": 0.5747,
+      "step": 129
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9925265826902936e-05,
+      "loss": 0.4749,
+      "step": 130
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9923178141830536e-05,
+      "loss": 0.5227,
+      "step": 131
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9921061809925777e-05,
+      "loss": 0.4497,
+      "step": 132
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.991891683729822e-05,
+      "loss": 0.5908,
+      "step": 133
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.99167432301401e-05,
+      "loss": 0.5381,
+      "step": 134
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.991454099472632e-05,
+      "loss": 0.6006,
+      "step": 135
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9912310137414426e-05,
+      "loss": 0.5664,
+      "step": 136
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.991005066464459e-05,
+      "loss": 0.6118,
+      "step": 137
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9907762582939596e-05,
+      "loss": 0.5234,
+      "step": 138
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9905445898904822e-05,
+      "loss": 0.6145,
+      "step": 139
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9903100619228202e-05,
+      "loss": 0.4875,
+      "step": 140
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.990072675068024e-05,
+      "loss": 0.541,
+      "step": 141
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9898324300113963e-05,
+      "loss": 0.53,
+      "step": 142
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 1.9895893274464907e-05,
+      "loss": 0.6978,
+      "step": 143
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9893433680751105e-05,
+      "loss": 0.5923,
+      "step": 144
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.989094552607306e-05,
+      "loss": 0.5837,
+      "step": 145
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9888428817613735e-05,
+      "loss": 0.5898,
+      "step": 146
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.988588356263851e-05,
+      "loss": 0.5996,
+      "step": 147
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9883309768495183e-05,
+      "loss": 0.5852,
+      "step": 148
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.988070744261394e-05,
+      "loss": 0.5464,
+      "step": 149
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.987807659250733e-05,
+      "loss": 0.4521,
+      "step": 150
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.987541722577025e-05,
+      "loss": 0.55,
+      "step": 151
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.987272935007993e-05,
+      "loss": 0.5466,
+      "step": 152
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9870012973195883e-05,
+      "loss": 0.6631,
+      "step": 153
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.986726810295992e-05,
+      "loss": 0.5215,
+      "step": 154
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9864494747296084e-05,
+      "loss": 0.5754,
+      "step": 155
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.986169291421068e-05,
+      "loss": 0.6035,
+      "step": 156
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9858862611792206e-05,
+      "loss": 0.5581,
+      "step": 157
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9856003848211345e-05,
+      "loss": 0.5442,
+      "step": 158
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.985311663172096e-05,
+      "loss": 0.6165,
+      "step": 159
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9850200970656032e-05,
+      "loss": 0.4795,
+      "step": 160
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9847256873433672e-05,
+      "loss": 0.5403,
+      "step": 161
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 1.9844284348553077e-05,
+      "loss": 0.5129,
+      "step": 162
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.984128340459552e-05,
+      "loss": 0.6267,
+      "step": 163
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9838254050224294e-05,
+      "loss": 0.554,
+      "step": 164
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9835196294184736e-05,
+      "loss": 0.4951,
+      "step": 165
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9832110145304152e-05,
+      "loss": 0.5786,
+      "step": 166
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.982899561249183e-05,
+      "loss": 0.584,
+      "step": 167
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9825852704738993e-05,
+      "loss": 0.4436,
+      "step": 168
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.982268143111877e-05,
+      "loss": 0.5303,
+      "step": 169
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9819481800786192e-05,
+      "loss": 0.6147,
+      "step": 170
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9816253822978147e-05,
+      "loss": 0.6025,
+      "step": 171
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.981299750701335e-05,
+      "loss": 0.5898,
+      "step": 172
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.980971286229234e-05,
+      "loss": 0.5249,
+      "step": 173
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9806399898297422e-05,
+      "loss": 0.5642,
+      "step": 174
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9803058624592665e-05,
+      "loss": 0.5918,
+      "step": 175
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.979968905082386e-05,
+      "loss": 0.5615,
+      "step": 176
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9796291186718494e-05,
+      "loss": 0.5024,
+      "step": 177
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9792865042085732e-05,
+      "loss": 0.5581,
+      "step": 178
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9789410626816373e-05,
+      "loss": 0.5264,
+      "step": 179
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.978592795088283e-05,
+      "loss": 0.5242,
+      "step": 180
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 1.9782417024339108e-05,
+      "loss": 0.5344,
+      "step": 181
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.977887785732076e-05,
+      "loss": 0.5933,
+      "step": 182
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.977531046004487e-05,
+      "loss": 0.6267,
+      "step": 183
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9771714842810014e-05,
+      "loss": 0.626,
+      "step": 184
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.976809101599624e-05,
+      "loss": 0.5605,
+      "step": 185
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.976443899006503e-05,
+      "loss": 0.5615,
+      "step": 186
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9760758775559275e-05,
+      "loss": 0.5896,
+      "step": 187
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9757050383103235e-05,
+      "loss": 0.5977,
+      "step": 188
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9753313823402536e-05,
+      "loss": 0.5261,
+      "step": 189
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.974954910724409e-05,
+      "loss": 0.4697,
+      "step": 190
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9745756245496118e-05,
+      "loss": 0.5627,
+      "step": 191
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.974193524910808e-05,
+      "loss": 0.4646,
+      "step": 192
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9738086129110655e-05,
+      "loss": 0.5371,
+      "step": 193
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973420889661573e-05,
+      "loss": 0.5386,
+      "step": 194
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.973030356281632e-05,
+      "loss": 0.5984,
+      "step": 195
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9726370138986583e-05,
+      "loss": 0.5742,
+      "step": 196
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.972240863648177e-05,
+      "loss": 0.4746,
+      "step": 197
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9718419066738186e-05,
+      "loss": 0.5134,
+      "step": 198
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.971440144127316e-05,
+      "loss": 0.4702,
+      "step": 199
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 1.9710355771685017e-05,
+      "loss": 0.4642,
+      "step": 200
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9706282069653043e-05,
+      "loss": 0.6436,
+      "step": 201
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.970218034693745e-05,
+      "loss": 0.5859,
+      "step": 202
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9698050615379333e-05,
+      "loss": 0.4788,
+      "step": 203
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9693892886900656e-05,
+      "loss": 0.5596,
+      "step": 204
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.96897071735042e-05,
+      "loss": 0.6055,
+      "step": 205
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9685493487273542e-05,
+      "loss": 0.4956,
+      "step": 206
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9681251840372998e-05,
+      "loss": 0.5662,
+      "step": 207
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.967698224504762e-05,
+      "loss": 0.5554,
+      "step": 208
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9672684713623125e-05,
+      "loss": 0.5825,
+      "step": 209
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9668359258505898e-05,
+      "loss": 0.5081,
+      "step": 210
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9664005892182917e-05,
+      "loss": 0.5281,
+      "step": 211
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.965962462722175e-05,
+      "loss": 0.4951,
+      "step": 212
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9655215476270494e-05,
+      "loss": 0.5801,
+      "step": 213
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9650778452057752e-05,
+      "loss": 0.6194,
+      "step": 214
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.96463135673926e-05,
+      "loss": 0.6279,
+      "step": 215
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9641820835164532e-05,
+      "loss": 0.4539,
+      "step": 216
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9637300268343442e-05,
+      "loss": 0.5894,
+      "step": 217
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.963275187997958e-05,
+      "loss": 0.5913,
+      "step": 218
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 1.9628175683203505e-05,
+      "loss": 0.4436,
+      "step": 219
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.962357169122606e-05,
+      "loss": 0.5229,
+      "step": 220
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9618939917338328e-05,
+      "loss": 0.5698,
+      "step": 221
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9614280374911597e-05,
+      "loss": 0.6255,
+      "step": 222
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9609593077397317e-05,
+      "loss": 0.5796,
+      "step": 223
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.960487803832706e-05,
+      "loss": 0.5601,
+      "step": 224
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9600135271312484e-05,
+      "loss": 0.5706,
+      "step": 225
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9595364790045303e-05,
+      "loss": 0.5808,
+      "step": 226
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.959056660829723e-05,
+      "loss": 0.4875,
+      "step": 227
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9585740739919946e-05,
+      "loss": 0.6111,
+      "step": 228
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9580887198845055e-05,
+      "loss": 0.5518,
+      "step": 229
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.957600599908406e-05,
+      "loss": 0.5408,
+      "step": 230
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9571097154728306e-05,
+      "loss": 0.5188,
+      "step": 231
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.956616067994894e-05,
+      "loss": 0.5471,
+      "step": 232
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9561196588996873e-05,
+      "loss": 0.5251,
+      "step": 233
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9556204896202748e-05,
+      "loss": 0.5212,
+      "step": 234
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9551185615976883e-05,
+      "loss": 0.5327,
+      "step": 235
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9546138762809246e-05,
+      "loss": 0.5193,
+      "step": 236
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9541064351269398e-05,
+      "loss": 0.6174,
+      "step": 237
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 1.9535962396006453e-05,
+      "loss": 0.5632,
+      "step": 238
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9530832911749046e-05,
+      "loss": 0.4832,
+      "step": 239
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.952567591330529e-05,
+      "loss": 0.5657,
+      "step": 240
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.952049141556271e-05,
+      "loss": 0.4272,
+      "step": 241
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9515279433488235e-05,
+      "loss": 0.6211,
+      "step": 242
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.951003998212813e-05,
+      "loss": 0.5833,
+      "step": 243
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9504773076607963e-05,
+      "loss": 0.571,
+      "step": 244
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.949947873213255e-05,
+      "loss": 0.5493,
+      "step": 245
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9494156963985934e-05,
+      "loss": 0.5928,
+      "step": 246
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.948880778753131e-05,
+      "loss": 0.5315,
+      "step": 247
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.948343121821101e-05,
+      "loss": 0.4658,
+      "step": 248
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.947802727154644e-05,
+      "loss": 0.5562,
+      "step": 249
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9472595963138037e-05,
+      "loss": 0.5127,
+      "step": 250
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9467137308665237e-05,
+      "loss": 0.4441,
+      "step": 251
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9461651323886408e-05,
+      "loss": 0.55,
+      "step": 252
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9456138024638832e-05,
+      "loss": 0.5601,
+      "step": 253
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9450597426838627e-05,
+      "loss": 0.5413,
+      "step": 254
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9445029546480735e-05,
+      "loss": 0.4785,
+      "step": 255
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.943943439963884e-05,
+      "loss": 0.5532,
+      "step": 256
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 1.9433812002465364e-05,
+      "loss": 0.4805,
+      "step": 257
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9428162371191373e-05,
+      "loss": 0.5483,
+      "step": 258
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9422485522126574e-05,
+      "loss": 0.4998,
+      "step": 259
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.941678147165923e-05,
+      "loss": 0.5327,
+      "step": 260
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9411050236256145e-05,
+      "loss": 0.4932,
+      "step": 261
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.94052918324626e-05,
+      "loss": 0.5146,
+      "step": 262
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9399506276902297e-05,
+      "loss": 0.4795,
+      "step": 263
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9393693586277332e-05,
+      "loss": 0.5518,
+      "step": 264
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9387853777368136e-05,
+      "loss": 0.5688,
+      "step": 265
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9381986867033418e-05,
+      "loss": 0.5015,
+      "step": 266
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9376092872210135e-05,
+      "loss": 0.6091,
+      "step": 267
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9370171809913426e-05,
+      "loss": 0.5676,
+      "step": 268
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.936422369723657e-05,
+      "loss": 0.5601,
+      "step": 269
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9358248551350945e-05,
+      "loss": 0.5593,
+      "step": 270
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9352246389505963e-05,
+      "loss": 0.4207,
+      "step": 271
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9346217229029026e-05,
+      "loss": 0.4026,
+      "step": 272
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9340161087325483e-05,
+      "loss": 0.4924,
+      "step": 273
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.933407798187857e-05,
+      "loss": 0.542,
+      "step": 274
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9327967930249362e-05,
+      "loss": 0.6052,
+      "step": 275
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 1.9321830950076732e-05,
+      "loss": 0.4976,
+      "step": 276
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.931566705907728e-05,
+      "loss": 0.5066,
+      "step": 277
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9309476275045306e-05,
+      "loss": 0.498,
+      "step": 278
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9303258615852737e-05,
+      "loss": 0.532,
+      "step": 279
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9297014099449095e-05,
+      "loss": 0.5591,
+      "step": 280
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.929074274386142e-05,
+      "loss": 0.5356,
+      "step": 281
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9284444567194246e-05,
+      "loss": 0.5974,
+      "step": 282
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.927811958762953e-05,
+      "loss": 0.499,
+      "step": 283
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9271767823426607e-05,
+      "loss": 0.5334,
+      "step": 284
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9265389292922133e-05,
+      "loss": 0.573,
+      "step": 285
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9258984014530043e-05,
+      "loss": 0.6228,
+      "step": 286
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.925255200674148e-05,
+      "loss": 0.4543,
+      "step": 287
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9246093288124748e-05,
+      "loss": 0.5518,
+      "step": 288
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9239607877325275e-05,
+      "loss": 0.5803,
+      "step": 289
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9233095793065534e-05,
+      "loss": 0.5205,
+      "step": 290
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9226557054145008e-05,
+      "loss": 0.5217,
+      "step": 291
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.921999167944012e-05,
+      "loss": 0.5776,
+      "step": 292
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.92133996879042e-05,
+      "loss": 0.5901,
+      "step": 293
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9206781098567403e-05,
+      "loss": 0.4946,
+      "step": 294
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.9200135930536672e-05,
+      "loss": 0.4978,
+      "step": 295
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9193464202995688e-05,
+      "loss": 0.5276,
+      "step": 296
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.918676593520479e-05,
+      "loss": 0.5881,
+      "step": 297
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.918004114650095e-05,
+      "loss": 0.5542,
+      "step": 298
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9173289856297693e-05,
+      "loss": 0.5618,
+      "step": 299
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9166512084085055e-05,
+      "loss": 0.5552,
+      "step": 300
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.915970784942952e-05,
+      "loss": 0.5046,
+      "step": 301
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.915287717197396e-05,
+      "loss": 0.4871,
+      "step": 302
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9146020071437597e-05,
+      "loss": 0.4529,
+      "step": 303
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.913913656761592e-05,
+      "loss": 0.4966,
+      "step": 304
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9132226680380642e-05,
+      "loss": 0.542,
+      "step": 305
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9125290429679652e-05,
+      "loss": 0.4097,
+      "step": 306
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.911832783553693e-05,
+      "loss": 0.5015,
+      "step": 307
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9111338918052525e-05,
+      "loss": 0.4641,
+      "step": 308
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9104323697402462e-05,
+      "loss": 0.5627,
+      "step": 309
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9097282193838704e-05,
+      "loss": 0.5186,
+      "step": 310
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9090214427689094e-05,
+      "loss": 0.4041,
+      "step": 311
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9083120419357286e-05,
+      "loss": 0.5173,
+      "step": 312
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9076000189322694e-05,
+      "loss": 0.5217,
+      "step": 313
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.9068853758140432e-05,
+      "loss": 0.4949,
+      "step": 314
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9061681146441245e-05,
+      "loss": 0.5012,
+      "step": 315
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.905448237493147e-05,
+      "loss": 0.5188,
+      "step": 316
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.904725746439295e-05,
+      "loss": 0.4326,
+      "step": 317
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9040006435683e-05,
+      "loss": 0.5562,
+      "step": 318
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.903272930973433e-05,
+      "loss": 0.5496,
+      "step": 319
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9025426107554987e-05,
+      "loss": 0.5166,
+      "step": 320
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9018096850228297e-05,
+      "loss": 0.5222,
+      "step": 321
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9010741558912803e-05,
+      "loss": 0.4668,
+      "step": 322
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.9003360254842207e-05,
+      "loss": 0.5735,
+      "step": 323
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8995952959325312e-05,
+      "loss": 0.5718,
+      "step": 324
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.898851969374594e-05,
+      "loss": 0.5928,
+      "step": 325
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8981060479562893e-05,
+      "loss": 0.4883,
+      "step": 326
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8973575338309884e-05,
+      "loss": 0.5156,
+      "step": 327
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8966064291595472e-05,
+      "loss": 0.5347,
+      "step": 328
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8958527361103005e-05,
+      "loss": 0.5669,
+      "step": 329
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8950964568590548e-05,
+      "loss": 0.5095,
+      "step": 330
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8943375935890834e-05,
+      "loss": 0.406,
+      "step": 331
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8935761484911183e-05,
+      "loss": 0.5049,
+      "step": 332
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.8928121237633454e-05,
+      "loss": 0.6616,
+      "step": 333
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8920455216113975e-05,
+      "loss": 0.4961,
+      "step": 334
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8912763442483484e-05,
+      "loss": 0.5432,
+      "step": 335
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8905045938947063e-05,
+      "loss": 0.5366,
+      "step": 336
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8897302727784063e-05,
+      "loss": 0.5029,
+      "step": 337
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8889533831348058e-05,
+      "loss": 0.4333,
+      "step": 338
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8881739272066772e-05,
+      "loss": 0.5208,
+      "step": 339
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8873919072442003e-05,
+      "loss": 0.4453,
+      "step": 340
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8866073255049588e-05,
+      "loss": 0.4832,
+      "step": 341
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8858201842539303e-05,
+      "loss": 0.5554,
+      "step": 342
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8850304857634817e-05,
+      "loss": 0.5388,
+      "step": 343
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.884238232313363e-05,
+      "loss": 0.5132,
+      "step": 344
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8834434261906992e-05,
+      "loss": 0.5657,
+      "step": 345
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8826460696899848e-05,
+      "loss": 0.5413,
+      "step": 346
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8818461651130767e-05,
+      "loss": 0.4563,
+      "step": 347
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8810437147691883e-05,
+      "loss": 0.4199,
+      "step": 348
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.880238720974882e-05,
+      "loss": 0.5559,
+      "step": 349
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8794311860540623e-05,
+      "loss": 0.5127,
+      "step": 350
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.87862111233797e-05,
+      "loss": 0.4189,
+      "step": 351
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.8778085021651756e-05,
+      "loss": 0.4304,
+      "step": 352
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8769933578815707e-05,
+      "loss": 0.4934,
+      "step": 353
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8761756818403634e-05,
+      "loss": 0.5442,
+      "step": 354
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8753554764020707e-05,
+      "loss": 0.4585,
+      "step": 355
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8745327439345113e-05,
+      "loss": 0.4546,
+      "step": 356
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8737074868127997e-05,
+      "loss": 0.4912,
+      "step": 357
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872879707419337e-05,
+      "loss": 0.448,
+      "step": 358
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.872049408143808e-05,
+      "loss": 0.4124,
+      "step": 359
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.87121659138317e-05,
+      "loss": 0.5342,
+      "step": 360
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8703812595416503e-05,
+      "loss": 0.5232,
+      "step": 361
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8695434150307342e-05,
+      "loss": 0.521,
+      "step": 362
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8687030602691628e-05,
+      "loss": 0.5071,
+      "step": 363
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.867860197682923e-05,
+      "loss": 0.5544,
+      "step": 364
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8670148297052416e-05,
+      "loss": 0.5161,
+      "step": 365
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.866166958776578e-05,
+      "loss": 0.4966,
+      "step": 366
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.865316587344618e-05,
+      "loss": 0.4746,
+      "step": 367
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8644637178642642e-05,
+      "loss": 0.4773,
+      "step": 368
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.863608352797633e-05,
+      "loss": 0.5391,
+      "step": 369
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.862750494614044e-05,
+      "loss": 0.5198,
+      "step": 370
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.8618901457900138e-05,
+      "loss": 0.3965,
+      "step": 371
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8610273088092497e-05,
+      "loss": 0.5288,
+      "step": 372
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8601619861626424e-05,
+      "loss": 0.4719,
+      "step": 373
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.859294180348257e-05,
+      "loss": 0.4128,
+      "step": 374
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.858423893871328e-05,
+      "loss": 0.4319,
+      "step": 375
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8575511292442518e-05,
+      "loss": 0.4822,
+      "step": 376
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8566758889865775e-05,
+      "loss": 0.4529,
+      "step": 377
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.855798175625002e-05,
+      "loss": 0.5144,
+      "step": 378
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.854917991693362e-05,
+      "loss": 0.4553,
+      "step": 379
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8540353397326253e-05,
+      "loss": 0.4558,
+      "step": 380
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.853150222290885e-05,
+      "loss": 0.5469,
+      "step": 381
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.852262641923352e-05,
+      "loss": 0.4849,
+      "step": 382
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8513726011923472e-05,
+      "loss": 0.5427,
+      "step": 383
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.850480102667294e-05,
+      "loss": 0.5376,
+      "step": 384
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8495851489247114e-05,
+      "loss": 0.416,
+      "step": 385
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8486877425482062e-05,
+      "loss": 0.5186,
+      "step": 386
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.847787886128465e-05,
+      "loss": 0.5881,
+      "step": 387
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8468855822632485e-05,
+      "loss": 0.4148,
+      "step": 388
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.8459808335573818e-05,
+      "loss": 0.5085,
+      "step": 389
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.845073642622749e-05,
+      "loss": 0.463,
+      "step": 390
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.844164012078283e-05,
+      "loss": 0.5513,
+      "step": 391
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8432519445499607e-05,
+      "loss": 0.4026,
+      "step": 392
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8423374426707944e-05,
+      "loss": 0.4739,
+      "step": 393
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.841420509080823e-05,
+      "loss": 0.522,
+      "step": 394
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.840501146427106e-05,
+      "loss": 0.584,
+      "step": 395
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.839579357363716e-05,
+      "loss": 0.5378,
+      "step": 396
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.838655144551729e-05,
+      "loss": 0.646,
+      "step": 397
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8377285106592185e-05,
+      "loss": 0.5391,
+      "step": 398
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8367994583612475e-05,
+      "loss": 0.457,
+      "step": 399
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8358679903398602e-05,
+      "loss": 0.5327,
+      "step": 400
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8349341092840757e-05,
+      "loss": 0.5212,
+      "step": 401
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.833997817889878e-05,
+      "loss": 0.4998,
+      "step": 402
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.83305911886021e-05,
+      "loss": 0.5913,
+      "step": 403
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.832118014904965e-05,
+      "loss": 0.4858,
+      "step": 404
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8311745087409785e-05,
+      "loss": 0.4556,
+      "step": 405
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8302286030920217e-05,
+      "loss": 0.4861,
+      "step": 406
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8292803006887923e-05,
+      "loss": 0.5417,
+      "step": 407
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8283296042689075e-05,
+      "loss": 0.5432,
+      "step": 408
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8273765165768954e-05,
+      "loss": 0.3877,
+      "step": 409
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.8264210403641866e-05,
+      "loss": 0.4636,
+      "step": 410
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.825463178389108e-05,
+      "loss": 0.532,
+      "step": 411
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8245029334168743e-05,
+      "loss": 0.4756,
+      "step": 412
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8235403082195782e-05,
+      "loss": 0.4282,
+      "step": 413
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.822575305576185e-05,
+      "loss": 0.5808,
+      "step": 414
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8216079282725225e-05,
+      "loss": 0.4351,
+      "step": 415
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.820638179101274e-05,
+      "loss": 0.4807,
+      "step": 416
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8196660608619704e-05,
+      "loss": 0.5681,
+      "step": 417
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.818691576360982e-05,
+      "loss": 0.5706,
+      "step": 418
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8177147284115092e-05,
+      "loss": 0.5127,
+      "step": 419
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8167355198335758e-05,
+      "loss": 0.4939,
+      "step": 420
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8157539534540213e-05,
+      "loss": 0.4507,
+      "step": 421
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8147700321064897e-05,
+      "loss": 0.4858,
+      "step": 422
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8137837586314264e-05,
+      "loss": 0.5237,
+      "step": 423
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.812795135876064e-05,
+      "loss": 0.5688,
+      "step": 424
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8118041666944202e-05,
+      "loss": 0.4485,
+      "step": 425
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8108108539472833e-05,
+      "loss": 0.4827,
+      "step": 426
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8098152005022097e-05,
+      "loss": 0.449,
+      "step": 427
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.808817209233512e-05,
+      "loss": 0.511,
+      "step": 428
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.8078168830222516e-05,
+      "loss": 0.5134,
+      "step": 429
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.806814224756231e-05,
+      "loss": 0.5806,
+      "step": 430
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8058092373299855e-05,
+      "loss": 0.5654,
+      "step": 431
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.804801923644773e-05,
+      "loss": 0.6431,
+      "step": 432
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.803792286608568e-05,
+      "loss": 0.5322,
+      "step": 433
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8027803291360515e-05,
+      "loss": 0.4814,
+      "step": 434
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.801766054148604e-05,
+      "loss": 0.377,
+      "step": 435
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.8007494645742967e-05,
+      "loss": 0.5122,
+      "step": 436
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.799730563347881e-05,
+      "loss": 0.562,
+      "step": 437
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.798709353410783e-05,
+      "loss": 0.4299,
+      "step": 438
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.797685837711094e-05,
+      "loss": 0.522,
+      "step": 439
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7966600192035605e-05,
+      "loss": 0.6135,
+      "step": 440
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.795631900849578e-05,
+      "loss": 0.5466,
+      "step": 441
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7946014856171805e-05,
+      "loss": 0.4539,
+      "step": 442
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7935687764810337e-05,
+      "loss": 0.426,
+      "step": 443
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7925337764224246e-05,
+      "loss": 0.5083,
+      "step": 444
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7914964884292543e-05,
+      "loss": 0.5308,
+      "step": 445
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7904569154960287e-05,
+      "loss": 0.5879,
+      "step": 446
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7894150606238498e-05,
+      "loss": 0.4614,
+      "step": 447
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.7883709268204073e-05,
+      "loss": 0.446,
+      "step": 448
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.78732451709997e-05,
+      "loss": 0.501,
+      "step": 449
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7862758344833773e-05,
+      "loss": 0.5305,
+      "step": 450
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7852248819980297e-05,
+      "loss": 0.5237,
+      "step": 451
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7841716626778803e-05,
+      "loss": 0.5054,
+      "step": 452
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.783116179563427e-05,
+      "loss": 0.5212,
+      "step": 453
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7820584357017018e-05,
+      "loss": 0.4338,
+      "step": 454
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7809984341462637e-05,
+      "loss": 0.4702,
+      "step": 455
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7799361779571904e-05,
+      "loss": 0.4946,
+      "step": 456
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7788716702010666e-05,
+      "loss": 0.4214,
+      "step": 457
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7778049139509785e-05,
+      "loss": 0.5007,
+      "step": 458
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.776735912286502e-05,
+      "loss": 0.5144,
+      "step": 459
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7756646682936964e-05,
+      "loss": 0.5037,
+      "step": 460
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7745911850650942e-05,
+      "loss": 0.4722,
+      "step": 461
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7735154656996908e-05,
+      "loss": 0.5432,
+      "step": 462
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7724375133029388e-05,
+      "loss": 0.5947,
+      "step": 463
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7713573309867365e-05,
+      "loss": 0.5254,
+      "step": 464
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7702749218694196e-05,
+      "loss": 0.4834,
+      "step": 465
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7691902890757528e-05,
+      "loss": 0.5486,
+      "step": 466
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.7681034357369193e-05,
+      "loss": 0.5054,
+      "step": 467
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.767014364990513e-05,
+      "loss": 0.5518,
+      "step": 468
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7659230799805298e-05,
+      "loss": 0.5576,
+      "step": 469
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7648295838573576e-05,
+      "loss": 0.49,
+      "step": 470
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.763733879777767e-05,
+      "loss": 0.4902,
+      "step": 471
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.762635970904903e-05,
+      "loss": 0.5256,
+      "step": 472
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7615358604082752e-05,
+      "loss": 0.3926,
+      "step": 473
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7604335514637502e-05,
+      "loss": 0.5217,
+      "step": 474
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7593290472535394e-05,
+      "loss": 0.5896,
+      "step": 475
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7582223509661932e-05,
+      "loss": 0.4246,
+      "step": 476
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7571134657965888e-05,
+      "loss": 0.4932,
+      "step": 477
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.756002394945924e-05,
+      "loss": 0.4717,
+      "step": 478
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7548891416217048e-05,
+      "loss": 0.5122,
+      "step": 479
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7537737090377383e-05,
+      "loss": 0.5938,
+      "step": 480
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7526561004141235e-05,
+      "loss": 0.5291,
+      "step": 481
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7515363189772406e-05,
+      "loss": 0.5005,
+      "step": 482
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7504143679597425e-05,
+      "loss": 0.5273,
+      "step": 483
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7492902506005457e-05,
+      "loss": 0.5437,
+      "step": 484
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7481639701448196e-05,
+      "loss": 0.5027,
+      "step": 485
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.7470355298439795e-05,
+      "loss": 0.4922,
+      "step": 486
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.745904932955675e-05,
+      "loss": 0.3771,
+      "step": 487
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.744772182743782e-05,
+      "loss": 0.4954,
+      "step": 488
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7436372824783924e-05,
+      "loss": 0.4922,
+      "step": 489
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.742500235435805e-05,
+      "loss": 0.5483,
+      "step": 490
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7413610448985165e-05,
+      "loss": 0.5752,
+      "step": 491
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.740219714155211e-05,
+      "loss": 0.5374,
+      "step": 492
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.739076246500751e-05,
+      "loss": 0.5754,
+      "step": 493
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7379306452361684e-05,
+      "loss": 0.4453,
+      "step": 494
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7367829136686546e-05,
+      "loss": 0.5498,
+      "step": 495
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7356330551115506e-05,
+      "loss": 0.512,
+      "step": 496
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.734481072884338e-05,
+      "loss": 0.501,
+      "step": 497
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7333269703126285e-05,
+      "loss": 0.4712,
+      "step": 498
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7321707507281557e-05,
+      "loss": 0.4502,
+      "step": 499
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.731012417468764e-05,
+      "loss": 0.4983,
+      "step": 500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7298519738784006e-05,
+      "loss": 0.5029,
+      "step": 501
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.728689423307105e-05,
+      "loss": 0.4187,
+      "step": 502
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.727524769110997e-05,
+      "loss": 0.4946,
+      "step": 503
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7263580146522718e-05,
+      "loss": 0.5317,
+      "step": 504
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.7251891632991874e-05,
+      "loss": 0.3818,
+      "step": 505
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7240182184260544e-05,
+      "loss": 0.4711,
+      "step": 506
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7228451834132275e-05,
+      "loss": 0.4727,
+      "step": 507
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7216700616470946e-05,
+      "loss": 0.5127,
+      "step": 508
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7204928565200693e-05,
+      "loss": 0.4185,
+      "step": 509
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.719313571430579e-05,
+      "loss": 0.5691,
+      "step": 510
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7181322097830545e-05,
+      "loss": 0.5149,
+      "step": 511
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7169487749879227e-05,
+      "loss": 0.4668,
+      "step": 512
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.715763270461595e-05,
+      "loss": 0.4722,
+      "step": 513
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7145756996264575e-05,
+      "loss": 0.5146,
+      "step": 514
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7133860659108618e-05,
+      "loss": 0.5688,
+      "step": 515
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7121943727491147e-05,
+      "loss": 0.4449,
+      "step": 516
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.711000623581468e-05,
+      "loss": 0.5205,
+      "step": 517
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.70980482185411e-05,
+      "loss": 0.5232,
+      "step": 518
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7086069710191526e-05,
+      "loss": 0.5525,
+      "step": 519
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7074070745346247e-05,
+      "loss": 0.3424,
+      "step": 520
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7062051358644605e-05,
+      "loss": 0.4895,
+      "step": 521
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.705001158478489e-05,
+      "loss": 0.4031,
+      "step": 522
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7037951458524255e-05,
+      "loss": 0.4792,
+      "step": 523
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7025871014678596e-05,
+      "loss": 0.5002,
+      "step": 524
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.701377028812248e-05,
+      "loss": 0.4946,
+      "step": 525
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.7001649313789014e-05,
+      "loss": 0.5068,
+      "step": 526
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.698950812666976e-05,
+      "loss": 0.4666,
+      "step": 527
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6977346761814636e-05,
+      "loss": 0.5127,
+      "step": 528
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.696516525433181e-05,
+      "loss": 0.4817,
+      "step": 529
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6952963639387588e-05,
+      "loss": 0.4556,
+      "step": 530
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6940741952206342e-05,
+      "loss": 0.5732,
+      "step": 531
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.692850022807037e-05,
+      "loss": 0.4929,
+      "step": 532
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.691623850231983e-05,
+      "loss": 0.4756,
+      "step": 533
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.690395681035261e-05,
+      "loss": 0.4578,
+      "step": 534
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6891655187624255e-05,
+      "loss": 0.469,
+      "step": 535
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.687933366964782e-05,
+      "loss": 0.5437,
+      "step": 536
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.686699229199382e-05,
+      "loss": 0.4949,
+      "step": 537
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6854631090290088e-05,
+      "loss": 0.5107,
+      "step": 538
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6842250100221693e-05,
+      "loss": 0.4351,
+      "step": 539
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.682984935753083e-05,
+      "loss": 0.5259,
+      "step": 540
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6817428898016707e-05,
+      "loss": 0.5127,
+      "step": 541
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6804988757535463e-05,
+      "loss": 0.4661,
+      "step": 542
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.6792528972000054e-05,
+      "loss": 0.5374,
+      "step": 543
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6780049577380143e-05,
+      "loss": 0.4871,
+      "step": 544
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6767550609702e-05,
+      "loss": 0.4978,
+      "step": 545
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.675503210504841e-05,
+      "loss": 0.4329,
+      "step": 546
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.674249409955854e-05,
+      "loss": 0.4546,
+      "step": 547
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.672993662942787e-05,
+      "loss": 0.4954,
+      "step": 548
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6717359730908072e-05,
+      "loss": 0.4353,
+      "step": 549
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.670476344030689e-05,
+      "loss": 0.5588,
+      "step": 550
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6692147793988062e-05,
+      "loss": 0.4351,
+      "step": 551
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6679512828371203e-05,
+      "loss": 0.4619,
+      "step": 552
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6666858579931702e-05,
+      "loss": 0.5442,
+      "step": 553
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6654185085200597e-05,
+      "loss": 0.4331,
+      "step": 554
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6641492380764516e-05,
+      "loss": 0.3738,
+      "step": 555
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.662878050326552e-05,
+      "loss": 0.5098,
+      "step": 556
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6616049489401026e-05,
+      "loss": 0.46,
+      "step": 557
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6603299375923706e-05,
+      "loss": 0.4753,
+      "step": 558
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.659053019964135e-05,
+      "loss": 0.4092,
+      "step": 559
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6577741997416797e-05,
+      "loss": 0.4778,
+      "step": 560
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6564934806167806e-05,
+      "loss": 0.4175,
+      "step": 561
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.6552108662866943e-05,
+      "loss": 0.5076,
+      "step": 562
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.653926360454151e-05,
+      "loss": 0.4116,
+      "step": 563
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6526399668273388e-05,
+      "loss": 0.4314,
+      "step": 564
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6513516891198974e-05,
+      "loss": 0.5569,
+      "step": 565
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.650061531050905e-05,
+      "loss": 0.5006,
+      "step": 566
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6487694963448683e-05,
+      "loss": 0.4319,
+      "step": 567
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.647475588731711e-05,
+      "loss": 0.4019,
+      "step": 568
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6461798119467648e-05,
+      "loss": 0.4868,
+      "step": 569
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6448821697307556e-05,
+      "loss": 0.4783,
+      "step": 570
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6435826658297964e-05,
+      "loss": 0.4714,
+      "step": 571
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6422813039953735e-05,
+      "loss": 0.5579,
+      "step": 572
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6409780879843377e-05,
+      "loss": 0.481,
+      "step": 573
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6396730215588913e-05,
+      "loss": 0.5488,
+      "step": 574
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.63836610848658e-05,
+      "loss": 0.5037,
+      "step": 575
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6370573525402785e-05,
+      "loss": 0.4749,
+      "step": 576
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.635746757498184e-05,
+      "loss": 0.5146,
+      "step": 577
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6344343271438013e-05,
+      "loss": 0.5112,
+      "step": 578
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6331200652659337e-05,
+      "loss": 0.5308,
+      "step": 579
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6318039756586724e-05,
+      "loss": 0.5596,
+      "step": 580
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.6304860621213846e-05,
+      "loss": 0.4517,
+      "step": 581
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6291663284587028e-05,
+      "loss": 0.5049,
+      "step": 582
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.627844778480514e-05,
+      "loss": 0.4131,
+      "step": 583
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.62652141600195e-05,
+      "loss": 0.4492,
+      "step": 584
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6251962448433722e-05,
+      "loss": 0.5776,
+      "step": 585
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6238692688303664e-05,
+      "loss": 0.4814,
+      "step": 586
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6225404917937265e-05,
+      "loss": 0.4751,
+      "step": 587
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.621209917569447e-05,
+      "loss": 0.5156,
+      "step": 588
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6198775499987106e-05,
+      "loss": 0.4504,
+      "step": 589
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.618543392927876e-05,
+      "loss": 0.4565,
+      "step": 590
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6172074502084695e-05,
+      "loss": 0.4265,
+      "step": 591
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.615869725697171e-05,
+      "loss": 0.4917,
+      "step": 592
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6145302232558047e-05,
+      "loss": 0.5327,
+      "step": 593
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6131889467513283e-05,
+      "loss": 0.3975,
+      "step": 594
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.611845900055819e-05,
+      "loss": 0.4875,
+      "step": 595
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.610501087046466e-05,
+      "loss": 0.4438,
+      "step": 596
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6091545116055572e-05,
+      "loss": 0.6289,
+      "step": 597
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.607806177620468e-05,
+      "loss": 0.3948,
+      "step": 598
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.606456088983651e-05,
+      "loss": 0.4385,
+      "step": 599
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.6051042495926235e-05,
+      "loss": 0.4441,
+      "step": 600
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.6037506633499573e-05,
+      "loss": 0.4761,
+      "step": 601
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.602395334163268e-05,
+      "loss": 0.4192,
+      "step": 602
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.601038265945201e-05,
+      "loss": 0.4041,
+      "step": 603
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5996794626134245e-05,
+      "loss": 0.5747,
+      "step": 604
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5983189280906134e-05,
+      "loss": 0.4561,
+      "step": 605
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5969566663044412e-05,
+      "loss": 0.4905,
+      "step": 606
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.595592681187568e-05,
+      "loss": 0.4941,
+      "step": 607
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.594226976677629e-05,
+      "loss": 0.4651,
+      "step": 608
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.592859556717223e-05,
+      "loss": 0.5488,
+      "step": 609
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5914904252539e-05,
+      "loss": 0.4768,
+      "step": 610
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5901195862401522e-05,
+      "loss": 0.4785,
+      "step": 611
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5887470436334008e-05,
+      "loss": 0.4597,
+      "step": 612
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5873728013959852e-05,
+      "loss": 0.4922,
+      "step": 613
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.585996863495151e-05,
+      "loss": 0.4221,
+      "step": 614
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5846192339030388e-05,
+      "loss": 0.4939,
+      "step": 615
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.583239916596674e-05,
+      "loss": 0.5679,
+      "step": 616
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.581858915557953e-05,
+      "loss": 0.5046,
+      "step": 617
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5804762347736333e-05,
+      "loss": 0.5479,
+      "step": 618
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.5790918782353217e-05,
+      "loss": 0.3948,
+      "step": 619
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5777058499394627e-05,
+      "loss": 0.5198,
+      "step": 620
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5763181538873267e-05,
+      "loss": 0.5574,
+      "step": 621
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.574928794084999e-05,
+      "loss": 0.5627,
+      "step": 622
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5735377745433672e-05,
+      "loss": 0.4937,
+      "step": 623
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5721450992781115e-05,
+      "loss": 0.4365,
+      "step": 624
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5707507723096913e-05,
+      "loss": 0.4312,
+      "step": 625
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5693547976633335e-05,
+      "loss": 0.479,
+      "step": 626
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5679571793690233e-05,
+      "loss": 0.5496,
+      "step": 627
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.566557921461489e-05,
+      "loss": 0.5608,
+      "step": 628
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5651570279801934e-05,
+      "loss": 0.5315,
+      "step": 629
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5637545029693218e-05,
+      "loss": 0.4355,
+      "step": 630
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.562350350477767e-05,
+      "loss": 0.5408,
+      "step": 631
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5609445745591224e-05,
+      "loss": 0.4241,
+      "step": 632
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.559537179271667e-05,
+      "loss": 0.3994,
+      "step": 633
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5581281686783544e-05,
+      "loss": 0.4812,
+      "step": 634
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5567175468468025e-05,
+      "loss": 0.4641,
+      "step": 635
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5553053178492797e-05,
+      "loss": 0.4487,
+      "step": 636
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.553891485762694e-05,
+      "loss": 0.4805,
+      "step": 637
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.5524760546685816e-05,
+      "loss": 0.4421,
+      "step": 638
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5510590286530954e-05,
+      "loss": 0.4978,
+      "step": 639
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5496404118069922e-05,
+      "loss": 0.5864,
+      "step": 640
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.54822020822562e-05,
+      "loss": 0.4666,
+      "step": 641
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.546798422008909e-05,
+      "loss": 0.5259,
+      "step": 642
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5453750572613586e-05,
+      "loss": 0.5034,
+      "step": 643
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5439501180920237e-05,
+      "loss": 0.5154,
+      "step": 644
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.542523608614505e-05,
+      "loss": 0.4409,
+      "step": 645
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5410955329469374e-05,
+      "loss": 0.4883,
+      "step": 646
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5396658952119757e-05,
+      "loss": 0.4805,
+      "step": 647
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.538234699536785e-05,
+      "loss": 0.4546,
+      "step": 648
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5368019500530272e-05,
+      "loss": 0.4844,
+      "step": 649
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5353676508968514e-05,
+      "loss": 0.3633,
+      "step": 650
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5339318062088785e-05,
+      "loss": 0.394,
+      "step": 651
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5324944201341924e-05,
+      "loss": 0.4795,
+      "step": 652
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5310554968223263e-05,
+      "loss": 0.4316,
+      "step": 653
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5296150404272513e-05,
+      "loss": 0.4365,
+      "step": 654
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.528173055107364e-05,
+      "loss": 0.4785,
+      "step": 655
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.526729545025476e-05,
+      "loss": 0.5205,
+      "step": 656
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.5252845143487981e-05,
+      "loss": 0.448,
+      "step": 657
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5238379672489342e-05,
+      "loss": 0.5259,
+      "step": 658
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5223899079018629e-05,
+      "loss": 0.4092,
+      "step": 659
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5209403404879305e-05,
+      "loss": 0.5068,
+      "step": 660
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5194892691918357e-05,
+      "loss": 0.4265,
+      "step": 661
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.518036698202619e-05,
+      "loss": 0.5776,
+      "step": 662
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5165826317136509e-05,
+      "loss": 0.5054,
+      "step": 663
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5151270739226185e-05,
+      "loss": 0.5427,
+      "step": 664
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.513670029031514e-05,
+      "loss": 0.4417,
+      "step": 665
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5122115012466234e-05,
+      "loss": 0.5234,
+      "step": 666
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5107514947785127e-05,
+      "loss": 0.439,
+      "step": 667
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5092900138420175e-05,
+      "loss": 0.469,
+      "step": 668
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5078270626562293e-05,
+      "loss": 0.4763,
+      "step": 669
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5063626454444844e-05,
+      "loss": 0.4287,
+      "step": 670
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5048967664343512e-05,
+      "loss": 0.4443,
+      "step": 671
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5034294298576183e-05,
+      "loss": 0.5049,
+      "step": 672
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.501960639950281e-05,
+      "loss": 0.45,
+      "step": 673
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.5004904009525328e-05,
+      "loss": 0.4749,
+      "step": 674
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.4990187171087469e-05,
+      "loss": 0.363,
+      "step": 675
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.497545592667471e-05,
+      "loss": 0.3958,
+      "step": 676
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4960710318814087e-05,
+      "loss": 0.5103,
+      "step": 677
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4945950390074124e-05,
+      "loss": 0.4653,
+      "step": 678
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4931176183064677e-05,
+      "loss": 0.4341,
+      "step": 679
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4916387740436822e-05,
+      "loss": 0.4675,
+      "step": 680
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.490158510488273e-05,
+      "loss": 0.4702,
+      "step": 681
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.488676831913555e-05,
+      "loss": 0.479,
+      "step": 682
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4871937425969271e-05,
+      "loss": 0.4453,
+      "step": 683
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4857092468198626e-05,
+      "loss": 0.5593,
+      "step": 684
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4842233488678927e-05,
+      "loss": 0.4204,
+      "step": 685
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4827360530305987e-05,
+      "loss": 0.4534,
+      "step": 686
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4812473636015958e-05,
+      "loss": 0.4644,
+      "step": 687
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.479757284878523e-05,
+      "loss": 0.5037,
+      "step": 688
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4782658211630301e-05,
+      "loss": 0.449,
+      "step": 689
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.476772976760765e-05,
+      "loss": 0.4521,
+      "step": 690
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4752787559813612e-05,
+      "loss": 0.4949,
+      "step": 691
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.473783163138426e-05,
+      "loss": 0.5129,
+      "step": 692
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4722862025495275e-05,
+      "loss": 0.5208,
+      "step": 693
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4707878785361822e-05,
+      "loss": 0.4705,
+      "step": 694
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4692881954238429e-05,
+      "loss": 0.4761,
+      "step": 695
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.4677871575418854e-05,
+      "loss": 0.4021,
+      "step": 696
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4662847692235973e-05,
+      "loss": 0.4675,
+      "step": 697
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4647810348061642e-05,
+      "loss": 0.4622,
+      "step": 698
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4632759586306575e-05,
+      "loss": 0.4858,
+      "step": 699
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4617695450420226e-05,
+      "loss": 0.4294,
+      "step": 700
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4602617983890657e-05,
+      "loss": 0.509,
+      "step": 701
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.458752723024441e-05,
+      "loss": 0.4382,
+      "step": 702
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4572423233046386e-05,
+      "loss": 0.4966,
+      "step": 703
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4557306035899728e-05,
+      "loss": 0.4866,
+      "step": 704
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.454217568244567e-05,
+      "loss": 0.4351,
+      "step": 705
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4527032216363432e-05,
+      "loss": 0.5117,
+      "step": 706
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.45118756813701e-05,
+      "loss": 0.4919,
+      "step": 707
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4496706121220467e-05,
+      "loss": 0.425,
+      "step": 708
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4481523579706947e-05,
+      "loss": 0.4639,
+      "step": 709
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4466328100659417e-05,
+      "loss": 0.4221,
+      "step": 710
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.445111972794511e-05,
+      "loss": 0.4644,
+      "step": 711
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.443589850546848e-05,
+      "loss": 0.3706,
+      "step": 712
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.442066447717107e-05,
+      "loss": 0.4758,
+      "step": 713
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4405417687031396e-05,
+      "loss": 0.4719,
+      "step": 714
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.4390158179064827e-05,
+      "loss": 0.4143,
+      "step": 715
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.437488599732342e-05,
+      "loss": 0.5071,
+      "step": 716
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4359601185895851e-05,
+      "loss": 0.533,
+      "step": 717
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4344303788907228e-05,
+      "loss": 0.51,
+      "step": 718
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4328993850519008e-05,
+      "loss": 0.4905,
+      "step": 719
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4313671414928854e-05,
+      "loss": 0.4871,
+      "step": 720
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4298336526370497e-05,
+      "loss": 0.54,
+      "step": 721
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4282989229113627e-05,
+      "loss": 0.4746,
+      "step": 722
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4267629567463748e-05,
+      "loss": 0.4558,
+      "step": 723
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4252257585762067e-05,
+      "loss": 0.499,
+      "step": 724
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4236873328385348e-05,
+      "loss": 0.3541,
+      "step": 725
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.42214768397458e-05,
+      "loss": 0.4888,
+      "step": 726
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4206068164290944e-05,
+      "loss": 0.4246,
+      "step": 727
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4190647346503475e-05,
+      "loss": 0.4766,
+      "step": 728
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4175214430901143e-05,
+      "loss": 0.4988,
+      "step": 729
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4159769462036628e-05,
+      "loss": 0.4619,
+      "step": 730
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.41443124844974e-05,
+      "loss": 0.4512,
+      "step": 731
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4128843542905597e-05,
+      "loss": 0.4585,
+      "step": 732
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.4113362681917908e-05,
+      "loss": 0.5164,
+      "step": 733
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.409786994622541e-05,
+      "loss": 0.4602,
+      "step": 734
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4082365380553475e-05,
+      "loss": 0.4283,
+      "step": 735
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4066849029661625e-05,
+      "loss": 0.47,
+      "step": 736
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4051320938343402e-05,
+      "loss": 0.4861,
+      "step": 737
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4035781151426242e-05,
+      "loss": 0.5144,
+      "step": 738
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4020229713771343e-05,
+      "loss": 0.5288,
+      "step": 739
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.4004666670273536e-05,
+      "loss": 0.4705,
+      "step": 740
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3989092065861165e-05,
+      "loss": 0.4685,
+      "step": 741
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3973505945495936e-05,
+      "loss": 0.5034,
+      "step": 742
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3957908354172811e-05,
+      "loss": 0.4856,
+      "step": 743
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3942299336919858e-05,
+      "loss": 0.5183,
+      "step": 744
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.392667893879814e-05,
+      "loss": 0.45,
+      "step": 745
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.391104720490156e-05,
+      "loss": 0.469,
+      "step": 746
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3895404180356766e-05,
+      "loss": 0.426,
+      "step": 747
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3879749910322986e-05,
+      "loss": 0.5857,
+      "step": 748
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3864084439991914e-05,
+      "loss": 0.4131,
+      "step": 749
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.384840781458758e-05,
+      "loss": 0.5393,
+      "step": 750
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.383272007936622e-05,
+      "loss": 0.4326,
+      "step": 751
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3817021279616132e-05,
+      "loss": 0.3905,
+      "step": 752
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.3801311460657574e-05,
+      "loss": 0.509,
+      "step": 753
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3785590667842592e-05,
+      "loss": 0.481,
+      "step": 754
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3769858946554933e-05,
+      "loss": 0.4861,
+      "step": 755
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.375411634220988e-05,
+      "loss": 0.4609,
+      "step": 756
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3738362900254138e-05,
+      "loss": 0.4353,
+      "step": 757
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3722598666165701e-05,
+      "loss": 0.5627,
+      "step": 758
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3706823685453712e-05,
+      "loss": 0.4705,
+      "step": 759
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3691038003658346e-05,
+      "loss": 0.3962,
+      "step": 760
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.367524166635066e-05,
+      "loss": 0.3701,
+      "step": 761
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3659434719132485e-05,
+      "loss": 0.3884,
+      "step": 762
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.364361720763627e-05,
+      "loss": 0.488,
+      "step": 763
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.362778917752497e-05,
+      "loss": 0.4658,
+      "step": 764
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3611950674491898e-05,
+      "loss": 0.4766,
+      "step": 765
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3596101744260608e-05,
+      "loss": 0.4929,
+      "step": 766
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3580242432584753e-05,
+      "loss": 0.385,
+      "step": 767
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3564372785247955e-05,
+      "loss": 0.4861,
+      "step": 768
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3548492848063675e-05,
+      "loss": 0.4778,
+      "step": 769
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3532602666875074e-05,
+      "loss": 0.4812,
+      "step": 770
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3516702287554896e-05,
+      "loss": 0.5017,
+      "step": 771
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.3500791756005318e-05,
+      "loss": 0.4543,
+      "step": 772
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3484871118157826e-05,
+      "loss": 0.4805,
+      "step": 773
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3468940419973086e-05,
+      "loss": 0.4404,
+      "step": 774
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.34529997074408e-05,
+      "loss": 0.4653,
+      "step": 775
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3437049026579587e-05,
+      "loss": 0.4771,
+      "step": 776
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3421088423436834e-05,
+      "loss": 0.3882,
+      "step": 777
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3405117944088584e-05,
+      "loss": 0.433,
+      "step": 778
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3389137634639381e-05,
+      "loss": 0.4148,
+      "step": 779
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3373147541222155e-05,
+      "loss": 0.4526,
+      "step": 780
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3357147709998075e-05,
+      "loss": 0.47,
+      "step": 781
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3341138187156423e-05,
+      "loss": 0.4031,
+      "step": 782
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3325119018914463e-05,
+      "loss": 0.5305,
+      "step": 783
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.33090902515173e-05,
+      "loss": 0.4749,
+      "step": 784
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.329305193123775e-05,
+      "loss": 0.5273,
+      "step": 785
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.327700410437621e-05,
+      "loss": 0.4722,
+      "step": 786
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.326094681726052e-05,
+      "loss": 0.4595,
+      "step": 787
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3244880116245829e-05,
+      "loss": 0.5415,
+      "step": 788
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3228804047714462e-05,
+      "loss": 0.5168,
+      "step": 789
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3212718658075794e-05,
+      "loss": 0.4312,
+      "step": 790
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.3196623993766097e-05,
+      "loss": 0.4871,
+      "step": 791
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.318052010124843e-05,
+      "loss": 0.399,
+      "step": 792
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.316440702701248e-05,
+      "loss": 0.4785,
+      "step": 793
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3148284817574455e-05,
+      "loss": 0.4922,
+      "step": 794
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3132153519476915e-05,
+      "loss": 0.5405,
+      "step": 795
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3116013179288677e-05,
+      "loss": 0.4858,
+      "step": 796
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3099863843604656e-05,
+      "loss": 0.4438,
+      "step": 797
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.308370555904573e-05,
+      "loss": 0.4954,
+      "step": 798
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3067538372258612e-05,
+      "loss": 0.4153,
+      "step": 799
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3051362329915722e-05,
+      "loss": 0.4355,
+      "step": 800
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3035177478715035e-05,
+      "loss": 0.5127,
+      "step": 801
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3018983865379968e-05,
+      "loss": 0.5105,
+      "step": 802
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.3002781536659216e-05,
+      "loss": 0.4573,
+      "step": 803
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2986570539326655e-05,
+      "loss": 0.4871,
+      "step": 804
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2970350920181164e-05,
+      "loss": 0.4237,
+      "step": 805
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2954122726046534e-05,
+      "loss": 0.4373,
+      "step": 806
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2937886003771293e-05,
+      "loss": 0.4814,
+      "step": 807
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2921640800228596e-05,
+      "loss": 0.418,
+      "step": 808
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.2905387162316081e-05,
+      "loss": 0.4199,
+      "step": 809
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.288912513695574e-05,
+      "loss": 0.3904,
+      "step": 810
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2872854771093766e-05,
+      "loss": 0.4058,
+      "step": 811
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.285657611170045e-05,
+      "loss": 0.5703,
+      "step": 812
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2840289205769998e-05,
+      "loss": 0.4395,
+      "step": 813
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2823994100320448e-05,
+      "loss": 0.4312,
+      "step": 814
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2807690842393494e-05,
+      "loss": 0.4331,
+      "step": 815
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2791379479054373e-05,
+      "loss": 0.4268,
+      "step": 816
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2775060057391716e-05,
+      "loss": 0.4888,
+      "step": 817
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2758732624517419e-05,
+      "loss": 0.3289,
+      "step": 818
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2742397227566503e-05,
+      "loss": 0.3782,
+      "step": 819
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2726053913696988e-05,
+      "loss": 0.4319,
+      "step": 820
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2709702730089737e-05,
+      "loss": 0.3967,
+      "step": 821
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2693343723948342e-05,
+      "loss": 0.5242,
+      "step": 822
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2676976942498971e-05,
+      "loss": 0.397,
+      "step": 823
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2660602432990245e-05,
+      "loss": 0.488,
+      "step": 824
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2644220242693088e-05,
+      "loss": 0.441,
+      "step": 825
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2627830418900601e-05,
+      "loss": 0.382,
+      "step": 826
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2611433008927925e-05,
+      "loss": 0.4426,
+      "step": 827
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2595028060112094e-05,
+      "loss": 0.4026,
+      "step": 828
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.2578615619811905e-05,
+      "loss": 0.4902,
+      "step": 829
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2562195735407792e-05,
+      "loss": 0.3938,
+      "step": 830
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.254576845430167e-05,
+      "loss": 0.498,
+      "step": 831
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2529333823916807e-05,
+      "loss": 0.4968,
+      "step": 832
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.25128918916977e-05,
+      "loss": 0.4446,
+      "step": 833
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2496442705109908e-05,
+      "loss": 0.4683,
+      "step": 834
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2479986311639943e-05,
+      "loss": 0.4924,
+      "step": 835
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2463522758795117e-05,
+      "loss": 0.4858,
+      "step": 836
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2447052094103416e-05,
+      "loss": 0.4998,
+      "step": 837
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2430574365113352e-05,
+      "loss": 0.4688,
+      "step": 838
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2414089619393832e-05,
+      "loss": 0.4436,
+      "step": 839
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.239759790453402e-05,
+      "loss": 0.4683,
+      "step": 840
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2381099268143199e-05,
+      "loss": 0.4775,
+      "step": 841
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.236459375785063e-05,
+      "loss": 0.5293,
+      "step": 842
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2348081421305427e-05,
+      "loss": 0.4592,
+      "step": 843
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2331562306176395e-05,
+      "loss": 0.4893,
+      "step": 844
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2315036460151926e-05,
+      "loss": 0.4421,
+      "step": 845
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2298503930939829e-05,
+      "loss": 0.4089,
+      "step": 846
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2281964766267216e-05,
+      "loss": 0.4661,
+      "step": 847
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.2265419013880346e-05,
+      "loss": 0.4858,
+      "step": 848
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2248866721544505e-05,
+      "loss": 0.5171,
+      "step": 849
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2232307937043848e-05,
+      "loss": 0.5017,
+      "step": 850
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2215742708181286e-05,
+      "loss": 0.5122,
+      "step": 851
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2199171082778318e-05,
+      "loss": 0.5869,
+      "step": 852
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2182593108674925e-05,
+      "loss": 0.4263,
+      "step": 853
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2166008833729399e-05,
+      "loss": 0.5693,
+      "step": 854
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2149418305818237e-05,
+      "loss": 0.4714,
+      "step": 855
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.213282157283598e-05,
+      "loss": 0.4226,
+      "step": 856
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2116218682695083e-05,
+      "loss": 0.4604,
+      "step": 857
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2099609683325775e-05,
+      "loss": 0.4204,
+      "step": 858
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2082994622675922e-05,
+      "loss": 0.5061,
+      "step": 859
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2066373548710887e-05,
+      "loss": 0.4827,
+      "step": 860
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2049746509413402e-05,
+      "loss": 0.4004,
+      "step": 861
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2033113552783402e-05,
+      "loss": 0.5757,
+      "step": 862
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.2016474726837921e-05,
+      "loss": 0.429,
+      "step": 863
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1999830079610928e-05,
+      "loss": 0.479,
+      "step": 864
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1983179659153203e-05,
+      "loss": 0.3772,
+      "step": 865
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.1966523513532187e-05,
+      "loss": 0.4521,
+      "step": 866
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.194986169083185e-05,
+      "loss": 0.5205,
+      "step": 867
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1933194239152553e-05,
+      "loss": 0.4956,
+      "step": 868
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.191652120661091e-05,
+      "loss": 0.4294,
+      "step": 869
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1899842641339635e-05,
+      "loss": 0.4049,
+      "step": 870
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1883158591487429e-05,
+      "loss": 0.3838,
+      "step": 871
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1866469105218811e-05,
+      "loss": 0.3931,
+      "step": 872
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1849774230714011e-05,
+      "loss": 0.4578,
+      "step": 873
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1833074016168798e-05,
+      "loss": 0.3694,
+      "step": 874
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1816368509794365e-05,
+      "loss": 0.5559,
+      "step": 875
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1799657759817184e-05,
+      "loss": 0.4978,
+      "step": 876
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1782941814478858e-05,
+      "loss": 0.3802,
+      "step": 877
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1766220722035993e-05,
+      "loss": 0.46,
+      "step": 878
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1749494530760052e-05,
+      "loss": 0.4495,
+      "step": 879
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1732763288937219e-05,
+      "loss": 0.457,
+      "step": 880
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1716027044868257e-05,
+      "loss": 0.4336,
+      "step": 881
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.169928584686837e-05,
+      "loss": 0.4224,
+      "step": 882
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1682539743267063e-05,
+      "loss": 0.5347,
+      "step": 883
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1665788782408006e-05,
+      "loss": 0.4463,
+      "step": 884
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1649033012648885e-05,
+      "loss": 0.5479,
+      "step": 885
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.1632272482361273e-05,
+      "loss": 0.4082,
+      "step": 886
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1615507239930487e-05,
+      "loss": 0.4846,
+      "step": 887
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1598737333755443e-05,
+      "loss": 0.5002,
+      "step": 888
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1581962812248524e-05,
+      "loss": 0.5603,
+      "step": 889
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1565183723835436e-05,
+      "loss": 0.4377,
+      "step": 890
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1548400116955065e-05,
+      "loss": 0.4893,
+      "step": 891
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1531612040059355e-05,
+      "loss": 0.4626,
+      "step": 892
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1514819541613133e-05,
+      "loss": 0.4075,
+      "step": 893
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1498022670094011e-05,
+      "loss": 0.4761,
+      "step": 894
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.148122147399221e-05,
+      "loss": 0.5571,
+      "step": 895
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1464416001810443e-05,
+      "loss": 0.4819,
+      "step": 896
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.144760630206377e-05,
+      "loss": 0.4714,
+      "step": 897
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1430792423279444e-05,
+      "loss": 0.468,
+      "step": 898
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1413974413996794e-05,
+      "loss": 0.5076,
+      "step": 899
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1397152322767073e-05,
+      "loss": 0.4985,
+      "step": 900
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1380326198153307e-05,
+      "loss": 0.3665,
+      "step": 901
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1363496088730177e-05,
+      "loss": 0.5176,
+      "step": 902
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.134666204308386e-05,
+      "loss": 0.467,
+      "step": 903
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.1329824109811903e-05,
+      "loss": 0.4902,
+      "step": 904
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.131298233752307e-05,
+      "loss": 0.4502,
+      "step": 905
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1296136774837214e-05,
+      "loss": 0.3838,
+      "step": 906
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1279287470385126e-05,
+      "loss": 0.4497,
+      "step": 907
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1262434472808396e-05,
+      "loss": 0.417,
+      "step": 908
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1245577830759283e-05,
+      "loss": 0.4438,
+      "step": 909
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1228717592900562e-05,
+      "loss": 0.4531,
+      "step": 910
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1211853807905388e-05,
+      "loss": 0.4426,
+      "step": 911
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1194986524457165e-05,
+      "loss": 0.4192,
+      "step": 912
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.117811579124938e-05,
+      "loss": 0.4873,
+      "step": 913
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1161241656985498e-05,
+      "loss": 0.4792,
+      "step": 914
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1144364170378784e-05,
+      "loss": 0.4504,
+      "step": 915
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1127483380152198e-05,
+      "loss": 0.5183,
+      "step": 916
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1110599335038221e-05,
+      "loss": 0.4888,
+      "step": 917
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1093712083778748e-05,
+      "loss": 0.4526,
+      "step": 918
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1076821675124906e-05,
+      "loss": 0.4194,
+      "step": 919
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1059928157836965e-05,
+      "loss": 0.3545,
+      "step": 920
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1043031580684147e-05,
+      "loss": 0.4646,
+      "step": 921
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1026131992444517e-05,
+      "loss": 0.5149,
+      "step": 922
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.1009229441904831e-05,
+      "loss": 0.4917,
+      "step": 923
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.0992323977860399e-05,
+      "loss": 0.3953,
+      "step": 924
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0975415649114938e-05,
+      "loss": 0.4358,
+      "step": 925
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0958504504480434e-05,
+      "loss": 0.356,
+      "step": 926
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0941590592777007e-05,
+      "loss": 0.4431,
+      "step": 927
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0924673962832768e-05,
+      "loss": 0.4082,
+      "step": 928
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.090775466348366e-05,
+      "loss": 0.3735,
+      "step": 929
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0890832743573351e-05,
+      "loss": 0.5083,
+      "step": 930
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.087390825195306e-05,
+      "loss": 0.4822,
+      "step": 931
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0856981237481439e-05,
+      "loss": 0.469,
+      "step": 932
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0840051749024416e-05,
+      "loss": 0.4849,
+      "step": 933
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0823119835455069e-05,
+      "loss": 0.4309,
+      "step": 934
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0806185545653472e-05,
+      "loss": 0.5039,
+      "step": 935
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0789248928506557e-05,
+      "loss": 0.4237,
+      "step": 936
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0772310032907976e-05,
+      "loss": 0.5054,
+      "step": 937
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0755368907757969e-05,
+      "loss": 0.4039,
+      "step": 938
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0738425601963192e-05,
+      "loss": 0.5164,
+      "step": 939
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0721480164436614e-05,
+      "loss": 0.4775,
+      "step": 940
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0704532644097351e-05,
+      "loss": 0.4724,
+      "step": 941
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.068758308987053e-05,
+      "loss": 0.4583,
+      "step": 942
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.0670631550687157e-05,
+      "loss": 0.3818,
+      "step": 943
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0653678075483955e-05,
+      "loss": 0.4258,
+      "step": 944
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.063672271320325e-05,
+      "loss": 0.5127,
+      "step": 945
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0619765512792812e-05,
+      "loss": 0.4954,
+      "step": 946
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0602806523205708e-05,
+      "loss": 0.4893,
+      "step": 947
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0585845793400184e-05,
+      "loss": 0.5564,
+      "step": 948
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0568883372339498e-05,
+      "loss": 0.4729,
+      "step": 949
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.05519193089918e-05,
+      "loss": 0.4912,
+      "step": 950
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0534953652329976e-05,
+      "loss": 0.4485,
+      "step": 951
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.051798645133151e-05,
+      "loss": 0.5298,
+      "step": 952
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0501017754978348e-05,
+      "loss": 0.4697,
+      "step": 953
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0484047612256751e-05,
+      "loss": 0.4512,
+      "step": 954
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0467076072157159e-05,
+      "loss": 0.5164,
+      "step": 955
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0450103183674038e-05,
+      "loss": 0.3916,
+      "step": 956
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0433128995805757e-05,
+      "loss": 0.4507,
+      "step": 957
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0416153557554425e-05,
+      "loss": 0.4204,
+      "step": 958
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0399176917925773e-05,
+      "loss": 0.4387,
+      "step": 959
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0382199125928984e-05,
+      "loss": 0.5112,
+      "step": 960
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0365220230576592e-05,
+      "loss": 0.5088,
+      "step": 961
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0348240280884285e-05,
+      "loss": 0.5271,
+      "step": 962
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.0331259325870824e-05,
+      "loss": 0.5212,
+      "step": 963
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0314277414557854e-05,
+      "loss": 0.4631,
+      "step": 964
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0297294595969789e-05,
+      "loss": 0.47,
+      "step": 965
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0280310919133657e-05,
+      "loss": 0.4692,
+      "step": 966
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0263326433078966e-05,
+      "loss": 0.4453,
+      "step": 967
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0246341186837558e-05,
+      "loss": 0.5039,
+      "step": 968
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0229355229443483e-05,
+      "loss": 0.5356,
+      "step": 969
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0212368609932816e-05,
+      "loss": 0.4175,
+      "step": 970
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0195381377343576e-05,
+      "loss": 0.3665,
+      "step": 971
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0178393580715524e-05,
+      "loss": 0.4993,
+      "step": 972
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0161405269090068e-05,
+      "loss": 0.4099,
+      "step": 973
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0144416491510095e-05,
+      "loss": 0.4722,
+      "step": 974
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0127427297019836e-05,
+      "loss": 0.4453,
+      "step": 975
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0110437734664728e-05,
+      "loss": 0.5403,
+      "step": 976
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0093447853491276e-05,
+      "loss": 0.5037,
+      "step": 977
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0076457702546889e-05,
+      "loss": 0.5159,
+      "step": 978
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0059467330879776e-05,
+      "loss": 0.4482,
+      "step": 979
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0042476787538763e-05,
+      "loss": 0.408,
+      "step": 980
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0025486121573183e-05,
+      "loss": 0.3334,
+      "step": 981
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.0008495382032724e-05,
+      "loss": 0.4006,
+      "step": 982
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.99150461796728e-06,
+      "loss": 0.4978,
+      "step": 983
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.974513878426819e-06,
+      "loss": 0.4497,
+      "step": 984
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.957523212461239e-06,
+      "loss": 0.5269,
+      "step": 985
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.94053266912023e-06,
+      "loss": 0.4375,
+      "step": 986
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.923542297453113e-06,
+      "loss": 0.4631,
+      "step": 987
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.906552146508727e-06,
+      "loss": 0.5374,
+      "step": 988
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.889562265335272e-06,
+      "loss": 0.4565,
+      "step": 989
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.872572702980166e-06,
+      "loss": 0.4937,
+      "step": 990
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.85558350848991e-06,
+      "loss": 0.4478,
+      "step": 991
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.838594730909936e-06,
+      "loss": 0.3938,
+      "step": 992
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.82160641928448e-06,
+      "loss": 0.3872,
+      "step": 993
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.804618622656427e-06,
+      "loss": 0.4543,
+      "step": 994
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.787631390067184e-06,
+      "loss": 0.4438,
+      "step": 995
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.770644770556524e-06,
+      "loss": 0.4358,
+      "step": 996
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.753658813162444e-06,
+      "loss": 0.4062,
+      "step": 997
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.736673566921036e-06,
+      "loss": 0.4524,
+      "step": 998
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.719689080866347e-06,
+      "loss": 0.375,
+      "step": 999
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.702705404030215e-06,
+      "loss": 0.4246,
+      "step": 1000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 9.68572258544215e-06,
+      "loss": 0.4683,
+      "step": 1001
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.668740674129178e-06,
+      "loss": 0.4541,
+      "step": 1002
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.651759719115716e-06,
+      "loss": 0.3955,
+      "step": 1003
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.634779769423412e-06,
+      "loss": 0.3677,
+      "step": 1004
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.617800874071014e-06,
+      "loss": 0.4954,
+      "step": 1005
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.600823082074232e-06,
+      "loss": 0.4963,
+      "step": 1006
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.583846442445578e-06,
+      "loss": 0.509,
+      "step": 1007
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.566871004194245e-06,
+      "loss": 0.4309,
+      "step": 1008
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.549896816325964e-06,
+      "loss": 0.4719,
+      "step": 1009
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.532923927842844e-06,
+      "loss": 0.4868,
+      "step": 1010
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.515952387743252e-06,
+      "loss": 0.3757,
+      "step": 1011
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.498982245021655e-06,
+      "loss": 0.4346,
+      "step": 1012
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.482013548668492e-06,
+      "loss": 0.4182,
+      "step": 1013
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.465046347670029e-06,
+      "loss": 0.4751,
+      "step": 1014
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.448080691008202e-06,
+      "loss": 0.4517,
+      "step": 1015
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.431116627660505e-06,
+      "loss": 0.5264,
+      "step": 1016
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.414154206599821e-06,
+      "loss": 0.415,
+      "step": 1017
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.397193476794293e-06,
+      "loss": 0.4932,
+      "step": 1018
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.380234487207192e-06,
+      "loss": 0.4824,
+      "step": 1019
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 9.363277286796751e-06,
+      "loss": 0.4297,
+      "step": 1020
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.346321924516045e-06,
+      "loss": 0.4929,
+      "step": 1021
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.329368449312847e-06,
+      "loss": 0.386,
+      "step": 1022
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.312416910129471e-06,
+      "loss": 0.4097,
+      "step": 1023
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.295467355902652e-06,
+      "loss": 0.4194,
+      "step": 1024
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.278519835563387e-06,
+      "loss": 0.4241,
+      "step": 1025
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.26157439803681e-06,
+      "loss": 0.4753,
+      "step": 1026
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.244631092242036e-06,
+      "loss": 0.4919,
+      "step": 1027
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.227689967092025e-06,
+      "loss": 0.4683,
+      "step": 1028
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.210751071493448e-06,
+      "loss": 0.4706,
+      "step": 1029
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.193814454346531e-06,
+      "loss": 0.4436,
+      "step": 1030
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.17688016454493e-06,
+      "loss": 0.4185,
+      "step": 1031
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.159948250975586e-06,
+      "loss": 0.5737,
+      "step": 1032
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.143018762518566e-06,
+      "loss": 0.4756,
+      "step": 1033
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.126091748046943e-06,
+      "loss": 0.4819,
+      "step": 1034
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.10916725642665e-06,
+      "loss": 0.5063,
+      "step": 1035
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.092245336516341e-06,
+      "loss": 0.489,
+      "step": 1036
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.075326037167239e-06,
+      "loss": 0.4666,
+      "step": 1037
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.058409407222996e-06,
+      "loss": 0.5295,
+      "step": 1038
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 9.041495495519567e-06,
+      "loss": 0.4536,
+      "step": 1039
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.024584350885067e-06,
+      "loss": 0.5066,
+      "step": 1040
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 9.007676022139603e-06,
+      "loss": 0.3989,
+      "step": 1041
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.990770558095174e-06,
+      "loss": 0.3857,
+      "step": 1042
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.973868007555486e-06,
+      "loss": 0.4236,
+      "step": 1043
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.956968419315858e-06,
+      "loss": 0.4551,
+      "step": 1044
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.940071842163038e-06,
+      "loss": 0.4949,
+      "step": 1045
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.923178324875095e-06,
+      "loss": 0.4353,
+      "step": 1046
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.906287916221259e-06,
+      "loss": 0.4595,
+      "step": 1047
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.889400664961782e-06,
+      "loss": 0.4609,
+      "step": 1048
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.872516619847805e-06,
+      "loss": 0.4526,
+      "step": 1049
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.855635829621217e-06,
+      "loss": 0.4324,
+      "step": 1050
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.838758343014504e-06,
+      "loss": 0.5498,
+      "step": 1051
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.821884208750623e-06,
+      "loss": 0.4392,
+      "step": 1052
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.80501347554284e-06,
+      "loss": 0.4702,
+      "step": 1053
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.788146192094615e-06,
+      "loss": 0.4421,
+      "step": 1054
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.771282407099441e-06,
+      "loss": 0.4912,
+      "step": 1055
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.754422169240719e-06,
+      "loss": 0.4351,
+      "step": 1056
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.737565527191609e-06,
+      "loss": 0.4707,
+      "step": 1057
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 8.72071252961488e-06,
+      "loss": 0.5066,
+      "step": 1058
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.703863225162788e-06,
+      "loss": 0.5125,
+      "step": 1059
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.687017662476933e-06,
+      "loss": 0.4727,
+      "step": 1060
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.670175890188099e-06,
+      "loss": 0.4558,
+      "step": 1061
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.653337956916141e-06,
+      "loss": 0.4888,
+      "step": 1062
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.636503911269828e-06,
+      "loss": 0.3813,
+      "step": 1063
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.619673801846697e-06,
+      "loss": 0.4589,
+      "step": 1064
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.60284767723293e-06,
+      "loss": 0.4919,
+      "step": 1065
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.586025586003206e-06,
+      "loss": 0.4749,
+      "step": 1066
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.569207576720558e-06,
+      "loss": 0.4736,
+      "step": 1067
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.552393697936236e-06,
+      "loss": 0.4224,
+      "step": 1068
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.535583998189558e-06,
+      "loss": 0.4578,
+      "step": 1069
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.518778526007793e-06,
+      "loss": 0.397,
+      "step": 1070
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.501977329905992e-06,
+      "loss": 0.4578,
+      "step": 1071
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.485180458386865e-06,
+      "loss": 0.3694,
+      "step": 1072
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.46838795994065e-06,
+      "loss": 0.4341,
+      "step": 1073
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.451599883044937e-06,
+      "loss": 0.3617,
+      "step": 1074
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.434816276164567e-06,
+      "loss": 0.4651,
+      "step": 1075
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.41803718775148e-06,
+      "loss": 0.4646,
+      "step": 1076
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 8.40126266624456e-06,
+      "loss": 0.501,
+      "step": 1077
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.384492760069518e-06,
+      "loss": 0.4022,
+      "step": 1078
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.367727517638729e-06,
+      "loss": 0.4597,
+      "step": 1079
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.35096698735112e-06,
+      "loss": 0.4863,
+      "step": 1080
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.334211217591997e-06,
+      "loss": 0.4539,
+      "step": 1081
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.317460256732937e-06,
+      "loss": 0.418,
+      "step": 1082
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.300714153131632e-06,
+      "loss": 0.3911,
+      "step": 1083
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.283972955131747e-06,
+      "loss": 0.4788,
+      "step": 1084
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.267236711062785e-06,
+      "loss": 0.4456,
+      "step": 1085
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.25050546923995e-06,
+      "loss": 0.3948,
+      "step": 1086
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.233779277964009e-06,
+      "loss": 0.4763,
+      "step": 1087
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.217058185521145e-06,
+      "loss": 0.4573,
+      "step": 1088
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.20034224018282e-06,
+      "loss": 0.4089,
+      "step": 1089
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.183631490205636e-06,
+      "loss": 0.3936,
+      "step": 1090
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.166925983831206e-06,
+      "loss": 0.4763,
+      "step": 1091
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.15022576928599e-06,
+      "loss": 0.4722,
+      "step": 1092
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.13353089478119e-06,
+      "loss": 0.4341,
+      "step": 1093
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.116841408512576e-06,
+      "loss": 0.458,
+      "step": 1094
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.100157358660367e-06,
+      "loss": 0.4508,
+      "step": 1095
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 8.083478793389092e-06,
+      "loss": 0.4241,
+      "step": 1096
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.066805760847449e-06,
+      "loss": 0.5107,
+      "step": 1097
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.05013830916815e-06,
+      "loss": 0.4207,
+      "step": 1098
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.033476486467818e-06,
+      "loss": 0.4263,
+      "step": 1099
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.0168203408468e-06,
+      "loss": 0.395,
+      "step": 1100
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 8.000169920389074e-06,
+      "loss": 0.4578,
+      "step": 1101
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.98352527316208e-06,
+      "loss": 0.4187,
+      "step": 1102
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.9668864472166e-06,
+      "loss": 0.4504,
+      "step": 1103
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.950253490586603e-06,
+      "loss": 0.4617,
+      "step": 1104
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.933626451289116e-06,
+      "loss": 0.4575,
+      "step": 1105
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.91700537732408e-06,
+      "loss": 0.4417,
+      "step": 1106
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.900390316674229e-06,
+      "loss": 0.4221,
+      "step": 1107
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.883781317304917e-06,
+      "loss": 0.4097,
+      "step": 1108
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.867178427164024e-06,
+      "loss": 0.4327,
+      "step": 1109
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.850581694181765e-06,
+      "loss": 0.4573,
+      "step": 1110
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.833991166270604e-06,
+      "loss": 0.4082,
+      "step": 1111
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.817406891325078e-06,
+      "loss": 0.4929,
+      "step": 1112
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.800828917221683e-06,
+      "loss": 0.4333,
+      "step": 1113
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.78425729181872e-06,
+      "loss": 0.4829,
+      "step": 1114
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 7.767692062956156e-06,
+      "loss": 0.4727,
+      "step": 1115
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.751133278455498e-06,
+      "loss": 0.594,
+      "step": 1116
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.734580986119656e-06,
+      "loss": 0.4368,
+      "step": 1117
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.718035233732786e-06,
+      "loss": 0.5374,
+      "step": 1118
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.701496069060177e-06,
+      "loss": 0.4397,
+      "step": 1119
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.684963539848077e-06,
+      "loss": 0.3706,
+      "step": 1120
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.668437693823609e-06,
+      "loss": 0.3735,
+      "step": 1121
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.651918578694578e-06,
+      "loss": 0.3784,
+      "step": 1122
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.635406242149372e-06,
+      "loss": 0.3943,
+      "step": 1123
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.618900731856806e-06,
+      "loss": 0.4038,
+      "step": 1124
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.602402095465984e-06,
+      "loss": 0.4282,
+      "step": 1125
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.5859103806061694e-06,
+      "loss": 0.4248,
+      "step": 1126
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.56942563488665e-06,
+      "loss": 0.4915,
+      "step": 1127
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.552947905896586e-06,
+      "loss": 0.4607,
+      "step": 1128
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.5364772412048885e-06,
+      "loss": 0.4385,
+      "step": 1129
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.5200136883600615e-06,
+      "loss": 0.4324,
+      "step": 1130
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.503557294890096e-06,
+      "loss": 0.4739,
+      "step": 1131
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.487108108302301e-06,
+      "loss": 0.4585,
+      "step": 1132
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.470666176083193e-06,
+      "loss": 0.4355,
+      "step": 1133
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 7.454231545698334e-06,
+      "loss": 0.4624,
+      "step": 1134
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.4378042645922145e-06,
+      "loss": 0.4438,
+      "step": 1135
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.421384380188098e-06,
+      "loss": 0.491,
+      "step": 1136
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.40497193988791e-06,
+      "loss": 0.4534,
+      "step": 1137
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.388566991072077e-06,
+      "loss": 0.4414,
+      "step": 1138
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.372169581099398e-06,
+      "loss": 0.5117,
+      "step": 1139
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.355779757306915e-06,
+      "loss": 0.489,
+      "step": 1140
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.339397567009758e-06,
+      "loss": 0.5061,
+      "step": 1141
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.323023057501033e-06,
+      "loss": 0.3522,
+      "step": 1142
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.3066562760516615e-06,
+      "loss": 0.5056,
+      "step": 1143
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.290297269910267e-06,
+      "loss": 0.4373,
+      "step": 1144
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.273946086303019e-06,
+      "loss": 0.4832,
+      "step": 1145
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.257602772433499e-06,
+      "loss": 0.5024,
+      "step": 1146
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.241267375482584e-06,
+      "loss": 0.4832,
+      "step": 1147
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.224939942608287e-06,
+      "loss": 0.4331,
+      "step": 1148
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.208620520945627e-06,
+      "loss": 0.4502,
+      "step": 1149
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.192309157606508e-06,
+      "loss": 0.4504,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.176005899679556e-06,
+      "loss": 0.4614,
+      "step": 1151
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.159710794230006e-06,
+      "loss": 0.4204,
+      "step": 1152
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 7.143423888299554e-06,
+      "loss": 0.3762,
+      "step": 1153
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.127145228906234e-06,
+      "loss": 0.4336,
+      "step": 1154
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.110874863044266e-06,
+      "loss": 0.5125,
+      "step": 1155
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.094612837683923e-06,
+      "loss": 0.428,
+      "step": 1156
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.078359199771407e-06,
+      "loss": 0.4741,
+      "step": 1157
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.06211399622871e-06,
+      "loss": 0.4048,
+      "step": 1158
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.045877273953467e-06,
+      "loss": 0.3982,
+      "step": 1159
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.029649079818836e-06,
+      "loss": 0.4548,
+      "step": 1160
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 7.013429460673349e-06,
+      "loss": 0.4482,
+      "step": 1161
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.997218463340786e-06,
+      "loss": 0.3818,
+      "step": 1162
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.981016134620036e-06,
+      "loss": 0.397,
+      "step": 1163
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.964822521284966e-06,
+      "loss": 0.4519,
+      "step": 1164
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.9486376700842835e-06,
+      "loss": 0.3687,
+      "step": 1165
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.932461627741393e-06,
+      "loss": 0.4199,
+      "step": 1166
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.9162944409542744e-06,
+      "loss": 0.36,
+      "step": 1167
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.900136156395347e-06,
+      "loss": 0.4561,
+      "step": 1168
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.883986820711321e-06,
+      "loss": 0.4233,
+      "step": 1169
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.86784648052309e-06,
+      "loss": 0.4268,
+      "step": 1170
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.851715182425551e-06,
+      "loss": 0.3761,
+      "step": 1171
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 6.8355929729875216e-06,
+      "loss": 0.498,
+      "step": 1172
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.819479898751571e-06,
+      "loss": 0.4626,
+      "step": 1173
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.803376006233903e-06,
+      "loss": 0.4373,
+      "step": 1174
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.7872813419242055e-06,
+      "loss": 0.4487,
+      "step": 1175
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.771195952285541e-06,
+      "loss": 0.4604,
+      "step": 1176
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.7551198837541746e-06,
+      "loss": 0.4751,
+      "step": 1177
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.739053182739484e-06,
+      "loss": 0.4114,
+      "step": 1178
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.722995895623791e-06,
+      "loss": 0.4387,
+      "step": 1179
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.706948068762251e-06,
+      "loss": 0.4941,
+      "step": 1180
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.690909748482704e-06,
+      "loss": 0.437,
+      "step": 1181
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.674880981085541e-06,
+      "loss": 0.4197,
+      "step": 1182
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.658861812843578e-06,
+      "loss": 0.415,
+      "step": 1183
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.642852290001927e-06,
+      "loss": 0.4082,
+      "step": 1184
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.626852458777846e-06,
+      "loss": 0.4299,
+      "step": 1185
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.610862365360622e-06,
+      "loss": 0.5244,
+      "step": 1186
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.594882055911419e-06,
+      "loss": 0.406,
+      "step": 1187
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.57891157656317e-06,
+      "loss": 0.4534,
+      "step": 1188
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.562950973420416e-06,
+      "loss": 0.4856,
+      "step": 1189
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.547000292559201e-06,
+      "loss": 0.4216,
+      "step": 1190
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 6.531059580026917e-06,
+      "loss": 0.4147,
+      "step": 1191
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.515128881842177e-06,
+      "loss": 0.4685,
+      "step": 1192
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.499208243994683e-06,
+      "loss": 0.4622,
+      "step": 1193
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.483297712445107e-06,
+      "loss": 0.3503,
+      "step": 1194
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.467397333124928e-06,
+      "loss": 0.5273,
+      "step": 1195
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.451507151936332e-06,
+      "loss": 0.397,
+      "step": 1196
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.435627214752047e-06,
+      "loss": 0.4009,
+      "step": 1197
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.419757567415249e-06,
+      "loss": 0.4141,
+      "step": 1198
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.403898255739394e-06,
+      "loss": 0.4268,
+      "step": 1199
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.3880493255081035e-06,
+      "loss": 0.4541,
+      "step": 1200
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.372210822475034e-06,
+      "loss": 0.5239,
+      "step": 1201
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.356382792363733e-06,
+      "loss": 0.4995,
+      "step": 1202
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.340565280867519e-06,
+      "loss": 0.4204,
+      "step": 1203
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.324758333649343e-06,
+      "loss": 0.3838,
+      "step": 1204
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.308961996341658e-06,
+      "loss": 0.4363,
+      "step": 1205
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.293176314546292e-06,
+      "loss": 0.3914,
+      "step": 1206
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.277401333834301e-06,
+      "loss": 0.5369,
+      "step": 1207
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.261637099745863e-06,
+      "loss": 0.4683,
+      "step": 1208
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.245883657790122e-06,
+      "loss": 0.4373,
+      "step": 1209
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 6.230141053445067e-06,
+      "loss": 0.4832,
+      "step": 1210
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.214409332157408e-06,
+      "loss": 0.4653,
+      "step": 1211
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.198688539342431e-06,
+      "loss": 0.4084,
+      "step": 1212
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.18297872038387e-06,
+      "loss": 0.4341,
+      "step": 1213
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.167279920633784e-06,
+      "loss": 0.469,
+      "step": 1214
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.1515921854124225e-06,
+      "loss": 0.5259,
+      "step": 1215
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.135915560008088e-06,
+      "loss": 0.3706,
+      "step": 1216
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.1202500896770175e-06,
+      "loss": 0.438,
+      "step": 1217
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.104595819643235e-06,
+      "loss": 0.4888,
+      "step": 1218
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.088952795098442e-06,
+      "loss": 0.4065,
+      "step": 1219
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.073321061201865e-06,
+      "loss": 0.5107,
+      "step": 1220
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.057700663080144e-06,
+      "loss": 0.4177,
+      "step": 1221
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.042091645827194e-06,
+      "loss": 0.5105,
+      "step": 1222
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.026494054504068e-06,
+      "loss": 0.4585,
+      "step": 1223
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 6.010907934138838e-06,
+      "loss": 0.4087,
+      "step": 1224
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.995333329726465e-06,
+      "loss": 0.396,
+      "step": 1225
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.979770286228659e-06,
+      "loss": 0.3523,
+      "step": 1226
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.964218848573762e-06,
+      "loss": 0.3921,
+      "step": 1227
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.9486790616566e-06,
+      "loss": 0.3959,
+      "step": 1228
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.933150970338377e-06,
+      "loss": 0.4622,
+      "step": 1229
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 5.917634619446526e-06,
+      "loss": 0.4268,
+      "step": 1230
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.902130053774593e-06,
+      "loss": 0.4082,
+      "step": 1231
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.8866373180820976e-06,
+      "loss": 0.4099,
+      "step": 1232
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.871156457094405e-06,
+      "loss": 0.3647,
+      "step": 1233
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.855687515502604e-06,
+      "loss": 0.3676,
+      "step": 1234
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.840230537963376e-06,
+      "loss": 0.4868,
+      "step": 1235
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.824785569098858e-06,
+      "loss": 0.5198,
+      "step": 1236
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.8093526534965315e-06,
+      "loss": 0.4519,
+      "step": 1237
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7939318357090615e-06,
+      "loss": 0.5012,
+      "step": 1238
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.778523160254199e-06,
+      "loss": 0.4734,
+      "step": 1239
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7631266716146535e-06,
+      "loss": 0.5002,
+      "step": 1240
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.747742414237936e-06,
+      "loss": 0.4866,
+      "step": 1241
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.732370432536258e-06,
+      "loss": 0.4314,
+      "step": 1242
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.7170107708863755e-06,
+      "loss": 0.3398,
+      "step": 1243
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.701663473629504e-06,
+      "loss": 0.4607,
+      "step": 1244
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.686328585071148e-06,
+      "loss": 0.4219,
+      "step": 1245
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.671006149480993e-06,
+      "loss": 0.4773,
+      "step": 1246
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.655696211092776e-06,
+      "loss": 0.4636,
+      "step": 1247
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.640398814104156e-06,
+      "loss": 0.3899,
+      "step": 1248
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 5.6251140026765815e-06,
+      "loss": 0.441,
+      "step": 1249
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.609841820935179e-06,
+      "loss": 0.4038,
+      "step": 1250
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.594582312968602e-06,
+      "loss": 0.3708,
+      "step": 1251
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.579335522828932e-06,
+      "loss": 0.4105,
+      "step": 1252
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.564101494531526e-06,
+      "loss": 0.4214,
+      "step": 1253
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.548880272054894e-06,
+      "loss": 0.4768,
+      "step": 1254
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.5336718993405825e-06,
+      "loss": 0.4639,
+      "step": 1255
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.518476420293055e-06,
+      "loss": 0.4404,
+      "step": 1256
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.503293878779534e-06,
+      "loss": 0.5127,
+      "step": 1257
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.488124318629908e-06,
+      "loss": 0.4312,
+      "step": 1258
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.4729677836365695e-06,
+      "loss": 0.4397,
+      "step": 1259
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.4578243175543344e-06,
+      "loss": 0.439,
+      "step": 1260
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.442693964100276e-06,
+      "loss": 0.4905,
+      "step": 1261
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.427576766953615e-06,
+      "loss": 0.4658,
+      "step": 1262
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.412472769755595e-06,
+      "loss": 0.4196,
+      "step": 1263
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.397382016109347e-06,
+      "loss": 0.4683,
+      "step": 1264
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.3823045495797775e-06,
+      "loss": 0.3975,
+      "step": 1265
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.367240413693429e-06,
+      "loss": 0.3726,
+      "step": 1266
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.352189651938362e-06,
+      "loss": 0.4333,
+      "step": 1267
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 5.337152307764031e-06,
+      "loss": 0.438,
+      "step": 1268
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.3221284245811485e-06,
+      "loss": 0.396,
+      "step": 1269
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.3071180457615755e-06,
+      "loss": 0.3857,
+      "step": 1270
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2921212146381815e-06,
+      "loss": 0.3975,
+      "step": 1271
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2771379745047255e-06,
+      "loss": 0.4453,
+      "step": 1272
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2621683686157445e-06,
+      "loss": 0.4741,
+      "step": 1273
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2472124401863936e-06,
+      "loss": 0.4373,
+      "step": 1274
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2322702323923544e-06,
+      "loss": 0.48,
+      "step": 1275
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.217341788369699e-06,
+      "loss": 0.4617,
+      "step": 1276
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.2024271512147706e-06,
+      "loss": 0.533,
+      "step": 1277
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.187526363984048e-06,
+      "loss": 0.3962,
+      "step": 1278
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.172639469694014e-06,
+      "loss": 0.4504,
+      "step": 1279
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.157766511321074e-06,
+      "loss": 0.5129,
+      "step": 1280
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.142907531801378e-06,
+      "loss": 0.3811,
+      "step": 1281
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.1280625740307296e-06,
+      "loss": 0.4482,
+      "step": 1282
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.113231680864454e-06,
+      "loss": 0.4072,
+      "step": 1283
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.098414895117273e-06,
+      "loss": 0.3972,
+      "step": 1284
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.083612259563182e-06,
+      "loss": 0.4485,
+      "step": 1285
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.0688238169353265e-06,
+      "loss": 0.5049,
+      "step": 1286
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 5.054049609925878e-06,
+      "loss": 0.4358,
+      "step": 1287
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.039289681185913e-06,
+      "loss": 0.5029,
+      "step": 1288
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.024544073325297e-06,
+      "loss": 0.3982,
+      "step": 1289
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 5.009812828912535e-06,
+      "loss": 0.489,
+      "step": 1290
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.995095990474679e-06,
+      "loss": 0.3986,
+      "step": 1291
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.9803936004971874e-06,
+      "loss": 0.3859,
+      "step": 1292
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.96570570142382e-06,
+      "loss": 0.4197,
+      "step": 1293
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.951032335656493e-06,
+      "loss": 0.4121,
+      "step": 1294
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.9363735455551606e-06,
+      "loss": 0.5396,
+      "step": 1295
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.921729373437708e-06,
+      "loss": 0.4067,
+      "step": 1296
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.907099861579827e-06,
+      "loss": 0.3831,
+      "step": 1297
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.892485052214873e-06,
+      "loss": 0.4875,
+      "step": 1298
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.8778849875337716e-06,
+      "loss": 0.4561,
+      "step": 1299
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.863299709684861e-06,
+      "loss": 0.5029,
+      "step": 1300
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.848729260773818e-06,
+      "loss": 0.4103,
+      "step": 1301
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.834173682863492e-06,
+      "loss": 0.4438,
+      "step": 1302
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.819633017973812e-06,
+      "loss": 0.3395,
+      "step": 1303
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.805107308081647e-06,
+      "loss": 0.4771,
+      "step": 1304
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.790596595120699e-06,
+      "loss": 0.4431,
+      "step": 1305
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.776100920981374e-06,
+      "loss": 0.4255,
+      "step": 1306
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.761620327510662e-06,
+      "loss": 0.415,
+      "step": 1307
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.747154856512016e-06,
+      "loss": 0.4196,
+      "step": 1308
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.732704549745247e-06,
+      "loss": 0.4281,
+      "step": 1309
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.718269448926361e-06,
+      "loss": 0.4236,
+      "step": 1310
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.703849595727491e-06,
+      "loss": 0.4299,
+      "step": 1311
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.689445031776737e-06,
+      "loss": 0.3752,
+      "step": 1312
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.675055798658077e-06,
+      "loss": 0.3809,
+      "step": 1313
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.66068193791122e-06,
+      "loss": 0.459,
+      "step": 1314
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.646323491031492e-06,
+      "loss": 0.3921,
+      "step": 1315
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.6319804994697284e-06,
+      "loss": 0.4126,
+      "step": 1316
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.6176530046321534e-06,
+      "loss": 0.4209,
+      "step": 1317
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.603341047880244e-06,
+      "loss": 0.4111,
+      "step": 1318
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.5890446705306315e-06,
+      "loss": 0.4514,
+      "step": 1319
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.574763913854949e-06,
+      "loss": 0.4202,
+      "step": 1320
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.560498819079766e-06,
+      "loss": 0.4895,
+      "step": 1321
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.546249427386417e-06,
+      "loss": 0.3835,
+      "step": 1322
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.532015779910911e-06,
+      "loss": 0.3662,
+      "step": 1323
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.5177979177438035e-06,
+      "loss": 0.3979,
+      "step": 1324
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.503595881930083e-06,
+      "loss": 0.4111,
+      "step": 1325
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.489409713469046e-06,
+      "loss": 0.4321,
+      "step": 1326
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.475239453314184e-06,
+      "loss": 0.4475,
+      "step": 1327
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.461085142373064e-06,
+      "loss": 0.4248,
+      "step": 1328
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.446946821507205e-06,
+      "loss": 0.4937,
+      "step": 1329
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.432824531531979e-06,
+      "loss": 0.4849,
+      "step": 1330
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.418718313216461e-06,
+      "loss": 0.5063,
+      "step": 1331
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.404628207283338e-06,
+      "loss": 0.408,
+      "step": 1332
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.390554254408778e-06,
+      "loss": 0.4675,
+      "step": 1333
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.376496495222332e-06,
+      "loss": 0.4045,
+      "step": 1334
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3624549703067895e-06,
+      "loss": 0.3701,
+      "step": 1335
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.348429720198065e-06,
+      "loss": 0.4529,
+      "step": 1336
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.334420785385113e-06,
+      "loss": 0.4355,
+      "step": 1337
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.320428206309771e-06,
+      "loss": 0.4604,
+      "step": 1338
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.3064520233666665e-06,
+      "loss": 0.428,
+      "step": 1339
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.292492276903091e-06,
+      "loss": 0.4648,
+      "step": 1340
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.278549007218885e-06,
+      "loss": 0.4614,
+      "step": 1341
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.264622254566328e-06,
+      "loss": 0.4094,
+      "step": 1342
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.2507120591500125e-06,
+      "loss": 0.4407,
+      "step": 1343
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.2368184611267345e-06,
+      "loss": 0.41,
+      "step": 1344
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.222941500605375e-06,
+      "loss": 0.4402,
+      "step": 1345
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.209081217646785e-06,
+      "loss": 0.4983,
+      "step": 1346
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.195237652263669e-06,
+      "loss": 0.3706,
+      "step": 1347
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.181410844420473e-06,
+      "loss": 0.384,
+      "step": 1348
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.1676008340332596e-06,
+      "loss": 0.5376,
+      "step": 1349
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.153807660969614e-06,
+      "loss": 0.4751,
+      "step": 1350
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.1400313650484955e-06,
+      "loss": 0.4399,
+      "step": 1351
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.126271986040152e-06,
+      "loss": 0.3992,
+      "step": 1352
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.112529563665992e-06,
+      "loss": 0.4044,
+      "step": 1353
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.0988041375984784e-06,
+      "loss": 0.4966,
+      "step": 1354
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.085095747461005e-06,
+      "loss": 0.4248,
+      "step": 1355
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.071404432827777e-06,
+      "loss": 0.4338,
+      "step": 1356
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.05773023322371e-06,
+      "loss": 0.4126,
+      "step": 1357
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.04407318812432e-06,
+      "loss": 0.4832,
+      "step": 1358
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.030433336955591e-06,
+      "loss": 0.4868,
+      "step": 1359
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.01681071909387e-06,
+      "loss": 0.4536,
+      "step": 1360
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.003205373865758e-06,
+      "loss": 0.4453,
+      "step": 1361
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.98961734054799e-06,
+      "loss": 0.4966,
+      "step": 1362
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.976046658367325e-06,
+      "loss": 0.3784,
+      "step": 1363
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.962493366500429e-06,
+      "loss": 0.4119,
+      "step": 1364
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.948957504073767e-06,
+      "loss": 0.3762,
+      "step": 1365
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.9354391101634946e-06,
+      "loss": 0.4475,
+      "step": 1366
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.921938223795323e-06,
+      "loss": 0.4714,
+      "step": 1367
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.908454883944431e-06,
+      "loss": 0.3926,
+      "step": 1368
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.894989129535339e-06,
+      "loss": 0.4412,
+      "step": 1369
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.881540999441811e-06,
+      "loss": 0.4314,
+      "step": 1370
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.868110532486723e-06,
+      "loss": 0.4365,
+      "step": 1371
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.854697767441955e-06,
+      "loss": 0.3994,
+      "step": 1372
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.841302743028291e-06,
+      "loss": 0.4506,
+      "step": 1373
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.827925497915305e-06,
+      "loss": 0.4104,
+      "step": 1374
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.8145660707212395e-06,
+      "loss": 0.4261,
+      "step": 1375
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.8012245000128987e-06,
+      "loss": 0.3829,
+      "step": 1376
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7879008243055292e-06,
+      "loss": 0.4167,
+      "step": 1377
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7745950820627355e-06,
+      "loss": 0.3782,
+      "step": 1378
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.761307311696337e-06,
+      "loss": 0.4319,
+      "step": 1379
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7480375515662783e-06,
+      "loss": 0.4026,
+      "step": 1380
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7347858399805035e-06,
+      "loss": 0.4551,
+      "step": 1381
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.721552215194859e-06,
+      "loss": 0.4595,
+      "step": 1382
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7083367154129756e-06,
+      "loss": 0.5386,
+      "step": 1383
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6951393787861587e-06,
+      "loss": 0.4702,
+      "step": 1384
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6819602434132796e-06,
+      "loss": 0.4495,
+      "step": 1385
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6687993473406667e-06,
+      "loss": 0.4978,
+      "step": 1386
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6556567285619917e-06,
+      "loss": 0.4241,
+      "step": 1387
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.642532425018164e-06,
+      "loss": 0.4609,
+      "step": 1388
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6294264745972176e-06,
+      "loss": 0.4094,
+      "step": 1389
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6163389151342033e-06,
+      "loss": 0.5364,
+      "step": 1390
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.6032697844110896e-06,
+      "loss": 0.4385,
+      "step": 1391
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5902191201566273e-06,
+      "loss": 0.4346,
+      "step": 1392
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.577186960046264e-06,
+      "loss": 0.3691,
+      "step": 1393
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5641733417020373e-06,
+      "loss": 0.4304,
+      "step": 1394
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.551178302692445e-06,
+      "loss": 0.4297,
+      "step": 1395
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5382018805323593e-06,
+      "loss": 0.4426,
+      "step": 1396
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.5252441126828895e-06,
+      "loss": 0.4409,
+      "step": 1397
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.512305036551319e-06,
+      "loss": 0.4639,
+      "step": 1398
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.499384689490951e-06,
+      "loss": 0.4919,
+      "step": 1399
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.4864831088010277e-06,
+      "loss": 0.4707,
+      "step": 1400
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.4736003317266155e-06,
+      "loss": 0.4563,
+      "step": 1401
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.460736395458495e-06,
+      "loss": 0.3398,
+      "step": 1402
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.447891337133059e-06,
+      "loss": 0.397,
+      "step": 1403
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.4350651938322e-06,
+      "loss": 0.4761,
+      "step": 1404
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.422258002583205e-06,
+      "loss": 0.4019,
+      "step": 1405
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.409469800358649e-06,
+      "loss": 0.4927,
+      "step": 1406
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.396700624076298e-06,
+      "loss": 0.4136,
+      "step": 1407
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3839505105989747e-06,
+      "loss": 0.3865,
+      "step": 1408
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3712194967344835e-06,
+      "loss": 0.4431,
+      "step": 1409
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.358507619235485e-06,
+      "loss": 0.4109,
+      "step": 1410
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3458149147994025e-06,
+      "loss": 0.5068,
+      "step": 1411
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.333141420068304e-06,
+      "loss": 0.4519,
+      "step": 1412
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.3204871716288e-06,
+      "loss": 0.4365,
+      "step": 1413
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.307852206011939e-06,
+      "loss": 0.4094,
+      "step": 1414
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.295236559693112e-06,
+      "loss": 0.3567,
+      "step": 1415
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2826402690919314e-06,
+      "loss": 0.5132,
+      "step": 1416
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.27006337057213e-06,
+      "loss": 0.408,
+      "step": 1417
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.257505900441462e-06,
+      "loss": 0.4683,
+      "step": 1418
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.2449678949515937e-06,
+      "loss": 0.4219,
+      "step": 1419
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.232449390298e-06,
+      "loss": 0.502,
+      "step": 1420
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.2199504226198587e-06,
+      "loss": 0.3687,
+      "step": 1421
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.2074710279999468e-06,
+      "loss": 0.4434,
+      "step": 1422
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1950112424645385e-06,
+      "loss": 0.4241,
+      "step": 1423
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.182571101983297e-06,
+      "loss": 0.437,
+      "step": 1424
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1701506424691763e-06,
+      "loss": 0.3689,
+      "step": 1425
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1577498997783074e-06,
+      "loss": 0.46,
+      "step": 1426
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.145368909709915e-06,
+      "loss": 0.4224,
+      "step": 1427
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1330077080061837e-06,
+      "loss": 0.368,
+      "step": 1428
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1206663303521833e-06,
+      "loss": 0.3928,
+      "step": 1429
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.1083448123757475e-06,
+      "loss": 0.3333,
+      "step": 1430
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0960431896473877e-06,
+      "loss": 0.4985,
+      "step": 1431
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0837614976801742e-06,
+      "loss": 0.4656,
+      "step": 1432
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0714997719296346e-06,
+      "loss": 0.4395,
+      "step": 1433
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0592580477936606e-06,
+      "loss": 0.4644,
+      "step": 1434
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0470363606124133e-06,
+      "loss": 0.4358,
+      "step": 1435
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0348347456681937e-06,
+      "loss": 0.4041,
+      "step": 1436
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0226532381853646e-06,
+      "loss": 0.406,
+      "step": 1437
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.0104918733302415e-06,
+      "loss": 0.446,
+      "step": 1438
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 2.998350686210989e-06,
+      "loss": 0.4521,
+      "step": 1439
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.986229711877523e-06,
+      "loss": 0.4368,
+      "step": 1440
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9741289853214063e-06,
+      "loss": 0.3918,
+      "step": 1441
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9620485414757504e-06,
+      "loss": 0.4873,
+      "step": 1442
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9499884152151126e-06,
+      "loss": 0.4124,
+      "step": 1443
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.937948641355398e-06,
+      "loss": 0.4395,
+      "step": 1444
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9259292546537543e-06,
+      "loss": 0.4373,
+      "step": 1445
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.913930289808474e-06,
+      "loss": 0.5002,
+      "step": 1446
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.9019517814589015e-06,
+      "loss": 0.4143,
+      "step": 1447
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8899937641853216e-06,
+      "loss": 0.4456,
+      "step": 1448
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.878056272508858e-06,
+      "loss": 0.4429,
+      "step": 1449
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.866139340891384e-06,
+      "loss": 0.4348,
+      "step": 1450
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8542430037354274e-06,
+      "loss": 0.4595,
+      "step": 1451
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.842367295384053e-06,
+      "loss": 0.3805,
+      "step": 1452
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8305122501207784e-06,
+      "loss": 0.4363,
+      "step": 1453
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.8186779021694577e-06,
+      "loss": 0.4155,
+      "step": 1454
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.806864285694213e-06,
+      "loss": 0.4543,
+      "step": 1455
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.795071434799307e-06,
+      "loss": 0.4666,
+      "step": 1456
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.7832993835290556e-06,
+      "loss": 0.4741,
+      "step": 1457
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 2.77154816586773e-06,
+      "loss": 0.4692,
+      "step": 1458
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7598178157394582e-06,
+      "loss": 0.3881,
+      "step": 1459
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7481083670081277e-06,
+      "loss": 0.4727,
+      "step": 1460
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7364198534772824e-06,
+      "loss": 0.467,
+      "step": 1461
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7247523088900342e-06,
+      "loss": 0.4155,
+      "step": 1462
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.7131057669289574e-06,
+      "loss": 0.4463,
+      "step": 1463
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.701480261215994e-06,
+      "loss": 0.5159,
+      "step": 1464
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6898758253123613e-06,
+      "loss": 0.4097,
+      "step": 1465
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6782924927184484e-06,
+      "loss": 0.4106,
+      "step": 1466
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.666730296873716e-06,
+      "loss": 0.5039,
+      "step": 1467
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.655189271156625e-06,
+      "loss": 0.459,
+      "step": 1468
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.6436694488844984e-06,
+      "loss": 0.3958,
+      "step": 1469
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.632170863313459e-06,
+      "loss": 0.479,
+      "step": 1470
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.620693547638318e-06,
+      "loss": 0.3989,
+      "step": 1471
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.609237534992494e-06,
+      "loss": 0.3655,
+      "step": 1472
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5978028584478974e-06,
+      "loss": 0.4558,
+      "step": 1473
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5863895510148373e-06,
+      "loss": 0.3813,
+      "step": 1474
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.574997645641951e-06,
+      "loss": 0.4585,
+      "step": 1475
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.563627175216078e-06,
+      "loss": 0.3684,
+      "step": 1476
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 2.5522781725621814e-06,
+      "loss": 0.3923,
+      "step": 1477
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.540950670443252e-06,
+      "loss": 0.4517,
+      "step": 1478
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.5296447015602087e-06,
+      "loss": 0.4868,
+      "step": 1479
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.5183602985518073e-06,
+      "loss": 0.4492,
+      "step": 1480
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.507097493994548e-06,
+      "loss": 0.3918,
+      "step": 1481
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4958563204025766e-06,
+      "loss": 0.3909,
+      "step": 1482
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.484636810227593e-06,
+      "loss": 0.4128,
+      "step": 1483
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.473438995858767e-06,
+      "loss": 0.3667,
+      "step": 1484
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.46226290962262e-06,
+      "loss": 0.5217,
+      "step": 1485
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4511085837829596e-06,
+      "loss": 0.3932,
+      "step": 1486
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4399760505407646e-06,
+      "loss": 0.4824,
+      "step": 1487
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.4288653420341134e-06,
+      "loss": 0.3982,
+      "step": 1488
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.417776490338074e-06,
+      "loss": 0.3557,
+      "step": 1489
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.40670952746461e-06,
+      "loss": 0.4907,
+      "step": 1490
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3956644853625e-06,
+      "loss": 0.3682,
+      "step": 1491
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3846413959172477e-06,
+      "loss": 0.4214,
+      "step": 1492
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.373640290950974e-06,
+      "loss": 0.4724,
+      "step": 1493
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.362661202222336e-06,
+      "loss": 0.3823,
+      "step": 1494
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.351704161426427e-06,
+      "loss": 0.4404,
+      "step": 1495
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 2.3407692001947034e-06,
+      "loss": 0.4602,
+      "step": 1496
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.329856350094872e-06,
+      "loss": 0.4023,
+      "step": 1497
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3189656426308115e-06,
+      "loss": 0.4492,
+      "step": 1498
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.3080971092424754e-06,
+      "loss": 0.4167,
+      "step": 1499
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2972507813058044e-06,
+      "loss": 0.4836,
+      "step": 1500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.286426690132636e-06,
+      "loss": 0.4749,
+      "step": 1501
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.275624866970614e-06,
+      "loss": 0.4302,
+      "step": 1502
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2648453430030927e-06,
+      "loss": 0.408,
+      "step": 1503
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2540881493490642e-06,
+      "loss": 0.4685,
+      "step": 1504
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2433533170630383e-06,
+      "loss": 0.4194,
+      "step": 1505
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2326408771349827e-06,
+      "loss": 0.365,
+      "step": 1506
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2219508604902175e-06,
+      "loss": 0.4492,
+      "step": 1507
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.211283297989334e-06,
+      "loss": 0.5,
+      "step": 1508
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.2006382204280997e-06,
+      "loss": 0.4299,
+      "step": 1509
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1900156585373654e-06,
+      "loss": 0.4164,
+      "step": 1510
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1794156429829847e-06,
+      "loss": 0.4211,
+      "step": 1511
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.168838204365733e-06,
+      "loss": 0.3759,
+      "step": 1512
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.158283373221196e-06,
+      "loss": 0.4111,
+      "step": 1513
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1477511800197036e-06,
+      "loss": 0.3538,
+      "step": 1514
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.137241655166227e-06,
+      "loss": 0.4443,
+      "step": 1515
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 2.1267548290003003e-06,
+      "loss": 0.3977,
+      "step": 1516
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.11629073179593e-06,
+      "loss": 0.4753,
+      "step": 1517
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.1058493937615065e-06,
+      "loss": 0.3772,
+      "step": 1518
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0954308450397177e-06,
+      "loss": 0.4524,
+      "step": 1519
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.08503511570746e-06,
+      "loss": 0.4165,
+      "step": 1520
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0746622357757574e-06,
+      "loss": 0.405,
+      "step": 1521
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0643122351896662e-06,
+      "loss": 0.4829,
+      "step": 1522
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0539851438281976e-06,
+      "loss": 0.4136,
+      "step": 1523
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.043680991504222e-06,
+      "loss": 0.4116,
+      "step": 1524
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0333998079643992e-06,
+      "loss": 0.344,
+      "step": 1525
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0231416228890654e-06,
+      "loss": 0.469,
+      "step": 1526
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.0129064658921737e-06,
+      "loss": 0.344,
+      "step": 1527
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 2.002694366521193e-06,
+      "loss": 0.49,
+      "step": 1528
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.992505354257036e-06,
+      "loss": 0.397,
+      "step": 1529
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.982339458513962e-06,
+      "loss": 0.4387,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9721967086394864e-06,
+      "loss": 0.3781,
+      "step": 1531
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.962077133914324e-06,
+      "loss": 0.4307,
+      "step": 1532
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9519807635522726e-06,
+      "loss": 0.4331,
+      "step": 1533
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9419076267001468e-06,
+      "loss": 0.415,
+      "step": 1534
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.931857752437689e-06,
+      "loss": 0.395,
+      "step": 1535
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.921831169777485e-06,
+      "loss": 0.4106,
+      "step": 1536
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9118279076648825e-06,
+      "loss": 0.3459,
+      "step": 1537
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9018479949779057e-06,
+      "loss": 0.4587,
+      "step": 1538
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8918914605271687e-06,
+      "loss": 0.4863,
+      "step": 1539
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8819583330558022e-06,
+      "loss": 0.5005,
+      "step": 1540
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.872048641239359e-06,
+      "loss": 0.4636,
+      "step": 1541
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8621624136857397e-06,
+      "loss": 0.4124,
+      "step": 1542
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8522996789351033e-06,
+      "loss": 0.4226,
+      "step": 1543
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.84246046545979e-06,
+      "loss": 0.446,
+      "step": 1544
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8326448016642452e-06,
+      "loss": 0.459,
+      "step": 1545
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.822852715884913e-06,
+      "loss": 0.4167,
+      "step": 1546
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8130842363901845e-06,
+      "loss": 0.406,
+      "step": 1547
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.8033393913802966e-06,
+      "loss": 0.5137,
+      "step": 1548
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7936182089872623e-06,
+      "loss": 0.4702,
+      "step": 1549
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7839207172747786e-06,
+      "loss": 0.3679,
+      "step": 1550
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7742469442381549e-06,
+      "loss": 0.439,
+      "step": 1551
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7645969178042188e-06,
+      "loss": 0.502,
+      "step": 1552
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7549706658312592e-06,
+      "loss": 0.4517,
+      "step": 1553
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.7453682161089203e-06,
+      "loss": 0.4033,
+      "step": 1554
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7357895963581373e-06,
+      "loss": 0.4565,
+      "step": 1555
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7262348342310497e-06,
+      "loss": 0.4502,
+      "step": 1556
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7167039573109258e-06,
+      "loss": 0.4775,
+      "step": 1557
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.7071969931120768e-06,
+      "loss": 0.3872,
+      "step": 1558
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.697713969079785e-06,
+      "loss": 0.4417,
+      "step": 1559
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6882549125902159e-06,
+      "loss": 0.4441,
+      "step": 1560
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.678819850950355e-06,
+      "loss": 0.3892,
+      "step": 1561
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6694088113979034e-06,
+      "loss": 0.375,
+      "step": 1562
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.660021821101222e-06,
+      "loss": 0.458,
+      "step": 1563
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6506589071592427e-06,
+      "loss": 0.3502,
+      "step": 1564
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6413200966013965e-06,
+      "loss": 0.3279,
+      "step": 1565
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6320054163875288e-06,
+      "loss": 0.3525,
+      "step": 1566
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6227148934078195e-06,
+      "loss": 0.4189,
+      "step": 1567
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.6134485544827116e-06,
+      "loss": 0.417,
+      "step": 1568
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.60420642636284e-06,
+      "loss": 0.4028,
+      "step": 1569
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5949885357289375e-06,
+      "loss": 0.4373,
+      "step": 1570
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5857949091917735e-06,
+      "loss": 0.4236,
+      "step": 1571
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5766255732920576e-06,
+      "loss": 0.4216,
+      "step": 1572
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.5674805545003934e-06,
+      "loss": 0.3462,
+      "step": 1573
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5583598792171727e-06,
+      "loss": 0.4377,
+      "step": 1574
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5492635737725138e-06,
+      "loss": 0.3854,
+      "step": 1575
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5401916644261816e-06,
+      "loss": 0.4148,
+      "step": 1576
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5311441773675173e-06,
+      "loss": 0.3767,
+      "step": 1577
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5221211387153522e-06,
+      "loss": 0.368,
+      "step": 1578
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5131225745179423e-06,
+      "loss": 0.4299,
+      "step": 1579
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.5041485107528885e-06,
+      "loss": 0.4282,
+      "step": 1580
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4951989733270622e-06,
+      "loss": 0.3455,
+      "step": 1581
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4862739880765309e-06,
+      "loss": 0.418,
+      "step": 1582
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4773735807664824e-06,
+      "loss": 0.4695,
+      "step": 1583
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4684977770911535e-06,
+      "loss": 0.4001,
+      "step": 1584
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4596466026737489e-06,
+      "loss": 0.4011,
+      "step": 1585
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4508200830663833e-06,
+      "loss": 0.4001,
+      "step": 1586
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4420182437499809e-06,
+      "loss": 0.4004,
+      "step": 1587
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4332411101342259e-06,
+      "loss": 0.543,
+      "step": 1588
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4244887075574844e-06,
+      "loss": 0.5068,
+      "step": 1589
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4157610612867202e-06,
+      "loss": 0.3323,
+      "step": 1590
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.4070581965174323e-06,
+      "loss": 0.3843,
+      "step": 1591
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.3983801383735784e-06,
+      "loss": 0.4343,
+      "step": 1592
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.389726911907502e-06,
+      "loss": 0.5251,
+      "step": 1593
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3810985420998624e-06,
+      "loss": 0.3505,
+      "step": 1594
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3724950538595616e-06,
+      "loss": 0.4338,
+      "step": 1595
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3639164720236708e-06,
+      "loss": 0.4746,
+      "step": 1596
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3553628213573589e-06,
+      "loss": 0.4568,
+      "step": 1597
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3468341265538243e-06,
+      "loss": 0.5518,
+      "step": 1598
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.338330412234221e-06,
+      "loss": 0.4675,
+      "step": 1599
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3298517029475854e-06,
+      "loss": 0.47,
+      "step": 1600
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3213980231707691e-06,
+      "loss": 0.4329,
+      "step": 1601
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3129693973083734e-06,
+      "loss": 0.499,
+      "step": 1602
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.3045658496926594e-06,
+      "loss": 0.4187,
+      "step": 1603
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2961874045835011e-06,
+      "loss": 0.4963,
+      "step": 1604
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2878340861682993e-06,
+      "loss": 0.3706,
+      "step": 1605
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.279505918561923e-06,
+      "loss": 0.4111,
+      "step": 1606
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2712029258066338e-06,
+      "loss": 0.3867,
+      "step": 1607
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2629251318720092e-06,
+      "loss": 0.3972,
+      "step": 1608
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2546725606548871e-06,
+      "loss": 0.4189,
+      "step": 1609
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.2464452359792932e-06,
+      "loss": 0.3977,
+      "step": 1610
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.238243181596367e-06,
+      "loss": 0.3831,
+      "step": 1611
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.230066421184296e-06,
+      "loss": 0.4022,
+      "step": 1612
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2219149783482464e-06,
+      "loss": 0.4739,
+      "step": 1613
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2137888766202998e-06,
+      "loss": 0.4236,
+      "step": 1614
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.2056881394593788e-06,
+      "loss": 0.4817,
+      "step": 1615
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1976127902511825e-06,
+      "loss": 0.4021,
+      "step": 1616
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.189562852308117e-06,
+      "loss": 0.4543,
+      "step": 1617
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1815383488692345e-06,
+      "loss": 0.4141,
+      "step": 1618
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1735393031001551e-06,
+      "loss": 0.4788,
+      "step": 1619
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1655657380930107e-06,
+      "loss": 0.4529,
+      "step": 1620
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1576176768663706e-06,
+      "loss": 0.4485,
+      "step": 1621
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1496951423651848e-06,
+      "loss": 0.4927,
+      "step": 1622
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1417981574607006e-06,
+      "loss": 0.3989,
+      "step": 1623
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1339267449504143e-06,
+      "loss": 0.3495,
+      "step": 1624
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1260809275579966e-06,
+      "loss": 0.4702,
+      "step": 1625
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1182607279332314e-06,
+      "loss": 0.4282,
+      "step": 1626
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1104661686519436e-06,
+      "loss": 0.3765,
+      "step": 1627
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.1026972722159412e-06,
+      "loss": 0.4148,
+      "step": 1628
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.0949540610529396e-06,
+      "loss": 0.4258,
+      "step": 1629
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.0872365575165168e-06,
+      "loss": 0.4075,
+      "step": 1630
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0795447838860273e-06,
+      "loss": 0.4456,
+      "step": 1631
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0718787623665504e-06,
+      "loss": 0.4043,
+      "step": 1632
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0642385150888213e-06,
+      "loss": 0.3792,
+      "step": 1633
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0566240641091686e-06,
+      "loss": 0.4836,
+      "step": 1634
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.049035431409452e-06,
+      "loss": 0.3889,
+      "step": 1635
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0414726388969964e-06,
+      "loss": 0.3473,
+      "step": 1636
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.03393570840453e-06,
+      "loss": 0.4326,
+      "step": 1637
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0264246616901197e-06,
+      "loss": 0.4175,
+      "step": 1638
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.018939520437111e-06,
+      "loss": 0.4006,
+      "step": 1639
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0114803062540647e-06,
+      "loss": 0.3745,
+      "step": 1640
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.0040470406746916e-06,
+      "loss": 0.3706,
+      "step": 1641
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.966397451577914e-07,
+      "loss": 0.4736,
+      "step": 1642
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.892584410871996e-07,
+      "loss": 0.4033,
+      "step": 1643
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.81903149771707e-07,
+      "loss": 0.4893,
+      "step": 1644
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.745738924450143e-07,
+      "loss": 0.3798,
+      "step": 1645
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.672706902656704e-07,
+      "loss": 0.4744,
+      "step": 1646
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.59993564317e-07,
+      "loss": 0.4592,
+      "step": 1647
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.52742535607053e-07,
+      "loss": 0.4717,
+      "step": 1648
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 9.455176250685338e-07,
+      "loss": 0.4243,
+      "step": 1649
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.383188535587573e-07,
+      "loss": 0.4739,
+      "step": 1650
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.311462418595718e-07,
+      "loss": 0.4705,
+      "step": 1651
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.239998106773074e-07,
+      "loss": 0.3538,
+      "step": 1652
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.168795806427156e-07,
+      "loss": 0.3894,
+      "step": 1653
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.097855723109083e-07,
+      "loss": 0.4475,
+      "step": 1654
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 9.027178061612984e-07,
+      "loss": 0.4391,
+      "step": 1655
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.956763025975413e-07,
+      "loss": 0.4392,
+      "step": 1656
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.886610819474772e-07,
+      "loss": 0.4949,
+      "step": 1657
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.81672164463071e-07,
+      "loss": 0.4058,
+      "step": 1658
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.747095703203523e-07,
+      "loss": 0.4602,
+      "step": 1659
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.677733196193605e-07,
+      "loss": 0.4158,
+      "step": 1660
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.608634323840848e-07,
+      "loss": 0.4448,
+      "step": 1661
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.53979928562404e-07,
+      "loss": 0.4121,
+      "step": 1662
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.471228280260401e-07,
+      "loss": 0.4626,
+      "step": 1663
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.402921505704842e-07,
+      "loss": 0.4106,
+      "step": 1664
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.334879159149478e-07,
+      "loss": 0.4468,
+      "step": 1665
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.267101437023084e-07,
+      "loss": 0.4714,
+      "step": 1666
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.199588534990521e-07,
+      "loss": 0.3948,
+      "step": 1667
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 8.132340647952119e-07,
+      "loss": 0.5496,
+      "step": 1668
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 8.065357970043163e-07,
+      "loss": 0.4161,
+      "step": 1669
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.998640694633297e-07,
+      "loss": 0.4509,
+      "step": 1670
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.932189014326008e-07,
+      "loss": 0.4741,
+      "step": 1671
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.866003120958022e-07,
+      "loss": 0.4147,
+      "step": 1672
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.800083205598796e-07,
+      "loss": 0.3967,
+      "step": 1673
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.734429458549941e-07,
+      "loss": 0.4312,
+      "step": 1674
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.669042069344678e-07,
+      "loss": 0.4666,
+      "step": 1675
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.60392122674728e-07,
+      "loss": 0.467,
+      "step": 1676
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.539067118752541e-07,
+      "loss": 0.4495,
+      "step": 1677
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.474479932585221e-07,
+      "loss": 0.5168,
+      "step": 1678
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.410159854699584e-07,
+      "loss": 0.3745,
+      "step": 1679
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.346107070778674e-07,
+      "loss": 0.4038,
+      "step": 1680
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.282321765733957e-07,
+      "loss": 0.5261,
+      "step": 1681
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.218804123704715e-07,
+      "loss": 0.4617,
+      "step": 1682
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.155554328057556e-07,
+      "loss": 0.4001,
+      "step": 1683
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.092572561385836e-07,
+      "loss": 0.428,
+      "step": 1684
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 7.029859005509087e-07,
+      "loss": 0.4373,
+      "step": 1685
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.967413841472625e-07,
+      "loss": 0.4512,
+      "step": 1686
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 6.905237249546947e-07,
+      "loss": 0.3931,
+      "step": 1687
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.843329409227206e-07,
+      "loss": 0.4832,
+      "step": 1688
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.781690499232729e-07,
+      "loss": 0.4431,
+      "step": 1689
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.720320697506399e-07,
+      "loss": 0.3324,
+      "step": 1690
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.659220181214343e-07,
+      "loss": 0.3894,
+      "step": 1691
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.598389126745209e-07,
+      "loss": 0.4226,
+      "step": 1692
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.537827709709765e-07,
+      "loss": 0.3892,
+      "step": 1693
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.477536104940407e-07,
+      "loss": 0.4175,
+      "step": 1694
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.417514486490572e-07,
+      "loss": 0.4326,
+      "step": 1695
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.357763027634323e-07,
+      "loss": 0.3799,
+      "step": 1696
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.298281900865788e-07,
+      "loss": 0.4622,
+      "step": 1697
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.239071277898678e-07,
+      "loss": 0.4773,
+      "step": 1698
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.18013132966585e-07,
+      "loss": 0.5129,
+      "step": 1699
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.121462226318675e-07,
+      "loss": 0.4282,
+      "step": 1700
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.063064137226693e-07,
+      "loss": 0.4951,
+      "step": 1701
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 6.004937230977048e-07,
+      "loss": 0.3821,
+      "step": 1702
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.947081675374033e-07,
+      "loss": 0.584,
+      "step": 1703
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.889497637438557e-07,
+      "loss": 0.3969,
+      "step": 1704
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.832185283407732e-07,
+      "loss": 0.4636,
+      "step": 1705
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 5.775144778734299e-07,
+      "loss": 0.3833,
+      "step": 1706
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.71837628808628e-07,
+      "loss": 0.4482,
+      "step": 1707
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.661879975346385e-07,
+      "loss": 0.5203,
+      "step": 1708
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.605656003611593e-07,
+      "loss": 0.3979,
+      "step": 1709
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.549704535192679e-07,
+      "loss": 0.3945,
+      "step": 1710
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.494025731613739e-07,
+      "loss": 0.4202,
+      "step": 1711
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.438619753611707e-07,
+      "loss": 0.3906,
+      "step": 1712
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.383486761135926e-07,
+      "loss": 0.4119,
+      "step": 1713
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.328626913347656e-07,
+      "loss": 0.4734,
+      "step": 1714
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.274040368619648e-07,
+      "loss": 0.405,
+      "step": 1715
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.219727284535625e-07,
+      "loss": 0.3682,
+      "step": 1716
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.165687817889919e-07,
+      "loss": 0.4353,
+      "step": 1717
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.111922124686919e-07,
+      "loss": 0.4309,
+      "step": 1718
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.058430360140676e-07,
+      "loss": 0.3561,
+      "step": 1719
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 5.00521267867451e-07,
+      "loss": 0.3647,
+      "step": 1720
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.952269233920404e-07,
+      "loss": 0.53,
+      "step": 1721
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.899600178718722e-07,
+      "loss": 0.4001,
+      "step": 1722
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.84720566511766e-07,
+      "loss": 0.428,
+      "step": 1723
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.795085844372927e-07,
+      "loss": 0.3624,
+      "step": 1724
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.7432408669471517e-07,
+      "loss": 0.4033,
+      "step": 1725
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.6916708825095467e-07,
+      "loss": 0.416,
+      "step": 1726
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.64037603993549e-07,
+      "loss": 0.4224,
+      "step": 1727
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.5893564873060514e-07,
+      "loss": 0.4429,
+      "step": 1728
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.538612371907547e-07,
+      "loss": 0.4678,
+      "step": 1729
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.488143840231174e-07,
+      "loss": 0.5139,
+      "step": 1730
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.4379510379725543e-07,
+      "loss": 0.429,
+      "step": 1731
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.388034110031303e-07,
+      "loss": 0.3962,
+      "step": 1732
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.3383932005106486e-07,
+      "loss": 0.5022,
+      "step": 1733
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.28902845271697e-07,
+      "loss": 0.4565,
+      "step": 1734
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.2399400091594154e-07,
+      "loss": 0.4475,
+      "step": 1735
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1911280115494726e-07,
+      "loss": 0.4297,
+      "step": 1736
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.1425926008005877e-07,
+      "loss": 0.3657,
+      "step": 1737
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0943339170277265e-07,
+      "loss": 0.4241,
+      "step": 1738
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0463520995469575e-07,
+      "loss": 0.4666,
+      "step": 1739
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.998647286875146e-07,
+      "loss": 0.4031,
+      "step": 1740
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.9512196167294293e-07,
+      "loss": 0.4128,
+      "step": 1741
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.904069226026852e-07,
+      "loss": 0.4133,
+      "step": 1742
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.857196250884021e-07,
+      "loss": 0.4065,
+      "step": 1743
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.810600826616706e-07,
+      "loss": 0.4189,
+      "step": 1744
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.7642830877394067e-07,
+      "loss": 0.4705,
+      "step": 1745
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.718243167964974e-07,
+      "loss": 0.4387,
+      "step": 1746
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.6724812002042145e-07,
+      "loss": 0.4116,
+      "step": 1747
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.626997316565584e-07,
+      "loss": 0.3556,
+      "step": 1748
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.5817916483547045e-07,
+      "loss": 0.4377,
+      "step": 1749
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.53686432607403e-07,
+      "loss": 0.3521,
+      "step": 1750
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4922154794225004e-07,
+      "loss": 0.3404,
+      "step": 1751
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4478452372950886e-07,
+      "loss": 0.408,
+      "step": 1752
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4037537277825215e-07,
+      "loss": 0.4888,
+      "step": 1753
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.3599410781708255e-07,
+      "loss": 0.4705,
+      "step": 1754
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.3164074149410143e-07,
+      "loss": 0.405,
+      "step": 1755
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.273152863768747e-07,
+      "loss": 0.4524,
+      "step": 1756
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.2301775495238277e-07,
+      "loss": 0.3845,
+      "step": 1757
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1874815962700254e-07,
+      "loss": 0.4351,
+      "step": 1758
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.1450651272645905e-07,
+      "loss": 0.4429,
+      "step": 1759
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.102928264957994e-07,
+      "loss": 0.4836,
+      "step": 1760
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.0610711309934647e-07,
+      "loss": 0.3982,
+      "step": 1761
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.0194938462067e-07,
+      "loss": 0.3975,
+      "step": 1762
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 2.9781965306255324e-07,
+      "loss": 0.4082,
+      "step": 1763
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.9371793034695726e-07,
+      "loss": 0.3806,
+      "step": 1764
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8964422831498363e-07,
+      "loss": 0.3835,
+      "step": 1765
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.855985587268428e-07,
+      "loss": 0.3975,
+      "step": 1766
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.8158093326181577e-07,
+      "loss": 0.4236,
+      "step": 1767
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.7759136351823033e-07,
+      "loss": 0.4485,
+      "step": 1768
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.7362986101341803e-07,
+      "loss": 0.3743,
+      "step": 1769
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.696964371836841e-07,
+      "loss": 0.4243,
+      "step": 1770
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.6579110338427416e-07,
+      "loss": 0.4546,
+      "step": 1771
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.6191387088934426e-07,
+      "loss": 0.4507,
+      "step": 1772
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.580647508919232e-07,
+      "loss": 0.4194,
+      "step": 1773
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.542437545038845e-07,
+      "loss": 0.4441,
+      "step": 1774
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.504508927559113e-07,
+      "loss": 0.4868,
+      "step": 1775
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.4668617659746707e-07,
+      "loss": 0.3638,
+      "step": 1776
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.4294961689676487e-07,
+      "loss": 0.4019,
+      "step": 1777
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.392412244407294e-07,
+      "loss": 0.4346,
+      "step": 1778
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.3556100993497255e-07,
+      "loss": 0.356,
+      "step": 1779
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.319089840037614e-07,
+      "loss": 0.3335,
+      "step": 1780
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.282851571899869e-07,
+      "loss": 0.3826,
+      "step": 1781
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.24689539955133e-07,
+      "loss": 0.4316,
+      "step": 1782
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 2.2112214267924093e-07,
+      "loss": 0.4246,
+      "step": 1783
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1758297566089382e-07,
+      "loss": 0.4666,
+      "step": 1784
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1407204911717105e-07,
+      "loss": 0.4585,
+      "step": 1785
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.1058937318363059e-07,
+      "loss": 0.4243,
+      "step": 1786
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.0713495791427006e-07,
+      "loss": 0.4446,
+      "step": 1787
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.0370881328150682e-07,
+      "loss": 0.4089,
+      "step": 1788
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 2.0031094917614235e-07,
+      "loss": 0.4121,
+      "step": 1789
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9694137540733682e-07,
+      "loss": 0.4631,
+      "step": 1790
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9360010170257903e-07,
+      "loss": 0.4614,
+      "step": 1791
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9028713770766205e-07,
+      "loss": 0.364,
+      "step": 1792
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8700249298665096e-07,
+      "loss": 0.3921,
+      "step": 1793
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8374617702185627e-07,
+      "loss": 0.3667,
+      "step": 1794
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.8051819921380942e-07,
+      "loss": 0.4683,
+      "step": 1795
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7731856888123068e-07,
+      "loss": 0.4592,
+      "step": 1796
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7414729526101015e-07,
+      "loss": 0.4788,
+      "step": 1797
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.7100438750817018e-07,
+      "loss": 0.4323,
+      "step": 1798
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.678898546958485e-07,
+      "loss": 0.438,
+      "step": 1799
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.6480370581526629e-07,
+      "loss": 0.4429,
+      "step": 1800
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.617459497757068e-07,
+      "loss": 0.4747,
+      "step": 1801
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.5871659540448448e-07,
+      "loss": 0.4851,
+      "step": 1802
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.557156514469238e-07,
+      "loss": 0.447,
+      "step": 1803
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.527431265663304e-07,
+      "loss": 0.3896,
+      "step": 1804
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4979902934397106e-07,
+      "loss": 0.4077,
+      "step": 1805
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.468833682790427e-07,
+      "loss": 0.4324,
+      "step": 1806
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4399615178865457e-07,
+      "loss": 0.3845,
+      "step": 1807
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.4113738820779598e-07,
+      "loss": 0.386,
+      "step": 1808
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3830708578931984e-07,
+      "loss": 0.4683,
+      "step": 1809
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3550525270391578e-07,
+      "loss": 0.4866,
+      "step": 1810
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.3273189704008371e-07,
+      "loss": 0.3402,
+      "step": 1811
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2998702680411702e-07,
+      "loss": 0.3894,
+      "step": 1812
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2727064992007154e-07,
+      "loss": 0.4409,
+      "step": 1813
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.2458277422974895e-07,
+      "loss": 0.3892,
+      "step": 1814
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.219234074926734e-07,
+      "loss": 0.4636,
+      "step": 1815
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.192925573860637e-07,
+      "loss": 0.4189,
+      "step": 1816
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1669023150481906e-07,
+      "loss": 0.415,
+      "step": 1817
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1411643736149336e-07,
+      "loss": 0.397,
+      "step": 1818
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.1157118238626752e-07,
+      "loss": 0.3877,
+      "step": 1819
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.0905447392694058e-07,
+      "loss": 0.4629,
+      "step": 1820
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.0656631924889749e-07,
+      "loss": 0.4973,
+      "step": 1821
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.0410672553509693e-07,
+      "loss": 0.5037,
+      "step": 1822
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.016756998860402e-07,
+      "loss": 0.3833,
+      "step": 1823
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.927324931976012e-08,
+      "loss": 0.4529,
+      "step": 1824
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.689938077179883e-08,
+      "loss": 0.4126,
+      "step": 1825
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.455410109518116e-08,
+      "loss": 0.4221,
+      "step": 1826
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 9.223741706040456e-08,
+      "loss": 0.4243,
+      "step": 1827
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.994933535541151e-08,
+      "loss": 0.4373,
+      "step": 1828
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.7689862585576e-08,
+      "loss": 0.4308,
+      "step": 1829
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.545900527368146e-08,
+      "loss": 0.4028,
+      "step": 1830
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.325676985990073e-08,
+      "loss": 0.426,
+      "step": 1831
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 8.108316270178162e-08,
+      "loss": 0.3273,
+      "step": 1832
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.893819007422476e-08,
+      "loss": 0.4265,
+      "step": 1833
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.682185816946686e-08,
+      "loss": 0.4094,
+      "step": 1834
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.47341730970641e-08,
+      "loss": 0.4031,
+      "step": 1835
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.267514088387107e-08,
+      "loss": 0.3726,
+      "step": 1836
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 7.064476747402626e-08,
+      "loss": 0.3958,
+      "step": 1837
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.864305872893773e-08,
+      "loss": 0.437,
+      "step": 1838
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.667002042725745e-08,
+      "loss": 0.4319,
+      "step": 1839
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 6.472565826487142e-08,
+      "loss": 0.4076,
+      "step": 1840
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.280997785488407e-08,
+      "loss": 0.3562,
+      "step": 1841
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 6.092298472759716e-08,
+      "loss": 0.4231,
+      "step": 1842
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.9064684330496505e-08,
+      "loss": 0.3882,
+      "step": 1843
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.7235082028235243e-08,
+      "loss": 0.4592,
+      "step": 1844
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.5434183102622827e-08,
+      "loss": 0.459,
+      "step": 1845
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.366199275260164e-08,
+      "loss": 0.3484,
+      "step": 1846
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.191851609423926e-08,
+      "loss": 0.4001,
+      "step": 1847
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 5.020375816070844e-08,
+      "loss": 0.3885,
+      "step": 1848
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.851772390227827e-08,
+      "loss": 0.4788,
+      "step": 1849
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.6860418186291945e-08,
+      "loss": 0.4003,
+      "step": 1850
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.5231845797163444e-08,
+      "loss": 0.4487,
+      "step": 1851
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.3632011436351984e-08,
+      "loss": 0.4031,
+      "step": 1852
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.2060919722356486e-08,
+      "loss": 0.4519,
+      "step": 1853
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.0518575190701124e-08,
+      "loss": 0.3608,
+      "step": 1854
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.900498229391869e-08,
+      "loss": 0.5022,
+      "step": 1855
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.7520145401540585e-08,
+      "loss": 0.5139,
+      "step": 1856
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.606406880008462e-08,
+      "loss": 0.3948,
+      "step": 1857
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.463675669304056e-08,
+      "loss": 0.3777,
+      "step": 1858
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.323821320086018e-08,
+      "loss": 0.4646,
+      "step": 1859
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.1868442360943886e-08,
+      "loss": 0.4524,
+      "step": 1860
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.052744812762742e-08,
+      "loss": 0.4043,
+      "step": 1861
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.9215234372176327e-08,
+      "loss": 0.3962,
+      "step": 1862
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.7931804882767033e-08,
+      "loss": 0.3826,
+      "step": 1863
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.667716336448356e-08,
+      "loss": 0.4426,
+      "step": 1864
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.545131343929974e-08,
+      "loss": 0.364,
+      "step": 1865
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.425425864607367e-08,
+      "loss": 0.3931,
+      "step": 1866
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.3086002440535493e-08,
+      "loss": 0.4241,
+      "step": 1867
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.194654819527853e-08,
+      "loss": 0.4873,
+      "step": 1868
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 2.0835899199748156e-08,
+      "loss": 0.4109,
+      "step": 1869
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.9754058660234053e-08,
+      "loss": 0.4077,
+      "step": 1870
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.8701029699857987e-08,
+      "loss": 0.3989,
+      "step": 1871
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.7676815358568244e-08,
+      "loss": 0.47,
+      "step": 1872
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.668141859312855e-08,
+      "loss": 0.4072,
+      "step": 1873
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.5714842277111396e-08,
+      "loss": 0.3821,
+      "step": 1874
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.477708920088694e-08,
+      "loss": 0.5063,
+      "step": 1875
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.3868162071618562e-08,
+      "loss": 0.4302,
+      "step": 1876
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.298806351325066e-08,
+      "loss": 0.4524,
+      "step": 1877
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.2136796066507527e-08,
+      "loss": 0.3899,
+      "step": 1878
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.1314362188877825e-08,
+      "loss": 0.4353,
+      "step": 1879
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.0520764254614568e-08,
+      "loss": 0.4512,
+      "step": 1880
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.75600455472292e-09,
+      "loss": 0.3689,
+      "step": 1881
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 9.02008529695908e-09,
+      "loss": 0.3838,
+      "step": 1882
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 8.313008605819183e-09,
+      "loss": 0.3376,
+      "step": 1883
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 7.634776522534858e-09,
+      "loss": 0.4338,
+      "step": 1884
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.985391005066566e-09,
+      "loss": 0.3958,
+      "step": 1885
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 6.364853928101378e-09,
+      "loss": 0.45,
+      "step": 1886
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.773167083041875e-09,
+      "loss": 0.45,
+      "step": 1887
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 5.21033217800726e-09,
+      "loss": 0.4592,
+      "step": 1888
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.676350837822252e-09,
+      "loss": 0.4739,
+      "step": 1889
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.171224604017088e-09,
+      "loss": 0.4634,
+      "step": 1890
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.69495493481975e-09,
+      "loss": 0.4463,
+      "step": 1891
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.2475432051559676e-09,
+      "loss": 0.3488,
+      "step": 1892
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.8289907066403334e-09,
+      "loss": 0.4026,
+      "step": 1893
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.4392986475751945e-09,
+      "loss": 0.4639,
+      "step": 1894
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 2.078468152948432e-09,
+      "loss": 0.4144,
+      "step": 1895
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.7465002644256878e-09,
+      "loss": 0.4202,
+      "step": 1896
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.4433959403525877e-09,
+      "loss": 0.4636,
+      "step": 1897
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.1691560557491876e-09,
+      "loss": 0.396,
+      "step": 1898
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 9.237814023088653e-10,
+      "loss": 0.4216,
+      "step": 1899
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 7.072726883927683e-10,
+      "loss": 0.4668,
+      "step": 1900
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.196305390320344e-10,
+      "loss": 0.3979,
+      "step": 1901
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.60855495922241e-10,
+      "loss": 0.4165,
+      "step": 1902
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 2.3094801742673534e-10,
+      "loss": 0.3585,
+      "step": 1903
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.2990847856886313e-10,
+      "loss": 0.4233,
+      "step": 1904
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 5.773717103529919e-11,
+      "loss": 0.4133,
+      "step": 1905
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.4434303176047437e-11,
+      "loss": 0.3691,
+      "step": 1906
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 0.0,
+      "loss": 0.4199,
+      "step": 1907
+    },
+    {
+      "epoch": 1.0,
+      "step": 1907,
+      "total_flos": 793307121844224.0,
+      "train_loss": 0.48644651463973193,
+      "train_runtime": 40253.9366,
+      "train_samples_per_second": 12.126,
+      "train_steps_per_second": 0.047
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 1907,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1100,
+  "total_flos": 793307121844224.0,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": null
+}