diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,12036 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 2.997752808988764,
+  "eval_steps": 500,
+  "global_step": 2001,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 8.19672131147541e-08,
+      "loss": 0.6761,
+      "step": 1
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.639344262295082e-07,
+      "loss": 0.6723,
+      "step": 2
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 2.459016393442623e-07,
+      "loss": 0.716,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.278688524590164e-07,
+      "loss": 0.6672,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.0983606557377047e-07,
+      "loss": 0.7021,
+      "step": 5
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.918032786885246e-07,
+      "loss": 0.7126,
+      "step": 6
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.737704918032787e-07,
+      "loss": 0.6171,
+      "step": 7
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.557377049180328e-07,
+      "loss": 0.674,
+      "step": 8
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 7.377049180327869e-07,
+      "loss": 0.7363,
+      "step": 9
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.196721311475409e-07,
+      "loss": 0.6252,
+      "step": 10
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.016393442622952e-07,
+      "loss": 0.701,
+      "step": 11
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 9.836065573770493e-07,
+      "loss": 0.6565,
+      "step": 12
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.0655737704918034e-06,
+      "loss": 0.6797,
+      "step": 13
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.1475409836065575e-06,
+      "loss": 0.6497,
+      "step": 14
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.2295081967213116e-06,
+      "loss": 0.6321,
+      "step": 15
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3114754098360657e-06,
+      "loss": 0.6629,
+      "step": 16
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.3934426229508196e-06,
+      "loss": 0.573,
+      "step": 17
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.4754098360655739e-06,
+      "loss": 0.6581,
+      "step": 18
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.557377049180328e-06,
+      "loss": 0.6611,
+      "step": 19
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6393442622950819e-06,
+      "loss": 0.6616,
+      "step": 20
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.7213114754098362e-06,
+      "loss": 0.7401,
+      "step": 21
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8032786885245903e-06,
+      "loss": 0.5323,
+      "step": 22
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8852459016393442e-06,
+      "loss": 0.5931,
+      "step": 23
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 1.9672131147540985e-06,
+      "loss": 0.5919,
+      "step": 24
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.0491803278688526e-06,
+      "loss": 0.6706,
+      "step": 25
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.1311475409836067e-06,
+      "loss": 0.669,
+      "step": 26
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.213114754098361e-06,
+      "loss": 0.6188,
+      "step": 27
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.295081967213115e-06,
+      "loss": 0.6978,
+      "step": 28
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.377049180327869e-06,
+      "loss": 0.7513,
+      "step": 29
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.459016393442623e-06,
+      "loss": 0.5814,
+      "step": 30
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.5409836065573773e-06,
+      "loss": 0.6299,
+      "step": 31
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.6229508196721314e-06,
+      "loss": 0.7074,
+      "step": 32
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.704918032786886e-06,
+      "loss": 0.6554,
+      "step": 33
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.786885245901639e-06,
+      "loss": 0.624,
+      "step": 34
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.8688524590163937e-06,
+      "loss": 0.6478,
+      "step": 35
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 2.9508196721311478e-06,
+      "loss": 0.5773,
+      "step": 36
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.032786885245902e-06,
+      "loss": 0.5796,
+      "step": 37
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.114754098360656e-06,
+      "loss": 0.6698,
+      "step": 38
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.1967213114754105e-06,
+      "loss": 0.7186,
+      "step": 39
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.2786885245901638e-06,
+      "loss": 0.5754,
+      "step": 40
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.3606557377049183e-06,
+      "loss": 0.6216,
+      "step": 41
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.4426229508196724e-06,
+      "loss": 0.5774,
+      "step": 42
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.5245901639344265e-06,
+      "loss": 0.6411,
+      "step": 43
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.6065573770491806e-06,
+      "loss": 0.6059,
+      "step": 44
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.6885245901639347e-06,
+      "loss": 0.6565,
+      "step": 45
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.7704918032786884e-06,
+      "loss": 0.6685,
+      "step": 46
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.852459016393443e-06,
+      "loss": 0.7742,
+      "step": 47
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 3.934426229508197e-06,
+      "loss": 0.628,
+      "step": 48
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.016393442622951e-06,
+      "loss": 0.6035,
+      "step": 49
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.098360655737705e-06,
+      "loss": 0.6134,
+      "step": 50
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.180327868852459e-06,
+      "loss": 0.5835,
+      "step": 51
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.2622950819672135e-06,
+      "loss": 0.6619,
+      "step": 52
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.3442622950819676e-06,
+      "loss": 0.6015,
+      "step": 53
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.426229508196722e-06,
+      "loss": 0.5854,
+      "step": 54
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.508196721311476e-06,
+      "loss": 0.5894,
+      "step": 55
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.59016393442623e-06,
+      "loss": 0.5941,
+      "step": 56
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.672131147540984e-06,
+      "loss": 0.5916,
+      "step": 57
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.754098360655738e-06,
+      "loss": 0.5683,
+      "step": 58
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.836065573770492e-06,
+      "loss": 0.5923,
+      "step": 59
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.918032786885246e-06,
+      "loss": 0.6848,
+      "step": 60
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5e-06,
+      "loss": 0.6484,
+      "step": 61
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.999996722020724e-06,
+      "loss": 0.5775,
+      "step": 62
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.99998688809149e-06,
+      "loss": 0.5868,
+      "step": 63
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999970498238088e-06,
+      "loss": 0.5943,
+      "step": 64
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.9999475525034974e-06,
+      "loss": 0.572,
+      "step": 65
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999918050947891e-06,
+      "loss": 0.6075,
+      "step": 66
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999881993648633e-06,
+      "loss": 0.6677,
+      "step": 67
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.999839380700281e-06,
+      "loss": 0.631,
+      "step": 68
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.99979021221458e-06,
+      "loss": 0.5494,
+      "step": 69
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.99973448832047e-06,
+      "loss": 0.6002,
+      "step": 70
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9996722091640805e-06,
+      "loss": 0.5477,
+      "step": 71
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999603374908732e-06,
+      "loss": 0.6094,
+      "step": 72
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999527985734932e-06,
+      "loss": 0.5381,
+      "step": 73
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999446041840382e-06,
+      "loss": 0.6069,
+      "step": 74
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.999357543439969e-06,
+      "loss": 0.6533,
+      "step": 75
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.99926249076577e-06,
+      "loss": 0.6675,
+      "step": 76
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.999160884067051e-06,
+      "loss": 0.6182,
+      "step": 77
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.999052723610262e-06,
+      "loss": 0.6043,
+      "step": 78
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9989380096790416e-06,
+      "loss": 0.61,
+      "step": 79
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998816742574214e-06,
+      "loss": 0.5656,
+      "step": 80
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998688922613788e-06,
+      "loss": 0.6517,
+      "step": 81
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998554550132956e-06,
+      "loss": 0.6521,
+      "step": 82
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.998413625484095e-06,
+      "loss": 0.6573,
+      "step": 83
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.998266149036763e-06,
+      "loss": 0.6796,
+      "step": 84
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9981121211777e-06,
+      "loss": 0.6498,
+      "step": 85
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9979515423108255e-06,
+      "loss": 0.6605,
+      "step": 86
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.997784412857239e-06,
+      "loss": 0.6944,
+      "step": 87
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.99761073325522e-06,
+      "loss": 0.6093,
+      "step": 88
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.99743050396022e-06,
+      "loss": 0.6733,
+      "step": 89
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.9972437254448705e-06,
+      "loss": 0.625,
+      "step": 90
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.997050398198977e-06,
+      "loss": 0.5802,
+      "step": 91
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9968505227295176e-06,
+      "loss": 0.5964,
+      "step": 92
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9966440995606415e-06,
+      "loss": 0.6095,
+      "step": 93
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.99643112923367e-06,
+      "loss": 0.6376,
+      "step": 94
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9962116123070925e-06,
+      "loss": 0.6364,
+      "step": 95
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.995985549356568e-06,
+      "loss": 0.6177,
+      "step": 96
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9957529409749185e-06,
+      "loss": 0.6,
+      "step": 97
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.995513787772132e-06,
+      "loss": 0.6116,
+      "step": 98
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.995268090375362e-06,
+      "loss": 0.6042,
+      "step": 99
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.995015849428921e-06,
+      "loss": 0.5356,
+      "step": 100
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.99475706559428e-06,
+      "loss": 0.7096,
+      "step": 101
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.994491739550069e-06,
+      "loss": 0.7266,
+      "step": 102
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.994219871992077e-06,
+      "loss": 0.6318,
+      "step": 103
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9939414636332434e-06,
+      "loss": 0.6112,
+      "step": 104
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.993656515203662e-06,
+      "loss": 0.6076,
+      "step": 105
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.993365027450576e-06,
+      "loss": 0.6624,
+      "step": 106
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.99306700113838e-06,
+      "loss": 0.651,
+      "step": 107
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.992762437048612e-06,
+      "loss": 0.6102,
+      "step": 108
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.9924513359799555e-06,
+      "loss": 0.7002,
+      "step": 109
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.992133698748237e-06,
+      "loss": 0.6259,
+      "step": 110
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.991809526186424e-06,
+      "loss": 0.7244,
+      "step": 111
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.99147881914462e-06,
+      "loss": 0.6634,
+      "step": 112
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.991141578490066e-06,
+      "loss": 0.6946,
+      "step": 113
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.990797805107138e-06,
+      "loss": 0.6544,
+      "step": 114
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.990447499897339e-06,
+      "loss": 0.6008,
+      "step": 115
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.990090663779305e-06,
+      "loss": 0.569,
+      "step": 116
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.989727297688797e-06,
+      "loss": 0.6246,
+      "step": 117
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.989357402578699e-06,
+      "loss": 0.6341,
+      "step": 118
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.98898097941902e-06,
+      "loss": 0.6458,
+      "step": 119
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.988598029196885e-06,
+      "loss": 0.6828,
+      "step": 120
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.988208552916535e-06,
+      "loss": 0.6724,
+      "step": 121
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.987812551599328e-06,
+      "loss": 0.6199,
+      "step": 122
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.98741002628373e-06,
+      "loss": 0.5968,
+      "step": 123
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9870009780253185e-06,
+      "loss": 0.6202,
+      "step": 124
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9865854078967715e-06,
+      "loss": 0.6136,
+      "step": 125
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.986163316987877e-06,
+      "loss": 0.651,
+      "step": 126
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.985734706405516e-06,
+      "loss": 0.6177,
+      "step": 127
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.985299577273672e-06,
+      "loss": 0.6689,
+      "step": 128
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.9848579307334195e-06,
+      "loss": 0.5994,
+      "step": 129
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.984409767942926e-06,
+      "loss": 0.6221,
+      "step": 130
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.983955090077445e-06,
+      "loss": 0.6143,
+      "step": 131
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.983493898329316e-06,
+      "loss": 0.6091,
+      "step": 132
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.983026193907962e-06,
+      "loss": 0.6061,
+      "step": 133
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.982551978039882e-06,
+      "loss": 0.5806,
+      "step": 134
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.982071251968653e-06,
+      "loss": 0.6395,
+      "step": 135
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.9815840169549216e-06,
+      "loss": 0.6702,
+      "step": 136
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.981090274276406e-06,
+      "loss": 0.7482,
+      "step": 137
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.980590025227888e-06,
+      "loss": 0.6407,
+      "step": 138
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.980083271121215e-06,
+      "loss": 0.671,
+      "step": 139
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.979570013285286e-06,
+      "loss": 0.522,
+      "step": 140
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.979050253066064e-06,
+      "loss": 0.6292,
+      "step": 141
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.978523991826556e-06,
+      "loss": 0.6229,
+      "step": 142
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.977991230946824e-06,
+      "loss": 0.6597,
+      "step": 143
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.97745197182397e-06,
+      "loss": 0.6406,
+      "step": 144
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.976906215872137e-06,
+      "loss": 0.5841,
+      "step": 145
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.976353964522509e-06,
+      "loss": 0.6574,
+      "step": 146
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.975795219223299e-06,
+      "loss": 0.5837,
+      "step": 147
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.97522998143975e-06,
+      "loss": 0.6075,
+      "step": 148
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.974658252654135e-06,
+      "loss": 0.5642,
+      "step": 149
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.974080034365746e-06,
+      "loss": 0.6342,
+      "step": 150
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.973495328090891e-06,
+      "loss": 0.5901,
+      "step": 151
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.972904135362895e-06,
+      "loss": 0.6706,
+      "step": 152
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.972306457732091e-06,
+      "loss": 0.632,
+      "step": 153
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.971702296765821e-06,
+      "loss": 0.6075,
+      "step": 154
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.971091654048427e-06,
+      "loss": 0.6302,
+      "step": 155
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.970474531181245e-06,
+      "loss": 0.5482,
+      "step": 156
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.96985092978261e-06,
+      "loss": 0.6682,
+      "step": 157
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.9692208514878445e-06,
+      "loss": 0.5676,
+      "step": 158
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.968584297949255e-06,
+      "loss": 0.7012,
+      "step": 159
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.967941270836128e-06,
+      "loss": 0.6688,
+      "step": 160
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.967291771834727e-06,
+      "loss": 0.6586,
+      "step": 161
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.966635802648289e-06,
+      "loss": 0.6643,
+      "step": 162
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.965973364997015e-06,
+      "loss": 0.6209,
+      "step": 163
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9653044606180725e-06,
+      "loss": 0.6063,
+      "step": 164
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.964629091265583e-06,
+      "loss": 0.726,
+      "step": 165
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.963947258710626e-06,
+      "loss": 0.604,
+      "step": 166
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.963258964741227e-06,
+      "loss": 0.565,
+      "step": 167
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.9625642111623555e-06,
+      "loss": 0.662,
+      "step": 168
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.961862999795923e-06,
+      "loss": 0.5711,
+      "step": 169
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.961155332480775e-06,
+      "loss": 0.5919,
+      "step": 170
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.960441211072686e-06,
+      "loss": 0.6247,
+      "step": 171
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.959720637444356e-06,
+      "loss": 0.5887,
+      "step": 172
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.958993613485406e-06,
+      "loss": 0.661,
+      "step": 173
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.95826014110237e-06,
+      "loss": 0.6851,
+      "step": 174
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.957520222218695e-06,
+      "loss": 0.6626,
+      "step": 175
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.9567738587747314e-06,
+      "loss": 0.5471,
+      "step": 176
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.956021052727731e-06,
+      "loss": 0.6451,
+      "step": 177
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9552618060518405e-06,
+      "loss": 0.6277,
+      "step": 178
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.954496120738094e-06,
+      "loss": 0.5807,
+      "step": 179
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.953723998794413e-06,
+      "loss": 0.5631,
+      "step": 180
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.952945442245598e-06,
+      "loss": 0.6253,
+      "step": 181
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.9521604531333215e-06,
+      "loss": 0.572,
+      "step": 182
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.951369033516127e-06,
+      "loss": 0.5782,
+      "step": 183
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.9505711854694185e-06,
+      "loss": 0.6427,
+      "step": 184
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.949766911085461e-06,
+      "loss": 0.5894,
+      "step": 185
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.948956212473371e-06,
+      "loss": 0.5976,
+      "step": 186
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.948139091759108e-06,
+      "loss": 0.5425,
+      "step": 187
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.947315551085479e-06,
+      "loss": 0.6366,
+      "step": 188
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.946485592612122e-06,
+      "loss": 0.6142,
+      "step": 189
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.945649218515507e-06,
+      "loss": 0.6244,
+      "step": 190
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.944806430988927e-06,
+      "loss": 0.5596,
+      "step": 191
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.943957232242494e-06,
+      "loss": 0.6424,
+      "step": 192
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.943101624503133e-06,
+      "loss": 0.6272,
+      "step": 193
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.942239610014575e-06,
+      "loss": 0.5966,
+      "step": 194
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.941371191037353e-06,
+      "loss": 0.6485,
+      "step": 195
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.940496369848795e-06,
+      "loss": 0.6308,
+      "step": 196
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.939615148743017e-06,
+      "loss": 0.6047,
+      "step": 197
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.938727530030919e-06,
+      "loss": 0.5873,
+      "step": 198
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.937833516040177e-06,
+      "loss": 0.6599,
+      "step": 199
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.936933109115239e-06,
+      "loss": 0.5939,
+      "step": 200
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.936026311617316e-06,
+      "loss": 0.6251,
+      "step": 201
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.935113125924379e-06,
+      "loss": 0.6417,
+      "step": 202
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.9341935544311536e-06,
+      "loss": 0.5841,
+      "step": 203
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.933267599549106e-06,
+      "loss": 0.6066,
+      "step": 204
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.932335263706446e-06,
+      "loss": 0.6134,
+      "step": 205
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.931396549348115e-06,
+      "loss": 0.5698,
+      "step": 206
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.930451458935783e-06,
+      "loss": 0.635,
+      "step": 207
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.929499994947839e-06,
+      "loss": 0.6165,
+      "step": 208
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.928542159879386e-06,
+      "loss": 0.6167,
+      "step": 209
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.9275779562422355e-06,
+      "loss": 0.6663,
+      "step": 210
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.926607386564898e-06,
+      "loss": 0.6235,
+      "step": 211
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.9256304533925804e-06,
+      "loss": 0.6469,
+      "step": 212
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.924647159287176e-06,
+      "loss": 0.5697,
+      "step": 213
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.923657506827259e-06,
+      "loss": 0.5889,
+      "step": 214
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.922661498608077e-06,
+      "loss": 0.6645,
+      "step": 215
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.921659137241544e-06,
+      "loss": 0.674,
+      "step": 216
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.920650425356239e-06,
+      "loss": 0.6413,
+      "step": 217
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.919635365597391e-06,
+      "loss": 0.6093,
+      "step": 218
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9186139606268735e-06,
+      "loss": 0.6794,
+      "step": 219
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9175862131232035e-06,
+      "loss": 0.6325,
+      "step": 220
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.916552125781529e-06,
+      "loss": 0.6077,
+      "step": 221
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.915511701313622e-06,
+      "loss": 0.5717,
+      "step": 222
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.9144649424478765e-06,
+      "loss": 0.6746,
+      "step": 223
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.913411851929295e-06,
+      "loss": 0.6442,
+      "step": 224
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.912352432519484e-06,
+      "loss": 0.5812,
+      "step": 225
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.911286686996648e-06,
+      "loss": 0.6622,
+      "step": 226
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.910214618155579e-06,
+      "loss": 0.5944,
+      "step": 227
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.909136228807655e-06,
+      "loss": 0.7183,
+      "step": 228
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.908051521780824e-06,
+      "loss": 0.6217,
+      "step": 229
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.906960499919606e-06,
+      "loss": 0.6871,
+      "step": 230
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.905863166085076e-06,
+      "loss": 0.6696,
+      "step": 231
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.904759523154866e-06,
+      "loss": 0.6337,
+      "step": 232
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.903649574023151e-06,
+      "loss": 0.5539,
+      "step": 233
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.902533321600641e-06,
+      "loss": 0.6833,
+      "step": 234
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.901410768814581e-06,
+      "loss": 0.6552,
+      "step": 235
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.900281918608732e-06,
+      "loss": 0.5629,
+      "step": 236
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.899146773943374e-06,
+      "loss": 0.5611,
+      "step": 237
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.898005337795291e-06,
+      "loss": 0.6746,
+      "step": 238
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.896857613157765e-06,
+      "loss": 0.5316,
+      "step": 239
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.895703603040572e-06,
+      "loss": 0.6485,
+      "step": 240
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.894543310469968e-06,
+      "loss": 0.6109,
+      "step": 241
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.893376738488685e-06,
+      "loss": 0.6111,
+      "step": 242
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.8922038901559225e-06,
+      "loss": 0.5975,
+      "step": 243
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.891024768547338e-06,
+      "loss": 0.678,
+      "step": 244
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.889839376755041e-06,
+      "loss": 0.6715,
+      "step": 245
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.888647717887582e-06,
+      "loss": 0.5975,
+      "step": 246
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.887449795069948e-06,
+      "loss": 0.6213,
+      "step": 247
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.886245611443554e-06,
+      "loss": 0.5558,
+      "step": 248
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.885035170166229e-06,
+      "loss": 0.5469,
+      "step": 249
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.883818474412214e-06,
+      "loss": 0.6462,
+      "step": 250
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.8825955273721524e-06,
+      "loss": 0.6745,
+      "step": 251
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.8813663322530806e-06,
+      "loss": 0.6691,
+      "step": 252
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.88013089227842e-06,
+      "loss": 0.5922,
+      "step": 253
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.878889210687966e-06,
+      "loss": 0.6481,
+      "step": 254
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.8776412907378845e-06,
+      "loss": 0.6175,
+      "step": 255
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.876387135700701e-06,
+      "loss": 0.5692,
+      "step": 256
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.87512674886529e-06,
+      "loss": 0.6829,
+      "step": 257
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.87386013353687e-06,
+      "loss": 0.605,
+      "step": 258
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.872587293036991e-06,
+      "loss": 0.6143,
+      "step": 259
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.871308230703529e-06,
+      "loss": 0.6172,
+      "step": 260
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.870022949890676e-06,
+      "loss": 0.6852,
+      "step": 261
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.868731453968932e-06,
+      "loss": 0.609,
+      "step": 262
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.867433746325093e-06,
+      "loss": 0.6801,
+      "step": 263
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.866129830362247e-06,
+      "loss": 0.5764,
+      "step": 264
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.864819709499762e-06,
+      "loss": 0.688,
+      "step": 265
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.863503387173276e-06,
+      "loss": 0.6252,
+      "step": 266
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.862180866834691e-06,
+      "loss": 0.6563,
+      "step": 267
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8608521519521625e-06,
+      "loss": 0.4993,
+      "step": 268
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.8595172460100914e-06,
+      "loss": 0.6069,
+      "step": 269
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.858176152509112e-06,
+      "loss": 0.608,
+      "step": 270
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.856828874966086e-06,
+      "loss": 0.5774,
+      "step": 271
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.855475416914091e-06,
+      "loss": 0.6014,
+      "step": 272
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.854115781902414e-06,
+      "loss": 0.6316,
+      "step": 273
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.852749973496539e-06,
+      "loss": 0.6048,
+      "step": 274
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.851377995278138e-06,
+      "loss": 0.7687,
+      "step": 275
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.849999850845066e-06,
+      "loss": 0.5891,
+      "step": 276
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.8486155438113455e-06,
+      "loss": 0.635,
+      "step": 277
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.847225077807159e-06,
+      "loss": 0.6205,
+      "step": 278
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.845828456478843e-06,
+      "loss": 0.5812,
+      "step": 279
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.844425683488873e-06,
+      "loss": 0.5342,
+      "step": 280
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.84301676251586e-06,
+      "loss": 0.6074,
+      "step": 281
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.841601697254532e-06,
+      "loss": 0.6138,
+      "step": 282
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.840180491415733e-06,
+      "loss": 0.6071,
+      "step": 283
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.838753148726411e-06,
+      "loss": 0.5794,
+      "step": 284
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.837319672929606e-06,
+      "loss": 0.6332,
+      "step": 285
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.835880067784441e-06,
+      "loss": 0.5647,
+      "step": 286
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.834434337066112e-06,
+      "loss": 0.5776,
+      "step": 287
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.832982484565879e-06,
+      "loss": 0.6861,
+      "step": 288
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.831524514091056e-06,
+      "loss": 0.6966,
+      "step": 289
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.830060429465002e-06,
+      "loss": 0.5352,
+      "step": 290
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.828590234527107e-06,
+      "loss": 0.5985,
+      "step": 291
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.827113933132785e-06,
+      "loss": 0.5687,
+      "step": 292
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.825631529153466e-06,
+      "loss": 0.5822,
+      "step": 293
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.824143026476581e-06,
+      "loss": 0.6897,
+      "step": 294
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8226484290055544e-06,
+      "loss": 0.5765,
+      "step": 295
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.821147740659795e-06,
+      "loss": 0.6701,
+      "step": 296
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.8196409653746815e-06,
+      "loss": 0.5659,
+      "step": 297
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.818128107101558e-06,
+      "loss": 0.6271,
+      "step": 298
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8166091698077165e-06,
+      "loss": 0.612,
+      "step": 299
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.815084157476395e-06,
+      "loss": 0.5545,
+      "step": 300
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.813553074106761e-06,
+      "loss": 0.6543,
+      "step": 301
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.812015923713901e-06,
+      "loss": 0.5932,
+      "step": 302
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.8104727103288125e-06,
+      "loss": 0.5789,
+      "step": 303
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.808923437998393e-06,
+      "loss": 0.5923,
+      "step": 304
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.80736811078543e-06,
+      "loss": 0.6777,
+      "step": 305
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.805806732768585e-06,
+      "loss": 0.6167,
+      "step": 306
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.804239308042392e-06,
+      "loss": 0.5634,
+      "step": 307
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.802665840717239e-06,
+      "loss": 0.611,
+      "step": 308
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.8010863349193605e-06,
+      "loss": 0.5364,
+      "step": 309
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.799500794790827e-06,
+      "loss": 0.5718,
+      "step": 310
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.797909224489531e-06,
+      "loss": 0.5419,
+      "step": 311
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.796311628189182e-06,
+      "loss": 0.5493,
+      "step": 312
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.794708010079288e-06,
+      "loss": 0.6358,
+      "step": 313
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.793098374365153e-06,
+      "loss": 0.6557,
+      "step": 314
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.791482725267858e-06,
+      "loss": 0.544,
+      "step": 315
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.789861067024253e-06,
+      "loss": 0.5988,
+      "step": 316
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.78823340388695e-06,
+      "loss": 0.6317,
+      "step": 317
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.786599740124303e-06,
+      "loss": 0.6068,
+      "step": 318
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.7849600800204075e-06,
+      "loss": 0.6514,
+      "step": 319
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.783314427875079e-06,
+      "loss": 0.5539,
+      "step": 320
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.781662788003851e-06,
+      "loss": 0.5888,
+      "step": 321
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.780005164737953e-06,
+      "loss": 0.6263,
+      "step": 322
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.778341562424312e-06,
+      "loss": 0.5975,
+      "step": 323
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.77667198542553e-06,
+      "loss": 0.5616,
+      "step": 324
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.774996438119876e-06,
+      "loss": 0.6529,
+      "step": 325
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.773314924901281e-06,
+      "loss": 0.6917,
+      "step": 326
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.771627450179315e-06,
+      "loss": 0.6655,
+      "step": 327
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.769934018379184e-06,
+      "loss": 0.561,
+      "step": 328
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.768234633941716e-06,
+      "loss": 0.6629,
+      "step": 329
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.766529301323348e-06,
+      "loss": 0.6153,
+      "step": 330
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.764818024996117e-06,
+      "loss": 0.6468,
+      "step": 331
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.763100809447645e-06,
+      "loss": 0.554,
+      "step": 332
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.76137765918113e-06,
+      "loss": 0.6316,
+      "step": 333
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.759648578715332e-06,
+      "loss": 0.6015,
+      "step": 334
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.757913572584564e-06,
+      "loss": 0.6326,
+      "step": 335
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.756172645338675e-06,
+      "loss": 0.5583,
+      "step": 336
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.754425801543047e-06,
+      "loss": 0.5641,
+      "step": 337
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.752673045778571e-06,
+      "loss": 0.6476,
+      "step": 338
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.750914382641647e-06,
+      "loss": 0.665,
+      "step": 339
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.7491498167441645e-06,
+      "loss": 0.5557,
+      "step": 340
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.747379352713489e-06,
+      "loss": 0.5964,
+      "step": 341
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.745602995192458e-06,
+      "loss": 0.6362,
+      "step": 342
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.743820748839362e-06,
+      "loss": 0.6329,
+      "step": 343
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.742032618327933e-06,
+      "loss": 0.6097,
+      "step": 344
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.740238608347337e-06,
+      "loss": 0.6248,
+      "step": 345
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.738438723602154e-06,
+      "loss": 0.588,
+      "step": 346
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.736632968812374e-06,
+      "loss": 0.6924,
+      "step": 347
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.734821348713375e-06,
+      "loss": 0.605,
+      "step": 348
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.733003868055923e-06,
+      "loss": 0.5942,
+      "step": 349
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.731180531606147e-06,
+      "loss": 0.6224,
+      "step": 350
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.729351344145536e-06,
+      "loss": 0.6019,
+      "step": 351
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.72751631047092e-06,
+      "loss": 0.5545,
+      "step": 352
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.725675435394461e-06,
+      "loss": 0.6235,
+      "step": 353
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.723828723743638e-06,
+      "loss": 0.6676,
+      "step": 354
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.721976180361239e-06,
+      "loss": 0.5874,
+      "step": 355
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.720117810105341e-06,
+      "loss": 0.624,
+      "step": 356
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.718253617849306e-06,
+      "loss": 0.6212,
+      "step": 357
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.7163836084817586e-06,
+      "loss": 0.5665,
+      "step": 358
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.7145077869065815e-06,
+      "loss": 0.7178,
+      "step": 359
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.712626158042898e-06,
+      "loss": 0.5714,
+      "step": 360
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.710738726825059e-06,
+      "loss": 0.5936,
+      "step": 361
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.708845498202635e-06,
+      "loss": 0.5382,
+      "step": 362
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.706946477140396e-06,
+      "loss": 0.7356,
+      "step": 363
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.705041668618304e-06,
+      "loss": 0.6052,
+      "step": 364
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.703131077631498e-06,
+      "loss": 0.6056,
+      "step": 365
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.701214709190277e-06,
+      "loss": 0.5881,
+      "step": 366
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.699292568320097e-06,
+      "loss": 0.616,
+      "step": 367
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.697364660061548e-06,
+      "loss": 0.6016,
+      "step": 368
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.6954309894703435e-06,
+      "loss": 0.61,
+      "step": 369
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.693491561617309e-06,
+      "loss": 0.5474,
+      "step": 370
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.69154638158837e-06,
+      "loss": 0.6722,
+      "step": 371
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.689595454484531e-06,
+      "loss": 0.5979,
+      "step": 372
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.687638785421875e-06,
+      "loss": 0.6345,
+      "step": 373
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.685676379531535e-06,
+      "loss": 0.6472,
+      "step": 374
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.683708241959694e-06,
+      "loss": 0.6943,
+      "step": 375
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.681734377867562e-06,
+      "loss": 0.6549,
+      "step": 376
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.679754792431368e-06,
+      "loss": 0.5948,
+      "step": 377
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.677769490842344e-06,
+      "loss": 0.5744,
+      "step": 378
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.675778478306712e-06,
+      "loss": 0.5946,
+      "step": 379
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.673781760045669e-06,
+      "loss": 0.585,
+      "step": 380
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.671779341295378e-06,
+      "loss": 0.5754,
+      "step": 381
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.6697712273069475e-06,
+      "loss": 0.6294,
+      "step": 382
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.667757423346423e-06,
+      "loss": 0.6437,
+      "step": 383
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.665737934694769e-06,
+      "loss": 0.6052,
+      "step": 384
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.663712766647862e-06,
+      "loss": 0.562,
+      "step": 385
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.661681924516466e-06,
+      "loss": 0.5709,
+      "step": 386
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.65964541362623e-06,
+      "loss": 0.6165,
+      "step": 387
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.657603239317665e-06,
+      "loss": 0.625,
+      "step": 388
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.655555406946135e-06,
+      "loss": 0.6402,
+      "step": 389
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.6535019218818435e-06,
+      "loss": 0.536,
+      "step": 390
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.651442789509813e-06,
+      "loss": 0.6343,
+      "step": 391
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.649378015229881e-06,
+      "loss": 0.5938,
+      "step": 392
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.647307604456675e-06,
+      "loss": 0.5436,
+      "step": 393
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.645231562619606e-06,
+      "loss": 0.517,
+      "step": 394
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.643149895162854e-06,
+      "loss": 0.5736,
+      "step": 395
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.641062607545347e-06,
+      "loss": 0.6032,
+      "step": 396
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.6389697052407535e-06,
+      "loss": 0.6507,
+      "step": 397
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.636871193737466e-06,
+      "loss": 0.5688,
+      "step": 398
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.634767078538589e-06,
+      "loss": 0.594,
+      "step": 399
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.6326573651619145e-06,
+      "loss": 0.6575,
+      "step": 400
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.630542059139923e-06,
+      "loss": 0.7322,
+      "step": 401
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.628421166019758e-06,
+      "loss": 0.5744,
+      "step": 402
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.626294691363213e-06,
+      "loss": 0.599,
+      "step": 403
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.624162640746722e-06,
+      "loss": 0.6485,
+      "step": 404
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.622025019761336e-06,
+      "loss": 0.6172,
+      "step": 405
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.61988183401272e-06,
+      "loss": 0.5888,
+      "step": 406
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.617733089121127e-06,
+      "loss": 0.5841,
+      "step": 407
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.615578790721391e-06,
+      "loss": 0.6132,
+      "step": 408
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.613418944462907e-06,
+      "loss": 0.5621,
+      "step": 409
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.611253556009621e-06,
+      "loss": 0.6287,
+      "step": 410
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.609082631040012e-06,
+      "loss": 0.6153,
+      "step": 411
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.606906175247077e-06,
+      "loss": 0.6265,
+      "step": 412
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.604724194338318e-06,
+      "loss": 0.5902,
+      "step": 413
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.602536694035725e-06,
+      "loss": 0.6881,
+      "step": 414
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.600343680075764e-06,
+      "loss": 0.6076,
+      "step": 415
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.598145158209356e-06,
+      "loss": 0.5271,
+      "step": 416
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.5959411342018715e-06,
+      "loss": 0.6298,
+      "step": 417
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.593731613833103e-06,
+      "loss": 0.518,
+      "step": 418
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.591516602897263e-06,
+      "loss": 0.6723,
+      "step": 419
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.589296107202957e-06,
+      "loss": 0.6282,
+      "step": 420
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.587070132573178e-06,
+      "loss": 0.6757,
+      "step": 421
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.5848386848452845e-06,
+      "loss": 0.6332,
+      "step": 422
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.582601769870988e-06,
+      "loss": 0.614,
+      "step": 423
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.580359393516337e-06,
+      "loss": 0.5615,
+      "step": 424
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.578111561661702e-06,
+      "loss": 0.5801,
+      "step": 425
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.575858280201761e-06,
+      "loss": 0.5845,
+      "step": 426
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.57359955504548e-06,
+      "loss": 0.573,
+      "step": 427
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.571335392116104e-06,
+      "loss": 0.6202,
+      "step": 428
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.569065797351135e-06,
+      "loss": 0.5797,
+      "step": 429
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.566790776702322e-06,
+      "loss": 0.6794,
+      "step": 430
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.564510336135642e-06,
+      "loss": 0.6543,
+      "step": 431
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.562224481631282e-06,
+      "loss": 0.5959,
+      "step": 432
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.559933219183631e-06,
+      "loss": 0.6523,
+      "step": 433
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.557636554801258e-06,
+      "loss": 0.5575,
+      "step": 434
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.555334494506895e-06,
+      "loss": 0.5857,
+      "step": 435
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.5530270443374305e-06,
+      "loss": 0.5401,
+      "step": 436
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.550714210343879e-06,
+      "loss": 0.5499,
+      "step": 437
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.548395998591383e-06,
+      "loss": 0.607,
+      "step": 438
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.546072415159179e-06,
+      "loss": 0.6222,
+      "step": 439
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.543743466140595e-06,
+      "loss": 0.5737,
+      "step": 440
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.541409157643027e-06,
+      "loss": 0.718,
+      "step": 441
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.53906949578793e-06,
+      "loss": 0.576,
+      "step": 442
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.5367244867107905e-06,
+      "loss": 0.6725,
+      "step": 443
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.534374136561125e-06,
+      "loss": 0.6788,
+      "step": 444
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.53201845150245e-06,
+      "loss": 0.5242,
+      "step": 445
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.5296574377122765e-06,
+      "loss": 0.588,
+      "step": 446
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.527291101382088e-06,
+      "loss": 0.6863,
+      "step": 447
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.524919448717324e-06,
+      "loss": 0.5085,
+      "step": 448
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.522542485937369e-06,
+      "loss": 0.6354,
+      "step": 449
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.52016021927553e-06,
+      "loss": 0.5679,
+      "step": 450
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.517772654979024e-06,
+      "loss": 0.5983,
+      "step": 451
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.515379799308959e-06,
+      "loss": 0.5498,
+      "step": 452
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.512981658540321e-06,
+      "loss": 0.6239,
+      "step": 453
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.510578238961955e-06,
+      "loss": 0.6092,
+      "step": 454
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.508169546876547e-06,
+      "loss": 0.5611,
+      "step": 455
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.505755588600613e-06,
+      "loss": 0.5784,
+      "step": 456
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.503336370464476e-06,
+      "loss": 0.6644,
+      "step": 457
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.500911898812254e-06,
+      "loss": 0.672,
+      "step": 458
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.49848218000184e-06,
+      "loss": 0.6009,
+      "step": 459
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.496047220404891e-06,
+      "loss": 0.6443,
+      "step": 460
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.493607026406802e-06,
+      "loss": 0.6516,
+      "step": 461
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.4911616044067e-06,
+      "loss": 0.6199,
+      "step": 462
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.488710960817416e-06,
+      "loss": 0.5599,
+      "step": 463
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.48625510206548e-06,
+      "loss": 0.6582,
+      "step": 464
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.483794034591092e-06,
+      "loss": 0.5751,
+      "step": 465
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.481327764848118e-06,
+      "loss": 0.5437,
+      "step": 466
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.4788562993040615e-06,
+      "loss": 0.5231,
+      "step": 467
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.476379644440052e-06,
+      "loss": 0.6338,
+      "step": 468
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.473897806750829e-06,
+      "loss": 0.5646,
+      "step": 469
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.471410792744723e-06,
+      "loss": 0.6362,
+      "step": 470
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.4689186089436365e-06,
+      "loss": 0.5566,
+      "step": 471
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.466421261883033e-06,
+      "loss": 0.6317,
+      "step": 472
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.463918758111912e-06,
+      "loss": 0.5598,
+      "step": 473
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.4614111041928e-06,
+      "loss": 0.5669,
+      "step": 474
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.4588983067017255e-06,
+      "loss": 0.6058,
+      "step": 475
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.456380372228208e-06,
+      "loss": 0.5063,
+      "step": 476
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.4538573073752365e-06,
+      "loss": 0.5848,
+      "step": 477
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.451329118759254e-06,
+      "loss": 0.5156,
+      "step": 478
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.448795813010142e-06,
+      "loss": 0.6578,
+      "step": 479
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.446257396771198e-06,
+      "loss": 0.6498,
+      "step": 480
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.443713876699124e-06,
+      "loss": 0.5344,
+      "step": 481
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.441165259464003e-06,
+      "loss": 0.6453,
+      "step": 482
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.438611551749288e-06,
+      "loss": 0.5904,
+      "step": 483
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.43605276025178e-06,
+      "loss": 0.6063,
+      "step": 484
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.4334888916816096e-06,
+      "loss": 0.6648,
+      "step": 485
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.430919952762226e-06,
+      "loss": 0.6272,
+      "step": 486
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.42834595023037e-06,
+      "loss": 0.6077,
+      "step": 487
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.425766890836063e-06,
+      "loss": 0.5405,
+      "step": 488
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.423182781342589e-06,
+      "loss": 0.652,
+      "step": 489
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.420593628526472e-06,
+      "loss": 0.5342,
+      "step": 490
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.417999439177465e-06,
+      "loss": 0.6104,
+      "step": 491
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.415400220098528e-06,
+      "loss": 0.6223,
+      "step": 492
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.412795978105807e-06,
+      "loss": 0.6396,
+      "step": 493
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.410186720028626e-06,
+      "loss": 0.5831,
+      "step": 494
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.407572452709459e-06,
+      "loss": 0.5691,
+      "step": 495
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.404953183003916e-06,
+      "loss": 0.5309,
+      "step": 496
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.402328917780728e-06,
+      "loss": 0.6451,
+      "step": 497
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.399699663921725e-06,
+      "loss": 0.5589,
+      "step": 498
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.397065428321818e-06,
+      "loss": 0.5466,
+      "step": 499
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.394426217888981e-06,
+      "loss": 0.567,
+      "step": 500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.391782039544239e-06,
+      "loss": 0.5423,
+      "step": 501
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.389132900221638e-06,
+      "loss": 0.6357,
+      "step": 502
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.386478806868242e-06,
+      "loss": 0.7513,
+      "step": 503
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.383819766444095e-06,
+      "loss": 0.6854,
+      "step": 504
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.381155785922226e-06,
+      "loss": 0.616,
+      "step": 505
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.378486872288611e-06,
+      "loss": 0.6471,
+      "step": 506
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.375813032542164e-06,
+      "loss": 0.6156,
+      "step": 507
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.37313427369472e-06,
+      "loss": 0.6519,
+      "step": 508
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.37045060277101e-06,
+      "loss": 0.6409,
+      "step": 509
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.36776202680865e-06,
+      "loss": 0.5925,
+      "step": 510
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.365068552858116e-06,
+      "loss": 0.6579,
+      "step": 511
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.362370187982729e-06,
+      "loss": 0.6236,
+      "step": 512
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.359666939258637e-06,
+      "loss": 0.6297,
+      "step": 513
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.356958813774793e-06,
+      "loss": 0.5612,
+      "step": 514
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.354245818632944e-06,
+      "loss": 0.5776,
+      "step": 515
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.3515279609476e-06,
+      "loss": 0.5681,
+      "step": 516
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.348805247846027e-06,
+      "loss": 0.5934,
+      "step": 517
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.346077686468224e-06,
+      "loss": 0.537,
+      "step": 518
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.343345283966901e-06,
+      "loss": 0.6193,
+      "step": 519
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.3406080475074656e-06,
+      "loss": 0.5566,
+      "step": 520
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.337865984268002e-06,
+      "loss": 0.6129,
+      "step": 521
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.33511910143925e-06,
+      "loss": 0.566,
+      "step": 522
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.33236740622459e-06,
+      "loss": 0.578,
+      "step": 523
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.329610905840023e-06,
+      "loss": 0.5835,
+      "step": 524
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.326849607514149e-06,
+      "loss": 0.542,
+      "step": 525
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.324083518488151e-06,
+      "loss": 0.5565,
+      "step": 526
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.321312646015775e-06,
+      "loss": 0.5528,
+      "step": 527
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.318536997363312e-06,
+      "loss": 0.6319,
+      "step": 528
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.315756579809575e-06,
+      "loss": 0.5557,
+      "step": 529
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.312971400645887e-06,
+      "loss": 0.5684,
+      "step": 530
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.3101814671760546e-06,
+      "loss": 0.5834,
+      "step": 531
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.3073867867163525e-06,
+      "loss": 0.6087,
+      "step": 532
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.304587366595506e-06,
+      "loss": 0.5746,
+      "step": 533
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.301783214154667e-06,
+      "loss": 0.6238,
+      "step": 534
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.298974336747397e-06,
+      "loss": 0.6575,
+      "step": 535
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.296160741739652e-06,
+      "loss": 0.5296,
+      "step": 536
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.2933424365097565e-06,
+      "loss": 0.6385,
+      "step": 537
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.290519428448387e-06,
+      "loss": 0.7025,
+      "step": 538
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.287691724958551e-06,
+      "loss": 0.6279,
+      "step": 539
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.284859333455576e-06,
+      "loss": 0.6559,
+      "step": 540
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.282022261367074e-06,
+      "loss": 0.5987,
+      "step": 541
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.279180516132937e-06,
+      "loss": 0.5819,
+      "step": 542
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.276334105205312e-06,
+      "loss": 0.579,
+      "step": 543
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.273483036048578e-06,
+      "loss": 0.5873,
+      "step": 544
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.270627316139333e-06,
+      "loss": 0.6081,
+      "step": 545
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.267766952966369e-06,
+      "loss": 0.6179,
+      "step": 546
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.264901954030655e-06,
+      "loss": 0.6371,
+      "step": 547
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.262032326845317e-06,
+      "loss": 0.6634,
+      "step": 548
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.259158078935616e-06,
+      "loss": 0.6294,
+      "step": 549
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.256279217838934e-06,
+      "loss": 0.5198,
+      "step": 550
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.2533957511047485e-06,
+      "loss": 0.6399,
+      "step": 551
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.250507686294614e-06,
+      "loss": 0.5324,
+      "step": 552
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.247615030982144e-06,
+      "loss": 0.5856,
+      "step": 553
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.244717792752991e-06,
+      "loss": 0.5626,
+      "step": 554
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.241815979204822e-06,
+      "loss": 0.604,
+      "step": 555
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.238909597947307e-06,
+      "loss": 0.6262,
+      "step": 556
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.235998656602091e-06,
+      "loss": 0.5808,
+      "step": 557
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.233083162802779e-06,
+      "loss": 0.5585,
+      "step": 558
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.230163124194913e-06,
+      "loss": 0.6306,
+      "step": 559
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.227238548435954e-06,
+      "loss": 0.573,
+      "step": 560
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.224309443195261e-06,
+      "loss": 0.5121,
+      "step": 561
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.2213758161540705e-06,
+      "loss": 0.6073,
+      "step": 562
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.218437675005479e-06,
+      "loss": 0.6018,
+      "step": 563
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.2154950274544175e-06,
+      "loss": 0.6426,
+      "step": 564
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.212547881217637e-06,
+      "loss": 0.7009,
+      "step": 565
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2095962440236846e-06,
+      "loss": 0.5819,
+      "step": 566
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.206640123612885e-06,
+      "loss": 0.6125,
+      "step": 567
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2036795277373185e-06,
+      "loss": 0.5404,
+      "step": 568
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.2007144641608035e-06,
+      "loss": 0.6136,
+      "step": 569
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.197744940658874e-06,
+      "loss": 0.5464,
+      "step": 570
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.194770965018758e-06,
+      "loss": 0.6098,
+      "step": 571
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.1917925450393606e-06,
+      "loss": 0.5808,
+      "step": 572
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.188809688531241e-06,
+      "loss": 0.5729,
+      "step": 573
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.185822403316593e-06,
+      "loss": 0.6157,
+      "step": 574
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.182830697229223e-06,
+      "loss": 0.6356,
+      "step": 575
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.179834578114531e-06,
+      "loss": 0.6285,
+      "step": 576
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.176834053829492e-06,
+      "loss": 0.6826,
+      "step": 577
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.173829132242629e-06,
+      "loss": 0.6885,
+      "step": 578
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.170819821234001e-06,
+      "loss": 0.6004,
+      "step": 579
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.167806128695173e-06,
+      "loss": 0.5933,
+      "step": 580
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.164788062529203e-06,
+      "loss": 0.586,
+      "step": 581
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.1617656306506185e-06,
+      "loss": 0.5679,
+      "step": 582
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.158738840985393e-06,
+      "loss": 0.6283,
+      "step": 583
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.155707701470932e-06,
+      "loss": 0.6267,
+      "step": 584
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.1526722200560445e-06,
+      "loss": 0.5994,
+      "step": 585
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.149632404700925e-06,
+      "loss": 0.5081,
+      "step": 586
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.146588263377137e-06,
+      "loss": 0.5055,
+      "step": 587
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.143539804067585e-06,
+      "loss": 0.526,
+      "step": 588
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.140487034766499e-06,
+      "loss": 0.5529,
+      "step": 589
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.13742996347941e-06,
+      "loss": 0.5759,
+      "step": 590
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.134368598223132e-06,
+      "loss": 0.6688,
+      "step": 591
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.131302947025737e-06,
+      "loss": 0.5315,
+      "step": 592
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.128233017926538e-06,
+      "loss": 0.5832,
+      "step": 593
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1251588189760685e-06,
+      "loss": 0.5271,
+      "step": 594
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.1220803582360545e-06,
+      "loss": 0.6289,
+      "step": 595
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.118997643779401e-06,
+      "loss": 0.5818,
+      "step": 596
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.115910683690167e-06,
+      "loss": 0.6281,
+      "step": 597
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1128194860635465e-06,
+      "loss": 0.6165,
+      "step": 598
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.109724059005844e-06,
+      "loss": 0.6239,
+      "step": 599
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1066244106344565e-06,
+      "loss": 0.6188,
+      "step": 600
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.1035205490778505e-06,
+      "loss": 0.5995,
+      "step": 601
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.10041248247554e-06,
+      "loss": 0.6094,
+      "step": 602
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.09730021897807e-06,
+      "loss": 0.6056,
+      "step": 603
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.094183766746985e-06,
+      "loss": 0.5832,
+      "step": 604
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.091063133954821e-06,
+      "loss": 0.5761,
+      "step": 605
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.087938328785071e-06,
+      "loss": 0.5014,
+      "step": 606
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.084809359432175e-06,
+      "loss": 0.5328,
+      "step": 607
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.081676234101488e-06,
+      "loss": 0.5917,
+      "step": 608
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.0785389610092684e-06,
+      "loss": 0.7309,
+      "step": 609
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.075397548382646e-06,
+      "loss": 0.5448,
+      "step": 610
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.072252004459612e-06,
+      "loss": 0.6734,
+      "step": 611
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.069102337488986e-06,
+      "loss": 0.5728,
+      "step": 612
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.065948555730405e-06,
+      "loss": 0.5664,
+      "step": 613
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.062790667454293e-06,
+      "loss": 0.6255,
+      "step": 614
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.059628680941843e-06,
+      "loss": 0.5642,
+      "step": 615
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.056462604484998e-06,
+      "loss": 0.6263,
+      "step": 616
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.053292446386422e-06,
+      "loss": 0.5899,
+      "step": 617
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.0501182149594865e-06,
+      "loss": 0.6601,
+      "step": 618
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.046939918528243e-06,
+      "loss": 0.6086,
+      "step": 619
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.043757565427404e-06,
+      "loss": 0.5684,
+      "step": 620
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.040571164002319e-06,
+      "loss": 0.6022,
+      "step": 621
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.037380722608954e-06,
+      "loss": 0.5628,
+      "step": 622
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.034186249613869e-06,
+      "loss": 0.6052,
+      "step": 623
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.030987753394199e-06,
+      "loss": 0.5342,
+      "step": 624
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.027785242337626e-06,
+      "loss": 0.6214,
+      "step": 625
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0245787248423614e-06,
+      "loss": 0.6991,
+      "step": 626
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.021368209317126e-06,
+      "loss": 0.5696,
+      "step": 627
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.01815370418112e-06,
+      "loss": 0.6142,
+      "step": 628
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.014935217864009e-06,
+      "loss": 0.6116,
+      "step": 629
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.0117127588058985e-06,
+      "loss": 0.6735,
+      "step": 630
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.008486335457312e-06,
+      "loss": 0.5799,
+      "step": 631
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.005255956279168e-06,
+      "loss": 0.6867,
+      "step": 632
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.002021629742759e-06,
+      "loss": 0.5983,
+      "step": 633
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.99878336432973e-06,
+      "loss": 0.5497,
+      "step": 634
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.995541168532055e-06,
+      "loss": 0.5928,
+      "step": 635
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.992295050852013e-06,
+      "loss": 0.6073,
+      "step": 636
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.989045019802171e-06,
+      "loss": 0.6163,
+      "step": 637
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.9857910839053545e-06,
+      "loss": 0.6317,
+      "step": 638
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.982533251694632e-06,
+      "loss": 0.5788,
+      "step": 639
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.97927153171329e-06,
+      "loss": 0.5851,
+      "step": 640
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.976005932514807e-06,
+      "loss": 0.5841,
+      "step": 641
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.972736462662837e-06,
+      "loss": 0.6002,
+      "step": 642
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.969463130731183e-06,
+      "loss": 0.6247,
+      "step": 643
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.966185945303778e-06,
+      "loss": 0.5453,
+      "step": 644
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.962904914974656e-06,
+      "loss": 0.6066,
+      "step": 645
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.959620048347938e-06,
+      "loss": 0.5291,
+      "step": 646
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.956331354037805e-06,
+      "loss": 0.5673,
+      "step": 647
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.953038840668474e-06,
+      "loss": 0.5611,
+      "step": 648
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.949742516874175e-06,
+      "loss": 0.6054,
+      "step": 649
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.946442391299136e-06,
+      "loss": 0.5625,
+      "step": 650
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.943138472597549e-06,
+      "loss": 0.6522,
+      "step": 651
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.939830769433558e-06,
+      "loss": 0.6329,
+      "step": 652
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.936519290481226e-06,
+      "loss": 0.5235,
+      "step": 653
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.933204044424524e-06,
+      "loss": 0.5808,
+      "step": 654
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.929885039957296e-06,
+      "loss": 0.5508,
+      "step": 655
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.9265622857832455e-06,
+      "loss": 0.5463,
+      "step": 656
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.923235790615907e-06,
+      "loss": 0.5845,
+      "step": 657
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.919905563178627e-06,
+      "loss": 0.6196,
+      "step": 658
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.916571612204538e-06,
+      "loss": 0.5859,
+      "step": 659
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.913233946436537e-06,
+      "loss": 0.5717,
+      "step": 660
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.909892574627267e-06,
+      "loss": 0.5689,
+      "step": 661
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.906547505539082e-06,
+      "loss": 0.5938,
+      "step": 662
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.903198747944037e-06,
+      "loss": 0.6675,
+      "step": 663
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.899846310623859e-06,
+      "loss": 0.5145,
+      "step": 664
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.896490202369924e-06,
+      "loss": 0.5914,
+      "step": 665
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.893130431983234e-06,
+      "loss": 0.574,
+      "step": 666
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.889767008274396e-06,
+      "loss": 0.5431,
+      "step": 667
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.886399940063595e-06,
+      "loss": 0.5941,
+      "step": 668
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.883029236180577e-06,
+      "loss": 0.5043,
+      "step": 669
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.8796549054646184e-06,
+      "loss": 0.5017,
+      "step": 670
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.876276956764509e-06,
+      "loss": 0.473,
+      "step": 671
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.8728953989385255e-06,
+      "loss": 0.4909,
+      "step": 672
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.869510240854408e-06,
+      "loss": 0.4671,
+      "step": 673
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.86612149138934e-06,
+      "loss": 0.5723,
+      "step": 674
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.862729159429921e-06,
+      "loss": 0.4804,
+      "step": 675
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.8593332538721465e-06,
+      "loss": 0.5114,
+      "step": 676
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.855933783621384e-06,
+      "loss": 0.467,
+      "step": 677
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.852530757592347e-06,
+      "loss": 0.5083,
+      "step": 678
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.849124184709073e-06,
+      "loss": 0.5073,
+      "step": 679
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.845714073904906e-06,
+      "loss": 0.4761,
+      "step": 680
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.84230043412246e-06,
+      "loss": 0.4761,
+      "step": 681
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.83888327431361e-06,
+      "loss": 0.4626,
+      "step": 682
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.835462603439458e-06,
+      "loss": 0.5056,
+      "step": 683
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.832038430470315e-06,
+      "loss": 0.4626,
+      "step": 684
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.828610764385676e-06,
+      "loss": 0.5288,
+      "step": 685
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.825179614174195e-06,
+      "loss": 0.4825,
+      "step": 686
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.821744988833664e-06,
+      "loss": 0.5162,
+      "step": 687
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.818306897370986e-06,
+      "loss": 0.511,
+      "step": 688
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.814865348802157e-06,
+      "loss": 0.4796,
+      "step": 689
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.8114203521522366e-06,
+      "loss": 0.3858,
+      "step": 690
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.807971916455325e-06,
+      "loss": 0.4366,
+      "step": 691
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.8045200507545453e-06,
+      "loss": 0.4443,
+      "step": 692
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.8010647641020116e-06,
+      "loss": 0.5227,
+      "step": 693
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.7976060655588102e-06,
+      "loss": 0.488,
+      "step": 694
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.794143964194976e-06,
+      "loss": 0.4265,
+      "step": 695
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.790678469089465e-06,
+      "loss": 0.4861,
+      "step": 696
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.7872095893301344e-06,
+      "loss": 0.5413,
+      "step": 697
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7837373340137164e-06,
+      "loss": 0.4536,
+      "step": 698
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7802617122457976e-06,
+      "loss": 0.5624,
+      "step": 699
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.776782733140788e-06,
+      "loss": 0.4944,
+      "step": 700
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.773300405821908e-06,
+      "loss": 0.521,
+      "step": 701
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.7698147394211525e-06,
+      "loss": 0.4256,
+      "step": 702
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.766325743079277e-06,
+      "loss": 0.4662,
+      "step": 703
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.762833425945767e-06,
+      "loss": 0.5025,
+      "step": 704
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7593377971788162e-06,
+      "loss": 0.5456,
+      "step": 705
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7558388659453052e-06,
+      "loss": 0.5057,
+      "step": 706
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.752336641420772e-06,
+      "loss": 0.5027,
+      "step": 707
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7488311327893922e-06,
+      "loss": 0.507,
+      "step": 708
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.7453223492439544e-06,
+      "loss": 0.4689,
+      "step": 709
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.741810299985833e-06,
+      "loss": 0.4349,
+      "step": 710
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7382949942249695e-06,
+      "loss": 0.4442,
+      "step": 711
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7347764411798422e-06,
+      "loss": 0.4545,
+      "step": 712
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.731254650077446e-06,
+      "loss": 0.5191,
+      "step": 713
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7277296301532685e-06,
+      "loss": 0.4767,
+      "step": 714
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.724201390651263e-06,
+      "loss": 0.4285,
+      "step": 715
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.720669940823827e-06,
+      "loss": 0.48,
+      "step": 716
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.7171352899317743e-06,
+      "loss": 0.4373,
+      "step": 717
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.7135974472443166e-06,
+      "loss": 0.4133,
+      "step": 718
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.710056422039033e-06,
+      "loss": 0.5456,
+      "step": 719
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.706512223601849e-06,
+      "loss": 0.4978,
+      "step": 720
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.702964861227013e-06,
+      "loss": 0.4953,
+      "step": 721
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.6994143442170683e-06,
+      "loss": 0.486,
+      "step": 722
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.695860681882832e-06,
+      "loss": 0.4799,
+      "step": 723
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.692303883543369e-06,
+      "loss": 0.5119,
+      "step": 724
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6887439585259693e-06,
+      "loss": 0.5357,
+      "step": 725
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6851809161661206e-06,
+      "loss": 0.5487,
+      "step": 726
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6816147658074864e-06,
+      "loss": 0.4881,
+      "step": 727
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.678045516801879e-06,
+      "loss": 0.4846,
+      "step": 728
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.6744731785092396e-06,
+      "loss": 0.5572,
+      "step": 729
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.670897760297608e-06,
+      "loss": 0.5725,
+      "step": 730
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6673192715431016e-06,
+      "loss": 0.4608,
+      "step": 731
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6637377216298907e-06,
+      "loss": 0.5474,
+      "step": 732
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6601531199501715e-06,
+      "loss": 0.4307,
+      "step": 733
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.6565654759041446e-06,
+      "loss": 0.4736,
+      "step": 734
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.652974798899988e-06,
+      "loss": 0.5021,
+      "step": 735
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.649381098353834e-06,
+      "loss": 0.4511,
+      "step": 736
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.645784383689742e-06,
+      "loss": 0.4818,
+      "step": 737
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6421846643396787e-06,
+      "loss": 0.553,
+      "step": 738
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6385819497434877e-06,
+      "loss": 0.5092,
+      "step": 739
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.634976249348867e-06,
+      "loss": 0.4806,
+      "step": 740
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.631367572611348e-06,
+      "loss": 0.4286,
+      "step": 741
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.627755928994262e-06,
+      "loss": 0.5189,
+      "step": 742
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6241413279687256e-06,
+      "loss": 0.5707,
+      "step": 743
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.6205237790136053e-06,
+      "loss": 0.5301,
+      "step": 744
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.616903291615506e-06,
+      "loss": 0.4593,
+      "step": 745
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.613279875268731e-06,
+      "loss": 0.5231,
+      "step": 746
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.609653539475268e-06,
+      "loss": 0.4621,
+      "step": 747
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.6060242937447594e-06,
+      "loss": 0.4795,
+      "step": 748
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.6023921475944795e-06,
+      "loss": 0.4324,
+      "step": 749
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.598757110549308e-06,
+      "loss": 0.5271,
+      "step": 750
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5951191921417063e-06,
+      "loss": 0.513,
+      "step": 751
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5914784019116905e-06,
+      "loss": 0.5479,
+      "step": 752
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5878347494068083e-06,
+      "loss": 0.5577,
+      "step": 753
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.584188244182115e-06,
+      "loss": 0.4831,
+      "step": 754
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.580538895800144e-06,
+      "loss": 0.4347,
+      "step": 755
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.5768867138308872e-06,
+      "loss": 0.4709,
+      "step": 756
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.573231707851765e-06,
+      "loss": 0.5782,
+      "step": 757
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5695738874476047e-06,
+      "loss": 0.4089,
+      "step": 758
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5659132622106152e-06,
+      "loss": 0.5439,
+      "step": 759
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5622498417403574e-06,
+      "loss": 0.4004,
+      "step": 760
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5585836356437266e-06,
+      "loss": 0.4893,
+      "step": 761
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5549146535349184e-06,
+      "loss": 0.5225,
+      "step": 762
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.551242905035412e-06,
+      "loss": 0.5373,
+      "step": 763
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.5475683997739403e-06,
+      "loss": 0.4327,
+      "step": 764
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.5438911473864633e-06,
+      "loss": 0.4531,
+      "step": 765
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.540211157516149e-06,
+      "loss": 0.5454,
+      "step": 766
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.5365284398133404e-06,
+      "loss": 0.5155,
+      "step": 767
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.5328430039355353e-06,
+      "loss": 0.5663,
+      "step": 768
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.52915485954736e-06,
+      "loss": 0.4588,
+      "step": 769
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.5254640163205434e-06,
+      "loss": 0.5731,
+      "step": 770
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.521770483933891e-06,
+      "loss": 0.576,
+      "step": 771
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.518074272073261e-06,
+      "loss": 0.4517,
+      "step": 772
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.514375390431539e-06,
+      "loss": 0.527,
+      "step": 773
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.510673848708609e-06,
+      "loss": 0.4884,
+      "step": 774
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.5069696566113347e-06,
+      "loss": 0.5657,
+      "step": 775
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.503262823853527e-06,
+      "loss": 0.4533,
+      "step": 776
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.499553360155923e-06,
+      "loss": 0.5213,
+      "step": 777
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4958412752461583e-06,
+      "loss": 0.472,
+      "step": 778
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4921265788587432e-06,
+      "loss": 0.4764,
+      "step": 779
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4884092807350367e-06,
+      "loss": 0.5878,
+      "step": 780
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.484689390623218e-06,
+      "loss": 0.498,
+      "step": 781
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4809669182782667e-06,
+      "loss": 0.4663,
+      "step": 782
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4772418734619325e-06,
+      "loss": 0.4843,
+      "step": 783
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.4735142659427102e-06,
+      "loss": 0.4357,
+      "step": 784
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4697841054958163e-06,
+      "loss": 0.5438,
+      "step": 785
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.466051401903162e-06,
+      "loss": 0.5156,
+      "step": 786
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4623161649533284e-06,
+      "loss": 0.482,
+      "step": 787
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4585784044415365e-06,
+      "loss": 0.4361,
+      "step": 788
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.4548381301696298e-06,
+      "loss": 0.5013,
+      "step": 789
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.45109535194604e-06,
+      "loss": 0.4834,
+      "step": 790
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4473500795857674e-06,
+      "loss": 0.4456,
+      "step": 791
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4436023229103517e-06,
+      "loss": 0.4587,
+      "step": 792
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4398520917478478e-06,
+      "loss": 0.5479,
+      "step": 793
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.436099395932801e-06,
+      "loss": 0.4881,
+      "step": 794
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4323442453062173e-06,
+      "loss": 0.4747,
+      "step": 795
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.428586649715542e-06,
+      "loss": 0.5291,
+      "step": 796
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.4248266190146307e-06,
+      "loss": 0.536,
+      "step": 797
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4210641630637264e-06,
+      "loss": 0.4948,
+      "step": 798
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.417299291729431e-06,
+      "loss": 0.5834,
+      "step": 799
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.41353201488468e-06,
+      "loss": 0.4333,
+      "step": 800
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4097623424087196e-06,
+      "loss": 0.5375,
+      "step": 801
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.405990284187074e-06,
+      "loss": 0.5037,
+      "step": 802
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.4022158501115283e-06,
+      "loss": 0.4861,
+      "step": 803
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.3984390500800932e-06,
+      "loss": 0.5502,
+      "step": 804
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.39465989399699e-06,
+      "loss": 0.437,
+      "step": 805
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3908783917726123e-06,
+      "loss": 0.4335,
+      "step": 806
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3870945533235104e-06,
+      "loss": 0.4678,
+      "step": 807
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.383308388572358e-06,
+      "loss": 0.4722,
+      "step": 808
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3795199074479312e-06,
+      "loss": 0.5446,
+      "step": 809
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3757291198850817e-06,
+      "loss": 0.5517,
+      "step": 810
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 3.3719360358247054e-06,
+      "loss": 0.6252,
+      "step": 811
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3681406652137255e-06,
+      "loss": 0.5173,
+      "step": 812
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3643430180050573e-06,
+      "loss": 0.5053,
+      "step": 813
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3605431041575887e-06,
+      "loss": 0.53,
+      "step": 814
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3567409336361502e-06,
+      "loss": 0.6014,
+      "step": 815
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3529365164114903e-06,
+      "loss": 0.4742,
+      "step": 816
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 3.3491298624602514e-06,
+      "loss": 0.4514,
+      "step": 817
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.345320981764938e-06,
+      "loss": 0.5592,
+      "step": 818
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3415098843138972e-06,
+      "loss": 0.5446,
+      "step": 819
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.337696580101286e-06,
+      "loss": 0.4876,
+      "step": 820
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.333881079127052e-06,
+      "loss": 0.5474,
+      "step": 821
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3300633913969005e-06,
+      "loss": 0.4983,
+      "step": 822
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.326243526922272e-06,
+      "loss": 0.5551,
+      "step": 823
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 3.3224214957203176e-06,
+      "loss": 0.3718,
+      "step": 824
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.3185973078138665e-06,
+      "loss": 0.4982,
+      "step": 825
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.314770973231408e-06,
+      "loss": 0.5327,
+      "step": 826
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.3109425020070564e-06,
+      "loss": 0.4849,
+      "step": 827
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.307111904180532e-06,
+      "loss": 0.4646,
+      "step": 828
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.3032791897971313e-06,
+      "loss": 0.5004,
+      "step": 829
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.2994443689077e-06,
+      "loss": 0.5764,
+      "step": 830
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 3.2956074515686105e-06,
+      "loss": 0.5672,
+      "step": 831
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.291768447841729e-06,
+      "loss": 0.4859,
+      "step": 832
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.2879273677943972e-06,
+      "loss": 0.4829,
+      "step": 833
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.2840842214993978e-06,
+      "loss": 0.4857,
+      "step": 834
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.2802390190349364e-06,
+      "loss": 0.5019,
+      "step": 835
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.276391770484606e-06,
+      "loss": 0.4825,
+      "step": 836
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 3.272542485937369e-06,
+      "loss": 0.4874,
+      "step": 837
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.268691175487524e-06,
+      "loss": 0.579,
+      "step": 838
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.264837849234685e-06,
+      "loss": 0.4404,
+      "step": 839
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.2609825172837495e-06,
+      "loss": 0.4423,
+      "step": 840
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.257125189744877e-06,
+      "loss": 0.4318,
+      "step": 841
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.2532658767334595e-06,
+      "loss": 0.5344,
+      "step": 842
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.249404588370095e-06,
+      "loss": 0.4019,
+      "step": 843
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 3.245541334780562e-06,
+      "loss": 0.542,
+      "step": 844
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2416761260957925e-06,
+      "loss": 0.5446,
+      "step": 845
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2378089724518464e-06,
+      "loss": 0.4628,
+      "step": 846
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.233939883989882e-06,
+      "loss": 0.4826,
+      "step": 847
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2300688708561345e-06,
+      "loss": 0.5077,
+      "step": 848
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2261959432018834e-06,
+      "loss": 0.5351,
+      "step": 849
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.2223211111834306e-06,
+      "loss": 0.5823,
+      "step": 850
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 3.218444384962071e-06,
+      "loss": 0.446,
+      "step": 851
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.2145657747040663e-06,
+      "loss": 0.4748,
+      "step": 852
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.210685290580622e-06,
+      "loss": 0.4909,
+      "step": 853
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.2068029427678525e-06,
+      "loss": 0.5022,
+      "step": 854
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.2029187414467645e-06,
+      "loss": 0.4789,
+      "step": 855
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.1990326968032225e-06,
+      "loss": 0.5023,
+      "step": 856
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 3.1951448190279256e-06,
+      "loss": 0.46,
+      "step": 857
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.19125511831638e-06,
+      "loss": 0.4693,
+      "step": 858
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1873636048688714e-06,
+      "loss": 0.4592,
+      "step": 859
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1834702888904424e-06,
+      "loss": 0.5009,
+      "step": 860
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1795751805908578e-06,
+      "loss": 0.4205,
+      "step": 861
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.175678290184586e-06,
+      "loss": 0.5392,
+      "step": 862
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.171779627890769e-06,
+      "loss": 0.5776,
+      "step": 863
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 3.1678792039331917e-06,
+      "loss": 0.5196,
+      "step": 864
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1639770285402632e-06,
+      "loss": 0.5299,
+      "step": 865
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.160073111944983e-06,
+      "loss": 0.5453,
+      "step": 866
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1561674643849173e-06,
+      "loss": 0.468,
+      "step": 867
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.152260096102171e-06,
+      "loss": 0.5196,
+      "step": 868
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.148351017343363e-06,
+      "loss": 0.5232,
+      "step": 869
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1444402383595955e-06,
+      "loss": 0.5114,
+      "step": 870
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 3.1405277694064306e-06,
+      "loss": 0.5567,
+      "step": 871
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1366136207438635e-06,
+      "loss": 0.5202,
+      "step": 872
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1326978026362907e-06,
+      "loss": 0.4576,
+      "step": 873
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.12878032535249e-06,
+      "loss": 0.5598,
+      "step": 874
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1248611991655885e-06,
+      "loss": 0.5501,
+      "step": 875
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1209404343530374e-06,
+      "loss": 0.475,
+      "step": 876
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 3.1170180411965854e-06,
+      "loss": 0.5219,
+      "step": 877
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.113094029982252e-06,
+      "loss": 0.4435,
+      "step": 878
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.109168411000299e-06,
+      "loss": 0.5005,
+      "step": 879
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.105241194545204e-06,
+      "loss": 0.5275,
+      "step": 880
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.1013123909156347e-06,
+      "loss": 0.4582,
+      "step": 881
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.0973820104144197e-06,
+      "loss": 0.6142,
+      "step": 882
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.093450063348525e-06,
+      "loss": 0.5646,
+      "step": 883
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 3.089516560029023e-06,
+      "loss": 0.5341,
+      "step": 884
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.085581510771067e-06,
+      "loss": 0.4985,
+      "step": 885
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.081644925893866e-06,
+      "loss": 0.5239,
+      "step": 886
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0777068157206535e-06,
+      "loss": 0.427,
+      "step": 887
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0737671905786653e-06,
+      "loss": 0.4801,
+      "step": 888
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.0698260607991094e-06,
+      "loss": 0.4978,
+      "step": 889
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.065883436717138e-06,
+      "loss": 0.4666,
+      "step": 890
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 3.061939328671824e-06,
+      "loss": 0.4736,
+      "step": 891
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0579937470061305e-06,
+      "loss": 0.4418,
+      "step": 892
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.054046702066886e-06,
+      "loss": 0.4579,
+      "step": 893
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0500982042047555e-06,
+      "loss": 0.4075,
+      "step": 894
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0461482637742133e-06,
+      "loss": 0.5333,
+      "step": 895
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0421968911335196e-06,
+      "loss": 0.5518,
+      "step": 896
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 3.0382440966446876e-06,
+      "loss": 0.5113,
+      "step": 897
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0342898906734597e-06,
+      "loss": 0.4723,
+      "step": 898
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0303342835892804e-06,
+      "loss": 0.4595,
+      "step": 899
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.026377285765268e-06,
+      "loss": 0.4464,
+      "step": 900
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0224189075781886e-06,
+      "loss": 0.5138,
+      "step": 901
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0184591594084263e-06,
+      "loss": 0.5555,
+      "step": 902
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.014498051639959e-06,
+      "loss": 0.4019,
+      "step": 903
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 3.0105355946603324e-06,
+      "loss": 0.4993,
+      "step": 904
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.006571798860626e-06,
+      "loss": 0.5812,
+      "step": 905
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 3.002606674635432e-06,
+      "loss": 0.5125,
+      "step": 906
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.9986402323828274e-06,
+      "loss": 0.4585,
+      "step": 907
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.994672482504345e-06,
+      "loss": 0.5517,
+      "step": 908
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.9907034354049443e-06,
+      "loss": 0.4642,
+      "step": 909
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.98673310149299e-06,
+      "loss": 0.4954,
+      "step": 910
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.9827614911802205e-06,
+      "loss": 0.5117,
+      "step": 911
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.978788614881719e-06,
+      "loss": 0.5103,
+      "step": 912
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9748144830158925e-06,
+      "loss": 0.527,
+      "step": 913
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9708391060044374e-06,
+      "loss": 0.4502,
+      "step": 914
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.966862494272316e-06,
+      "loss": 0.5163,
+      "step": 915
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9628846582477305e-06,
+      "loss": 0.4853,
+      "step": 916
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.9589056083620902e-06,
+      "loss": 0.5106,
+      "step": 917
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9549253550499924e-06,
+      "loss": 0.5099,
+      "step": 918
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9509439087491837e-06,
+      "loss": 0.5553,
+      "step": 919
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.946961279900546e-06,
+      "loss": 0.4385,
+      "step": 920
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9429774789480576e-06,
+      "loss": 0.47,
+      "step": 921
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9389925163387718e-06,
+      "loss": 0.4715,
+      "step": 922
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.93500640252279e-06,
+      "loss": 0.5703,
+      "step": 923
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.9310191479532298e-06,
+      "loss": 0.532,
+      "step": 924
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.927030763086201e-06,
+      "loss": 0.5145,
+      "step": 925
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.923041258380779e-06,
+      "loss": 0.571,
+      "step": 926
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9190506442989753e-06,
+      "loss": 0.4615,
+      "step": 927
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9150589313057086e-06,
+      "loss": 0.4007,
+      "step": 928
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9110661298687824e-06,
+      "loss": 0.4269,
+      "step": 929
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9070722504588513e-06,
+      "loss": 0.5819,
+      "step": 930
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.9030773035493997e-06,
+      "loss": 0.5739,
+      "step": 931
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8990812996167102e-06,
+      "loss": 0.4984,
+      "step": 932
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8950842491398358e-06,
+      "loss": 0.4627,
+      "step": 933
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8910861626005774e-06,
+      "loss": 0.5121,
+      "step": 934
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.8870870504834497e-06,
+      "loss": 0.4897,
+      "step": 935
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.883086923275658e-06,
+      "loss": 0.5266,
+      "step": 936
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.87908579146707e-06,
+      "loss": 0.429,
+      "step": 937
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8750836655501874e-06,
+      "loss": 0.5851,
+      "step": 938
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8710805560201184e-06,
+      "loss": 0.4728,
+      "step": 939
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.867076473374551e-06,
+      "loss": 0.5167,
+      "step": 940
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8630714281137263e-06,
+      "loss": 0.5019,
+      "step": 941
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8590654307404068e-06,
+      "loss": 0.4635,
+      "step": 942
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.8550584917598558e-06,
+      "loss": 0.4756,
+      "step": 943
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.851050621679801e-06,
+      "loss": 0.5008,
+      "step": 944
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8470418310104175e-06,
+      "loss": 0.5307,
+      "step": 945
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8430321302642887e-06,
+      "loss": 0.5341,
+      "step": 946
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.839021529956388e-06,
+      "loss": 0.5312,
+      "step": 947
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8350100406040492e-06,
+      "loss": 0.562,
+      "step": 948
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8309976727269335e-06,
+      "loss": 0.4367,
+      "step": 949
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8269844368470085e-06,
+      "loss": 0.4559,
+      "step": 950
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.8229703434885165e-06,
+      "loss": 0.5784,
+      "step": 951
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8189554031779508e-06,
+      "loss": 0.4789,
+      "step": 952
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.814939626444023e-06,
+      "loss": 0.5634,
+      "step": 953
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.810923023817641e-06,
+      "loss": 0.4415,
+      "step": 954
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8069056058318754e-06,
+      "loss": 0.5008,
+      "step": 955
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.8028873830219373e-06,
+      "loss": 0.5114,
+      "step": 956
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.7988683659251475e-06,
+      "loss": 0.5364,
+      "step": 957
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7948485650809096e-06,
+      "loss": 0.4791,
+      "step": 958
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7908279910306834e-06,
+      "loss": 0.4378,
+      "step": 959
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.786806654317955e-06,
+      "loss": 0.4795,
+      "step": 960
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7827845654882112e-06,
+      "loss": 0.5101,
+      "step": 961
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7787617350889107e-06,
+      "loss": 0.552,
+      "step": 962
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7747381736694573e-06,
+      "loss": 0.4923,
+      "step": 963
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.7707138917811733e-06,
+      "loss": 0.5065,
+      "step": 964
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.766688899977266e-06,
+      "loss": 0.5146,
+      "step": 965
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.76266320881281e-06,
+      "loss": 0.4569,
+      "step": 966
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7586368288447094e-06,
+      "loss": 0.494,
+      "step": 967
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7546097706316773e-06,
+      "loss": 0.5223,
+      "step": 968
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.750582044734203e-06,
+      "loss": 0.478,
+      "step": 969
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7465536617145287e-06,
+      "loss": 0.4938,
+      "step": 970
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.7425246321366205e-06,
+      "loss": 0.5392,
+      "step": 971
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.738494966566137e-06,
+      "loss": 0.5138,
+      "step": 972
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7344646755704078e-06,
+      "loss": 0.4357,
+      "step": 973
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7304337697183997e-06,
+      "loss": 0.6013,
+      "step": 974
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.726402259580695e-06,
+      "loss": 0.4476,
+      "step": 975
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.7223701557294574e-06,
+      "loss": 0.4629,
+      "step": 976
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.71833746873841e-06,
+      "loss": 0.505,
+      "step": 977
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.7143042091828053e-06,
+      "loss": 0.5201,
+      "step": 978
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.7102703876393942e-06,
+      "loss": 0.5117,
+      "step": 979
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.706236014686406e-06,
+      "loss": 0.5353,
+      "step": 980
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.702201100903511e-06,
+      "loss": 0.5682,
+      "step": 981
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.6981656568718017e-06,
+      "loss": 0.4621,
+      "step": 982
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.694129693173759e-06,
+      "loss": 0.5545,
+      "step": 983
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.690093220393226e-06,
+      "loss": 0.4588,
+      "step": 984
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6860562491153854e-06,
+      "loss": 0.5528,
+      "step": 985
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6820187899267203e-06,
+      "loss": 0.4689,
+      "step": 986
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6779808534149986e-06,
+      "loss": 0.4699,
+      "step": 987
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6739424501692376e-06,
+      "loss": 0.5171,
+      "step": 988
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6699035907796796e-06,
+      "loss": 0.534,
+      "step": 989
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.665864285837761e-06,
+      "loss": 0.4538,
+      "step": 990
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.6618245459360896e-06,
+      "loss": 0.4738,
+      "step": 991
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6577843816684117e-06,
+      "loss": 0.4636,
+      "step": 992
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6537438036295876e-06,
+      "loss": 0.5225,
+      "step": 993
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.649702822415561e-06,
+      "loss": 0.4756,
+      "step": 994
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6456614486233344e-06,
+      "loss": 0.4719,
+      "step": 995
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.641619692850941e-06,
+      "loss": 0.4759,
+      "step": 996
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.6375775656974124e-06,
+      "loss": 0.4395,
+      "step": 997
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6335350777627565e-06,
+      "loss": 0.4886,
+      "step": 998
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6294922396479263e-06,
+      "loss": 0.5329,
+      "step": 999
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.625449061954794e-06,
+      "loss": 0.4866,
+      "step": 1000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6214055552861213e-06,
+      "loss": 0.5342,
+      "step": 1001
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6173617302455324e-06,
+      "loss": 0.3964,
+      "step": 1002
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.613317597437489e-06,
+      "loss": 0.5626,
+      "step": 1003
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.6092731674672563e-06,
+      "loss": 0.5611,
+      "step": 1004
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.6052284509408805e-06,
+      "loss": 0.3921,
+      "step": 1005
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.6011834584651597e-06,
+      "loss": 0.505,
+      "step": 1006
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.5971382006476153e-06,
+      "loss": 0.4751,
+      "step": 1007
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.5930926880964634e-06,
+      "loss": 0.4358,
+      "step": 1008
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.5890469314205898e-06,
+      "loss": 0.4749,
+      "step": 1009
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.5850009412295194e-06,
+      "loss": 0.4952,
+      "step": 1010
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.5809547281333904e-06,
+      "loss": 0.5613,
+      "step": 1011
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5769083027429242e-06,
+      "loss": 0.5262,
+      "step": 1012
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5728616756693995e-06,
+      "loss": 0.483,
+      "step": 1013
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.5688148575246254e-06,
+      "loss": 0.4952,
+      "step": 1014
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.564767858920909e-06,
+      "loss": 0.4913,
+      "step": 1015
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.560720690471033e-06,
+      "loss": 0.4466,
+      "step": 1016
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.556673362788225e-06,
+      "loss": 0.5003,
+      "step": 1017
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5526258864861304e-06,
+      "loss": 0.6677,
+      "step": 1018
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5485782721787837e-06,
+      "loss": 0.5516,
+      "step": 1019
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.544530530480581e-06,
+      "loss": 0.5296,
+      "step": 1020
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5404826720062544e-06,
+      "loss": 0.4941,
+      "step": 1021
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5364347073708384e-06,
+      "loss": 0.572,
+      "step": 1022
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.532386647189651e-06,
+      "loss": 0.4584,
+      "step": 1023
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.5283385020782562e-06,
+      "loss": 0.5087,
+      "step": 1024
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5242902826524435e-06,
+      "loss": 0.5094,
+      "step": 1025
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5202419995281966e-06,
+      "loss": 0.4697,
+      "step": 1026
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5161936633216656e-06,
+      "loss": 0.5397,
+      "step": 1027
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5121452846491406e-06,
+      "loss": 0.571,
+      "step": 1028
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5080968741270224e-06,
+      "loss": 0.5727,
+      "step": 1029
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.504048442371795e-06,
+      "loss": 0.4273,
+      "step": 1030
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.5e-06,
+      "loss": 0.4893,
+      "step": 1031
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.495951557628205e-06,
+      "loss": 0.5545,
+      "step": 1032
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.491903125872979e-06,
+      "loss": 0.4403,
+      "step": 1033
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.48785471535086e-06,
+      "loss": 0.515,
+      "step": 1034
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4838063366783353e-06,
+      "loss": 0.4715,
+      "step": 1035
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4797580004718038e-06,
+      "loss": 0.4971,
+      "step": 1036
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4757097173475574e-06,
+      "loss": 0.4762,
+      "step": 1037
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.471661497921744e-06,
+      "loss": 0.5492,
+      "step": 1038
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4676133528103497e-06,
+      "loss": 0.5093,
+      "step": 1039
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.463565292629162e-06,
+      "loss": 0.5479,
+      "step": 1040
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4595173279937464e-06,
+      "loss": 0.4562,
+      "step": 1041
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4554694695194194e-06,
+      "loss": 0.5205,
+      "step": 1042
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4514217278212167e-06,
+      "loss": 0.4652,
+      "step": 1043
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4473741135138704e-06,
+      "loss": 0.5105,
+      "step": 1044
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.4433266372117755e-06,
+      "loss": 0.462,
+      "step": 1045
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.4392793095289677e-06,
+      "loss": 0.4759,
+      "step": 1046
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.435232141079092e-06,
+      "loss": 0.496,
+      "step": 1047
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.431185142475376e-06,
+      "loss": 0.5108,
+      "step": 1048
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.4271383243306017e-06,
+      "loss": 0.5185,
+      "step": 1049
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.423091697257077e-06,
+      "loss": 0.4515,
+      "step": 1050
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.419045271866611e-06,
+      "loss": 0.5399,
+      "step": 1051
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.4149990587704814e-06,
+      "loss": 0.4567,
+      "step": 1052
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.410953068579411e-06,
+      "loss": 0.4839,
+      "step": 1053
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.406907311903538e-06,
+      "loss": 0.5399,
+      "step": 1054
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.402861799352386e-06,
+      "loss": 0.5397,
+      "step": 1055
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.3988165415348416e-06,
+      "loss": 0.4597,
+      "step": 1056
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 2.3947715490591207e-06,
+      "loss": 0.4526,
+      "step": 1057
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3907268325327454e-06,
+      "loss": 0.4773,
+      "step": 1058
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3866824025625123e-06,
+      "loss": 0.4824,
+      "step": 1059
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.382638269754468e-06,
+      "loss": 0.4885,
+      "step": 1060
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3785944447138804e-06,
+      "loss": 0.4886,
+      "step": 1061
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.374550938045207e-06,
+      "loss": 0.5711,
+      "step": 1062
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.370507760352074e-06,
+      "loss": 0.4374,
+      "step": 1063
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 2.3664649222372443e-06,
+      "loss": 0.5125,
+      "step": 1064
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.362422434302588e-06,
+      "loss": 0.5322,
+      "step": 1065
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.358380307149059e-06,
+      "loss": 0.5098,
+      "step": 1066
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3543385513766656e-06,
+      "loss": 0.4873,
+      "step": 1067
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.350297177584439e-06,
+      "loss": 0.5108,
+      "step": 1068
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3462561963704132e-06,
+      "loss": 0.4766,
+      "step": 1069
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3422156183315887e-06,
+      "loss": 0.5204,
+      "step": 1070
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 2.3381754540639108e-06,
+      "loss": 0.482,
+      "step": 1071
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3341357141622394e-06,
+      "loss": 0.5545,
+      "step": 1072
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.330096409220321e-06,
+      "loss": 0.5057,
+      "step": 1073
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3260575498307624e-06,
+      "loss": 0.5294,
+      "step": 1074
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3220191465850014e-06,
+      "loss": 0.5575,
+      "step": 1075
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.31798121007328e-06,
+      "loss": 0.4529,
+      "step": 1076
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3139437508846155e-06,
+      "loss": 0.443,
+      "step": 1077
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 2.3099067796067744e-06,
+      "loss": 0.4621,
+      "step": 1078
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.3058703068262424e-06,
+      "loss": 0.5196,
+      "step": 1079
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.301834343128199e-06,
+      "loss": 0.5207,
+      "step": 1080
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.29779889909649e-06,
+      "loss": 0.4993,
+      "step": 1081
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.293763985313595e-06,
+      "loss": 0.4522,
+      "step": 1082
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.289729612360606e-06,
+      "loss": 0.4189,
+      "step": 1083
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 2.2856957908171955e-06,
+      "loss": 0.5737,
+      "step": 1084
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2816625312615903e-06,
+      "loss": 0.5188,
+      "step": 1085
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2776298442705434e-06,
+      "loss": 0.5018,
+      "step": 1086
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.273597740419306e-06,
+      "loss": 0.4492,
+      "step": 1087
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.269566230281601e-06,
+      "loss": 0.5001,
+      "step": 1088
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.265535324429593e-06,
+      "loss": 0.4987,
+      "step": 1089
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.2615050334338636e-06,
+      "loss": 0.5669,
+      "step": 1090
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 2.25747536786338e-06,
+      "loss": 0.4715,
+      "step": 1091
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2534463382854718e-06,
+      "loss": 0.4903,
+      "step": 1092
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2494179552657977e-06,
+      "loss": 0.5823,
+      "step": 1093
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.245390229368323e-06,
+      "loss": 0.4431,
+      "step": 1094
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.241363171155291e-06,
+      "loss": 0.5309,
+      "step": 1095
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2373367911871904e-06,
+      "loss": 0.5351,
+      "step": 1096
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2333111000227343e-06,
+      "loss": 0.4597,
+      "step": 1097
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 2.2292861082188276e-06,
+      "loss": 0.4519,
+      "step": 1098
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.225261826330543e-06,
+      "loss": 0.48,
+      "step": 1099
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.22123826491109e-06,
+      "loss": 0.4742,
+      "step": 1100
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2172154345117896e-06,
+      "loss": 0.4557,
+      "step": 1101
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.213193345682046e-06,
+      "loss": 0.4929,
+      "step": 1102
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.209172008969317e-06,
+      "loss": 0.463,
+      "step": 1103
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 2.2051514349190908e-06,
+      "loss": 0.5553,
+      "step": 1104
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.2011316340748533e-06,
+      "loss": 0.5215,
+      "step": 1105
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1971126169780636e-06,
+      "loss": 0.4804,
+      "step": 1106
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1930943941681255e-06,
+      "loss": 0.5164,
+      "step": 1107
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.18907697618236e-06,
+      "loss": 0.5613,
+      "step": 1108
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.185060373555978e-06,
+      "loss": 0.563,
+      "step": 1109
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1810445968220505e-06,
+      "loss": 0.4743,
+      "step": 1110
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 2.1770296565114847e-06,
+      "loss": 0.572,
+      "step": 1111
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1730155631529932e-06,
+      "loss": 0.5581,
+      "step": 1112
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.169002327273068e-06,
+      "loss": 0.4895,
+      "step": 1113
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.164989959395952e-06,
+      "loss": 0.5519,
+      "step": 1114
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1609784700436122e-06,
+      "loss": 0.5032,
+      "step": 1115
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1569678697357126e-06,
+      "loss": 0.4841,
+      "step": 1116
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1529581689895838e-06,
+      "loss": 0.594,
+      "step": 1117
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 2.1489493783201997e-06,
+      "loss": 0.5565,
+      "step": 1118
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.1449415082401455e-06,
+      "loss": 0.4721,
+      "step": 1119
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.140934569259594e-06,
+      "loss": 0.4983,
+      "step": 1120
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.136928571886275e-06,
+      "loss": 0.4783,
+      "step": 1121
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.132923526625449e-06,
+      "loss": 0.4311,
+      "step": 1122
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.128919443979882e-06,
+      "loss": 0.53,
+      "step": 1123
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 2.1249163344498134e-06,
+      "loss": 0.4366,
+      "step": 1124
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.12091420853293e-06,
+      "loss": 0.5647,
+      "step": 1125
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.1169130767243424e-06,
+      "loss": 0.5462,
+      "step": 1126
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.1129129495165508e-06,
+      "loss": 0.4361,
+      "step": 1127
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.1089138373994226e-06,
+      "loss": 0.5007,
+      "step": 1128
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.104915750860164e-06,
+      "loss": 0.4496,
+      "step": 1129
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.10091870038329e-06,
+      "loss": 0.441,
+      "step": 1130
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 2.0969226964506007e-06,
+      "loss": 0.4813,
+      "step": 1131
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0929277495411487e-06,
+      "loss": 0.5451,
+      "step": 1132
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0889338701312184e-06,
+      "loss": 0.5516,
+      "step": 1133
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0849410686942914e-06,
+      "loss": 0.4876,
+      "step": 1134
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.080949355701025e-06,
+      "loss": 0.4676,
+      "step": 1135
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0769587416192212e-06,
+      "loss": 0.4829,
+      "step": 1136
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.072969236913799e-06,
+      "loss": 0.5243,
+      "step": 1137
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 2.0689808520467715e-06,
+      "loss": 0.5115,
+      "step": 1138
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0649935974772104e-06,
+      "loss": 0.5241,
+      "step": 1139
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.061007483661229e-06,
+      "loss": 0.5045,
+      "step": 1140
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0570225210519433e-06,
+      "loss": 0.4717,
+      "step": 1141
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0530387200994543e-06,
+      "loss": 0.5838,
+      "step": 1142
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0490560912508167e-06,
+      "loss": 0.5112,
+      "step": 1143
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 2.0450746449500085e-06,
+      "loss": 0.4187,
+      "step": 1144
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.04109439163791e-06,
+      "loss": 0.428,
+      "step": 1145
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0371153417522703e-06,
+      "loss": 0.5858,
+      "step": 1146
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0331375057276844e-06,
+      "loss": 0.4573,
+      "step": 1147
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0291608939955635e-06,
+      "loss": 0.5734,
+      "step": 1148
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.025185516984108e-06,
+      "loss": 0.5282,
+      "step": 1149
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0212113851182817e-06,
+      "loss": 0.543,
+      "step": 1150
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 2.0172385088197804e-06,
+      "loss": 0.5294,
+      "step": 1151
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.013266898507011e-06,
+      "loss": 0.492,
+      "step": 1152
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.0092965645950565e-06,
+      "loss": 0.4996,
+      "step": 1153
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.005327517495656e-06,
+      "loss": 0.4802,
+      "step": 1154
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 2.001359767617173e-06,
+      "loss": 0.5033,
+      "step": 1155
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.9973933253645684e-06,
+      "loss": 0.5131,
+      "step": 1156
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.993428201139375e-06,
+      "loss": 0.5275,
+      "step": 1157
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.9894644053396684e-06,
+      "loss": 0.4588,
+      "step": 1158
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9855019483600413e-06,
+      "loss": 0.5581,
+      "step": 1159
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.981540840591575e-06,
+      "loss": 0.4628,
+      "step": 1160
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.9775810924218126e-06,
+      "loss": 0.4002,
+      "step": 1161
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.973622714234733e-06,
+      "loss": 0.4387,
+      "step": 1162
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.96966571641072e-06,
+      "loss": 0.4966,
+      "step": 1163
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.965710109326541e-06,
+      "loss": 0.5013,
+      "step": 1164
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9617559033553128e-06,
+      "loss": 0.4942,
+      "step": 1165
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9578031088664812e-06,
+      "loss": 0.4623,
+      "step": 1166
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.953851736225787e-06,
+      "loss": 0.4075,
+      "step": 1167
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9499017957952454e-06,
+      "loss": 0.5121,
+      "step": 1168
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.945953297933115e-06,
+      "loss": 0.4191,
+      "step": 1169
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9420062529938704e-06,
+      "loss": 0.4992,
+      "step": 1170
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.9380606713281773e-06,
+      "loss": 0.4606,
+      "step": 1171
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9341165632828627e-06,
+      "loss": 0.5246,
+      "step": 1172
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9301739392008923e-06,
+      "loss": 0.452,
+      "step": 1173
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9262328094213356e-06,
+      "loss": 0.4503,
+      "step": 1174
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9222931842793473e-06,
+      "loss": 0.4951,
+      "step": 1175
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9183550741061354e-06,
+      "loss": 0.5651,
+      "step": 1176
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.914418489228934e-06,
+      "loss": 0.498,
+      "step": 1177
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.9104834399709784e-06,
+      "loss": 0.4609,
+      "step": 1178
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.9065499366514759e-06,
+      "loss": 0.42,
+      "step": 1179
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.9026179895855815e-06,
+      "loss": 0.4847,
+      "step": 1180
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8986876090843668e-06,
+      "loss": 0.4996,
+      "step": 1181
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.894758805454797e-06,
+      "loss": 0.5302,
+      "step": 1182
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8908315889997009e-06,
+      "loss": 0.4738,
+      "step": 1183
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.886905970017748e-06,
+      "loss": 0.4987,
+      "step": 1184
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8829819588034143e-06,
+      "loss": 0.5145,
+      "step": 1185
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8790595656469628e-06,
+      "loss": 0.4861,
+      "step": 1186
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8751388008344117e-06,
+      "loss": 0.512,
+      "step": 1187
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.87121967464751e-06,
+      "loss": 0.4811,
+      "step": 1188
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8673021973637095e-06,
+      "loss": 0.5819,
+      "step": 1189
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8633863792561369e-06,
+      "loss": 0.4865,
+      "step": 1190
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8594722305935691e-06,
+      "loss": 0.4814,
+      "step": 1191
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8555597616404047e-06,
+      "loss": 0.4654,
+      "step": 1192
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8516489826566375e-06,
+      "loss": 0.5213,
+      "step": 1193
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8477399038978288e-06,
+      "loss": 0.683,
+      "step": 1194
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8438325356150827e-06,
+      "loss": 0.5486,
+      "step": 1195
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8399268880550174e-06,
+      "loss": 0.5408,
+      "step": 1196
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8360229714597372e-06,
+      "loss": 0.5271,
+      "step": 1197
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.832120796066809e-06,
+      "loss": 0.5622,
+      "step": 1198
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.828220372109232e-06,
+      "loss": 0.4903,
+      "step": 1199
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8243217098154148e-06,
+      "loss": 0.51,
+      "step": 1200
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8204248194091429e-06,
+      "loss": 0.4562,
+      "step": 1201
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8165297111095586e-06,
+      "loss": 0.585,
+      "step": 1202
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8126363951311288e-06,
+      "loss": 0.463,
+      "step": 1203
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8087448816836206e-06,
+      "loss": 0.4834,
+      "step": 1204
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8048551809720752e-06,
+      "loss": 0.4431,
+      "step": 1205
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.800967303196778e-06,
+      "loss": 0.5032,
+      "step": 1206
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.797081258553236e-06,
+      "loss": 0.4728,
+      "step": 1207
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7931970572321479e-06,
+      "loss": 0.529,
+      "step": 1208
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7893147094193786e-06,
+      "loss": 0.4668,
+      "step": 1209
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7854342252959339e-06,
+      "loss": 0.4923,
+      "step": 1210
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.7815556150379298e-06,
+      "loss": 0.5365,
+      "step": 1211
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7776788888165703e-06,
+      "loss": 0.5372,
+      "step": 1212
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7738040567981168e-06,
+      "loss": 0.5019,
+      "step": 1213
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7699311291438664e-06,
+      "loss": 0.476,
+      "step": 1214
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.766060116010118e-06,
+      "loss": 0.4739,
+      "step": 1215
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7621910275481544e-06,
+      "loss": 0.495,
+      "step": 1216
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.7583238739042086e-06,
+      "loss": 0.3598,
+      "step": 1217
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.754458665219439e-06,
+      "loss": 0.5047,
+      "step": 1218
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7505954116299062e-06,
+      "loss": 0.3628,
+      "step": 1219
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7467341232665413e-06,
+      "loss": 0.4657,
+      "step": 1220
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7428748102551237e-06,
+      "loss": 0.5016,
+      "step": 1221
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7390174827162512e-06,
+      "loss": 0.6096,
+      "step": 1222
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7351621507653157e-06,
+      "loss": 0.483,
+      "step": 1223
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.7313088245124764e-06,
+      "loss": 0.4911,
+      "step": 1224
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7274575140626318e-06,
+      "loss": 0.5463,
+      "step": 1225
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7236082295153948e-06,
+      "loss": 0.4316,
+      "step": 1226
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7197609809650644e-06,
+      "loss": 0.5521,
+      "step": 1227
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7159157785006028e-06,
+      "loss": 0.4866,
+      "step": 1228
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7120726322056042e-06,
+      "loss": 0.5667,
+      "step": 1229
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7082315521582716e-06,
+      "loss": 0.5053,
+      "step": 1230
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.7043925484313911e-06,
+      "loss": 0.5308,
+      "step": 1231
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.7005556310923005e-06,
+      "loss": 0.5292,
+      "step": 1232
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6967208102028698e-06,
+      "loss": 0.4672,
+      "step": 1233
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6928880958194688e-06,
+      "loss": 0.5538,
+      "step": 1234
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6890574979929447e-06,
+      "loss": 0.5385,
+      "step": 1235
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.685229026768593e-06,
+      "loss": 0.5469,
+      "step": 1236
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6814026921861337e-06,
+      "loss": 0.4525,
+      "step": 1237
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.6775785042796835e-06,
+      "loss": 0.5065,
+      "step": 1238
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6737564730777283e-06,
+      "loss": 0.4933,
+      "step": 1239
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6699366086031011e-06,
+      "loss": 0.3978,
+      "step": 1240
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6661189208729492e-06,
+      "loss": 0.4701,
+      "step": 1241
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6623034198987148e-06,
+      "loss": 0.5661,
+      "step": 1242
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.6584901156861038e-06,
+      "loss": 0.4971,
+      "step": 1243
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.654679018235062e-06,
+      "loss": 0.4487,
+      "step": 1244
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6508701375397488e-06,
+      "loss": 0.4797,
+      "step": 1245
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6470634835885097e-06,
+      "loss": 0.5188,
+      "step": 1246
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6432590663638504e-06,
+      "loss": 0.4896,
+      "step": 1247
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6394568958424121e-06,
+      "loss": 0.5022,
+      "step": 1248
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.635656981994943e-06,
+      "loss": 0.4992,
+      "step": 1249
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6318593347862743e-06,
+      "loss": 0.6003,
+      "step": 1250
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.6280639641752944e-06,
+      "loss": 0.4516,
+      "step": 1251
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6242708801149185e-06,
+      "loss": 0.4766,
+      "step": 1252
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6204800925520685e-06,
+      "loss": 0.4659,
+      "step": 1253
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.616691611427642e-06,
+      "loss": 0.4996,
+      "step": 1254
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6129054466764904e-06,
+      "loss": 0.5539,
+      "step": 1255
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6091216082273875e-06,
+      "loss": 0.4851,
+      "step": 1256
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6053401060030098e-06,
+      "loss": 0.4972,
+      "step": 1257
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.6015609499199063e-06,
+      "loss": 0.4929,
+      "step": 1258
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5977841498884725e-06,
+      "loss": 0.5388,
+      "step": 1259
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5940097158129264e-06,
+      "loss": 0.4236,
+      "step": 1260
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5902376575912815e-06,
+      "loss": 0.4647,
+      "step": 1261
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5864679851153206e-06,
+      "loss": 0.5081,
+      "step": 1262
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.58270070827057e-06,
+      "loss": 0.5796,
+      "step": 1263
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.5789358369362744e-06,
+      "loss": 0.522,
+      "step": 1264
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5751733809853703e-06,
+      "loss": 0.5283,
+      "step": 1265
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5714133502844591e-06,
+      "loss": 0.4921,
+      "step": 1266
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5676557546937838e-06,
+      "loss": 0.4463,
+      "step": 1267
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5639006040671998e-06,
+      "loss": 0.5656,
+      "step": 1268
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5601479082521526e-06,
+      "loss": 0.498,
+      "step": 1269
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.556397677089649e-06,
+      "loss": 0.4458,
+      "step": 1270
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.5526499204142332e-06,
+      "loss": 0.5699,
+      "step": 1271
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5489046480539611e-06,
+      "loss": 0.4624,
+      "step": 1272
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.545161869830371e-06,
+      "loss": 0.5133,
+      "step": 1273
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.541421595558464e-06,
+      "loss": 0.464,
+      "step": 1274
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5376838350466722e-06,
+      "loss": 0.5143,
+      "step": 1275
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5339485980968383e-06,
+      "loss": 0.4651,
+      "step": 1276
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.530215894504184e-06,
+      "loss": 0.4097,
+      "step": 1277
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.5264857340572904e-06,
+      "loss": 0.458,
+      "step": 1278
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5227581265380686e-06,
+      "loss": 0.4851,
+      "step": 1279
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5190330817217335e-06,
+      "loss": 0.5386,
+      "step": 1280
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5153106093767827e-06,
+      "loss": 0.4921,
+      "step": 1281
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5115907192649644e-06,
+      "loss": 0.4945,
+      "step": 1282
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5078734211412574e-06,
+      "loss": 0.4188,
+      "step": 1283
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.5041587247538423e-06,
+      "loss": 0.5197,
+      "step": 1284
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.5004466398440776e-06,
+      "loss": 0.4665,
+      "step": 1285
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4967371761464738e-06,
+      "loss": 0.5364,
+      "step": 1286
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4930303433886661e-06,
+      "loss": 0.4525,
+      "step": 1287
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4893261512913918e-06,
+      "loss": 0.4955,
+      "step": 1288
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4856246095684623e-06,
+      "loss": 0.5192,
+      "step": 1289
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4819257279267402e-06,
+      "loss": 0.5071,
+      "step": 1290
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.4782295160661103e-06,
+      "loss": 0.49,
+      "step": 1291
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4745359836794577e-06,
+      "loss": 0.5344,
+      "step": 1292
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4708451404526409e-06,
+      "loss": 0.4754,
+      "step": 1293
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4671569960644655e-06,
+      "loss": 0.5211,
+      "step": 1294
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4634715601866607e-06,
+      "loss": 0.5277,
+      "step": 1295
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4597888424838519e-06,
+      "loss": 0.4491,
+      "step": 1296
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4561088526135376e-06,
+      "loss": 0.5786,
+      "step": 1297
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.4524316002260605e-06,
+      "loss": 0.5032,
+      "step": 1298
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4487570949645888e-06,
+      "loss": 0.4815,
+      "step": 1299
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4450853464650833e-06,
+      "loss": 0.4062,
+      "step": 1300
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4414163643562755e-06,
+      "loss": 0.4631,
+      "step": 1301
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4377501582596432e-06,
+      "loss": 0.4175,
+      "step": 1302
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.434086737789386e-06,
+      "loss": 0.4886,
+      "step": 1303
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.4304261125523952e-06,
+      "loss": 0.5014,
+      "step": 1304
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4267682921482356e-06,
+      "loss": 0.4935,
+      "step": 1305
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4231132861691128e-06,
+      "loss": 0.5028,
+      "step": 1306
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.419461104199856e-06,
+      "loss": 0.4207,
+      "step": 1307
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4158117558178857e-06,
+      "loss": 0.5218,
+      "step": 1308
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4121652505931922e-06,
+      "loss": 0.551,
+      "step": 1309
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.40852159808831e-06,
+      "loss": 0.5425,
+      "step": 1310
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.4048808078582943e-06,
+      "loss": 0.5284,
+      "step": 1311
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.4012428894506923e-06,
+      "loss": 0.5098,
+      "step": 1312
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3976078524055203e-06,
+      "loss": 0.4785,
+      "step": 1313
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3939757062552405e-06,
+      "loss": 0.4344,
+      "step": 1314
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3903464605247325e-06,
+      "loss": 0.6089,
+      "step": 1315
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3867201247312697e-06,
+      "loss": 0.4436,
+      "step": 1316
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3830967083844944e-06,
+      "loss": 0.5209,
+      "step": 1317
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.3794762209863942e-06,
+      "loss": 0.5131,
+      "step": 1318
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.375858672031276e-06,
+      "loss": 0.4902,
+      "step": 1319
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3722440710057383e-06,
+      "loss": 0.4911,
+      "step": 1320
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3686324273886531e-06,
+      "loss": 0.4647,
+      "step": 1321
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3650237506511333e-06,
+      "loss": 0.5925,
+      "step": 1322
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3614180502565136e-06,
+      "loss": 0.5658,
+      "step": 1323
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.3578153356603217e-06,
+      "loss": 0.4705,
+      "step": 1324
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3542156163102582e-06,
+      "loss": 0.4746,
+      "step": 1325
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3506189016461674e-06,
+      "loss": 0.5419,
+      "step": 1326
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3470252011000124e-06,
+      "loss": 0.5226,
+      "step": 1327
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3434345240958558e-06,
+      "loss": 0.484,
+      "step": 1328
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3398468800498293e-06,
+      "loss": 0.4949,
+      "step": 1329
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3362622783701104e-06,
+      "loss": 0.4514,
+      "step": 1330
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.3326807284568984e-06,
+      "loss": 0.5136,
+      "step": 1331
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3291022397023926e-06,
+      "loss": 0.4269,
+      "step": 1332
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3255268214907612e-06,
+      "loss": 0.5337,
+      "step": 1333
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3219544831981212e-06,
+      "loss": 0.4548,
+      "step": 1334
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3183852341925147e-06,
+      "loss": 0.5338,
+      "step": 1335
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3148190838338804e-06,
+      "loss": 0.4665,
+      "step": 1336
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3112560414740315e-06,
+      "loss": 0.4307,
+      "step": 1337
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.3076961164566309e-06,
+      "loss": 0.4937,
+      "step": 1338
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.3041393181171688e-06,
+      "loss": 0.476,
+      "step": 1339
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.3005856557829327e-06,
+      "loss": 0.4017,
+      "step": 1340
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2970351387729875e-06,
+      "loss": 0.4383,
+      "step": 1341
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2934877763981513e-06,
+      "loss": 0.4089,
+      "step": 1342
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2899435779609681e-06,
+      "loss": 0.3659,
+      "step": 1343
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2864025527556845e-06,
+      "loss": 0.5362,
+      "step": 1344
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.2828647100682263e-06,
+      "loss": 0.4324,
+      "step": 1345
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2793300591761742e-06,
+      "loss": 0.4442,
+      "step": 1346
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.275798609348738e-06,
+      "loss": 0.5234,
+      "step": 1347
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.272270369846732e-06,
+      "loss": 0.5057,
+      "step": 1348
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2687453499225547e-06,
+      "loss": 0.4878,
+      "step": 1349
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2652235588201592e-06,
+      "loss": 0.4159,
+      "step": 1350
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.2617050057750322e-06,
+      "loss": 0.363,
+      "step": 1351
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2581897000141675e-06,
+      "loss": 0.5059,
+      "step": 1352
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2546776507560468e-06,
+      "loss": 0.463,
+      "step": 1353
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.251168867210609e-06,
+      "loss": 0.5327,
+      "step": 1354
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2476633585792287e-06,
+      "loss": 0.4702,
+      "step": 1355
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2441611340546958e-06,
+      "loss": 0.3838,
+      "step": 1356
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2406622028211846e-06,
+      "loss": 0.4579,
+      "step": 1357
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.2371665740542345e-06,
+      "loss": 0.3739,
+      "step": 1358
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2336742569207235e-06,
+      "loss": 0.4152,
+      "step": 1359
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2301852605788481e-06,
+      "loss": 0.4507,
+      "step": 1360
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2266995941780934e-06,
+      "loss": 0.5148,
+      "step": 1361
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2232172668592124e-06,
+      "loss": 0.4519,
+      "step": 1362
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.219738287754204e-06,
+      "loss": 0.4443,
+      "step": 1363
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2162626659862846e-06,
+      "loss": 0.3719,
+      "step": 1364
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.2127904106698665e-06,
+      "loss": 0.4679,
+      "step": 1365
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.2093215309105352e-06,
+      "loss": 0.4799,
+      "step": 1366
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.2058560358050242e-06,
+      "loss": 0.4253,
+      "step": 1367
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.2023939344411897e-06,
+      "loss": 0.3983,
+      "step": 1368
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.1989352358979888e-06,
+      "loss": 0.4969,
+      "step": 1369
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.1954799492454545e-06,
+      "loss": 0.4198,
+      "step": 1370
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.192028083544675e-06,
+      "loss": 0.563,
+      "step": 1371
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1885796478477643e-06,
+      "loss": 0.3931,
+      "step": 1372
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1851346511978427e-06,
+      "loss": 0.4782,
+      "step": 1373
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1816931026290137e-06,
+      "loss": 0.4498,
+      "step": 1374
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.178255011166337e-06,
+      "loss": 0.4177,
+      "step": 1375
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.1748203858258056e-06,
+      "loss": 0.4668,
+      "step": 1376
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.171389235614324e-06,
+      "loss": 0.5175,
+      "step": 1377
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.167961569529685e-06,
+      "loss": 0.4483,
+      "step": 1378
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1645373965605424e-06,
+      "loss": 0.5034,
+      "step": 1379
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1611167256863904e-06,
+      "loss": 0.4341,
+      "step": 1380
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1576995658775405e-06,
+      "loss": 0.4655,
+      "step": 1381
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1542859260950954e-06,
+      "loss": 0.4351,
+      "step": 1382
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1508758152909276e-06,
+      "loss": 0.4388,
+      "step": 1383
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.147469242407654e-06,
+      "loss": 0.5119,
+      "step": 1384
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.1440662163786168e-06,
+      "loss": 0.5246,
+      "step": 1385
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.140666746127854e-06,
+      "loss": 0.4905,
+      "step": 1386
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1372708405700794e-06,
+      "loss": 0.4278,
+      "step": 1387
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.133878508610661e-06,
+      "loss": 0.5272,
+      "step": 1388
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.130489759145593e-06,
+      "loss": 0.4667,
+      "step": 1389
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1271046010614758e-06,
+      "loss": 0.4303,
+      "step": 1390
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.1237230432354912e-06,
+      "loss": 0.43,
+      "step": 1391
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.120345094535382e-06,
+      "loss": 0.4072,
+      "step": 1392
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.116970763819424e-06,
+      "loss": 0.424,
+      "step": 1393
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1136000599364048e-06,
+      "loss": 0.4646,
+      "step": 1394
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1102329917256047e-06,
+      "loss": 0.4779,
+      "step": 1395
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1068695680167665e-06,
+      "loss": 0.4854,
+      "step": 1396
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.103509797630077e-06,
+      "loss": 0.4583,
+      "step": 1397
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.1001536893761412e-06,
+      "loss": 0.4938,
+      "step": 1398
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0968012520559634e-06,
+      "loss": 0.4551,
+      "step": 1399
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.093452494460919e-06,
+      "loss": 0.4492,
+      "step": 1400
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0901074253727338e-06,
+      "loss": 0.4131,
+      "step": 1401
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0867660535634625e-06,
+      "loss": 0.4242,
+      "step": 1402
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0834283877954629e-06,
+      "loss": 0.473,
+      "step": 1403
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0800944368213742e-06,
+      "loss": 0.4399,
+      "step": 1404
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.0767642093840933e-06,
+      "loss": 0.4823,
+      "step": 1405
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0734377142167549e-06,
+      "loss": 0.4108,
+      "step": 1406
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0701149600427044e-06,
+      "loss": 0.389,
+      "step": 1407
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0667959555754758e-06,
+      "loss": 0.5169,
+      "step": 1408
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0634807095187739e-06,
+      "loss": 0.4313,
+      "step": 1409
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0601692305664436e-06,
+      "loss": 0.4564,
+      "step": 1410
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.0568615274024521e-06,
+      "loss": 0.4819,
+      "step": 1411
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.053557608700865e-06,
+      "loss": 0.4822,
+      "step": 1412
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0502574831258259e-06,
+      "loss": 0.4398,
+      "step": 1413
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0469611593315276e-06,
+      "loss": 0.4748,
+      "step": 1414
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.043668645962195e-06,
+      "loss": 0.4937,
+      "step": 1415
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.0403799516520619e-06,
+      "loss": 0.4754,
+      "step": 1416
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.037095085025345e-06,
+      "loss": 0.3886,
+      "step": 1417
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.033814054696224e-06,
+      "loss": 0.4171,
+      "step": 1418
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0305368692688175e-06,
+      "loss": 0.4557,
+      "step": 1419
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0272635373371639e-06,
+      "loss": 0.4212,
+      "step": 1420
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0239940674851943e-06,
+      "loss": 0.4779,
+      "step": 1421
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0207284682867104e-06,
+      "loss": 0.4611,
+      "step": 1422
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0174667483053682e-06,
+      "loss": 0.5081,
+      "step": 1423
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0142089160946463e-06,
+      "loss": 0.4348,
+      "step": 1424
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.0109549801978306e-06,
+      "loss": 0.4838,
+      "step": 1425
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.0077049491479874e-06,
+      "loss": 0.4837,
+      "step": 1426
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.0044588314679452e-06,
+      "loss": 0.5279,
+      "step": 1427
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.0012166356702698e-06,
+      "loss": 0.4449,
+      "step": 1428
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.979783702572413e-07,
+      "loss": 0.4564,
+      "step": 1429
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.947440437208322e-07,
+      "loss": 0.4056,
+      "step": 1430
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 9.915136645426885e-07,
+      "loss": 0.4451,
+      "step": 1431
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.88287241194102e-07,
+      "loss": 0.4577,
+      "step": 1432
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.850647821359918e-07,
+      "loss": 0.5392,
+      "step": 1433
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.818462958188805e-07,
+      "loss": 0.4537,
+      "step": 1434
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.786317906828745e-07,
+      "loss": 0.4697,
+      "step": 1435
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.754212751576386e-07,
+      "loss": 0.466,
+      "step": 1436
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.722147576623745e-07,
+      "loss": 0.4242,
+      "step": 1437
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 9.690122466058017e-07,
+      "loss": 0.4562,
+      "step": 1438
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.658137503861314e-07,
+      "loss": 0.5334,
+      "step": 1439
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.626192773910476e-07,
+      "loss": 0.4609,
+      "step": 1440
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.594288359976817e-07,
+      "loss": 0.4432,
+      "step": 1441
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.562424345725964e-07,
+      "loss": 0.5096,
+      "step": 1442
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.530600814717575e-07,
+      "loss": 0.4612,
+      "step": 1443
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.498817850405137e-07,
+      "loss": 0.4408,
+      "step": 1444
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.467075536135787e-07,
+      "loss": 0.3795,
+      "step": 1445
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.435373955150032e-07,
+      "loss": 0.4764,
+      "step": 1446
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.403713190581576e-07,
+      "loss": 0.4324,
+      "step": 1447
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.372093325457074e-07,
+      "loss": 0.4194,
+      "step": 1448
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.340514442695953e-07,
+      "loss": 0.4972,
+      "step": 1449
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.308976625110144e-07,
+      "loss": 0.4386,
+      "step": 1450
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 9.277479955403887e-07,
+      "loss": 0.5053,
+      "step": 1451
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.246024516173544e-07,
+      "loss": 0.5108,
+      "step": 1452
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.214610389907327e-07,
+      "loss": 0.4599,
+      "step": 1453
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.183237658985122e-07,
+      "loss": 0.5028,
+      "step": 1454
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.151906405678251e-07,
+      "loss": 0.4876,
+      "step": 1455
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.120616712149291e-07,
+      "loss": 0.4122,
+      "step": 1456
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.0893686604518e-07,
+      "loss": 0.4667,
+      "step": 1457
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 9.058162332530152e-07,
+      "loss": 0.5324,
+      "step": 1458
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 9.026997810219313e-07,
+      "loss": 0.4661,
+      "step": 1459
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.995875175244606e-07,
+      "loss": 0.4138,
+      "step": 1460
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.964794509221508e-07,
+      "loss": 0.5155,
+      "step": 1461
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.933755893655441e-07,
+      "loss": 0.4465,
+      "step": 1462
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.902759409941567e-07,
+      "loss": 0.4247,
+      "step": 1463
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.871805139364545e-07,
+      "loss": 0.4801,
+      "step": 1464
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 8.840893163098332e-07,
+      "loss": 0.4326,
+      "step": 1465
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.810023562206e-07,
+      "loss": 0.4958,
+      "step": 1466
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.779196417639465e-07,
+      "loss": 0.5173,
+      "step": 1467
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.748411810239327e-07,
+      "loss": 0.425,
+      "step": 1468
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.71766982073462e-07,
+      "loss": 0.4145,
+      "step": 1469
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.686970529742641e-07,
+      "loss": 0.4673,
+      "step": 1470
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 8.656314017768694e-07,
+      "loss": 0.4612,
+      "step": 1471
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.625700365205902e-07,
+      "loss": 0.4258,
+      "step": 1472
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.595129652335019e-07,
+      "loss": 0.4608,
+      "step": 1473
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.564601959324159e-07,
+      "loss": 0.4926,
+      "step": 1474
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.534117366228645e-07,
+      "loss": 0.3865,
+      "step": 1475
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.503675952990756e-07,
+      "loss": 0.3896,
+      "step": 1476
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.473277799439569e-07,
+      "loss": 0.4522,
+      "step": 1477
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 8.442922985290689e-07,
+      "loss": 0.4293,
+      "step": 1478
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.412611590146069e-07,
+      "loss": 0.5035,
+      "step": 1479
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.382343693493825e-07,
+      "loss": 0.4195,
+      "step": 1480
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.352119374707979e-07,
+      "loss": 0.4237,
+      "step": 1481
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.321938713048286e-07,
+      "loss": 0.5307,
+      "step": 1482
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.291801787660001e-07,
+      "loss": 0.446,
+      "step": 1483
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.261708677573712e-07,
+      "loss": 0.4329,
+      "step": 1484
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 8.231659461705092e-07,
+      "loss": 0.4505,
+      "step": 1485
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.20165421885469e-07,
+      "loss": 0.3765,
+      "step": 1486
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.171693027707772e-07,
+      "loss": 0.5111,
+      "step": 1487
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.141775966834073e-07,
+      "loss": 0.43,
+      "step": 1488
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.111903114687591e-07,
+      "loss": 0.5395,
+      "step": 1489
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.082074549606389e-07,
+      "loss": 0.491,
+      "step": 1490
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 8.052290349812419e-07,
+      "loss": 0.4452,
+      "step": 1491
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 8.022550593411263e-07,
+      "loss": 0.4599,
+      "step": 1492
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.992855358391968e-07,
+      "loss": 0.4507,
+      "step": 1493
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.963204722626814e-07,
+      "loss": 0.3951,
+      "step": 1494
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.933598763871156e-07,
+      "loss": 0.514,
+      "step": 1495
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.904037559763162e-07,
+      "loss": 0.4778,
+      "step": 1496
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.874521187823631e-07,
+      "loss": 0.5329,
+      "step": 1497
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 7.845049725455828e-07,
+      "loss": 0.4519,
+      "step": 1498
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.815623249945215e-07,
+      "loss": 0.3927,
+      "step": 1499
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.786241838459299e-07,
+      "loss": 0.451,
+      "step": 1500
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.756905568047393e-07,
+      "loss": 0.4598,
+      "step": 1501
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.727614515640466e-07,
+      "loss": 0.4241,
+      "step": 1502
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.698368758050878e-07,
+      "loss": 0.4182,
+      "step": 1503
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.669168371972214e-07,
+      "loss": 0.4599,
+      "step": 1504
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 7.640013433979093e-07,
+      "loss": 0.4796,
+      "step": 1505
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.610904020526938e-07,
+      "loss": 0.4542,
+      "step": 1506
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.58184020795179e-07,
+      "loss": 0.4307,
+      "step": 1507
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.552822072470101e-07,
+      "loss": 0.408,
+      "step": 1508
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.523849690178567e-07,
+      "loss": 0.4883,
+      "step": 1509
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.49492313705387e-07,
+      "loss": 0.4884,
+      "step": 1510
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 7.466042488952521e-07,
+      "loss": 0.4791,
+      "step": 1511
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.437207821610665e-07,
+      "loss": 0.4492,
+      "step": 1512
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.408419210643847e-07,
+      "loss": 0.5333,
+      "step": 1513
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.379676731546847e-07,
+      "loss": 0.4531,
+      "step": 1514
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.350980459693455e-07,
+      "loss": 0.4612,
+      "step": 1515
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.322330470336314e-07,
+      "loss": 0.4807,
+      "step": 1516
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.293726838606674e-07,
+      "loss": 0.5121,
+      "step": 1517
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 7.265169639514219e-07,
+      "loss": 0.489,
+      "step": 1518
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.236658947946886e-07,
+      "loss": 0.5103,
+      "step": 1519
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.208194838670635e-07,
+      "loss": 0.4678,
+      "step": 1520
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.179777386329276e-07,
+      "loss": 0.5257,
+      "step": 1521
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.151406665444249e-07,
+      "loss": 0.4131,
+      "step": 1522
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.123082750414487e-07,
+      "loss": 0.4954,
+      "step": 1523
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.094805715516143e-07,
+      "loss": 0.4988,
+      "step": 1524
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 7.066575634902437e-07,
+      "loss": 0.5195,
+      "step": 1525
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 7.038392582603481e-07,
+      "loss": 0.4416,
+      "step": 1526
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 7.010256632526036e-07,
+      "loss": 0.457,
+      "step": 1527
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.982167858453348e-07,
+      "loss": 0.4461,
+      "step": 1528
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.95412633404495e-07,
+      "loss": 0.4322,
+      "step": 1529
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.926132132836483e-07,
+      "loss": 0.4114,
+      "step": 1530
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 6.898185328239468e-07,
+      "loss": 0.4219,
+      "step": 1531
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.870285993541137e-07,
+      "loss": 0.4751,
+      "step": 1532
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.842434201904255e-07,
+      "loss": 0.4236,
+      "step": 1533
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.814630026366891e-07,
+      "loss": 0.4776,
+      "step": 1534
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.78687353984226e-07,
+      "loss": 0.4358,
+      "step": 1535
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.759164815118493e-07,
+      "loss": 0.4614,
+      "step": 1536
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.731503924858518e-07,
+      "loss": 0.3945,
+      "step": 1537
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 6.703890941599781e-07,
+      "loss": 0.5037,
+      "step": 1538
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.676325937754102e-07,
+      "loss": 0.4205,
+      "step": 1539
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.648808985607511e-07,
+      "loss": 0.4381,
+      "step": 1540
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.621340157319998e-07,
+      "loss": 0.4694,
+      "step": 1541
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.593919524925357e-07,
+      "loss": 0.4127,
+      "step": 1542
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.566547160331e-07,
+      "loss": 0.488,
+      "step": 1543
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.539223135317771e-07,
+      "loss": 0.4316,
+      "step": 1544
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 6.511947521539738e-07,
+      "loss": 0.4441,
+      "step": 1545
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.484720390524008e-07,
+      "loss": 0.4642,
+      "step": 1546
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.457541813670565e-07,
+      "loss": 0.4684,
+      "step": 1547
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.430411862252067e-07,
+      "loss": 0.4608,
+      "step": 1548
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.403330607413643e-07,
+      "loss": 0.4528,
+      "step": 1549
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.376298120172716e-07,
+      "loss": 0.4635,
+      "step": 1550
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 6.349314471418849e-07,
+      "loss": 0.4574,
+      "step": 1551
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.322379731913505e-07,
+      "loss": 0.4016,
+      "step": 1552
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.295493972289904e-07,
+      "loss": 0.5174,
+      "step": 1553
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.268657263052805e-07,
+      "loss": 0.4636,
+      "step": 1554
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.241869674578363e-07,
+      "loss": 0.4174,
+      "step": 1555
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.2151312771139e-07,
+      "loss": 0.4968,
+      "step": 1556
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.188442140777742e-07,
+      "loss": 0.3709,
+      "step": 1557
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 6.161802335559048e-07,
+      "loss": 0.4102,
+      "step": 1558
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.135211931317594e-07,
+      "loss": 0.4896,
+      "step": 1559
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.108670997783619e-07,
+      "loss": 0.5289,
+      "step": 1560
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.082179604557617e-07,
+      "loss": 0.4658,
+      "step": 1561
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.055737821110194e-07,
+      "loss": 0.4265,
+      "step": 1562
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.029345716781837e-07,
+      "loss": 0.5053,
+      "step": 1563
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 6.003003360782755e-07,
+      "loss": 0.4435,
+      "step": 1564
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 5.976710822192722e-07,
+      "loss": 0.5288,
+      "step": 1565
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.950468169960846e-07,
+      "loss": 0.417,
+      "step": 1566
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.924275472905425e-07,
+      "loss": 0.4274,
+      "step": 1567
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.898132799713744e-07,
+      "loss": 0.4498,
+      "step": 1568
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.872040218941929e-07,
+      "loss": 0.5019,
+      "step": 1569
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.845997799014729e-07,
+      "loss": 0.4685,
+      "step": 1570
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.820005608225345e-07,
+      "loss": 0.4612,
+      "step": 1571
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.794063714735282e-07,
+      "loss": 0.4605,
+      "step": 1572
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.768172186574123e-07,
+      "loss": 0.5002,
+      "step": 1573
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.74233109163938e-07,
+      "loss": 0.5094,
+      "step": 1574
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.716540497696307e-07,
+      "loss": 0.4116,
+      "step": 1575
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.690800472377747e-07,
+      "loss": 0.4677,
+      "step": 1576
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.665111083183905e-07,
+      "loss": 0.4638,
+      "step": 1577
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 5.639472397482204e-07,
+      "loss": 0.4888,
+      "step": 1578
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.613884482507123e-07,
+      "loss": 0.5159,
+      "step": 1579
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.588347405359975e-07,
+      "loss": 0.4944,
+      "step": 1580
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.562861233008774e-07,
+      "loss": 0.4584,
+      "step": 1581
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.537426032288021e-07,
+      "loss": 0.4258,
+      "step": 1582
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.512041869898585e-07,
+      "loss": 0.4376,
+      "step": 1583
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.486708812407465e-07,
+      "loss": 0.5021,
+      "step": 1584
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 5.46142692624764e-07,
+      "loss": 0.4504,
+      "step": 1585
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.436196277717928e-07,
+      "loss": 0.3974,
+      "step": 1586
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.411016932982751e-07,
+      "loss": 0.527,
+      "step": 1587
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.385888958072014e-07,
+      "loss": 0.4177,
+      "step": 1588
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.360812418880884e-07,
+      "loss": 0.5192,
+      "step": 1589
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.335787381169683e-07,
+      "loss": 0.4776,
+      "step": 1590
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 5.310813910563645e-07,
+      "loss": 0.4739,
+      "step": 1591
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.285892072552779e-07,
+      "loss": 0.3835,
+      "step": 1592
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.261021932491714e-07,
+      "loss": 0.4951,
+      "step": 1593
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.236203555599487e-07,
+      "loss": 0.4454,
+      "step": 1594
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.211437006959396e-07,
+      "loss": 0.4613,
+      "step": 1595
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.186722351518822e-07,
+      "loss": 0.4251,
+      "step": 1596
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.162059654089083e-07,
+      "loss": 0.5011,
+      "step": 1597
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 5.137448979345217e-07,
+      "loss": 0.4768,
+      "step": 1598
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.112890391825844e-07,
+      "loss": 0.4279,
+      "step": 1599
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.088383955933013e-07,
+      "loss": 0.4581,
+      "step": 1600
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.063929735931985e-07,
+      "loss": 0.3997,
+      "step": 1601
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.039527795951102e-07,
+      "loss": 0.4709,
+      "step": 1602
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 5.015178199981602e-07,
+      "loss": 0.426,
+      "step": 1603
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.990881011877474e-07,
+      "loss": 0.4718,
+      "step": 1604
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 4.966636295355254e-07,
+      "loss": 0.3648,
+      "step": 1605
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.94244411399388e-07,
+      "loss": 0.4958,
+      "step": 1606
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.918304531234533e-07,
+      "loss": 0.3856,
+      "step": 1607
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.894217610380458e-07,
+      "loss": 0.4167,
+      "step": 1608
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.870183414596794e-07,
+      "loss": 0.4367,
+      "step": 1609
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.846202006910417e-07,
+      "loss": 0.4473,
+      "step": 1610
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.822273450209767e-07,
+      "loss": 0.4636,
+      "step": 1611
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 4.798397807244704e-07,
+      "loss": 0.4194,
+      "step": 1612
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.774575140626317e-07,
+      "loss": 0.4624,
+      "step": 1613
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.7508055128267615e-07,
+      "loss": 0.448,
+      "step": 1614
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.727088986179129e-07,
+      "loss": 0.4646,
+      "step": 1615
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.703425622877239e-07,
+      "loss": 0.4448,
+      "step": 1616
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.679815484975506e-07,
+      "loss": 0.5379,
+      "step": 1617
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 4.656258634388755e-07,
+      "loss": 0.4341,
+      "step": 1618
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.632755132892094e-07,
+      "loss": 0.4793,
+      "step": 1619
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.6093050421207096e-07,
+      "loss": 0.4053,
+      "step": 1620
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.5859084235697236e-07,
+      "loss": 0.4879,
+      "step": 1621
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.562565338594055e-07,
+      "loss": 0.5142,
+      "step": 1622
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.539275848408217e-07,
+      "loss": 0.4694,
+      "step": 1623
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.5160400140861785e-07,
+      "loss": 0.482,
+      "step": 1624
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 4.4928578965612034e-07,
+      "loss": 0.4565,
+      "step": 1625
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.469729556625704e-07,
+      "loss": 0.4221,
+      "step": 1626
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.446655054931051e-07,
+      "loss": 0.4285,
+      "step": 1627
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.423634451987427e-07,
+      "loss": 0.439,
+      "step": 1628
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.400667808163689e-07,
+      "loss": 0.4277,
+      "step": 1629
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.3777551836871825e-07,
+      "loss": 0.4622,
+      "step": 1630
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.354896638643591e-07,
+      "loss": 0.4539,
+      "step": 1631
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 4.3320922329767786e-07,
+      "loss": 0.4835,
+      "step": 1632
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.3093420264886525e-07,
+      "loss": 0.4471,
+      "step": 1633
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.2866460788389716e-07,
+      "loss": 0.4233,
+      "step": 1634
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.264004449545206e-07,
+      "loss": 0.3734,
+      "step": 1635
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.2414171979824e-07,
+      "loss": 0.4646,
+      "step": 1636
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.2188843833829874e-07,
+      "loss": 0.4503,
+      "step": 1637
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 4.196406064836639e-07,
+      "loss": 0.4595,
+      "step": 1638
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.1739823012901223e-07,
+      "loss": 0.4595,
+      "step": 1639
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.15161315154716e-07,
+      "loss": 0.4944,
+      "step": 1640
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.129298674268226e-07,
+      "loss": 0.4209,
+      "step": 1641
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.1070389279704305e-07,
+      "loss": 0.4408,
+      "step": 1642
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.08483397102738e-07,
+      "loss": 0.49,
+      "step": 1643
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.0626838616689786e-07,
+      "loss": 0.4216,
+      "step": 1644
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 4.040588657981301e-07,
+      "loss": 0.3884,
+      "step": 1645
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.0185484179064427e-07,
+      "loss": 0.448,
+      "step": 1646
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9965631992423704e-07,
+      "loss": 0.4997,
+      "step": 1647
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9746330596427574e-07,
+      "loss": 0.4953,
+      "step": 1648
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.952758056616826e-07,
+      "loss": 0.4843,
+      "step": 1649
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.930938247529237e-07,
+      "loss": 0.5593,
+      "step": 1650
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.9091736895998907e-07,
+      "loss": 0.4444,
+      "step": 1651
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.887464439903799e-07,
+      "loss": 0.391,
+      "step": 1652
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.8658105553709356e-07,
+      "loss": 0.5002,
+      "step": 1653
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.844212092786101e-07,
+      "loss": 0.4191,
+      "step": 1654
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.822669108788737e-07,
+      "loss": 0.4839,
+      "step": 1655
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.801181659872805e-07,
+      "loss": 0.4973,
+      "step": 1656
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.77974980238664e-07,
+      "loss": 0.4744,
+      "step": 1657
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.7583735925327906e-07,
+      "loss": 0.4576,
+      "step": 1658
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.737053086367873e-07,
+      "loss": 0.4942,
+      "step": 1659
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.715788339802423e-07,
+      "loss": 0.4899,
+      "step": 1660
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.6945794086007706e-07,
+      "loss": 0.4517,
+      "step": 1661
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.6734263483808625e-07,
+      "loss": 0.4461,
+      "step": 1662
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.6523292146141225e-07,
+      "loss": 0.5024,
+      "step": 1663
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.6312880626253406e-07,
+      "loss": 0.3888,
+      "step": 1664
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.610302947592473e-07,
+      "loss": 0.4584,
+      "step": 1665
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5893739245465465e-07,
+      "loss": 0.4431,
+      "step": 1666
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.56850104837147e-07,
+      "loss": 0.3447,
+      "step": 1667
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5476843738039364e-07,
+      "loss": 0.4098,
+      "step": 1668
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5269239554332565e-07,
+      "loss": 0.4417,
+      "step": 1669
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.5062198477011976e-07,
+      "loss": 0.4716,
+      "step": 1670
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.485572104901869e-07,
+      "loss": 0.4482,
+      "step": 1671
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.464980781181571e-07,
+      "loss": 0.5336,
+      "step": 1672
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.4444459305386507e-07,
+      "loss": 0.4907,
+      "step": 1673
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.423967606823353e-07,
+      "loss": 0.4517,
+      "step": 1674
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.403545863737706e-07,
+      "loss": 0.4158,
+      "step": 1675
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.383180754835344e-07,
+      "loss": 0.422,
+      "step": 1676
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.362872333521389e-07,
+      "loss": 0.5188,
+      "step": 1677
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.342620653052306e-07,
+      "loss": 0.4751,
+      "step": 1678
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.322425766535778e-07,
+      "loss": 0.4583,
+      "step": 1679
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.302287726930531e-07,
+      "loss": 0.5281,
+      "step": 1680
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.2822065870462216e-07,
+      "loss": 0.4336,
+      "step": 1681
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.262182399543312e-07,
+      "loss": 0.4727,
+      "step": 1682
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.2422152169328925e-07,
+      "loss": 0.4382,
+      "step": 1683
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.222305091576569e-07,
+      "loss": 0.473,
+      "step": 1684
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.2024520756863244e-07,
+      "loss": 0.4129,
+      "step": 1685
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.182656221324384e-07,
+      "loss": 0.5152,
+      "step": 1686
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.162917580403066e-07,
+      "loss": 0.4341,
+      "step": 1687
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.143236204684647e-07,
+      "loss": 0.4266,
+      "step": 1688
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.123612145781255e-07,
+      "loss": 0.4737,
+      "step": 1689
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.104045455154686e-07,
+      "loss": 0.4006,
+      "step": 1690
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.08453618411631e-07,
+      "loss": 0.5161,
+      "step": 1691
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.0650843838269076e-07,
+      "loss": 0.3798,
+      "step": 1692
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.0456901052965726e-07,
+      "loss": 0.445,
+      "step": 1693
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.026353399384527e-07,
+      "loss": 0.4455,
+      "step": 1694
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.0070743167990275e-07,
+      "loss": 0.5413,
+      "step": 1695
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.98785290809723e-07,
+      "loss": 0.4069,
+      "step": 1696
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.968689223685034e-07,
+      "loss": 0.45,
+      "step": 1697
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 2.949583313816964e-07,
+      "loss": 0.4663,
+      "step": 1698
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.9305352285960404e-07,
+      "loss": 0.4606,
+      "step": 1699
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.911545017973652e-07,
+      "loss": 0.4903,
+      "step": 1700
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.892612731749414e-07,
+      "loss": 0.4516,
+      "step": 1701
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.873738419571026e-07,
+      "loss": 0.4323,
+      "step": 1702
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.8549221309341903e-07,
+      "loss": 0.4244,
+      "step": 1703
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.836163915182419e-07,
+      "loss": 0.4434,
+      "step": 1704
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 2.8174638215069494e-07,
+      "loss": 0.4311,
+      "step": 1705
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.798821898946588e-07,
+      "loss": 0.3808,
+      "step": 1706
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.780238196387619e-07,
+      "loss": 0.4246,
+      "step": 1707
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.761712762563626e-07,
+      "loss": 0.4996,
+      "step": 1708
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.743245646055398e-07,
+      "loss": 0.4018,
+      "step": 1709
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.7248368952908055e-07,
+      "loss": 0.4356,
+      "step": 1710
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.706486558544644e-07,
+      "loss": 0.4619,
+      "step": 1711
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 2.6881946839385303e-07,
+      "loss": 0.4792,
+      "step": 1712
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.6699613194407724e-07,
+      "loss": 0.4993,
+      "step": 1713
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.651786512866253e-07,
+      "loss": 0.4701,
+      "step": 1714
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.633670311876277e-07,
+      "loss": 0.5283,
+      "step": 1715
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.615612763978462e-07,
+      "loss": 0.4505,
+      "step": 1716
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.5976139165266367e-07,
+      "loss": 0.4625,
+      "step": 1717
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 2.579673816720674e-07,
+      "loss": 0.4437,
+      "step": 1718
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.5617925116063926e-07,
+      "loss": 0.4563,
+      "step": 1719
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.5439700480754283e-07,
+      "loss": 0.5358,
+      "step": 1720
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.52620647286512e-07,
+      "loss": 0.488,
+      "step": 1721
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.5085018325583717e-07,
+      "loss": 0.4858,
+      "step": 1722
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.4908561735835306e-07,
+      "loss": 0.4603,
+      "step": 1723
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.4732695422142934e-07,
+      "loss": 0.4556,
+      "step": 1724
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 2.455741984569543e-07,
+      "loss": 0.4241,
+      "step": 1725
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.438273546613257e-07,
+      "loss": 0.498,
+      "step": 1726
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.420864274154372e-07,
+      "loss": 0.5554,
+      "step": 1727
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.403514212846686e-07,
+      "loss": 0.5224,
+      "step": 1728
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.386223408188704e-07,
+      "loss": 0.442,
+      "step": 1729
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.3689919055235517e-07,
+      "loss": 0.4045,
+      "step": 1730
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.3518197500388278e-07,
+      "loss": 0.431,
+      "step": 1731
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 2.334706986766519e-07,
+      "loss": 0.4654,
+      "step": 1732
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.3176536605828443e-07,
+      "loss": 0.4327,
+      "step": 1733
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.3006598162081673e-07,
+      "loss": 0.4774,
+      "step": 1734
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2837254982068568e-07,
+      "loss": 0.4721,
+      "step": 1735
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2668507509871957e-07,
+      "loss": 0.4697,
+      "step": 1736
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2500356188012413e-07,
+      "loss": 0.4773,
+      "step": 1737
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 2.2332801457447069e-07,
+      "loss": 0.5123,
+      "step": 1738
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.2165843757568807e-07,
+      "loss": 0.494,
+      "step": 1739
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1999483526204678e-07,
+      "loss": 0.4742,
+      "step": 1740
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1833721199614992e-07,
+      "loss": 0.5579,
+      "step": 1741
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1668557212492057e-07,
+      "loss": 0.4546,
+      "step": 1742
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.15039919979593e-07,
+      "loss": 0.4432,
+      "step": 1743
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1340025987569763e-07,
+      "loss": 0.4744,
+      "step": 1744
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 2.1176659611305133e-07,
+      "loss": 0.5185,
+      "step": 1745
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.1013893297574777e-07,
+      "loss": 0.3917,
+      "step": 1746
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0851727473214317e-07,
+      "loss": 0.4716,
+      "step": 1747
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0690162563484772e-07,
+      "loss": 0.4696,
+      "step": 1748
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0529198992071202e-07,
+      "loss": 0.4697,
+      "step": 1749
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0368837181081902e-07,
+      "loss": 0.5113,
+      "step": 1750
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.020907755104698e-07,
+      "loss": 0.4334,
+      "step": 1751
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 2.0049920520917394e-07,
+      "loss": 0.4544,
+      "step": 1752
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9891366508064003e-07,
+      "loss": 0.3962,
+      "step": 1753
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9733415928276157e-07,
+      "loss": 0.4291,
+      "step": 1754
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9576069195760883e-07,
+      "loss": 0.388,
+      "step": 1755
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9419326723141534e-07,
+      "loss": 0.4841,
+      "step": 1756
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.926318892145712e-07,
+      "loss": 0.3906,
+      "step": 1757
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.9107656200160752e-07,
+      "loss": 0.3973,
+      "step": 1758
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8952728967118805e-07,
+      "loss": 0.4184,
+      "step": 1759
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.879840762860999e-07,
+      "loss": 0.4955,
+      "step": 1760
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.864469258932397e-07,
+      "loss": 0.469,
+      "step": 1761
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.84915842523605e-07,
+      "loss": 0.474,
+      "step": 1762
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8339083019228405e-07,
+      "loss": 0.4643,
+      "step": 1763
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.8187189289844365e-07,
+      "loss": 0.429,
+      "step": 1764
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.803590346253195e-07,
+      "loss": 0.4823,
+      "step": 1765
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.788522593402059e-07,
+      "loss": 0.4334,
+      "step": 1766
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7735157099444594e-07,
+      "loss": 0.4092,
+      "step": 1767
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7585697352341974e-07,
+      "loss": 0.4562,
+      "step": 1768
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7436847084653458e-07,
+      "loss": 0.4879,
+      "step": 1769
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7288606686721582e-07,
+      "loss": 0.6055,
+      "step": 1770
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7140976547289438e-07,
+      "loss": 0.4479,
+      "step": 1771
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.6993957053499932e-07,
+      "loss": 0.4629,
+      "step": 1772
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6847548590894435e-07,
+      "loss": 0.4924,
+      "step": 1773
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6701751543412192e-07,
+      "loss": 0.4842,
+      "step": 1774
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6556566293388893e-07,
+      "loss": 0.5075,
+      "step": 1775
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6411993221555928e-07,
+      "loss": 0.5157,
+      "step": 1776
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6268032707039362e-07,
+      "loss": 0.5047,
+      "step": 1777
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.6124685127358873e-07,
+      "loss": 0.4941,
+      "step": 1778
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5981950858426715e-07,
+      "loss": 0.4716,
+      "step": 1779
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5839830274546885e-07,
+      "loss": 0.4319,
+      "step": 1780
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5698323748414123e-07,
+      "loss": 0.4837,
+      "step": 1781
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5557431651112705e-07,
+      "loss": 0.5027,
+      "step": 1782
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5417154352115742e-07,
+      "loss": 0.4557,
+      "step": 1783
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.527749221928415e-07,
+      "loss": 0.5527,
+      "step": 1784
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.5138445618865543e-07,
+      "loss": 0.4749,
+      "step": 1785
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.5000014915493467e-07,
+      "loss": 0.4007,
+      "step": 1786
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.48622004721862e-07,
+      "loss": 0.4084,
+      "step": 1787
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.472500265034621e-07,
+      "loss": 0.5062,
+      "step": 1788
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.458842180975864e-07,
+      "loss": 0.4239,
+      "step": 1789
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.4452458308590934e-07,
+      "loss": 0.4942,
+      "step": 1790
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.4317112503391433e-07,
+      "loss": 0.456,
+      "step": 1791
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.418238474908881e-07,
+      "loss": 0.451,
+      "step": 1792
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.4048275398990896e-07,
+      "loss": 0.4592,
+      "step": 1793
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.3914784804783783e-07,
+      "loss": 0.5063,
+      "step": 1794
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.378191331653095e-07,
+      "loss": 0.4276,
+      "step": 1795
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.3649661282672478e-07,
+      "loss": 0.4764,
+      "step": 1796
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.3518029050023862e-07,
+      "loss": 0.5557,
+      "step": 1797
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.3387016963775284e-07,
+      "loss": 0.4417,
+      "step": 1798
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.32566253674907e-07,
+      "loss": 0.3875,
+      "step": 1799
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.3126854603106838e-07,
+      "loss": 0.4576,
+      "step": 1800
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.2997705010932394e-07,
+      "loss": 0.5061,
+      "step": 1801
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.286917692964712e-07,
+      "loss": 0.4698,
+      "step": 1802
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.2741270696300962e-07,
+      "loss": 0.5052,
+      "step": 1803
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.261398664631308e-07,
+      "loss": 0.5296,
+      "step": 1804
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.2487325113471034e-07,
+      "loss": 0.4879,
+      "step": 1805
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.2361286429929953e-07,
+      "loss": 0.4908,
+      "step": 1806
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.223587092621162e-07,
+      "loss": 0.4461,
+      "step": 1807
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.211107893120353e-07,
+      "loss": 0.4612,
+      "step": 1808
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1986910772158106e-07,
+      "loss": 0.4199,
+      "step": 1809
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1863366774691965e-07,
+      "loss": 0.514,
+      "step": 1810
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1740447262784782e-07,
+      "loss": 0.466,
+      "step": 1811
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.1618152558778628e-07,
+      "loss": 0.4351,
+      "step": 1812
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1496482983377189e-07,
+      "loss": 0.4838,
+      "step": 1813
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1375438855644655e-07,
+      "loss": 0.4312,
+      "step": 1814
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.125502049300517e-07,
+      "loss": 0.4659,
+      "step": 1815
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1135228211241827e-07,
+      "loss": 0.4421,
+      "step": 1816
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.1016062324496007e-07,
+      "loss": 0.4452,
+      "step": 1817
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.0897523145266264e-07,
+      "loss": 0.4307,
+      "step": 1818
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0779610984407773e-07,
+      "loss": 0.3859,
+      "step": 1819
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0662326151131525e-07,
+      "loss": 0.4389,
+      "step": 1820
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.054566895300324e-07,
+      "loss": 0.4846,
+      "step": 1821
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.042963969594285e-07,
+      "loss": 0.5028,
+      "step": 1822
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0314238684223515e-07,
+      "loss": 0.415,
+      "step": 1823
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0199466220470994e-07,
+      "loss": 0.4629,
+      "step": 1824
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.0085322605662668e-07,
+      "loss": 0.441,
+      "step": 1825
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.97180813912682e-08,
+      "loss": 0.4204,
+      "step": 1826
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.858923118542003e-08,
+      "loss": 0.53,
+      "step": 1827
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.746667839935942e-08,
+      "loss": 0.5464,
+      "step": 1828
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.635042597685024e-08,
+      "loss": 0.4922,
+      "step": 1829
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.524047684513454e-08,
+      "loss": 0.5165,
+      "step": 1830
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.413683391492456e-08,
+      "loss": 0.4368,
+      "step": 1831
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 9.30395000803952e-08,
+      "loss": 0.4782,
+      "step": 1832
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.194847821917624e-08,
+      "loss": 0.5486,
+      "step": 1833
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 9.086377119234574e-08,
+      "loss": 0.5369,
+      "step": 1834
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.978538184442137e-08,
+      "loss": 0.53,
+      "step": 1835
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.871331300335322e-08,
+      "loss": 0.4626,
+      "step": 1836
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.764756748051661e-08,
+      "loss": 0.381,
+      "step": 1837
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 8.658814807070564e-08,
+      "loss": 0.529,
+      "step": 1838
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.553505755212383e-08,
+      "loss": 0.5135,
+      "step": 1839
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.44882986863782e-08,
+      "loss": 0.5368,
+      "step": 1840
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.344787421847216e-08,
+      "loss": 0.488,
+      "step": 1841
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.241378687679741e-08,
+      "loss": 0.4704,
+      "step": 1842
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.138603937312722e-08,
+      "loss": 0.4313,
+      "step": 1843
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 8.036463440260961e-08,
+      "loss": 0.4611,
+      "step": 1844
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 7.934957464376059e-08,
+      "loss": 0.4247,
+      "step": 1845
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.834086275845587e-08,
+      "loss": 0.394,
+      "step": 1846
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.733850139192395e-08,
+      "loss": 0.5262,
+      "step": 1847
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.63424931727419e-08,
+      "loss": 0.5004,
+      "step": 1848
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.535284071282456e-08,
+      "loss": 0.4194,
+      "step": 1849
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.436954660741957e-08,
+      "loss": 0.486,
+      "step": 1850
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.339261343510207e-08,
+      "loss": 0.4799,
+      "step": 1851
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 7.242204375776496e-08,
+      "loss": 0.4785,
+      "step": 1852
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 7.145784012061424e-08,
+      "loss": 0.4359,
+      "step": 1853
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 7.050000505216147e-08,
+      "loss": 0.4175,
+      "step": 1854
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.954854106421715e-08,
+      "loss": 0.4403,
+      "step": 1855
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.860345065188512e-08,
+      "loss": 0.4604,
+      "step": 1856
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.766473629355453e-08,
+      "loss": 0.495,
+      "step": 1857
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 6.67324004508943e-08,
+      "loss": 0.3993,
+      "step": 1858
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.580644556884703e-08,
+      "loss": 0.4357,
+      "step": 1859
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.488687407562089e-08,
+      "loss": 0.3662,
+      "step": 1860
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.397368838268497e-08,
+      "loss": 0.4428,
+      "step": 1861
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.306689088476231e-08,
+      "loss": 0.4796,
+      "step": 1862
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.216648395982377e-08,
+      "loss": 0.4271,
+      "step": 1863
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.12724699690817e-08,
+      "loss": 0.3873,
+      "step": 1864
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 6.038485125698296e-08,
+      "loss": 0.478,
+      "step": 1865
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.9503630151205025e-08,
+      "loss": 0.3726,
+      "step": 1866
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.862880896264689e-08,
+      "loss": 0.4879,
+      "step": 1867
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.7760389985425105e-08,
+      "loss": 0.463,
+      "step": 1868
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.6898375496867444e-08,
+      "loss": 0.4491,
+      "step": 1869
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.60427677575065e-08,
+      "loss": 0.5362,
+      "step": 1870
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.519356901107359e-08,
+      "loss": 0.416,
+      "step": 1871
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 5.4350781484493185e-08,
+      "loss": 0.3947,
+      "step": 1872
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.3514407387877946e-08,
+      "loss": 0.4217,
+      "step": 1873
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.268444891452118e-08,
+      "loss": 0.5164,
+      "step": 1874
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.186090824089218e-08,
+      "loss": 0.4489,
+      "step": 1875
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.104378752663008e-08,
+      "loss": 0.473,
+      "step": 1876
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 5.023308891453915e-08,
+      "loss": 0.3804,
+      "step": 1877
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.942881453058185e-08,
+      "loss": 0.489,
+      "step": 1878
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 4.8630966483873834e-08,
+      "loss": 0.4997,
+      "step": 1879
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.783954686667897e-08,
+      "loss": 0.3987,
+      "step": 1880
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.705455775440237e-08,
+      "loss": 0.4184,
+      "step": 1881
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.627600120558712e-08,
+      "loss": 0.4361,
+      "step": 1882
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.5503879261906134e-08,
+      "loss": 0.4311,
+      "step": 1883
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.473819394816004e-08,
+      "loss": 0.4554,
+      "step": 1884
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 4.397894727226931e-08,
+      "loss": 0.4761,
+      "step": 1885
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.3226141225268804e-08,
+      "loss": 0.4911,
+      "step": 1886
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.247977778130602e-08,
+      "loss": 0.4621,
+      "step": 1887
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.173985889763116e-08,
+      "loss": 0.4578,
+      "step": 1888
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.100638651459543e-08,
+      "loss": 0.384,
+      "step": 1889
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 4.027936255564468e-08,
+      "loss": 0.4359,
+      "step": 1890
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.955878892731441e-08,
+      "loss": 0.4653,
+      "step": 1891
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.884466751922528e-08,
+      "loss": 0.521,
+      "step": 1892
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.813700020407707e-08,
+      "loss": 0.4942,
+      "step": 1893
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.743578883764504e-08,
+      "loss": 0.5376,
+      "step": 1894
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.67410352587741e-08,
+      "loss": 0.4394,
+      "step": 1895
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.605274128937464e-08,
+      "loss": 0.4877,
+      "step": 1896
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.537090873441701e-08,
+      "loss": 0.4405,
+      "step": 1897
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.469553938192816e-08,
+      "loss": 0.5071,
+      "step": 1898
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.4026635002984974e-08,
+      "loss": 0.3802,
+      "step": 1899
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.336419735171154e-08,
+      "loss": 0.432,
+      "step": 1900
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.270822816527325e-08,
+      "loss": 0.4542,
+      "step": 1901
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.205872916387298e-08,
+      "loss": 0.4438,
+      "step": 1902
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.141570205074607e-08,
+      "loss": 0.4798,
+      "step": 1903
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.077914851215585e-08,
+      "loss": 0.46,
+      "step": 1904
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.014907021739011e-08,
+      "loss": 0.4533,
+      "step": 1905
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.9525468818755455e-08,
+      "loss": 0.4092,
+      "step": 1906
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.8908345951574045e-08,
+      "loss": 0.4099,
+      "step": 1907
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.8297703234178852e-08,
+      "loss": 0.4481,
+      "step": 1908
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7693542267908934e-08,
+      "loss": 0.427,
+      "step": 1909
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.7095864637106118e-08,
+      "loss": 0.4482,
+      "step": 1910
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.6504671909109993e-08,
+      "loss": 0.3983,
+      "step": 1911
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 2.591996563425486e-08,
+      "loss": 0.5271,
+      "step": 1912
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.534174734586503e-08,
+      "loss": 0.3952,
+      "step": 1913
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.4770018560250076e-08,
+      "loss": 0.3829,
+      "step": 1914
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.4204780776702075e-08,
+      "loss": 0.4242,
+      "step": 1915
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.3646035477491726e-08,
+      "loss": 0.4314,
+      "step": 1916
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.309378412786306e-08,
+      "loss": 0.4524,
+      "step": 1917
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.2548028176030688e-08,
+      "loss": 0.4234,
+      "step": 1918
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 2.200876905317645e-08,
+      "loss": 0.4744,
+      "step": 1919
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.147600817344414e-08,
+      "loss": 0.4194,
+      "step": 1920
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.094974693393731e-08,
+      "loss": 0.461,
+      "step": 1921
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 2.0429986714714245e-08,
+      "loss": 0.3801,
+      "step": 1922
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.9916728878786318e-08,
+      "loss": 0.4418,
+      "step": 1923
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.9409974772111872e-08,
+      "loss": 0.5143,
+      "step": 1924
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.890972572359456e-08,
+      "loss": 0.4935,
+      "step": 1925
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.841598304507891e-08,
+      "loss": 0.4361,
+      "step": 1926
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.792874803134781e-08,
+      "loss": 0.425,
+      "step": 1927
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.7448021960118634e-08,
+      "loss": 0.5806,
+      "step": 1928
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6973806092038525e-08,
+      "loss": 0.4666,
+      "step": 1929
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6506101670684395e-08,
+      "loss": 0.4305,
+      "step": 1930
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6044909922555973e-08,
+      "loss": 0.4814,
+      "step": 1931
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.5590232057074718e-08,
+      "loss": 0.4068,
+      "step": 1932
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.5142069266580462e-08,
+      "loss": 0.487,
+      "step": 1933
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.4700422726328378e-08,
+      "loss": 0.4531,
+      "step": 1934
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.4265293594484254e-08,
+      "loss": 0.4999,
+      "step": 1935
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.383668301212393e-08,
+      "loss": 0.4617,
+      "step": 1936
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.3414592103228597e-08,
+      "loss": 0.4821,
+      "step": 1937
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.2999021974682835e-08,
+      "loss": 0.4842,
+      "step": 1938
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.2589973716270188e-08,
+      "loss": 0.4488,
+      "step": 1939
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.218744840067232e-08,
+      "loss": 0.4407,
+      "step": 1940
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.1791447083465136e-08,
+      "loss": 0.5049,
+      "step": 1941
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.1401970803115725e-08,
+      "loss": 0.4363,
+      "step": 1942
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.1019020580980144e-08,
+      "loss": 0.4832,
+      "step": 1943
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.0642597421300915e-08,
+      "loss": 0.4957,
+      "step": 1944
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.0272702311203698e-08,
+      "loss": 0.4369,
+      "step": 1945
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.90933622069562e-09,
+      "loss": 0.4358,
+      "step": 1946
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.5525001026614e-09,
+      "loss": 0.536,
+      "step": 1947
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 9.20219489286278e-09,
+      "loss": 0.4679,
+      "step": 1948
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.858421509933823e-09,
+      "loss": 0.4478,
+      "step": 1949
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.52118085538034e-09,
+      "loss": 0.4516,
+      "step": 1950
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 8.190473813576571e-09,
+      "loss": 0.4522,
+      "step": 1951
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 7.866301251763241e-09,
+      "loss": 0.5201,
+      "step": 1952
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.54866402004506e-09,
+      "loss": 0.4544,
+      "step": 1953
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 7.237562951388499e-09,
+      "loss": 0.4873,
+      "step": 1954
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.93299886162041e-09,
+      "loss": 0.416,
+      "step": 1955
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.634972549423857e-09,
+      "loss": 0.4464,
+      "step": 1956
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.343484796338395e-09,
+      "loss": 0.3759,
+      "step": 1957
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 6.0585363667570175e-09,
+      "loss": 0.4408,
+      "step": 1958
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 5.78012800792338e-09,
+      "loss": 0.4414,
+      "step": 1959
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 5.508260449930969e-09,
+      "loss": 0.4462,
+      "step": 1960
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 5.242934405720879e-09,
+      "loss": 0.4841,
+      "step": 1961
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.984150571079594e-09,
+      "loss": 0.4402,
+      "step": 1962
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.7319096246378756e-09,
+      "loss": 0.5433,
+      "step": 1963
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.486212227867714e-09,
+      "loss": 0.4766,
+      "step": 1964
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 4.247059025082323e-09,
+      "loss": 0.4264,
+      "step": 1965
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 4.01445064343281e-09,
+      "loss": 0.4197,
+      "step": 1966
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.7883876929076245e-09,
+      "loss": 0.4925,
+      "step": 1967
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.5688707663303346e-09,
+      "loss": 0.5531,
+      "step": 1968
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.355900439359072e-09,
+      "loss": 0.5361,
+      "step": 1969
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.1494772704829256e-09,
+      "loss": 0.4193,
+      "step": 1970
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.9496018010233275e-09,
+      "loss": 0.5082,
+      "step": 1971
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 2.7562745551296123e-09,
+      "loss": 0.4359,
+      "step": 1972
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.5694960397806834e-09,
+      "loss": 0.4684,
+      "step": 1973
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.3892667447808495e-09,
+      "loss": 0.4809,
+      "step": 1974
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.215587142760933e-09,
+      "loss": 0.38,
+      "step": 1975
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 2.048457689174943e-09,
+      "loss": 0.4667,
+      "step": 1976
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.887878822300904e-09,
+      "loss": 0.4969,
+      "step": 1977
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.7338509632375267e-09,
+      "loss": 0.458,
+      "step": 1978
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.5863745159055977e-09,
+      "loss": 0.4676,
+      "step": 1979
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.4454498670443683e-09,
+      "loss": 0.4732,
+      "step": 1980
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.3110773862126669e-09,
+      "loss": 0.4644,
+      "step": 1981
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.183257425786677e-09,
+      "loss": 0.4624,
+      "step": 1982
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.0619903209588278e-09,
+      "loss": 0.4494,
+      "step": 1983
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 9.472763897386272e-10,
+      "loss": 0.4832,
+      "step": 1984
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 8.391159329496079e-10,
+      "loss": 0.3719,
+      "step": 1985
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 7.375092342298828e-10,
+      "loss": 0.4567,
+      "step": 1986
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 6.424565600315902e-10,
+      "loss": 0.4153,
+      "step": 1987
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 5.539581596186727e-10,
+      "loss": 0.4565,
+      "step": 1988
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 4.720142650685433e-10,
+      "loss": 0.4211,
+      "step": 1989
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.966250912690317e-10,
+      "loss": 0.4943,
+      "step": 1990
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.277908359194948e-10,
+      "loss": 0.428,
+      "step": 1991
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 2.655116795302615e-10,
+      "loss": 0.4849,
+      "step": 1992
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.0978778542041223e-10,
+      "loss": 0.463,
+      "step": 1993
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.6061929971972202e-10,
+      "loss": 0.4291,
+      "step": 1994
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.180063513669949e-10,
+      "loss": 0.4876,
+      "step": 1995
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 8.194905210923143e-11,
+      "loss": 0.4476,
+      "step": 1996
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 5.2447496503016395e-11,
+      "loss": 0.4585,
+      "step": 1997
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 2.950176191257592e-11,
+      "loss": 0.4161,
+      "step": 1998
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.3111908510332616e-11,
+      "loss": 0.4831,
+      "step": 1999
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.277979276627985e-12,
+      "loss": 0.4811,
+      "step": 2000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 0.0,
+      "loss": 0.4922,
+      "step": 2001
+    },
+    {
+      "epoch": 3.0,
+      "step": 2001,
+      "total_flos": 2.797426555804975e+18,
+      "train_loss": 0.5236783300203898,
+      "train_runtime": 21192.766,
+      "train_samples_per_second": 12.09,
+      "train_steps_per_second": 0.094
+    }
+  ],
+  "logging_steps": 1.0,
+  "max_steps": 2001,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 10000,
+  "total_flos": 2.797426555804975e+18,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}