diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,23485 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "global_step": 3910,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 1.6949152542372883e-07,
+      "loss": 2.5469,
+      "step": 1
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 3.3898305084745766e-07,
+      "loss": 2.29,
+      "step": 2
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 5.084745762711865e-07,
+      "loss": 2.4199,
+      "step": 3
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 6.779661016949153e-07,
+      "loss": 2.2871,
+      "step": 4
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 8.474576271186441e-07,
+      "loss": 2.1475,
+      "step": 5
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.016949152542373e-06,
+      "loss": 2.0771,
+      "step": 6
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.186440677966102e-06,
+      "loss": 2.584,
+      "step": 7
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.3559322033898307e-06,
+      "loss": 2.2402,
+      "step": 8
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 1.5254237288135596e-06,
+      "loss": 2.1982,
+      "step": 9
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.6949152542372882e-06,
+      "loss": 2.5576,
+      "step": 10
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 1.8644067796610171e-06,
+      "loss": 1.9814,
+      "step": 11
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.033898305084746e-06,
+      "loss": 2.2354,
+      "step": 12
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 2.203389830508475e-06,
+      "loss": 2.3203,
+      "step": 13
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.372881355932204e-06,
+      "loss": 2.0732,
+      "step": 14
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.5423728813559323e-06,
+      "loss": 2.1055,
+      "step": 15
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.7118644067796613e-06,
+      "loss": 2.1602,
+      "step": 16
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 2.8813559322033903e-06,
+      "loss": 1.8057,
+      "step": 17
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.0508474576271192e-06,
+      "loss": 1.7764,
+      "step": 18
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.2203389830508473e-06,
+      "loss": 1.7314,
+      "step": 19
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.3898305084745763e-06,
+      "loss": 1.833,
+      "step": 20
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 3.5593220338983053e-06,
+      "loss": 2.0586,
+      "step": 21
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.7288135593220342e-06,
+      "loss": 1.7988,
+      "step": 22
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 3.898305084745763e-06,
+      "loss": 1.8496,
+      "step": 23
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.067796610169492e-06,
+      "loss": 1.8389,
+      "step": 24
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.23728813559322e-06,
+      "loss": 1.6221,
+      "step": 25
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.40677966101695e-06,
+      "loss": 1.6924,
+      "step": 26
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.576271186440678e-06,
+      "loss": 1.6602,
+      "step": 27
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.745762711864408e-06,
+      "loss": 1.5,
+      "step": 28
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.915254237288136e-06,
+      "loss": 1.5273,
+      "step": 29
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.084745762711865e-06,
+      "loss": 1.6875,
+      "step": 30
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.254237288135594e-06,
+      "loss": 1.541,
+      "step": 31
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.423728813559323e-06,
+      "loss": 1.5605,
+      "step": 32
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 5.593220338983051e-06,
+      "loss": 1.4658,
+      "step": 33
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.7627118644067805e-06,
+      "loss": 1.7441,
+      "step": 34
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 5.932203389830509e-06,
+      "loss": 1.5146,
+      "step": 35
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 6.1016949152542385e-06,
+      "loss": 1.5273,
+      "step": 36
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 6.271186440677966e-06,
+      "loss": 1.4209,
+      "step": 37
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 6.440677966101695e-06,
+      "loss": 1.5928,
+      "step": 38
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 6.610169491525424e-06,
+      "loss": 1.5,
+      "step": 39
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 6.779661016949153e-06,
+      "loss": 1.5,
+      "step": 40
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 6.949152542372882e-06,
+      "loss": 1.457,
+      "step": 41
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 7.1186440677966106e-06,
+      "loss": 1.4707,
+      "step": 42
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 7.288135593220339e-06,
+      "loss": 1.5195,
+      "step": 43
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 7.4576271186440685e-06,
+      "loss": 1.4629,
+      "step": 44
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 7.627118644067797e-06,
+      "loss": 1.5391,
+      "step": 45
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 7.796610169491526e-06,
+      "loss": 1.4209,
+      "step": 46
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 7.966101694915255e-06,
+      "loss": 1.4844,
+      "step": 47
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 8.135593220338983e-06,
+      "loss": 1.418,
+      "step": 48
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 8.305084745762712e-06,
+      "loss": 1.6787,
+      "step": 49
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 8.47457627118644e-06,
+      "loss": 1.457,
+      "step": 50
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 8.64406779661017e-06,
+      "loss": 1.3203,
+      "step": 51
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 8.8135593220339e-06,
+      "loss": 1.4463,
+      "step": 52
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 8.983050847457628e-06,
+      "loss": 1.3652,
+      "step": 53
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.152542372881356e-06,
+      "loss": 1.5049,
+      "step": 54
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.322033898305085e-06,
+      "loss": 1.2676,
+      "step": 55
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 9.491525423728815e-06,
+      "loss": 1.6191,
+      "step": 56
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.661016949152544e-06,
+      "loss": 1.4629,
+      "step": 57
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 9.830508474576272e-06,
+      "loss": 1.5049,
+      "step": 58
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1e-05,
+      "loss": 1.4033,
+      "step": 59
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 1.016949152542373e-05,
+      "loss": 1.3916,
+      "step": 60
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.0338983050847458e-05,
+      "loss": 1.3955,
+      "step": 61
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.0508474576271188e-05,
+      "loss": 1.3286,
+      "step": 62
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.0677966101694917e-05,
+      "loss": 1.4629,
+      "step": 63
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.0847457627118645e-05,
+      "loss": 1.4131,
+      "step": 64
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.1016949152542374e-05,
+      "loss": 1.2529,
+      "step": 65
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.1186440677966102e-05,
+      "loss": 1.4209,
+      "step": 66
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.1355932203389833e-05,
+      "loss": 1.499,
+      "step": 67
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 1.1525423728813561e-05,
+      "loss": 1.5303,
+      "step": 68
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.169491525423729e-05,
+      "loss": 1.3662,
+      "step": 69
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.1864406779661018e-05,
+      "loss": 1.6152,
+      "step": 70
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.2033898305084745e-05,
+      "loss": 1.375,
+      "step": 71
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 1.2203389830508477e-05,
+      "loss": 1.4893,
+      "step": 72
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.2372881355932205e-05,
+      "loss": 1.248,
+      "step": 73
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.2542372881355932e-05,
+      "loss": 1.5488,
+      "step": 74
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.2711864406779661e-05,
+      "loss": 1.2793,
+      "step": 75
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 1.288135593220339e-05,
+      "loss": 1.3789,
+      "step": 76
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.305084745762712e-05,
+      "loss": 1.46,
+      "step": 77
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.3220338983050848e-05,
+      "loss": 1.291,
+      "step": 78
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.3389830508474577e-05,
+      "loss": 1.417,
+      "step": 79
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.3559322033898305e-05,
+      "loss": 1.4971,
+      "step": 80
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.3728813559322034e-05,
+      "loss": 1.3789,
+      "step": 81
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.3898305084745764e-05,
+      "loss": 1.4268,
+      "step": 82
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.4067796610169493e-05,
+      "loss": 1.4111,
+      "step": 83
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 1.4237288135593221e-05,
+      "loss": 1.4473,
+      "step": 84
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.440677966101695e-05,
+      "loss": 1.415,
+      "step": 85
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.4576271186440678e-05,
+      "loss": 1.459,
+      "step": 86
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 1.4745762711864408e-05,
+      "loss": 1.4336,
+      "step": 87
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.4915254237288137e-05,
+      "loss": 1.3809,
+      "step": 88
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5084745762711865e-05,
+      "loss": 1.4229,
+      "step": 89
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5254237288135594e-05,
+      "loss": 1.4043,
+      "step": 90
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 1.5423728813559326e-05,
+      "loss": 1.4277,
+      "step": 91
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.5593220338983053e-05,
+      "loss": 1.4473,
+      "step": 92
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.576271186440678e-05,
+      "loss": 1.5117,
+      "step": 93
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.593220338983051e-05,
+      "loss": 1.415,
+      "step": 94
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 1.6101694915254237e-05,
+      "loss": 1.2861,
+      "step": 95
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.6271186440677967e-05,
+      "loss": 1.3975,
+      "step": 96
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.6440677966101697e-05,
+      "loss": 1.5508,
+      "step": 97
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.6610169491525424e-05,
+      "loss": 1.4316,
+      "step": 98
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 1.6779661016949154e-05,
+      "loss": 1.4062,
+      "step": 99
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.694915254237288e-05,
+      "loss": 1.374,
+      "step": 100
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.711864406779661e-05,
+      "loss": 1.333,
+      "step": 101
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.728813559322034e-05,
+      "loss": 1.4209,
+      "step": 102
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 1.745762711864407e-05,
+      "loss": 1.3672,
+      "step": 103
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.76271186440678e-05,
+      "loss": 1.4258,
+      "step": 104
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7796610169491526e-05,
+      "loss": 1.3379,
+      "step": 105
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.7966101694915256e-05,
+      "loss": 1.3867,
+      "step": 106
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 1.8135593220338986e-05,
+      "loss": 1.2861,
+      "step": 107
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8305084745762713e-05,
+      "loss": 1.5645,
+      "step": 108
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.8474576271186443e-05,
+      "loss": 1.4102,
+      "step": 109
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.864406779661017e-05,
+      "loss": 1.4199,
+      "step": 110
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 1.88135593220339e-05,
+      "loss": 1.4434,
+      "step": 111
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.898305084745763e-05,
+      "loss": 1.3418,
+      "step": 112
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9152542372881357e-05,
+      "loss": 1.3672,
+      "step": 113
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9322033898305087e-05,
+      "loss": 1.3184,
+      "step": 114
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 1.9491525423728814e-05,
+      "loss": 1.3545,
+      "step": 115
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9661016949152545e-05,
+      "loss": 1.373,
+      "step": 116
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9830508474576275e-05,
+      "loss": 1.3516,
+      "step": 117
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 2e-05,
+      "loss": 1.4268,
+      "step": 118
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 1.9999996568112307e-05,
+      "loss": 1.4297,
+      "step": 119
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.999998627245158e-05,
+      "loss": 1.3008,
+      "step": 120
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.999996911302488e-05,
+      "loss": 1.3857,
+      "step": 121
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.9999945089843996e-05,
+      "loss": 1.2881,
+      "step": 122
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 1.999991420292541e-05,
+      "loss": 1.4053,
+      "step": 123
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.999987645229032e-05,
+      "loss": 1.3926,
+      "step": 124
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9999831837964646e-05,
+      "loss": 1.4404,
+      "step": 125
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9999780359979e-05,
+      "loss": 1.2744,
+      "step": 126
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 1.9999722018368724e-05,
+      "loss": 1.4121,
+      "step": 127
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9999656813173856e-05,
+      "loss": 1.2754,
+      "step": 128
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9999584744439155e-05,
+      "loss": 1.2832,
+      "step": 129
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.9999505812214087e-05,
+      "loss": 1.418,
+      "step": 130
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9999420016552828e-05,
+      "loss": 1.3438,
+      "step": 131
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9999327357514268e-05,
+      "loss": 1.333,
+      "step": 132
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9999227835162e-05,
+      "loss": 1.3184,
+      "step": 133
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 1.9999121449564347e-05,
+      "loss": 1.3301,
+      "step": 134
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.9999008200794315e-05,
+      "loss": 1.4434,
+      "step": 135
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.9998888088929647e-05,
+      "loss": 1.3799,
+      "step": 136
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.9998761114052775e-05,
+      "loss": 1.4482,
+      "step": 137
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 1.999862727625086e-05,
+      "loss": 1.2354,
+      "step": 138
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.999848657561576e-05,
+      "loss": 1.4297,
+      "step": 139
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9998339012244052e-05,
+      "loss": 1.291,
+      "step": 140
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9998184586237023e-05,
+      "loss": 1.3398,
+      "step": 141
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 1.9998023297700656e-05,
+      "loss": 1.3193,
+      "step": 142
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9997855146745667e-05,
+      "loss": 1.3076,
+      "step": 143
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9997680133487468e-05,
+      "loss": 1.2568,
+      "step": 144
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9997498258046183e-05,
+      "loss": 1.293,
+      "step": 145
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 1.9997309520546646e-05,
+      "loss": 1.2939,
+      "step": 146
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9997113921118404e-05,
+      "loss": 1.3291,
+      "step": 147
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9996911459895717e-05,
+      "loss": 1.4082,
+      "step": 148
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9996702137017537e-05,
+      "loss": 1.3926,
+      "step": 149
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 1.9996485952627554e-05,
+      "loss": 1.3428,
+      "step": 150
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.999626290687414e-05,
+      "loss": 1.4707,
+      "step": 151
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9996032999910394e-05,
+      "loss": 1.3789,
+      "step": 152
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.999579623189412e-05,
+      "loss": 1.3389,
+      "step": 153
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9995552602987828e-05,
+      "loss": 1.3735,
+      "step": 154
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.9995302113358735e-05,
+      "loss": 1.334,
+      "step": 155
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.999504476317878e-05,
+      "loss": 1.4492,
+      "step": 156
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.9994780552624595e-05,
+      "loss": 1.4736,
+      "step": 157
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 1.9994509481877536e-05,
+      "loss": 1.3711,
+      "step": 158
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9994231551123655e-05,
+      "loss": 1.417,
+      "step": 159
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9993946760553717e-05,
+      "loss": 1.3809,
+      "step": 160
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9993655110363197e-05,
+      "loss": 1.3457,
+      "step": 161
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 1.9993356600752277e-05,
+      "loss": 1.2822,
+      "step": 162
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9993051231925847e-05,
+      "loss": 1.3145,
+      "step": 163
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9992739004093503e-05,
+      "loss": 1.4189,
+      "step": 164
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.9992419917469558e-05,
+      "loss": 1.2695,
+      "step": 165
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 1.999209397227302e-05,
+      "loss": 1.2725,
+      "step": 166
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9991761168727612e-05,
+      "loss": 1.3174,
+      "step": 167
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9991421507061763e-05,
+      "loss": 1.3623,
+      "step": 168
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9991074987508613e-05,
+      "loss": 1.3584,
+      "step": 169
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 1.9990721610305998e-05,
+      "loss": 1.4746,
+      "step": 170
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9990361375696467e-05,
+      "loss": 1.3633,
+      "step": 171
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9989994283927287e-05,
+      "loss": 1.3701,
+      "step": 172
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 1.9989620335250413e-05,
+      "loss": 1.2725,
+      "step": 173
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9989239529922523e-05,
+      "loss": 1.3359,
+      "step": 174
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9988851868204983e-05,
+      "loss": 1.3789,
+      "step": 175
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9988457350363885e-05,
+      "loss": 1.3818,
+      "step": 176
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 1.9988055976670008e-05,
+      "loss": 1.3135,
+      "step": 177
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9987647747398854e-05,
+      "loss": 1.2754,
+      "step": 178
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9987232662830618e-05,
+      "loss": 1.3457,
+      "step": 179
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9986810723250206e-05,
+      "loss": 1.4043,
+      "step": 180
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 1.9986381928947227e-05,
+      "loss": 1.2832,
+      "step": 181
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.9985946280215996e-05,
+      "loss": 1.3438,
+      "step": 182
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.9985503777355534e-05,
+      "loss": 1.248,
+      "step": 183
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.9985054420669562e-05,
+      "loss": 1.3203,
+      "step": 184
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 1.998459821046651e-05,
+      "loss": 1.3291,
+      "step": 185
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.9984135147059515e-05,
+      "loss": 1.5117,
+      "step": 186
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.9983665230766406e-05,
+      "loss": 1.3145,
+      "step": 187
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.998318846190973e-05,
+      "loss": 1.2529,
+      "step": 188
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 1.998270484081672e-05,
+      "loss": 1.3594,
+      "step": 189
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.998221436781933e-05,
+      "loss": 1.3828,
+      "step": 190
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.998171704325421e-05,
+      "loss": 1.4746,
+      "step": 191
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.9981212867462715e-05,
+      "loss": 1.3984,
+      "step": 192
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.9980701840790893e-05,
+      "loss": 1.3994,
+      "step": 193
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9980183963589505e-05,
+      "loss": 1.4277,
+      "step": 194
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9979659236214005e-05,
+      "loss": 1.4141,
+      "step": 195
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9979127659024563e-05,
+      "loss": 1.3838,
+      "step": 196
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 1.9978589232386036e-05,
+      "loss": 1.2607,
+      "step": 197
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.997804395666799e-05,
+      "loss": 1.2734,
+      "step": 198
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.997749183224469e-05,
+      "loss": 1.2715,
+      "step": 199
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.99769328594951e-05,
+      "loss": 1.3408,
+      "step": 200
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 1.9976367038802888e-05,
+      "loss": 1.3027,
+      "step": 201
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.997579437055642e-05,
+      "loss": 1.3818,
+      "step": 202
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.997521485514876e-05,
+      "loss": 1.332,
+      "step": 203
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.9974628492977678e-05,
+      "loss": 1.3418,
+      "step": 204
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 1.9974035284445642e-05,
+      "loss": 1.2451,
+      "step": 205
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.9973435229959813e-05,
+      "loss": 1.2373,
+      "step": 206
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.997282832993205e-05,
+      "loss": 1.3301,
+      "step": 207
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.997221458477893e-05,
+      "loss": 1.2852,
+      "step": 208
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 1.99715939949217e-05,
+      "loss": 1.2783,
+      "step": 209
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.9970966560786324e-05,
+      "loss": 1.2549,
+      "step": 210
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.997033228280346e-05,
+      "loss": 1.3584,
+      "step": 211
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.996969116140846e-05,
+      "loss": 1.2754,
+      "step": 212
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 1.9969043197041377e-05,
+      "loss": 1.3027,
+      "step": 213
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.996838839014696e-05,
+      "loss": 1.2192,
+      "step": 214
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9967726741174652e-05,
+      "loss": 1.3828,
+      "step": 215
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.9967058250578593e-05,
+      "loss": 1.3965,
+      "step": 216
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 1.996638291881762e-05,
+      "loss": 1.3496,
+      "step": 217
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.996570074635527e-05,
+      "loss": 1.4541,
+      "step": 218
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.9965011733659762e-05,
+      "loss": 1.3525,
+      "step": 219
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 1.996431588120403e-05,
+      "loss": 1.1855,
+      "step": 220
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.9963613189465684e-05,
+      "loss": 1.4053,
+      "step": 221
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.9962903658927038e-05,
+      "loss": 1.417,
+      "step": 222
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.9962187290075097e-05,
+      "loss": 1.2324,
+      "step": 223
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 1.996146408340156e-05,
+      "loss": 1.2539,
+      "step": 224
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.9960734039402824e-05,
+      "loss": 1.4639,
+      "step": 225
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.995999715857997e-05,
+      "loss": 1.2354,
+      "step": 226
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.9959253441438778e-05,
+      "loss": 1.1719,
+      "step": 227
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 1.995850288848972e-05,
+      "loss": 1.248,
+      "step": 228
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9957745500247955e-05,
+      "loss": 1.333,
+      "step": 229
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9956981277233342e-05,
+      "loss": 1.3096,
+      "step": 230
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9956210219970422e-05,
+      "loss": 1.418,
+      "step": 231
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9955432328988437e-05,
+      "loss": 1.3906,
+      "step": 232
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9954647604821307e-05,
+      "loss": 1.4463,
+      "step": 233
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9953856048007654e-05,
+      "loss": 1.2075,
+      "step": 234
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9953057659090783e-05,
+      "loss": 1.334,
+      "step": 235
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 1.9952252438618692e-05,
+      "loss": 1.3945,
+      "step": 236
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9951440387144065e-05,
+      "loss": 1.3164,
+      "step": 237
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9950621505224276e-05,
+      "loss": 1.2842,
+      "step": 238
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9949795793421386e-05,
+      "loss": 1.292,
+      "step": 239
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 1.9948963252302147e-05,
+      "loss": 1.3848,
+      "step": 240
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9948123882437995e-05,
+      "loss": 1.2646,
+      "step": 241
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9947277684405056e-05,
+      "loss": 1.3037,
+      "step": 242
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9946424658784142e-05,
+      "loss": 1.3574,
+      "step": 243
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 1.9945564806160744e-05,
+      "loss": 1.2598,
+      "step": 244
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9944698127125053e-05,
+      "loss": 1.3135,
+      "step": 245
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9943824622271934e-05,
+      "loss": 1.3506,
+      "step": 246
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9942944292200948e-05,
+      "loss": 1.2725,
+      "step": 247
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 1.9942057137516323e-05,
+      "loss": 1.3701,
+      "step": 248
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9941163158826992e-05,
+      "loss": 1.3281,
+      "step": 249
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9940262356746553e-05,
+      "loss": 1.2754,
+      "step": 250
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9939354731893304e-05,
+      "loss": 1.3467,
+      "step": 251
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 1.9938440284890216e-05,
+      "loss": 1.2256,
+      "step": 252
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9937519016364943e-05,
+      "loss": 1.2871,
+      "step": 253
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9936590926949822e-05,
+      "loss": 1.3623,
+      "step": 254
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.9935656017281877e-05,
+      "loss": 1.3379,
+      "step": 255
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.993471428800281e-05,
+      "loss": 1.25,
+      "step": 256
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9933765739758992e-05,
+      "loss": 1.3301,
+      "step": 257
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9932810373201498e-05,
+      "loss": 1.2539,
+      "step": 258
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9931848188986063e-05,
+      "loss": 1.3359,
+      "step": 259
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 1.9930879187773108e-05,
+      "loss": 1.2773,
+      "step": 260
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.9929903370227742e-05,
+      "loss": 1.3994,
+      "step": 261
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.9928920737019735e-05,
+      "loss": 1.2559,
+      "step": 262
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 1.992793128882355e-05,
+      "loss": 1.415,
+      "step": 263
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.992693502631832e-05,
+      "loss": 1.3506,
+      "step": 264
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9925931950187852e-05,
+      "loss": 1.2266,
+      "step": 265
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.9924922061120644e-05,
+      "loss": 1.3359,
+      "step": 266
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 1.992390535980986e-05,
+      "loss": 1.2705,
+      "step": 267
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.9922881846953333e-05,
+      "loss": 1.3135,
+      "step": 268
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.992185152325359e-05,
+      "loss": 1.3174,
+      "step": 269
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.992081438941781e-05,
+      "loss": 1.2939,
+      "step": 270
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 1.9919770446157867e-05,
+      "loss": 1.1699,
+      "step": 271
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.99187196941903e-05,
+      "loss": 1.2158,
+      "step": 272
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9917662134236318e-05,
+      "loss": 1.3467,
+      "step": 273
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.991659776702181e-05,
+      "loss": 1.1895,
+      "step": 274
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 1.9915526593277327e-05,
+      "loss": 1.3691,
+      "step": 275
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9914448613738107e-05,
+      "loss": 1.4355,
+      "step": 276
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9913363829144042e-05,
+      "loss": 1.1318,
+      "step": 277
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9912272240239715e-05,
+      "loss": 1.1865,
+      "step": 278
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 1.9911173847774365e-05,
+      "loss": 1.2686,
+      "step": 279
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9910068652501894e-05,
+      "loss": 1.187,
+      "step": 280
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9908956655180894e-05,
+      "loss": 1.207,
+      "step": 281
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9907837856574608e-05,
+      "loss": 1.2529,
+      "step": 282
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 1.9906712257450958e-05,
+      "loss": 1.2891,
+      "step": 283
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.990557985858253e-05,
+      "loss": 1.2256,
+      "step": 284
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.990444066074658e-05,
+      "loss": 1.3135,
+      "step": 285
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.9903294664725023e-05,
+      "loss": 1.2686,
+      "step": 286
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 1.9902141871304442e-05,
+      "loss": 1.2246,
+      "step": 287
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.9900982281276096e-05,
+      "loss": 1.1279,
+      "step": 288
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.98998158954359e-05,
+      "loss": 1.3076,
+      "step": 289
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.989864271458443e-05,
+      "loss": 1.1777,
+      "step": 290
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 1.9897462739526936e-05,
+      "loss": 1.2891,
+      "step": 291
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.9896275971073326e-05,
+      "loss": 1.3076,
+      "step": 292
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.9895082410038166e-05,
+      "loss": 1.1133,
+      "step": 293
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.9893882057240698e-05,
+      "loss": 1.1523,
+      "step": 294
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 1.989267491350481e-05,
+      "loss": 1.25,
+      "step": 295
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.989146097965906e-05,
+      "loss": 1.3135,
+      "step": 296
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.989024025653667e-05,
+      "loss": 1.1865,
+      "step": 297
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.9889012744975508e-05,
+      "loss": 1.2607,
+      "step": 298
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 1.988777844581812e-05,
+      "loss": 1.2422,
+      "step": 299
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.9886537359911694e-05,
+      "loss": 1.2373,
+      "step": 300
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.9885289488108086e-05,
+      "loss": 1.2148,
+      "step": 301
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.9884034831263807e-05,
+      "loss": 1.2705,
+      "step": 302
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 1.9882773390240027e-05,
+      "loss": 1.2373,
+      "step": 303
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.9881505165902566e-05,
+      "loss": 1.2822,
+      "step": 304
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.988023015912191e-05,
+      "loss": 1.2344,
+      "step": 305
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 1.9878948370773194e-05,
+      "loss": 1.2188,
+      "step": 306
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9877659801736206e-05,
+      "loss": 1.2979,
+      "step": 307
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9876364452895394e-05,
+      "loss": 1.1763,
+      "step": 308
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9875062325139853e-05,
+      "loss": 1.2168,
+      "step": 309
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9873753419363336e-05,
+      "loss": 1.2612,
+      "step": 310
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.987243773646425e-05,
+      "loss": 1.3779,
+      "step": 311
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.987111527734564e-05,
+      "loss": 1.249,
+      "step": 312
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9869786042915225e-05,
+      "loss": 1.2334,
+      "step": 313
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 1.9868450034085352e-05,
+      "loss": 1.2529,
+      "step": 314
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9867107251773033e-05,
+      "loss": 1.1909,
+      "step": 315
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.986575769689992e-05,
+      "loss": 1.251,
+      "step": 316
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.986440137039232e-05,
+      "loss": 1.314,
+      "step": 317
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.9863038273181187e-05,
+      "loss": 1.1963,
+      "step": 318
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.9861668406202116e-05,
+      "loss": 1.2285,
+      "step": 319
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.986029177039535e-05,
+      "loss": 1.1958,
+      "step": 320
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.985890836670579e-05,
+      "loss": 1.168,
+      "step": 321
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 1.9857518196082964e-05,
+      "loss": 1.2529,
+      "step": 322
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.985612125948106e-05,
+      "loss": 1.1768,
+      "step": 323
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9854717557858906e-05,
+      "loss": 1.2207,
+      "step": 324
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9853307092179964e-05,
+      "loss": 1.2266,
+      "step": 325
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 1.9851889863412347e-05,
+      "loss": 1.1689,
+      "step": 326
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9850465872528812e-05,
+      "loss": 1.167,
+      "step": 327
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9849035120506756e-05,
+      "loss": 1.1689,
+      "step": 328
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9847597608328212e-05,
+      "loss": 1.2471,
+      "step": 329
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 1.9846153336979855e-05,
+      "loss": 1.1323,
+      "step": 330
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.984470230745301e-05,
+      "loss": 1.2129,
+      "step": 331
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.984324452074361e-05,
+      "loss": 1.1636,
+      "step": 332
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9841779977852273e-05,
+      "loss": 1.1641,
+      "step": 333
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 1.9840308679784207e-05,
+      "loss": 1.2021,
+      "step": 334
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.983883062754929e-05,
+      "loss": 1.1357,
+      "step": 335
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.983734582216202e-05,
+      "loss": 1.1611,
+      "step": 336
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9835854264641538e-05,
+      "loss": 1.1445,
+      "step": 337
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 1.9834355956011604e-05,
+      "loss": 1.208,
+      "step": 338
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.983285089730064e-05,
+      "loss": 1.2168,
+      "step": 339
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9831339089541673e-05,
+      "loss": 1.2402,
+      "step": 340
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9829820533772376e-05,
+      "loss": 1.2529,
+      "step": 341
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 1.9828295231035054e-05,
+      "loss": 1.1709,
+      "step": 342
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.9826763182376636e-05,
+      "loss": 1.1826,
+      "step": 343
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.9825224388848693e-05,
+      "loss": 1.248,
+      "step": 344
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.982367885150741e-05,
+      "loss": 1.1143,
+      "step": 345
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 1.9822126571413616e-05,
+      "loss": 1.2822,
+      "step": 346
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9820567549632755e-05,
+      "loss": 1.1338,
+      "step": 347
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9819001787234912e-05,
+      "loss": 1.2109,
+      "step": 348
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 1.9817429285294785e-05,
+      "loss": 1.2578,
+      "step": 349
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.981585004489171e-05,
+      "loss": 1.1104,
+      "step": 350
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.9814264067109634e-05,
+      "loss": 1.2686,
+      "step": 351
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.981267135303714e-05,
+      "loss": 1.1963,
+      "step": 352
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 1.9811071903767436e-05,
+      "loss": 1.187,
+      "step": 353
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.980946572039834e-05,
+      "loss": 1.1465,
+      "step": 354
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.9807852804032306e-05,
+      "loss": 1.2319,
+      "step": 355
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.9806233155776403e-05,
+      "loss": 1.1846,
+      "step": 356
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 1.9804606776742318e-05,
+      "loss": 1.2412,
+      "step": 357
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.9802973668046364e-05,
+      "loss": 1.2173,
+      "step": 358
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.9801333830809468e-05,
+      "loss": 1.3389,
+      "step": 359
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.979968726615718e-05,
+      "loss": 1.2119,
+      "step": 360
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 1.979803397521966e-05,
+      "loss": 1.1465,
+      "step": 361
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.97963739591317e-05,
+      "loss": 1.2148,
+      "step": 362
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.9794707219032686e-05,
+      "loss": 1.1543,
+      "step": 363
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.9793033756066635e-05,
+      "loss": 1.1289,
+      "step": 364
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 1.9791353571382178e-05,
+      "loss": 1.2031,
+      "step": 365
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9789666666132553e-05,
+      "loss": 1.1523,
+      "step": 366
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9787973041475616e-05,
+      "loss": 1.2046,
+      "step": 367
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.978627269857383e-05,
+      "loss": 1.2725,
+      "step": 368
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 1.9784565638594274e-05,
+      "loss": 1.1729,
+      "step": 369
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9782851862708633e-05,
+      "loss": 1.168,
+      "step": 370
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9781131372093206e-05,
+      "loss": 1.1992,
+      "step": 371
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9779404167928903e-05,
+      "loss": 1.2402,
+      "step": 372
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 1.9777670251401232e-05,
+      "loss": 1.2759,
+      "step": 373
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.9775929623700318e-05,
+      "loss": 1.2666,
+      "step": 374
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.9774182286020887e-05,
+      "loss": 1.1357,
+      "step": 375
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.9772428239562273e-05,
+      "loss": 1.2715,
+      "step": 376
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 1.9770667485528413e-05,
+      "loss": 1.1768,
+      "step": 377
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.9768900025127852e-05,
+      "loss": 1.2188,
+      "step": 378
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.9767125859573735e-05,
+      "loss": 1.1343,
+      "step": 379
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.9765344990083807e-05,
+      "loss": 1.209,
+      "step": 380
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 1.9763557417880412e-05,
+      "loss": 1.1924,
+      "step": 381
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.976176314419051e-05,
+      "loss": 1.1235,
+      "step": 382
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.9759962170245647e-05,
+      "loss": 1.2129,
+      "step": 383
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.975815449728196e-05,
+      "loss": 1.2485,
+      "step": 384
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.9756340126540216e-05,
+      "loss": 1.2329,
+      "step": 385
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.9754519059265737e-05,
+      "loss": 1.2568,
+      "step": 386
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.9752691296708476e-05,
+      "loss": 1.1226,
+      "step": 387
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.9750856840122968e-05,
+      "loss": 1.2363,
+      "step": 388
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.9749015690768333e-05,
+      "loss": 1.1123,
+      "step": 389
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.9747167849908305e-05,
+      "loss": 1.0693,
+      "step": 390
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.9745313318811195e-05,
+      "loss": 1.1333,
+      "step": 391
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 1.974345209874991e-05,
+      "loss": 1.1899,
+      "step": 392
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.9741584191001956e-05,
+      "loss": 1.2773,
+      "step": 393
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.9739709596849418e-05,
+      "loss": 1.1792,
+      "step": 394
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.9737828317578977e-05,
+      "loss": 1.1299,
+      "step": 395
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 1.97359403544819e-05,
+      "loss": 1.2842,
+      "step": 396
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.9734045708854044e-05,
+      "loss": 1.21,
+      "step": 397
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.9732144381995848e-05,
+      "loss": 1.2197,
+      "step": 398
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.973023637521234e-05,
+      "loss": 1.1606,
+      "step": 399
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 1.9728321689813142e-05,
+      "loss": 1.0957,
+      "step": 400
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.9726400327112437e-05,
+      "loss": 1.166,
+      "step": 401
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.972447228842902e-05,
+      "loss": 1.2246,
+      "step": 402
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.972253757508624e-05,
+      "loss": 1.1748,
+      "step": 403
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 1.972059618841205e-05,
+      "loss": 1.1655,
+      "step": 404
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.971864812973897e-05,
+      "loss": 1.0781,
+      "step": 405
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.97166934004041e-05,
+      "loss": 1.1807,
+      "step": 406
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.9714732001749134e-05,
+      "loss": 1.2007,
+      "step": 407
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 1.971276393512032e-05,
+      "loss": 1.1348,
+      "step": 408
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.97107892018685e-05,
+      "loss": 1.1973,
+      "step": 409
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.970880780334909e-05,
+      "loss": 1.166,
+      "step": 410
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.9706819740922068e-05,
+      "loss": 1.0918,
+      "step": 411
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 1.9704825015952005e-05,
+      "loss": 1.1553,
+      "step": 412
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.970282362980803e-05,
+      "loss": 1.1646,
+      "step": 413
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.9700815583863853e-05,
+      "loss": 1.1182,
+      "step": 414
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.9698800879497747e-05,
+      "loss": 1.2402,
+      "step": 415
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 1.9696779518092564e-05,
+      "loss": 1.0674,
+      "step": 416
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.9694751501035715e-05,
+      "loss": 1.2256,
+      "step": 417
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.9692716829719197e-05,
+      "loss": 1.2197,
+      "step": 418
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.969067550553955e-05,
+      "loss": 1.062,
+      "step": 419
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 1.96886275298979e-05,
+      "loss": 1.2051,
+      "step": 420
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.9686572904199928e-05,
+      "loss": 1.1694,
+      "step": 421
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.968451162985589e-05,
+      "loss": 1.2266,
+      "step": 422
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.968244370828059e-05,
+      "loss": 1.1353,
+      "step": 423
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 1.9680369140893405e-05,
+      "loss": 1.1216,
+      "step": 424
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.9678287929118275e-05,
+      "loss": 1.3252,
+      "step": 425
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.9676200074383692e-05,
+      "loss": 1.2168,
+      "step": 426
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.967410557812272e-05,
+      "loss": 1.1445,
+      "step": 427
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 1.9672004441772966e-05,
+      "loss": 1.1299,
+      "step": 428
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.9669896666776606e-05,
+      "loss": 1.2197,
+      "step": 429
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.9667782254580373e-05,
+      "loss": 1.1362,
+      "step": 430
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.9665661206635548e-05,
+      "loss": 1.1963,
+      "step": 431
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 1.9663533524397968e-05,
+      "loss": 1.1035,
+      "step": 432
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.966139920932803e-05,
+      "loss": 1.1011,
+      "step": 433
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.9659258262890683e-05,
+      "loss": 1.2314,
+      "step": 434
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 1.965711068655542e-05,
+      "loss": 1.1357,
+      "step": 435
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.965495648179629e-05,
+      "loss": 1.0732,
+      "step": 436
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.965279565009189e-05,
+      "loss": 1.1353,
+      "step": 437
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.9650628192925372e-05,
+      "loss": 1.1602,
+      "step": 438
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 1.964845411178442e-05,
+      "loss": 1.1748,
+      "step": 439
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.9646273408161286e-05,
+      "loss": 1.2031,
+      "step": 440
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.9644086083552745e-05,
+      "loss": 1.1514,
+      "step": 441
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.9641892139460133e-05,
+      "loss": 1.1528,
+      "step": 442
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 1.9639691577389322e-05,
+      "loss": 1.1621,
+      "step": 443
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.963748439885073e-05,
+      "loss": 1.1797,
+      "step": 444
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.963527060535932e-05,
+      "loss": 1.0884,
+      "step": 445
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.9633050198434578e-05,
+      "loss": 1.1021,
+      "step": 446
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.9630823179600548e-05,
+      "loss": 1.1587,
+      "step": 447
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.962858955038581e-05,
+      "loss": 1.1089,
+      "step": 448
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.962634931232347e-05,
+      "loss": 1.1592,
+      "step": 449
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.962410246695118e-05,
+      "loss": 1.1816,
+      "step": 450
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 1.9621849015811123e-05,
+      "loss": 1.0522,
+      "step": 451
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.961958896045002e-05,
+      "loss": 1.1167,
+      "step": 452
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.961732230241912e-05,
+      "loss": 1.1787,
+      "step": 453
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.9615049043274207e-05,
+      "loss": 1.2173,
+      "step": 454
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 1.9612769184575592e-05,
+      "loss": 1.1157,
+      "step": 455
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.9610482727888126e-05,
+      "loss": 1.0713,
+      "step": 456
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.960818967478117e-05,
+      "loss": 1.2104,
+      "step": 457
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.9605890026828634e-05,
+      "loss": 1.1758,
+      "step": 458
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 1.9603583785608947e-05,
+      "loss": 1.1533,
+      "step": 459
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.960127095270505e-05,
+      "loss": 1.127,
+      "step": 460
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9598951529704426e-05,
+      "loss": 1.125,
+      "step": 461
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9596625518199077e-05,
+      "loss": 1.2842,
+      "step": 462
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.9594292919785523e-05,
+      "loss": 1.2256,
+      "step": 463
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.9591953736064804e-05,
+      "loss": 1.2109,
+      "step": 464
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.9589607968642483e-05,
+      "loss": 1.1455,
+      "step": 465
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.9587255619128648e-05,
+      "loss": 1.1572,
+      "step": 466
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 1.9584896689137897e-05,
+      "loss": 1.1367,
+      "step": 467
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.9582531180289342e-05,
+      "loss": 1.3022,
+      "step": 468
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.9580159094206622e-05,
+      "loss": 1.2266,
+      "step": 469
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.957778043251788e-05,
+      "loss": 1.0752,
+      "step": 470
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 1.957539519685577e-05,
+      "loss": 1.1255,
+      "step": 471
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.9573003388857476e-05,
+      "loss": 1.1714,
+      "step": 472
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.9570605010164676e-05,
+      "loss": 1.1323,
+      "step": 473
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.9568200062423557e-05,
+      "loss": 1.1641,
+      "step": 474
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 1.956578854728483e-05,
+      "loss": 1.0718,
+      "step": 475
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.95633704664037e-05,
+      "loss": 1.1421,
+      "step": 476
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.9560945821439885e-05,
+      "loss": 1.1792,
+      "step": 477
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 1.955851461405761e-05,
+      "loss": 1.082,
+      "step": 478
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.9556076845925596e-05,
+      "loss": 1.1567,
+      "step": 479
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.9553632518717074e-05,
+      "loss": 1.1099,
+      "step": 480
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.9551181634109774e-05,
+      "loss": 1.1436,
+      "step": 481
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 1.9548724193785933e-05,
+      "loss": 1.1616,
+      "step": 482
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.9546260199432278e-05,
+      "loss": 1.1719,
+      "step": 483
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.9543789652740043e-05,
+      "loss": 1.1138,
+      "step": 484
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.954131255540495e-05,
+      "loss": 1.1963,
+      "step": 485
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 1.953882890912723e-05,
+      "loss": 1.2754,
+      "step": 486
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.9536338715611593e-05,
+      "loss": 1.2432,
+      "step": 487
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.9533841976567265e-05,
+      "loss": 1.0869,
+      "step": 488
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.953133869370794e-05,
+      "loss": 1.1733,
+      "step": 489
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 1.952882886875182e-05,
+      "loss": 1.0376,
+      "step": 490
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.952631250342159e-05,
+      "loss": 1.1035,
+      "step": 491
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.952378959944443e-05,
+      "loss": 1.1709,
+      "step": 492
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.9521260158552004e-05,
+      "loss": 1.1597,
+      "step": 493
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 1.951872418248046e-05,
+      "loss": 1.1421,
+      "step": 494
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.9516181672970444e-05,
+      "loss": 1.1333,
+      "step": 495
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.9513632631767062e-05,
+      "loss": 1.0488,
+      "step": 496
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.9511077060619933e-05,
+      "loss": 1.1392,
+      "step": 497
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 1.950851496128314e-05,
+      "loss": 1.1479,
+      "step": 498
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.950594633551524e-05,
+      "loss": 1.1504,
+      "step": 499
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.9503371185079295e-05,
+      "loss": 1.0332,
+      "step": 500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.9500789511742824e-05,
+      "loss": 1.1562,
+      "step": 501
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 1.949820131727783e-05,
+      "loss": 1.1318,
+      "step": 502
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.949560660346079e-05,
+      "loss": 1.0327,
+      "step": 503
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.9493005372072654e-05,
+      "loss": 1.1182,
+      "step": 504
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.949039762489886e-05,
+      "loss": 1.1035,
+      "step": 505
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 1.9487783363729297e-05,
+      "loss": 1.1934,
+      "step": 506
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.948516259035834e-05,
+      "loss": 1.229,
+      "step": 507
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.9482535306584826e-05,
+      "loss": 1.1621,
+      "step": 508
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.9479901514212068e-05,
+      "loss": 1.167,
+      "step": 509
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.9477261215047836e-05,
+      "loss": 1.1641,
+      "step": 510
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.9474614410904373e-05,
+      "loss": 1.1357,
+      "step": 511
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.947196110359839e-05,
+      "loss": 1.0962,
+      "step": 512
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.946930129495106e-05,
+      "loss": 1.1108,
+      "step": 513
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 1.9466634986788007e-05,
+      "loss": 1.1045,
+      "step": 514
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.9463962180939327e-05,
+      "loss": 1.0405,
+      "step": 515
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.946128287923958e-05,
+      "loss": 1.1699,
+      "step": 516
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.9458597083527774e-05,
+      "loss": 1.2217,
+      "step": 517
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 1.945590479564738e-05,
+      "loss": 1.1147,
+      "step": 518
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.9453206017446326e-05,
+      "loss": 1.0801,
+      "step": 519
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.9450500750776987e-05,
+      "loss": 1.1147,
+      "step": 520
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 1.94477889974962e-05,
+      "loss": 1.1758,
+      "step": 521
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.9445070759465254e-05,
+      "loss": 1.085,
+      "step": 522
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.9442346038549883e-05,
+      "loss": 1.1099,
+      "step": 523
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.9439614836620277e-05,
+      "loss": 1.1748,
+      "step": 524
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 1.943687715555107e-05,
+      "loss": 1.0947,
+      "step": 525
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.9434132997221347e-05,
+      "loss": 1.1606,
+      "step": 526
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.9431382363514634e-05,
+      "loss": 1.0762,
+      "step": 527
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.9428625256318906e-05,
+      "loss": 1.1494,
+      "step": 528
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 1.9425861677526577e-05,
+      "loss": 1.1792,
+      "step": 529
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.9423091629034506e-05,
+      "loss": 1.1011,
+      "step": 530
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.9420315112743997e-05,
+      "loss": 1.1758,
+      "step": 531
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.9417532130560784e-05,
+      "loss": 1.0859,
+      "step": 532
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 1.941474268439504e-05,
+      "loss": 1.0732,
+      "step": 533
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.9411946776161388e-05,
+      "loss": 1.0454,
+      "step": 534
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.9409144407778866e-05,
+      "loss": 1.1572,
+      "step": 535
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.9406335581170962e-05,
+      "loss": 1.0654,
+      "step": 536
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 1.9403520298265596e-05,
+      "loss": 1.0166,
+      "step": 537
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.9400698560995105e-05,
+      "loss": 1.0864,
+      "step": 538
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.9397870371296273e-05,
+      "loss": 1.1992,
+      "step": 539
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.93950357311103e-05,
+      "loss": 1.1455,
+      "step": 540
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.939219464238283e-05,
+      "loss": 1.106,
+      "step": 541
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.938934710706391e-05,
+      "loss": 1.1953,
+      "step": 542
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.9386493127108033e-05,
+      "loss": 1.0576,
+      "step": 543
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.9383632704474103e-05,
+      "loss": 1.0215,
+      "step": 544
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 1.938076584112545e-05,
+      "loss": 1.1323,
+      "step": 545
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.9377892539029827e-05,
+      "loss": 1.0522,
+      "step": 546
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.9375012800159404e-05,
+      "loss": 1.1816,
+      "step": 547
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.9372126626490763e-05,
+      "loss": 1.0405,
+      "step": 548
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 1.9369234020004916e-05,
+      "loss": 1.1133,
+      "step": 549
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.936633498268728e-05,
+      "loss": 1.0112,
+      "step": 550
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.936342951652769e-05,
+      "loss": 1.1172,
+      "step": 551
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.936051762352039e-05,
+      "loss": 1.1187,
+      "step": 552
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 1.935759930566404e-05,
+      "loss": 1.0923,
+      "step": 553
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.935467456496171e-05,
+      "loss": 1.1182,
+      "step": 554
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9351743403420875e-05,
+      "loss": 1.0181,
+      "step": 555
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.9348805823053415e-05,
+      "loss": 1.0962,
+      "step": 556
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 1.934586182587562e-05,
+      "loss": 0.9932,
+      "step": 557
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.934291141390819e-05,
+      "loss": 1.0669,
+      "step": 558
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9339954589176215e-05,
+      "loss": 1.1113,
+      "step": 559
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9336991353709195e-05,
+      "loss": 1.1504,
+      "step": 560
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 1.9334021709541027e-05,
+      "loss": 1.0308,
+      "step": 561
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.933104565871001e-05,
+      "loss": 1.0952,
+      "step": 562
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.932806320325884e-05,
+      "loss": 1.1147,
+      "step": 563
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 1.93250743452346e-05,
+      "loss": 1.0781,
+      "step": 564
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9322079086688786e-05,
+      "loss": 0.9878,
+      "step": 565
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.931907742967727e-05,
+      "loss": 1.0967,
+      "step": 566
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9316069376260317e-05,
+      "loss": 1.1475,
+      "step": 567
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 1.9313054928502596e-05,
+      "loss": 1.0127,
+      "step": 568
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9310034088473153e-05,
+      "loss": 1.0684,
+      "step": 569
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9307006858245424e-05,
+      "loss": 1.0444,
+      "step": 570
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9303973239897233e-05,
+      "loss": 1.0713,
+      "step": 571
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 1.9300933235510786e-05,
+      "loss": 1.1587,
+      "step": 572
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.9297886847172672e-05,
+      "loss": 1.1006,
+      "step": 573
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.9294834076973872e-05,
+      "loss": 1.0811,
+      "step": 574
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.9291774927009733e-05,
+      "loss": 1.1406,
+      "step": 575
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.928870939937998e-05,
+      "loss": 1.1309,
+      "step": 576
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.9285637496188734e-05,
+      "loss": 1.1465,
+      "step": 577
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.9282559219544477e-05,
+      "loss": 1.1719,
+      "step": 578
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.9279474571560065e-05,
+      "loss": 1.062,
+      "step": 579
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 1.9276383554352734e-05,
+      "loss": 1.0801,
+      "step": 580
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.927328617004409e-05,
+      "loss": 0.9795,
+      "step": 581
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.9270182420760104e-05,
+      "loss": 1.1929,
+      "step": 582
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.9267072308631122e-05,
+      "loss": 0.9927,
+      "step": 583
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 1.926395583579186e-05,
+      "loss": 1.1519,
+      "step": 584
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.9260833004381384e-05,
+      "loss": 1.0415,
+      "step": 585
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.9257703816543144e-05,
+      "loss": 0.9946,
+      "step": 586
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.9254568274424942e-05,
+      "loss": 1.0776,
+      "step": 587
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 1.9251426380178945e-05,
+      "loss": 1.0649,
+      "step": 588
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9248278135961676e-05,
+      "loss": 1.0439,
+      "step": 589
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9245123543934017e-05,
+      "loss": 0.9951,
+      "step": 590
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.924196260626122e-05,
+      "loss": 1.1045,
+      "step": 591
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 1.9238795325112867e-05,
+      "loss": 1.0811,
+      "step": 592
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.923562170266292e-05,
+      "loss": 1.0488,
+      "step": 593
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.923244174108968e-05,
+      "loss": 1.1514,
+      "step": 594
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.9229255442575795e-05,
+      "loss": 1.1079,
+      "step": 595
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 1.922606280930827e-05,
+      "loss": 1.1118,
+      "step": 596
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.9222863843478462e-05,
+      "loss": 1.0371,
+      "step": 597
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.921965854728207e-05,
+      "loss": 1.0103,
+      "step": 598
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.921644692291913e-05,
+      "loss": 1.0703,
+      "step": 599
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 1.9213228972594032e-05,
+      "loss": 1.1729,
+      "step": 600
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.921000469851551e-05,
+      "loss": 1.0654,
+      "step": 601
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.9206774102896627e-05,
+      "loss": 1.0713,
+      "step": 602
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.9203537187954796e-05,
+      "loss": 1.1289,
+      "step": 603
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 1.920029395591176e-05,
+      "loss": 1.1992,
+      "step": 604
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.9197044408993595e-05,
+      "loss": 1.104,
+      "step": 605
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.9193788549430726e-05,
+      "loss": 1.0845,
+      "step": 606
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.91905263794579e-05,
+      "loss": 1.1338,
+      "step": 607
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 1.9187257901314194e-05,
+      "loss": 1.0942,
+      "step": 608
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.9183983117243022e-05,
+      "loss": 1.0225,
+      "step": 609
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.9180702029492118e-05,
+      "loss": 1.0854,
+      "step": 610
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 1.917741464031355e-05,
+      "loss": 1.0913,
+      "step": 611
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.9174120951963705e-05,
+      "loss": 0.9414,
+      "step": 612
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.91708209667033e-05,
+      "loss": 1.1147,
+      "step": 613
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.916751468679737e-05,
+      "loss": 1.1226,
+      "step": 614
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 1.916420211451527e-05,
+      "loss": 1.063,
+      "step": 615
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.9160883252130678e-05,
+      "loss": 1.0625,
+      "step": 616
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.9157558101921583e-05,
+      "loss": 1.0249,
+      "step": 617
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.9154226666170296e-05,
+      "loss": 1.041,
+      "step": 618
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.9150888947163436e-05,
+      "loss": 0.9956,
+      "step": 619
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.9147544947191943e-05,
+      "loss": 1.1221,
+      "step": 620
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.9144194668551063e-05,
+      "loss": 0.9712,
+      "step": 621
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.9140838113540347e-05,
+      "loss": 1.0596,
+      "step": 622
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 1.9137475284463668e-05,
+      "loss": 0.9624,
+      "step": 623
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.9134106183629184e-05,
+      "loss": 1.0894,
+      "step": 624
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.913073081334938e-05,
+      "loss": 1.083,
+      "step": 625
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.912734917594103e-05,
+      "loss": 0.9932,
+      "step": 626
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 1.9123961273725218e-05,
+      "loss": 1.0391,
+      "step": 627
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.912056710902732e-05,
+      "loss": 1.0171,
+      "step": 628
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9117166684177017e-05,
+      "loss": 1.1455,
+      "step": 629
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.911376000150828e-05,
+      "loss": 1.0356,
+      "step": 630
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 1.9110347063359383e-05,
+      "loss": 0.9946,
+      "step": 631
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.910692787207289e-05,
+      "loss": 1.0742,
+      "step": 632
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.910350242999566e-05,
+      "loss": 1.0112,
+      "step": 633
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.9100070739478833e-05,
+      "loss": 1.0151,
+      "step": 634
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 1.9096632802877845e-05,
+      "loss": 0.9976,
+      "step": 635
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.9093188622552424e-05,
+      "loss": 1.1514,
+      "step": 636
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.9089738200866574e-05,
+      "loss": 0.9756,
+      "step": 637
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.908628154018859e-05,
+      "loss": 1.0801,
+      "step": 638
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.908281864289104e-05,
+      "loss": 1.0718,
+      "step": 639
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.9079349511350784e-05,
+      "loss": 0.9946,
+      "step": 640
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.9075874147948955e-05,
+      "loss": 1.0225,
+      "step": 641
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.9072392555070965e-05,
+      "loss": 1.0435,
+      "step": 642
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 1.9068904735106502e-05,
+      "loss": 1.0571,
+      "step": 643
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.9065410690449525e-05,
+      "loss": 1.0044,
+      "step": 644
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.9061910423498268e-05,
+      "loss": 0.9668,
+      "step": 645
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.9058403936655235e-05,
+      "loss": 1.0459,
+      "step": 646
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 1.90548912323272e-05,
+      "loss": 1.0518,
+      "step": 647
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.9051372312925208e-05,
+      "loss": 1.0645,
+      "step": 648
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.904784718086456e-05,
+      "loss": 1.0283,
+      "step": 649
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.9044315838564835e-05,
+      "loss": 0.9878,
+      "step": 650
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 1.9040778288449862e-05,
+      "loss": 1.0239,
+      "step": 651
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.9037234532947737e-05,
+      "loss": 1.0562,
+      "step": 652
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.9033684574490813e-05,
+      "loss": 1.063,
+      "step": 653
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 1.9030128415515705e-05,
+      "loss": 1.0952,
+      "step": 654
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.9026566058463277e-05,
+      "loss": 0.9648,
+      "step": 655
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.9022997505778652e-05,
+      "loss": 1.0063,
+      "step": 656
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.9019422759911204e-05,
+      "loss": 1.0923,
+      "step": 657
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 1.901584182331456e-05,
+      "loss": 1.1685,
+      "step": 658
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.9012254698446595e-05,
+      "loss": 0.9941,
+      "step": 659
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.9008661387769427e-05,
+      "loss": 1.0654,
+      "step": 660
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.900506189374943e-05,
+      "loss": 1.001,
+      "step": 661
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 1.9001456218857207e-05,
+      "loss": 1.105,
+      "step": 662
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.8997844365567623e-05,
+      "loss": 1.0869,
+      "step": 663
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.899422633635976e-05,
+      "loss": 1.0801,
+      "step": 664
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.8990602133716968e-05,
+      "loss": 1.0884,
+      "step": 665
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 1.8986971760126805e-05,
+      "loss": 1.001,
+      "step": 666
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.898333521808108e-05,
+      "loss": 1.0571,
+      "step": 667
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.897969251007584e-05,
+      "loss": 1.0547,
+      "step": 668
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.8976043638611353e-05,
+      "loss": 1.0151,
+      "step": 669
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 1.8972388606192124e-05,
+      "loss": 0.9312,
+      "step": 670
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.8968727415326885e-05,
+      "loss": 1.1328,
+      "step": 671
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.8965060068528594e-05,
+      "loss": 1.0127,
+      "step": 672
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.8961386568314438e-05,
+      "loss": 1.0107,
+      "step": 673
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 1.895770691720582e-05,
+      "loss": 1.0498,
+      "step": 674
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.8954021117728377e-05,
+      "loss": 1.0186,
+      "step": 675
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.8950329172411954e-05,
+      "loss": 1.0103,
+      "step": 676
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.894663108379062e-05,
+      "loss": 1.0059,
+      "step": 677
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 1.894292685440266e-05,
+      "loss": 1.0854,
+      "step": 678
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.8939216486790577e-05,
+      "loss": 1.02,
+      "step": 679
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.893549998350108e-05,
+      "loss": 1.0483,
+      "step": 680
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.8931777347085094e-05,
+      "loss": 0.9883,
+      "step": 681
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 1.8928048580097758e-05,
+      "loss": 1.0137,
+      "step": 682
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.8924313685098405e-05,
+      "loss": 1.0786,
+      "step": 683
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.892057266465059e-05,
+      "loss": 0.9824,
+      "step": 684
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.8916825521322064e-05,
+      "loss": 0.9736,
+      "step": 685
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 1.8913072257684777e-05,
+      "loss": 1.063,
+      "step": 686
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.890931287631489e-05,
+      "loss": 0.9844,
+      "step": 687
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.890554737979276e-05,
+      "loss": 1.0698,
+      "step": 688
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.8901775770702933e-05,
+      "loss": 1.063,
+      "step": 689
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 1.8897998051634165e-05,
+      "loss": 1.0591,
+      "step": 690
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.889421422517939e-05,
+      "loss": 1.0771,
+      "step": 691
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.8890424293935744e-05,
+      "loss": 1.0767,
+      "step": 692
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.888662826050455e-05,
+      "loss": 0.9507,
+      "step": 693
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.888282612749132e-05,
+      "loss": 1.0098,
+      "step": 694
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8879017897505757e-05,
+      "loss": 1.0776,
+      "step": 695
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.8875203573161737e-05,
+      "loss": 1.0737,
+      "step": 696
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 1.887138315707733e-05,
+      "loss": 1.0391,
+      "step": 697
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.886755665187479e-05,
+      "loss": 1.0093,
+      "step": 698
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.886372406018053e-05,
+      "loss": 1.1055,
+      "step": 699
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.885988538462517e-05,
+      "loss": 1.1113,
+      "step": 700
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.8856040627843484e-05,
+      "loss": 1.1123,
+      "step": 701
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8852189792474426e-05,
+      "loss": 1.0083,
+      "step": 702
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8848332881161126e-05,
+      "loss": 1.0283,
+      "step": 703
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8844469896550874e-05,
+      "loss": 1.0815,
+      "step": 704
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 1.8840600841295143e-05,
+      "loss": 1.002,
+      "step": 705
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8836725718049562e-05,
+      "loss": 1.1001,
+      "step": 706
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8832844529473933e-05,
+      "loss": 1.0317,
+      "step": 707
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8828957278232215e-05,
+      "loss": 1.041,
+      "step": 708
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 1.8825063966992524e-05,
+      "loss": 1.0327,
+      "step": 709
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.8821164598427148e-05,
+      "loss": 1.0317,
+      "step": 710
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.8817259175212524e-05,
+      "loss": 0.9683,
+      "step": 711
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.8813347700029244e-05,
+      "loss": 1.0454,
+      "step": 712
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 1.8809430175562068e-05,
+      "loss": 0.9746,
+      "step": 713
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.880550660449988e-05,
+      "loss": 1.0986,
+      "step": 714
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.8801576989535745e-05,
+      "loss": 0.9966,
+      "step": 715
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.8797641333366852e-05,
+      "loss": 0.9893,
+      "step": 716
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 1.8793699638694555e-05,
+      "loss": 1.0625,
+      "step": 717
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.878975190822434e-05,
+      "loss": 0.9766,
+      "step": 718
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.878579814466584e-05,
+      "loss": 1.0688,
+      "step": 719
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.8781838350732834e-05,
+      "loss": 1.0386,
+      "step": 720
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 1.8777872529143236e-05,
+      "loss": 0.9912,
+      "step": 721
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.8773900682619092e-05,
+      "loss": 0.999,
+      "step": 722
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.8769922813886588e-05,
+      "loss": 1.0376,
+      "step": 723
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.8765938925676046e-05,
+      "loss": 1.0435,
+      "step": 724
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 1.876194902072192e-05,
+      "loss": 1.0732,
+      "step": 725
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.8757953101762786e-05,
+      "loss": 1.0957,
+      "step": 726
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.875395117154136e-05,
+      "loss": 1.0654,
+      "step": 727
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.8749943232804467e-05,
+      "loss": 1.0078,
+      "step": 728
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 1.8745929288303075e-05,
+      "loss": 1.0415,
+      "step": 729
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.874190934079226e-05,
+      "loss": 1.0542,
+      "step": 730
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.873788339303123e-05,
+      "loss": 1.021,
+      "step": 731
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.8733851447783296e-05,
+      "loss": 1.002,
+      "step": 732
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 1.8729813507815903e-05,
+      "loss": 1.0396,
+      "step": 733
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.87257695759006e-05,
+      "loss": 1.0444,
+      "step": 734
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8721719654813043e-05,
+      "loss": 0.9922,
+      "step": 735
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.871766374733302e-05,
+      "loss": 0.9351,
+      "step": 736
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 1.8713601856244406e-05,
+      "loss": 1.0845,
+      "step": 737
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.8709533984335194e-05,
+      "loss": 0.9336,
+      "step": 738
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.870546013439748e-05,
+      "loss": 1.0854,
+      "step": 739
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 1.8701380309227465e-05,
+      "loss": 1.0088,
+      "step": 740
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.869729451162545e-05,
+      "loss": 1.0513,
+      "step": 741
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.869320274439583e-05,
+      "loss": 1.0298,
+      "step": 742
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.86891050103471e-05,
+      "loss": 1.0239,
+      "step": 743
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 1.8685001312291863e-05,
+      "loss": 1.0977,
+      "step": 744
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.8680891653046798e-05,
+      "loss": 1.0718,
+      "step": 745
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.867677603543268e-05,
+      "loss": 1.0483,
+      "step": 746
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.8672654462274383e-05,
+      "loss": 1.0146,
+      "step": 747
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 1.8668526936400863e-05,
+      "loss": 1.0449,
+      "step": 748
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.8664393460645154e-05,
+      "loss": 1.0503,
+      "step": 749
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.866025403784439e-05,
+      "loss": 0.9692,
+      "step": 750
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.8656108670839768e-05,
+      "loss": 1.1475,
+      "step": 751
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 1.8651957362476582e-05,
+      "loss": 1.0225,
+      "step": 752
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.864780011560419e-05,
+      "loss": 1.0366,
+      "step": 753
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.864363693307604e-05,
+      "loss": 0.9697,
+      "step": 754
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.8639467817749642e-05,
+      "loss": 1.0273,
+      "step": 755
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 1.863529277248659e-05,
+      "loss": 1.0684,
+      "step": 756
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.8631111800152532e-05,
+      "loss": 0.9751,
+      "step": 757
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.86269249036172e-05,
+      "loss": 0.9458,
+      "step": 758
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.8622732085754386e-05,
+      "loss": 0.9697,
+      "step": 759
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 1.861853334944194e-05,
+      "loss": 1.0483,
+      "step": 760
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.8614328697561785e-05,
+      "loss": 1.0273,
+      "step": 761
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.8610118132999896e-05,
+      "loss": 1.0205,
+      "step": 762
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.8605901658646318e-05,
+      "loss": 1.0381,
+      "step": 763
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.8601679277395136e-05,
+      "loss": 1.0488,
+      "step": 764
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.85974509921445e-05,
+      "loss": 0.9419,
+      "step": 765
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.8593216805796612e-05,
+      "loss": 0.9341,
+      "step": 766
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.858897672125772e-05,
+      "loss": 1.0254,
+      "step": 767
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 1.8584730741438128e-05,
+      "loss": 1.0444,
+      "step": 768
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.858047886925217e-05,
+      "loss": 0.9888,
+      "step": 769
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.8576221107618244e-05,
+      "loss": 0.9087,
+      "step": 770
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.8571957459458785e-05,
+      "loss": 0.9067,
+      "step": 771
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.8567687927700255e-05,
+      "loss": 0.9473,
+      "step": 772
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.8563412515273173e-05,
+      "loss": 1.0122,
+      "step": 773
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.8559131225112085e-05,
+      "loss": 1.0059,
+      "step": 774
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.855484406015557e-05,
+      "loss": 1.0244,
+      "step": 775
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 1.8550551023346245e-05,
+      "loss": 0.9785,
+      "step": 776
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.8546252117630746e-05,
+      "loss": 1.0767,
+      "step": 777
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.8541947345959753e-05,
+      "loss": 0.9893,
+      "step": 778
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.8537636711287968e-05,
+      "loss": 0.9937,
+      "step": 779
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 1.8533320216574103e-05,
+      "loss": 1.0259,
+      "step": 780
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.852899786478091e-05,
+      "loss": 0.9438,
+      "step": 781
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.8524669658875153e-05,
+      "loss": 0.9658,
+      "step": 782
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 1.8520335601827614e-05,
+      "loss": 0.9683,
+      "step": 783
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.8515995696613096e-05,
+      "loss": 0.9429,
+      "step": 784
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.8511649946210407e-05,
+      "loss": 0.9727,
+      "step": 785
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.8507298353602373e-05,
+      "loss": 1.0068,
+      "step": 786
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 1.8502940921775837e-05,
+      "loss": 0.9282,
+      "step": 787
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.8498577653721638e-05,
+      "loss": 0.9312,
+      "step": 788
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.849420855243462e-05,
+      "loss": 1.0068,
+      "step": 789
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.8489833620913644e-05,
+      "loss": 0.9985,
+      "step": 790
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 1.8485452862161555e-05,
+      "loss": 0.9624,
+      "step": 791
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.848106627918522e-05,
+      "loss": 1.0547,
+      "step": 792
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.847667387499548e-05,
+      "loss": 1.0049,
+      "step": 793
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.8472275652607184e-05,
+      "loss": 0.9712,
+      "step": 794
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 1.8467871615039182e-05,
+      "loss": 0.9707,
+      "step": 795
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.8463461765314302e-05,
+      "loss": 0.9893,
+      "step": 796
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.845904610645936e-05,
+      "loss": 0.9419,
+      "step": 797
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.845462464150517e-05,
+      "loss": 0.9316,
+      "step": 798
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 1.8450197373486527e-05,
+      "loss": 1.0239,
+      "step": 799
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.8445764305442205e-05,
+      "loss": 1.0371,
+      "step": 800
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.8441325440414965e-05,
+      "loss": 0.9404,
+      "step": 801
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.8436880781451545e-05,
+      "loss": 0.9946,
+      "step": 802
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 1.8432430331602654e-05,
+      "loss": 0.9829,
+      "step": 803
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.842797409392299e-05,
+      "loss": 1.0283,
+      "step": 804
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.8423512071471207e-05,
+      "loss": 0.9531,
+      "step": 805
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.841904426730994e-05,
+      "loss": 1.0493,
+      "step": 806
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 1.8414570684505783e-05,
+      "loss": 0.9199,
+      "step": 807
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.841009132612931e-05,
+      "loss": 0.9912,
+      "step": 808
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.8405606195255052e-05,
+      "loss": 1.0938,
+      "step": 809
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.84011152949615e-05,
+      "loss": 0.9258,
+      "step": 810
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 1.8396618628331106e-05,
+      "loss": 0.9653,
+      "step": 811
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.8392116198450282e-05,
+      "loss": 1.0112,
+      "step": 812
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.8387608008409396e-05,
+      "loss": 1.0068,
+      "step": 813
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.8383094061302767e-05,
+      "loss": 0.9956,
+      "step": 814
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 1.8378574360228668e-05,
+      "loss": 0.9795,
+      "step": 815
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.8374048908289318e-05,
+      "loss": 0.895,
+      "step": 816
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.8369517708590888e-05,
+      "loss": 0.9717,
+      "step": 817
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.836498076424349e-05,
+      "loss": 0.9297,
+      "step": 818
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 1.836043807836118e-05,
+      "loss": 0.9478,
+      "step": 819
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.835588965406196e-05,
+      "loss": 0.9736,
+      "step": 820
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.8351335494467762e-05,
+      "loss": 1.0186,
+      "step": 821
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.8346775602704464e-05,
+      "loss": 0.8989,
+      "step": 822
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 1.8342209981901863e-05,
+      "loss": 0.9385,
+      "step": 823
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.833763863519371e-05,
+      "loss": 0.9688,
+      "step": 824
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.8333061565717668e-05,
+      "loss": 1.0059,
+      "step": 825
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 1.8328478776615336e-05,
+      "loss": 0.9893,
+      "step": 826
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.832389027103224e-05,
+      "loss": 1.0903,
+      "step": 827
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.831929605211782e-05,
+      "loss": 0.9907,
+      "step": 828
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.8314696123025456e-05,
+      "loss": 0.939,
+      "step": 829
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.8310090486912423e-05,
+      "loss": 0.9927,
+      "step": 830
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.8305479146939937e-05,
+      "loss": 0.9409,
+      "step": 831
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.8300862106273113e-05,
+      "loss": 0.8989,
+      "step": 832
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.829623936808099e-05,
+      "loss": 1.0098,
+      "step": 833
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 1.8291610935536503e-05,
+      "loss": 1.0029,
+      "step": 834
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.8286976811816506e-05,
+      "loss": 0.979,
+      "step": 835
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.8282337000101763e-05,
+      "loss": 0.9961,
+      "step": 836
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.8277691503576928e-05,
+      "loss": 0.9019,
+      "step": 837
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 1.8273040325430575e-05,
+      "loss": 0.8809,
+      "step": 838
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.826838346885516e-05,
+      "loss": 1.0107,
+      "step": 839
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.826372093704705e-05,
+      "loss": 1.0239,
+      "step": 840
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.8259052733206503e-05,
+      "loss": 0.9292,
+      "step": 841
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 1.8254378860537664e-05,
+      "loss": 1.0234,
+      "step": 842
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.824969932224858e-05,
+      "loss": 0.897,
+      "step": 843
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.8245014121551173e-05,
+      "loss": 0.9946,
+      "step": 844
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.8240323261661266e-05,
+      "loss": 0.958,
+      "step": 845
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 1.823562674579856e-05,
+      "loss": 0.9102,
+      "step": 846
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.8230924577186635e-05,
+      "loss": 0.9253,
+      "step": 847
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.822621675905296e-05,
+      "loss": 1.0049,
+      "step": 848
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.822150329462886e-05,
+      "loss": 1.019,
+      "step": 849
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.821678418714957e-05,
+      "loss": 0.8135,
+      "step": 850
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.821205943985417e-05,
+      "loss": 0.9307,
+      "step": 851
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.8207329055985622e-05,
+      "loss": 1.0107,
+      "step": 852
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.8202593038790753e-05,
+      "loss": 0.9468,
+      "step": 853
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 1.8197851391520265e-05,
+      "loss": 0.9219,
+      "step": 854
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8193104117428712e-05,
+      "loss": 0.9639,
+      "step": 855
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.818835121977452e-05,
+      "loss": 1.0581,
+      "step": 856
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8183592701819965e-05,
+      "loss": 1.0181,
+      "step": 857
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 1.8178828566831192e-05,
+      "loss": 0.8828,
+      "step": 858
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.81740588180782e-05,
+      "loss": 0.9858,
+      "step": 859
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.816928345883483e-05,
+      "loss": 0.9937,
+      "step": 860
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.8164502492378783e-05,
+      "loss": 0.9463,
+      "step": 861
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 1.8159715921991612e-05,
+      "loss": 0.9556,
+      "step": 862
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.8154923750958704e-05,
+      "loss": 1.0059,
+      "step": 863
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.8150125982569304e-05,
+      "loss": 0.9536,
+      "step": 864
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.814532262011649e-05,
+      "loss": 0.9087,
+      "step": 865
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 1.814051366689718e-05,
+      "loss": 0.9209,
+      "step": 866
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.8135699126212135e-05,
+      "loss": 1.0332,
+      "step": 867
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.8130879001365944e-05,
+      "loss": 0.9453,
+      "step": 868
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 1.8126053295667036e-05,
+      "loss": 0.9009,
+      "step": 869
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.8121222012427666e-05,
+      "loss": 1.0044,
+      "step": 870
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.8116385154963913e-05,
+      "loss": 0.939,
+      "step": 871
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.81115427265957e-05,
+      "loss": 0.939,
+      "step": 872
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 1.8106694730646745e-05,
+      "loss": 0.8911,
+      "step": 873
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.8101841170444617e-05,
+      "loss": 0.9751,
+      "step": 874
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.8096982049320683e-05,
+      "loss": 0.98,
+      "step": 875
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.8092117370610137e-05,
+      "loss": 0.9297,
+      "step": 876
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 1.8087247137651985e-05,
+      "loss": 0.9316,
+      "step": 877
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.8082371353789046e-05,
+      "loss": 0.9033,
+      "step": 878
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.8077490022367948e-05,
+      "loss": 0.9365,
+      "step": 879
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.8072603146739126e-05,
+      "loss": 0.9419,
+      "step": 880
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 1.8067710730256825e-05,
+      "loss": 0.9653,
+      "step": 881
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.8062812776279085e-05,
+      "loss": 0.9131,
+      "step": 882
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.8057909288167758e-05,
+      "loss": 0.9055,
+      "step": 883
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.8053000269288486e-05,
+      "loss": 0.9951,
+      "step": 884
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 1.80480857230107e-05,
+      "loss": 0.8818,
+      "step": 885
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.804316565270765e-05,
+      "loss": 0.8862,
+      "step": 886
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.8038240061756348e-05,
+      "loss": 0.9502,
+      "step": 887
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.8033308953537617e-05,
+      "loss": 0.9189,
+      "step": 888
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 1.802837233143606e-05,
+      "loss": 0.9082,
+      "step": 889
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.8023430198840056e-05,
+      "loss": 0.9995,
+      "step": 890
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.801848255914178e-05,
+      "loss": 0.9077,
+      "step": 891
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.801352941573718e-05,
+      "loss": 1.0161,
+      "step": 892
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.8008570772025978e-05,
+      "loss": 0.8633,
+      "step": 893
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.800360663141168e-05,
+      "loss": 0.8306,
+      "step": 894
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.799863699730156e-05,
+      "loss": 0.9663,
+      "step": 895
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.799366187310666e-05,
+      "loss": 0.9526,
+      "step": 896
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 1.7988681262241798e-05,
+      "loss": 0.9756,
+      "step": 897
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.7983695168125552e-05,
+      "loss": 0.9561,
+      "step": 898
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.7978703594180266e-05,
+      "loss": 0.8682,
+      "step": 899
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.797370654383204e-05,
+      "loss": 1.0278,
+      "step": 900
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 1.796870402051074e-05,
+      "loss": 0.9385,
+      "step": 901
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.796369602764999e-05,
+      "loss": 0.8784,
+      "step": 902
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.7958682568687152e-05,
+      "loss": 0.8726,
+      "step": 903
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.7953663647063365e-05,
+      "loss": 0.9482,
+      "step": 904
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 1.7948639266223495e-05,
+      "loss": 0.9629,
+      "step": 905
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.794360942961617e-05,
+      "loss": 0.9258,
+      "step": 906
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.793857414069375e-05,
+      "loss": 0.9087,
+      "step": 907
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.7933533402912354e-05,
+      "loss": 1.0249,
+      "step": 908
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 1.792848721973182e-05,
+      "loss": 0.9062,
+      "step": 909
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.7923435594615744e-05,
+      "loss": 0.9175,
+      "step": 910
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.7918378531031444e-05,
+      "loss": 0.916,
+      "step": 911
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 1.7913316032449975e-05,
+      "loss": 0.9478,
+      "step": 912
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.790824810234612e-05,
+      "loss": 0.9478,
+      "step": 913
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.7903174744198398e-05,
+      "loss": 0.8394,
+      "step": 914
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.7898095961489047e-05,
+      "loss": 0.8867,
+      "step": 915
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 1.7893011757704022e-05,
+      "loss": 0.9619,
+      "step": 916
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.7887922136333015e-05,
+      "loss": 0.9692,
+      "step": 917
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.788282710086942e-05,
+      "loss": 0.9414,
+      "step": 918
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.7877726654810364e-05,
+      "loss": 0.876,
+      "step": 919
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 1.7872620801656674e-05,
+      "loss": 0.8452,
+      "step": 920
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.7867509544912893e-05,
+      "loss": 0.9326,
+      "step": 921
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.786239288808727e-05,
+      "loss": 0.9033,
+      "step": 922
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.7857270834691766e-05,
+      "loss": 0.9556,
+      "step": 923
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 1.7852143388242046e-05,
+      "loss": 0.9478,
+      "step": 924
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.784701055225747e-05,
+      "loss": 0.9702,
+      "step": 925
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.7841872330261103e-05,
+      "loss": 0.8926,
+      "step": 926
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.7836728725779705e-05,
+      "loss": 0.9453,
+      "step": 927
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.783157974234373e-05,
+      "loss": 0.9634,
+      "step": 928
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.7826425383487327e-05,
+      "loss": 0.897,
+      "step": 929
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.782126565274833e-05,
+      "loss": 0.9072,
+      "step": 930
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.781610055366826e-05,
+      "loss": 0.8701,
+      "step": 931
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 1.7810930089792328e-05,
+      "loss": 0.9238,
+      "step": 932
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.7805754264669426e-05,
+      "loss": 0.9038,
+      "step": 933
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.7800573081852124e-05,
+      "loss": 1.0576,
+      "step": 934
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.779538654489666e-05,
+      "loss": 0.8979,
+      "step": 935
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 1.7790194657362973e-05,
+      "loss": 0.8755,
+      "step": 936
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.7784997422814644e-05,
+      "loss": 0.832,
+      "step": 937
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.7779794844818944e-05,
+      "loss": 0.9443,
+      "step": 938
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.7774586926946805e-05,
+      "loss": 0.8696,
+      "step": 939
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 1.7769373672772824e-05,
+      "loss": 0.9102,
+      "step": 940
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.7764155085875262e-05,
+      "loss": 1.0078,
+      "step": 941
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.775893116983604e-05,
+      "loss": 0.9141,
+      "step": 942
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.7753701928240735e-05,
+      "loss": 0.7603,
+      "step": 943
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 1.7748467364678585e-05,
+      "loss": 0.917,
+      "step": 944
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.7743227482742475e-05,
+      "loss": 0.9575,
+      "step": 945
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.7737982286028938e-05,
+      "loss": 0.8179,
+      "step": 946
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.773273177813817e-05,
+      "loss": 0.8743,
+      "step": 947
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 1.7727475962673986e-05,
+      "loss": 0.7693,
+      "step": 948
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.7722214843243874e-05,
+      "loss": 0.9746,
+      "step": 949
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.771694842345894e-05,
+      "loss": 1.0098,
+      "step": 950
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.771167670693394e-05,
+      "loss": 0.9233,
+      "step": 951
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 1.770639969728726e-05,
+      "loss": 0.9692,
+      "step": 952
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.770111739814092e-05,
+      "loss": 0.8823,
+      "step": 953
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.7695829813120574e-05,
+      "loss": 0.9048,
+      "step": 954
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.76905369458555e-05,
+      "loss": 0.8608,
+      "step": 955
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.7685238799978605e-05,
+      "loss": 0.959,
+      "step": 956
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.7679935379126417e-05,
+      "loss": 0.9165,
+      "step": 957
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.767462668693908e-05,
+      "loss": 0.938,
+      "step": 958
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 1.766931272706037e-05,
+      "loss": 0.8276,
+      "step": 959
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.766399350313766e-05,
+      "loss": 0.8579,
+      "step": 960
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.7658669018821956e-05,
+      "loss": 0.8801,
+      "step": 961
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.7653339277767852e-05,
+      "loss": 0.9561,
+      "step": 962
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 1.7648004283633575e-05,
+      "loss": 0.9639,
+      "step": 963
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.764266404008094e-05,
+      "loss": 0.9028,
+      "step": 964
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.7637318550775368e-05,
+      "loss": 0.8447,
+      "step": 965
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.7631967819385883e-05,
+      "loss": 0.8872,
+      "step": 966
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 1.7626611849585113e-05,
+      "loss": 0.876,
+      "step": 967
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.7621250645049267e-05,
+      "loss": 0.915,
+      "step": 968
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.7615884209458164e-05,
+      "loss": 0.8787,
+      "step": 969
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.7610512546495195e-05,
+      "loss": 0.9155,
+      "step": 970
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 1.7605135659847357e-05,
+      "loss": 0.8462,
+      "step": 971
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.759975355320522e-05,
+      "loss": 0.9146,
+      "step": 972
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.759436623026294e-05,
+      "loss": 0.8682,
+      "step": 973
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.7588973694718263e-05,
+      "loss": 0.8306,
+      "step": 974
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 1.7583575950272495e-05,
+      "loss": 0.8828,
+      "step": 975
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.7578173000630528e-05,
+      "loss": 0.9043,
+      "step": 976
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.757276484950083e-05,
+      "loss": 0.9375,
+      "step": 977
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.7567351500595428e-05,
+      "loss": 0.8706,
+      "step": 978
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 1.7561932957629927e-05,
+      "loss": 0.8477,
+      "step": 979
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.7556509224323495e-05,
+      "loss": 0.918,
+      "step": 980
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.7551080304398858e-05,
+      "loss": 0.9609,
+      "step": 981
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.7545646201582304e-05,
+      "loss": 0.9067,
+      "step": 982
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 1.754020691960368e-05,
+      "loss": 0.9761,
+      "step": 983
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.7534762462196385e-05,
+      "loss": 0.8687,
+      "step": 984
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.752931283309738e-05,
+      "loss": 0.8564,
+      "step": 985
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.7523858036047156e-05,
+      "loss": 0.8882,
+      "step": 986
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 1.7518398074789776e-05,
+      "loss": 0.98,
+      "step": 987
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.7512932953072826e-05,
+      "loss": 0.9146,
+      "step": 988
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.7507462674647445e-05,
+      "loss": 0.8848,
+      "step": 989
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.7501987243268307e-05,
+      "loss": 0.9419,
+      "step": 990
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 1.7496506662693632e-05,
+      "loss": 0.9116,
+      "step": 991
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.7491020936685158e-05,
+      "loss": 0.8833,
+      "step": 992
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.7485530069008173e-05,
+      "loss": 0.9292,
+      "step": 993
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.748003406343148e-05,
+      "loss": 0.9629,
+      "step": 994
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 1.7474532923727418e-05,
+      "loss": 0.9238,
+      "step": 995
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.7469026653671835e-05,
+      "loss": 0.8999,
+      "step": 996
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.746351525704413e-05,
+      "loss": 0.938,
+      "step": 997
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.7457998737627183e-05,
+      "loss": 0.9053,
+      "step": 998
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 1.7452477099207416e-05,
+      "loss": 0.9419,
+      "step": 999
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.7446950345574762e-05,
+      "loss": 0.9116,
+      "step": 1000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.7441418480522656e-05,
+      "loss": 0.9043,
+      "step": 1001
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.7435881507848048e-05,
+      "loss": 0.9546,
+      "step": 1002
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.7430339431351388e-05,
+      "loss": 0.9043,
+      "step": 1003
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.7424792254836636e-05,
+      "loss": 0.9131,
+      "step": 1004
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.741923998211125e-05,
+      "loss": 0.9233,
+      "step": 1005
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 1.7413682616986185e-05,
+      "loss": 0.8359,
+      "step": 1006
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.740812016327589e-05,
+      "loss": 0.8833,
+      "step": 1007
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.7402552624798306e-05,
+      "loss": 0.9722,
+      "step": 1008
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.739698000537487e-05,
+      "loss": 0.8804,
+      "step": 1009
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 1.7391402308830506e-05,
+      "loss": 0.8921,
+      "step": 1010
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.738581953899361e-05,
+      "loss": 0.8726,
+      "step": 1011
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.738023169969608e-05,
+      "loss": 0.9658,
+      "step": 1012
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.7374638794773276e-05,
+      "loss": 0.9082,
+      "step": 1013
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 1.7369040828064046e-05,
+      "loss": 0.9111,
+      "step": 1014
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.7363437803410708e-05,
+      "loss": 0.895,
+      "step": 1015
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.7357829724659056e-05,
+      "loss": 0.876,
+      "step": 1016
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.7352216595658343e-05,
+      "loss": 0.8994,
+      "step": 1017
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 1.7346598420261294e-05,
+      "loss": 0.8774,
+      "step": 1018
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.7340975202324103e-05,
+      "loss": 0.9907,
+      "step": 1019
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.733534694570642e-05,
+      "loss": 0.8975,
+      "step": 1020
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.7329713654271356e-05,
+      "loss": 0.8027,
+      "step": 1021
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 1.7324075331885466e-05,
+      "loss": 0.855,
+      "step": 1022
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.731843198241878e-05,
+      "loss": 0.9351,
+      "step": 1023
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.7312783609744753e-05,
+      "loss": 0.9482,
+      "step": 1024
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.7307130217740317e-05,
+      "loss": 0.8154,
+      "step": 1025
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 1.7301471810285823e-05,
+      "loss": 0.7422,
+      "step": 1026
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.7295808391265074e-05,
+      "loss": 0.8438,
+      "step": 1027
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.729013996456532e-05,
+      "loss": 0.8916,
+      "step": 1028
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.728446653407724e-05,
+      "loss": 0.9521,
+      "step": 1029
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 1.7278788103694944e-05,
+      "loss": 0.7896,
+      "step": 1030
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.727310467731598e-05,
+      "loss": 0.8589,
+      "step": 1031
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.7267416258841333e-05,
+      "loss": 0.8516,
+      "step": 1032
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.7261722852175394e-05,
+      "loss": 0.8506,
+      "step": 1033
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 1.7256024461226002e-05,
+      "loss": 0.7795,
+      "step": 1034
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7250321089904393e-05,
+      "loss": 0.8599,
+      "step": 1035
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7244612742125237e-05,
+      "loss": 0.8789,
+      "step": 1036
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.7238899421806618e-05,
+      "loss": 0.9248,
+      "step": 1037
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 1.723318113287003e-05,
+      "loss": 0.855,
+      "step": 1038
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.7227457879240375e-05,
+      "loss": 0.8047,
+      "step": 1039
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.7221729664845966e-05,
+      "loss": 0.8369,
+      "step": 1040
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.7215996493618525e-05,
+      "loss": 0.8169,
+      "step": 1041
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 1.721025836949317e-05,
+      "loss": 0.8696,
+      "step": 1042
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.7204515296408418e-05,
+      "loss": 0.9707,
+      "step": 1043
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.719876727830619e-05,
+      "loss": 0.8535,
+      "step": 1044
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 1.7193014319131794e-05,
+      "loss": 0.8091,
+      "step": 1045
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.7187256422833928e-05,
+      "loss": 0.8511,
+      "step": 1046
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.7181493593364688e-05,
+      "loss": 1.0049,
+      "step": 1047
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.7175725834679552e-05,
+      "loss": 0.8315,
+      "step": 1048
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 1.7169953150737375e-05,
+      "loss": 0.8833,
+      "step": 1049
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.71641755455004e-05,
+      "loss": 0.8545,
+      "step": 1050
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.7158393022934246e-05,
+      "loss": 0.9102,
+      "step": 1051
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.7152605587007906e-05,
+      "loss": 0.7983,
+      "step": 1052
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 1.7146813241693742e-05,
+      "loss": 0.8647,
+      "step": 1053
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.71410159909675e-05,
+      "loss": 0.8994,
+      "step": 1054
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.7135213838808275e-05,
+      "loss": 0.7964,
+      "step": 1055
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.712940678919853e-05,
+      "loss": 1.0054,
+      "step": 1056
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 1.7123594846124105e-05,
+      "loss": 0.8599,
+      "step": 1057
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7117778013574174e-05,
+      "loss": 0.8672,
+      "step": 1058
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7111956295541296e-05,
+      "loss": 0.8872,
+      "step": 1059
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7106129696021353e-05,
+      "loss": 0.8247,
+      "step": 1060
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 1.7100298219013597e-05,
+      "loss": 0.8286,
+      "step": 1061
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.7094461868520625e-05,
+      "loss": 0.894,
+      "step": 1062
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.7088620648548376e-05,
+      "loss": 0.9043,
+      "step": 1063
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.708277456310613e-05,
+      "loss": 0.854,
+      "step": 1064
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 1.707692361620651e-05,
+      "loss": 0.9316,
+      "step": 1065
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7071067811865477e-05,
+      "loss": 0.8716,
+      "step": 1066
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.706520715410232e-05,
+      "loss": 0.9111,
+      "step": 1067
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7059341646939665e-05,
+      "loss": 0.834,
+      "step": 1068
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 1.7053471294403464e-05,
+      "loss": 0.8628,
+      "step": 1069
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7047596100522992e-05,
+      "loss": 0.8149,
+      "step": 1070
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.7041716069330853e-05,
+      "loss": 0.8994,
+      "step": 1071
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.703583120486297e-05,
+      "loss": 0.8643,
+      "step": 1072
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 1.702994151115858e-05,
+      "loss": 0.8052,
+      "step": 1073
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.702404699226024e-05,
+      "loss": 0.8569,
+      "step": 1074
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.7018147652213805e-05,
+      "loss": 0.7974,
+      "step": 1075
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.701224349506846e-05,
+      "loss": 0.8169,
+      "step": 1076
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 1.700633452487668e-05,
+      "loss": 0.9468,
+      "step": 1077
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.7000420745694256e-05,
+      "loss": 0.8716,
+      "step": 1078
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6994502161580263e-05,
+      "loss": 0.8892,
+      "step": 1079
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6988578776597094e-05,
+      "loss": 0.9033,
+      "step": 1080
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6982650594810424e-05,
+      "loss": 0.7939,
+      "step": 1081
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.697671762028922e-05,
+      "loss": 0.897,
+      "step": 1082
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.6970779857105747e-05,
+      "loss": 0.8384,
+      "step": 1083
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.6964837309335553e-05,
+      "loss": 0.8066,
+      "step": 1084
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 1.6958889981057467e-05,
+      "loss": 0.833,
+      "step": 1085
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.6952937876353595e-05,
+      "loss": 0.8687,
+      "step": 1086
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.6946980999309343e-05,
+      "loss": 0.8501,
+      "step": 1087
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 1.694101935401337e-05,
+      "loss": 0.8191,
+      "step": 1088
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.693505294455761e-05,
+      "loss": 0.7803,
+      "step": 1089
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.6929081775037276e-05,
+      "loss": 0.9282,
+      "step": 1090
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.692310584955085e-05,
+      "loss": 0.7559,
+      "step": 1091
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 1.6917125172200066e-05,
+      "loss": 0.8555,
+      "step": 1092
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.6911139747089932e-05,
+      "loss": 0.8857,
+      "step": 1093
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.6905149578328705e-05,
+      "loss": 0.8389,
+      "step": 1094
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.6899154670027903e-05,
+      "loss": 0.9092,
+      "step": 1095
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 1.68931550263023e-05,
+      "loss": 0.7739,
+      "step": 1096
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6887150651269918e-05,
+      "loss": 0.8765,
+      "step": 1097
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6881141549052014e-05,
+      "loss": 0.8037,
+      "step": 1098
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.6875127723773114e-05,
+      "loss": 0.9639,
+      "step": 1099
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 1.686910917956096e-05,
+      "loss": 0.8779,
+      "step": 1100
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.686308592054656e-05,
+      "loss": 0.9673,
+      "step": 1101
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.6857057950864134e-05,
+      "loss": 0.8826,
+      "step": 1102
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.6851025274651145e-05,
+      "loss": 0.8887,
+      "step": 1103
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 1.6844987896048287e-05,
+      "loss": 0.8916,
+      "step": 1104
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.6838945819199487e-05,
+      "loss": 0.8179,
+      "step": 1105
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.6832899048251886e-05,
+      "loss": 0.7422,
+      "step": 1106
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.682684758735585e-05,
+      "loss": 0.8315,
+      "step": 1107
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 1.682079144066497e-05,
+      "loss": 0.96,
+      "step": 1108
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6814730612336044e-05,
+      "loss": 0.8652,
+      "step": 1109
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6808665106529096e-05,
+      "loss": 0.9077,
+      "step": 1110
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.6802594927407346e-05,
+      "loss": 0.9019,
+      "step": 1111
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 1.679652007913723e-05,
+      "loss": 0.9102,
+      "step": 1112
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.679044056588839e-05,
+      "loss": 0.8457,
+      "step": 1113
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.6784356391833665e-05,
+      "loss": 0.9033,
+      "step": 1114
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.6778267561149097e-05,
+      "loss": 0.7881,
+      "step": 1115
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 1.677217407801392e-05,
+      "loss": 0.9575,
+      "step": 1116
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.6766075946610568e-05,
+      "loss": 0.873,
+      "step": 1117
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.675997317112466e-05,
+      "loss": 0.7927,
+      "step": 1118
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.6753865755745e-05,
+      "loss": 0.8989,
+      "step": 1119
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 1.6747753704663586e-05,
+      "loss": 0.8066,
+      "step": 1120
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.6741637022075593e-05,
+      "loss": 0.8628,
+      "step": 1121
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.673551571217937e-05,
+      "loss": 0.8394,
+      "step": 1122
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.6729389779176446e-05,
+      "loss": 0.7759,
+      "step": 1123
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 1.6723259227271527e-05,
+      "loss": 0.7695,
+      "step": 1124
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.671712406067249e-05,
+      "loss": 0.751,
+      "step": 1125
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.671098428359037e-05,
+      "loss": 0.8379,
+      "step": 1126
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.6704839900239372e-05,
+      "loss": 0.8779,
+      "step": 1127
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 1.6698690914836863e-05,
+      "loss": 0.9229,
+      "step": 1128
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6692537331603374e-05,
+      "loss": 0.8535,
+      "step": 1129
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.6686379154762575e-05,
+      "loss": 0.8765,
+      "step": 1130
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 1.668021638854131e-05,
+      "loss": 0.8145,
+      "step": 1131
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.6674049037169565e-05,
+      "loss": 0.8232,
+      "step": 1132
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.666787710488046e-05,
+      "loss": 0.8032,
+      "step": 1133
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.6661700595910284e-05,
+      "loss": 0.7954,
+      "step": 1134
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 1.6655519514498443e-05,
+      "loss": 0.8066,
+      "step": 1135
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.6649333864887497e-05,
+      "loss": 0.896,
+      "step": 1136
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.664314365132314e-05,
+      "loss": 0.8545,
+      "step": 1137
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.6636948878054192e-05,
+      "loss": 0.7417,
+      "step": 1138
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 1.6630749549332607e-05,
+      "loss": 0.8267,
+      "step": 1139
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.662454566941346e-05,
+      "loss": 0.8154,
+      "step": 1140
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.6618337242554963e-05,
+      "loss": 0.9321,
+      "step": 1141
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.661212427301844e-05,
+      "loss": 0.9009,
+      "step": 1142
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 1.6605906765068326e-05,
+      "loss": 0.8979,
+      "step": 1143
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.6599684722972188e-05,
+      "loss": 0.832,
+      "step": 1144
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.659345815100069e-05,
+      "loss": 0.8774,
+      "step": 1145
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.6587227053427613e-05,
+      "loss": 0.7622,
+      "step": 1146
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 1.658099143452984e-05,
+      "loss": 0.7654,
+      "step": 1147
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.657475129858737e-05,
+      "loss": 0.793,
+      "step": 1148
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.6568506649883276e-05,
+      "loss": 0.8745,
+      "step": 1149
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.6562257492703756e-05,
+      "loss": 0.957,
+      "step": 1150
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 1.655600383133809e-05,
+      "loss": 0.811,
+      "step": 1151
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.654974567007865e-05,
+      "loss": 0.7012,
+      "step": 1152
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.654348301322089e-05,
+      "loss": 0.772,
+      "step": 1153
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.6537215865063364e-05,
+      "loss": 0.936,
+      "step": 1154
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 1.6530944229907705e-05,
+      "loss": 0.8452,
+      "step": 1155
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6524668112058617e-05,
+      "loss": 0.832,
+      "step": 1156
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6518387515823887e-05,
+      "loss": 0.874,
+      "step": 1157
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6512102445514376e-05,
+      "loss": 0.835,
+      "step": 1158
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6505812905444014e-05,
+      "loss": 0.7183,
+      "step": 1159
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.6499518899929802e-05,
+      "loss": 0.8374,
+      "step": 1160
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.64932204332918e-05,
+      "loss": 0.7891,
+      "step": 1161
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.648691750985314e-05,
+      "loss": 0.8438,
+      "step": 1162
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 1.6480610133940005e-05,
+      "loss": 0.7944,
+      "step": 1163
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.6474298309881634e-05,
+      "loss": 0.8765,
+      "step": 1164
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.6467982042010323e-05,
+      "loss": 0.6978,
+      "step": 1165
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.6461661334661415e-05,
+      "loss": 0.8472,
+      "step": 1166
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 1.64553361921733e-05,
+      "loss": 0.8408,
+      "step": 1167
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.644900661888742e-05,
+      "loss": 0.7935,
+      "step": 1168
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.6442672619148248e-05,
+      "loss": 0.877,
+      "step": 1169
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.6436334197303298e-05,
+      "loss": 0.8809,
+      "step": 1170
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 1.642999135770312e-05,
+      "loss": 0.7605,
+      "step": 1171
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.64236441047013e-05,
+      "loss": 0.8911,
+      "step": 1172
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.6417292442654445e-05,
+      "loss": 0.9565,
+      "step": 1173
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 1.64109363759222e-05,
+      "loss": 0.854,
+      "step": 1174
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.6404575908867224e-05,
+      "loss": 0.8579,
+      "step": 1175
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.6398211045855198e-05,
+      "loss": 0.8594,
+      "step": 1176
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.6391841791254818e-05,
+      "loss": 0.8755,
+      "step": 1177
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 1.63854681494378e-05,
+      "loss": 0.8267,
+      "step": 1178
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.6379090124778866e-05,
+      "loss": 0.813,
+      "step": 1179
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.6372707721655756e-05,
+      "loss": 0.7183,
+      "step": 1180
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.63663209444492e-05,
+      "loss": 0.8291,
+      "step": 1181
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 1.6359929797542945e-05,
+      "loss": 0.8496,
+      "step": 1182
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.6353534285323724e-05,
+      "loss": 0.7837,
+      "step": 1183
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.634713441218128e-05,
+      "loss": 0.9233,
+      "step": 1184
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.6340730182508332e-05,
+      "loss": 0.8738,
+      "step": 1185
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 1.6334321600700612e-05,
+      "loss": 0.8052,
+      "step": 1186
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.632790867115682e-05,
+      "loss": 0.7864,
+      "step": 1187
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.6321491398278647e-05,
+      "loss": 0.7563,
+      "step": 1188
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.6315069786470767e-05,
+      "loss": 0.7039,
+      "step": 1189
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 1.630864384014083e-05,
+      "loss": 0.8628,
+      "step": 1190
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.630221356369945e-05,
+      "loss": 0.7744,
+      "step": 1191
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.6295778961560242e-05,
+      "loss": 0.8481,
+      "step": 1192
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.6289340038139765e-05,
+      "loss": 0.8359,
+      "step": 1193
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 1.6282896797857547e-05,
+      "loss": 0.8711,
+      "step": 1194
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.627644924513609e-05,
+      "loss": 0.79,
+      "step": 1195
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.6269997384400845e-05,
+      "loss": 0.8706,
+      "step": 1196
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.6263541220080227e-05,
+      "loss": 0.9092,
+      "step": 1197
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 1.62570807566056e-05,
+      "loss": 0.6714,
+      "step": 1198
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.6250615998411285e-05,
+      "loss": 0.9668,
+      "step": 1199
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.624414694993454e-05,
+      "loss": 0.7788,
+      "step": 1200
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.6237673615615582e-05,
+      "loss": 0.8335,
+      "step": 1201
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 1.6231195999897557e-05,
+      "loss": 0.7358,
+      "step": 1202
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.6224714107226557e-05,
+      "loss": 0.8228,
+      "step": 1203
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.6218227942051606e-05,
+      "loss": 0.8186,
+      "step": 1204
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.6211737508824662e-05,
+      "loss": 0.8516,
+      "step": 1205
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 1.620524281200062e-05,
+      "loss": 0.832,
+      "step": 1206
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.6198743856037286e-05,
+      "loss": 0.7444,
+      "step": 1207
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.6192240645395396e-05,
+      "loss": 0.8364,
+      "step": 1208
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.6185733184538613e-05,
+      "loss": 0.8579,
+      "step": 1209
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 1.617922147793351e-05,
+      "loss": 0.8564,
+      "step": 1210
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.6172705530049575e-05,
+      "loss": 0.7773,
+      "step": 1211
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.616618534535921e-05,
+      "loss": 0.7939,
+      "step": 1212
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.6159660928337723e-05,
+      "loss": 0.866,
+      "step": 1213
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 1.6153132283463327e-05,
+      "loss": 0.7878,
+      "step": 1214
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.614659941521714e-05,
+      "loss": 0.7676,
+      "step": 1215
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.614006232808317e-05,
+      "loss": 0.73,
+      "step": 1216
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 1.613352102654833e-05,
+      "loss": 0.8384,
+      "step": 1217
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.6126975515102424e-05,
+      "loss": 0.8657,
+      "step": 1218
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.6120425798238145e-05,
+      "loss": 0.7578,
+      "step": 1219
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.6113871880451062e-05,
+      "loss": 0.8442,
+      "step": 1220
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 1.610731376623965e-05,
+      "loss": 0.7861,
+      "step": 1221
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.6100751460105244e-05,
+      "loss": 0.8032,
+      "step": 1222
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.6094184966552066e-05,
+      "loss": 0.8164,
+      "step": 1223
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.608761429008721e-05,
+      "loss": 0.7839,
+      "step": 1224
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 1.6081039435220637e-05,
+      "loss": 0.6987,
+      "step": 1225
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.607446040646518e-05,
+      "loss": 0.7471,
+      "step": 1226
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.606787720833654e-05,
+      "loss": 0.7285,
+      "step": 1227
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.6061289845353276e-05,
+      "loss": 0.8232,
+      "step": 1228
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 1.6054698322036806e-05,
+      "loss": 0.8086,
+      "step": 1229
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.60481026429114e-05,
+      "loss": 0.8843,
+      "step": 1230
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.6041502812504185e-05,
+      "loss": 0.8091,
+      "step": 1231
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.6034898835345142e-05,
+      "loss": 0.8281,
+      "step": 1232
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.6028290715967087e-05,
+      "loss": 0.7832,
+      "step": 1233
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.6021678458905686e-05,
+      "loss": 0.8091,
+      "step": 1234
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.6015062068699445e-05,
+      "loss": 0.8428,
+      "step": 1235
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.6008441549889703e-05,
+      "loss": 0.7095,
+      "step": 1236
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 1.6001816907020634e-05,
+      "loss": 0.7104,
+      "step": 1237
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.599518814463925e-05,
+      "loss": 0.8999,
+      "step": 1238
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.5988555267295383e-05,
+      "loss": 0.7412,
+      "step": 1239
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.598191827954169e-05,
+      "loss": 0.8037,
+      "step": 1240
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 1.597527718593365e-05,
+      "loss": 0.7939,
+      "step": 1241
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.5968631991029554e-05,
+      "loss": 0.6157,
+      "step": 1242
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.596198269939053e-05,
+      "loss": 0.8257,
+      "step": 1243
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.5955329315580487e-05,
+      "loss": 0.7573,
+      "step": 1244
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.5948671844166167e-05,
+      "loss": 0.7778,
+      "step": 1245
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.5942010289717108e-05,
+      "loss": 0.8374,
+      "step": 1246
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.5935344656805648e-05,
+      "loss": 0.6758,
+      "step": 1247
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.5928674950006927e-05,
+      "loss": 0.8599,
+      "step": 1248
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 1.592200117389889e-05,
+      "loss": 0.7627,
+      "step": 1249
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.5915323333062255e-05,
+      "loss": 0.7566,
+      "step": 1250
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.590864143208055e-05,
+      "loss": 0.8611,
+      "step": 1251
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.5901955475540087e-05,
+      "loss": 0.8374,
+      "step": 1252
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 1.5895265468029943e-05,
+      "loss": 0.7822,
+      "step": 1253
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.5888571414141997e-05,
+      "loss": 0.8325,
+      "step": 1254
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.5881873318470894e-05,
+      "loss": 0.8237,
+      "step": 1255
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.587517118561406e-05,
+      "loss": 0.6975,
+      "step": 1256
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 1.586846502017169e-05,
+      "loss": 0.8188,
+      "step": 1257
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.5861754826746737e-05,
+      "loss": 0.895,
+      "step": 1258
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.5855040609944935e-05,
+      "loss": 0.7163,
+      "step": 1259
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 1.5848322374374765e-05,
+      "loss": 0.8477,
+      "step": 1260
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.584160012464748e-05,
+      "loss": 0.8076,
+      "step": 1261
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.5834873865377075e-05,
+      "loss": 0.7153,
+      "step": 1262
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.5828143601180306e-05,
+      "loss": 0.6936,
+      "step": 1263
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 1.5821409336676676e-05,
+      "loss": 0.783,
+      "step": 1264
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.5814671076488434e-05,
+      "loss": 0.9185,
+      "step": 1265
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.5807928825240566e-05,
+      "loss": 0.7715,
+      "step": 1266
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.5801182587560807e-05,
+      "loss": 0.832,
+      "step": 1267
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 1.5794432368079617e-05,
+      "loss": 0.7622,
+      "step": 1268
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.57876781714302e-05,
+      "loss": 0.7705,
+      "step": 1269
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.5780920002248484e-05,
+      "loss": 0.7554,
+      "step": 1270
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.5774157865173125e-05,
+      "loss": 0.7642,
+      "step": 1271
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 1.5767391764845503e-05,
+      "loss": 0.7163,
+      "step": 1272
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.5760621705909706e-05,
+      "loss": 0.749,
+      "step": 1273
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.5753847693012566e-05,
+      "loss": 0.8594,
+      "step": 1274
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.5747069730803607e-05,
+      "loss": 0.8091,
+      "step": 1275
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 1.5740287823935067e-05,
+      "loss": 0.7168,
+      "step": 1276
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.5733501977061904e-05,
+      "loss": 0.7251,
+      "step": 1277
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.5726712194841758e-05,
+      "loss": 0.7866,
+      "step": 1278
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.571991848193499e-05,
+      "loss": 0.7783,
+      "step": 1279
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 1.5713120843004653e-05,
+      "loss": 0.8428,
+      "step": 1280
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.5706319282716492e-05,
+      "loss": 0.8252,
+      "step": 1281
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.5699513805738942e-05,
+      "loss": 0.7964,
+      "step": 1282
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.5692704416743138e-05,
+      "loss": 0.7905,
+      "step": 1283
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 1.568589112040288e-05,
+      "loss": 0.7778,
+      "step": 1284
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 1.567907392139467e-05,
+      "loss": 0.6895,
+      "step": 1285
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 1.5672252824397683e-05,
+      "loss": 0.7832,
+      "step": 1286
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 1.5665427834093755e-05,
+      "loss": 0.8027,
+      "step": 1287
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 1.565859895516742e-05,
+      "loss": 0.8506,
+      "step": 1288
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 1.5651766192305853e-05,
+      "loss": 0.7642,
+      "step": 1289
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 1.564492955019892e-05,
+      "loss": 0.7024,
+      "step": 1290
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 1.5638089033539135e-05,
+      "loss": 0.8364,
+      "step": 1291
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 1.5631244647021672e-05,
+      "loss": 0.7473,
+      "step": 1292
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 1.562439639534437e-05,
+      "loss": 0.7788,
+      "step": 1293
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 1.561754428320771e-05,
+      "loss": 0.7993,
+      "step": 1294
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 1.561068831531483e-05,
+      "loss": 0.718,
+      "step": 1295
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 1.560382849637151e-05,
+      "loss": 0.7837,
+      "step": 1296
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 1.5596964831086185e-05,
+      "loss": 0.8154,
+      "step": 1297
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 1.5590097324169907e-05,
+      "loss": 0.8223,
+      "step": 1298
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 1.558322598033639e-05,
+      "loss": 0.812,
+      "step": 1299
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 1.557635080430196e-05,
+      "loss": 0.7817,
+      "step": 1300
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 1.556947180078559e-05,
+      "loss": 0.7476,
+      "step": 1301
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 1.556258897450887e-05,
+      "loss": 0.8169,
+      "step": 1302
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 1.5555702330196024e-05,
+      "loss": 0.8804,
+      "step": 1303
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 1.5548811872573883e-05,
+      "loss": 0.7959,
+      "step": 1304
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 1.5541917606371904e-05,
+      "loss": 0.8911,
+      "step": 1305
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 1.5535019536322158e-05,
+      "loss": 0.7607,
+      "step": 1306
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 1.5528117667159325e-05,
+      "loss": 0.7827,
+      "step": 1307
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.5521212003620692e-05,
+      "loss": 0.7219,
+      "step": 1308
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.5514302550446153e-05,
+      "loss": 0.8174,
+      "step": 1309
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.55073893123782e-05,
+      "loss": 0.7856,
+      "step": 1310
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.550047229416192e-05,
+      "loss": 0.7922,
+      "step": 1311
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 1.549355150054501e-05,
+      "loss": 0.7065,
+      "step": 1312
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 1.5486626936277737e-05,
+      "loss": 0.8335,
+      "step": 1313
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 1.547969860611297e-05,
+      "loss": 0.8364,
+      "step": 1314
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 1.547276651480616e-05,
+      "loss": 0.7998,
+      "step": 1315
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 1.546583066711534e-05,
+      "loss": 0.8062,
+      "step": 1316
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 1.545889106780112e-05,
+      "loss": 0.7502,
+      "step": 1317
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 1.5451947721626676e-05,
+      "loss": 0.8716,
+      "step": 1318
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 1.544500063335778e-05,
+      "loss": 0.7891,
+      "step": 1319
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 1.5438049807762746e-05,
+      "loss": 0.7512,
+      "step": 1320
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 1.5431095249612466e-05,
+      "loss": 0.7268,
+      "step": 1321
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 1.5424136963680398e-05,
+      "loss": 0.7891,
+      "step": 1322
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 1.541717495474255e-05,
+      "loss": 0.7251,
+      "step": 1323
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 1.5410209227577487e-05,
+      "loss": 0.7297,
+      "step": 1324
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 1.540323978696633e-05,
+      "loss": 0.71,
+      "step": 1325
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 1.5396266637692746e-05,
+      "loss": 0.8359,
+      "step": 1326
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 1.5389289784542944e-05,
+      "loss": 0.7979,
+      "step": 1327
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.5382309232305685e-05,
+      "loss": 0.7668,
+      "step": 1328
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.5375324985772262e-05,
+      "loss": 0.8032,
+      "step": 1329
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.5368337049736505e-05,
+      "loss": 0.8589,
+      "step": 1330
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 1.5361345428994772e-05,
+      "loss": 0.8291,
+      "step": 1331
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 1.535435012834596e-05,
+      "loss": 0.856,
+      "step": 1332
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 1.5347351152591485e-05,
+      "loss": 0.7417,
+      "step": 1333
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 1.534034850653528e-05,
+      "loss": 0.7866,
+      "step": 1334
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 1.533334219498382e-05,
+      "loss": 0.7615,
+      "step": 1335
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 1.532633222274606e-05,
+      "loss": 0.7573,
+      "step": 1336
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 1.53193185946335e-05,
+      "loss": 0.8438,
+      "step": 1337
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 1.531230131546014e-05,
+      "loss": 0.7378,
+      "step": 1338
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 1.530528039004247e-05,
+      "loss": 0.8198,
+      "step": 1339
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 1.5298255823199503e-05,
+      "loss": 0.7983,
+      "step": 1340
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 1.5291227619752748e-05,
+      "loss": 0.8198,
+      "step": 1341
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 1.5284195784526196e-05,
+      "loss": 0.6978,
+      "step": 1342
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 1.527716032234635e-05,
+      "loss": 0.7412,
+      "step": 1343
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 1.5270121238042186e-05,
+      "loss": 0.7612,
+      "step": 1344
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 1.5263078536445176e-05,
+      "loss": 0.7551,
+      "step": 1345
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 1.5256032222389276e-05,
+      "loss": 0.8027,
+      "step": 1346
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 1.5248982300710913e-05,
+      "loss": 0.7271,
+      "step": 1347
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 1.5241928776248993e-05,
+      "loss": 0.7617,
+      "step": 1348
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 1.5234871653844904e-05,
+      "loss": 0.7104,
+      "step": 1349
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 1.5227810938342493e-05,
+      "loss": 0.7573,
+      "step": 1350
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 1.5220746634588076e-05,
+      "loss": 0.689,
+      "step": 1351
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 1.5213678747430429e-05,
+      "loss": 0.6653,
+      "step": 1352
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 1.5206607281720797e-05,
+      "loss": 0.7759,
+      "step": 1353
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 1.5199532242312872e-05,
+      "loss": 0.718,
+      "step": 1354
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.51924536340628e-05,
+      "loss": 0.7773,
+      "step": 1355
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.5185371461829186e-05,
+      "loss": 0.7412,
+      "step": 1356
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.5178285730473068e-05,
+      "loss": 0.7717,
+      "step": 1357
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 1.5171196444857933e-05,
+      "loss": 0.8047,
+      "step": 1358
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 1.516410360984971e-05,
+      "loss": 0.7383,
+      "step": 1359
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 1.5157007230316757e-05,
+      "loss": 0.7852,
+      "step": 1360
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 1.5149907311129877e-05,
+      "loss": 0.8149,
+      "step": 1361
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 1.5142803857162289e-05,
+      "loss": 0.6367,
+      "step": 1362
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 1.5135696873289648e-05,
+      "loss": 0.6924,
+      "step": 1363
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 1.5128586364390021e-05,
+      "loss": 0.7419,
+      "step": 1364
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 1.5121472335343912e-05,
+      "loss": 0.8442,
+      "step": 1365
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 1.5114354791034225e-05,
+      "loss": 0.8228,
+      "step": 1366
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 1.5107233736346278e-05,
+      "loss": 0.7349,
+      "step": 1367
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 1.5100109176167814e-05,
+      "loss": 0.7754,
+      "step": 1368
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 1.5092981115388963e-05,
+      "loss": 0.835,
+      "step": 1369
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 1.5085849558902265e-05,
+      "loss": 0.8809,
+      "step": 1370
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 1.5078714511602667e-05,
+      "loss": 0.8647,
+      "step": 1371
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 1.5071575978387505e-05,
+      "loss": 0.8022,
+      "step": 1372
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 1.5064433964156497e-05,
+      "loss": 0.728,
+      "step": 1373
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 1.5057288473811773e-05,
+      "loss": 0.6687,
+      "step": 1374
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 1.505013951225783e-05,
+      "loss": 0.6501,
+      "step": 1375
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 1.5042987084401559e-05,
+      "loss": 0.8389,
+      "step": 1376
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 1.5035831195152221e-05,
+      "loss": 0.7341,
+      "step": 1377
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 1.5028671849421463e-05,
+      "loss": 0.7852,
+      "step": 1378
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 1.5021509052123295e-05,
+      "loss": 0.7432,
+      "step": 1379
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 1.5014342808174102e-05,
+      "loss": 0.8086,
+      "step": 1380
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 1.5007173122492635e-05,
+      "loss": 0.6887,
+      "step": 1381
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 1.5000000000000002e-05,
+      "loss": 0.7532,
+      "step": 1382
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.4992823445619672e-05,
+      "loss": 0.7427,
+      "step": 1383
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.4985643464277475e-05,
+      "loss": 0.7324,
+      "step": 1384
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.4978460060901587e-05,
+      "loss": 0.7568,
+      "step": 1385
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 1.4971273240422534e-05,
+      "loss": 0.7478,
+      "step": 1386
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.4964083007773189e-05,
+      "loss": 0.7969,
+      "step": 1387
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.4956889367888765e-05,
+      "loss": 0.7385,
+      "step": 1388
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.4949692325706818e-05,
+      "loss": 0.8105,
+      "step": 1389
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.494249188616723e-05,
+      "loss": 0.7476,
+      "step": 1390
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 1.4935288054212228e-05,
+      "loss": 0.8843,
+      "step": 1391
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 1.4928080834786356e-05,
+      "loss": 0.7651,
+      "step": 1392
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 1.4920870232836487e-05,
+      "loss": 0.8452,
+      "step": 1393
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.491365625331182e-05,
+      "loss": 0.748,
+      "step": 1394
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.490643890116387e-05,
+      "loss": 0.7842,
+      "step": 1395
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.4899218181346453e-05,
+      "loss": 0.7134,
+      "step": 1396
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 1.4891994098815725e-05,
+      "loss": 0.7759,
+      "step": 1397
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 1.4884766658530126e-05,
+      "loss": 0.6924,
+      "step": 1398
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 1.4877535865450405e-05,
+      "loss": 0.7349,
+      "step": 1399
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 1.4870301724539627e-05,
+      "loss": 0.6887,
+      "step": 1400
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 1.4863064240763132e-05,
+      "loss": 0.7371,
+      "step": 1401
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 1.4855823419088576e-05,
+      "loss": 0.6956,
+      "step": 1402
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 1.4848579264485887e-05,
+      "loss": 0.7993,
+      "step": 1403
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 1.4841331781927297e-05,
+      "loss": 0.7751,
+      "step": 1404
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 1.4834080976387316e-05,
+      "loss": 0.657,
+      "step": 1405
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 1.4826826852842725e-05,
+      "loss": 0.7332,
+      "step": 1406
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 1.4819569416272604e-05,
+      "loss": 0.8135,
+      "step": 1407
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 1.4812308671658285e-05,
+      "loss": 0.7744,
+      "step": 1408
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 1.4805044623983382e-05,
+      "loss": 0.7458,
+      "step": 1409
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.4797777278233779e-05,
+      "loss": 0.7559,
+      "step": 1410
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.4790506639397613e-05,
+      "loss": 0.7393,
+      "step": 1411
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.4783232712465287e-05,
+      "loss": 0.645,
+      "step": 1412
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 1.4775955502429467e-05,
+      "loss": 0.7639,
+      "step": 1413
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 1.4768675014285063e-05,
+      "loss": 0.7549,
+      "step": 1414
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 1.4761391253029236e-05,
+      "loss": 0.7715,
+      "step": 1415
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 1.47541042236614e-05,
+      "loss": 0.708,
+      "step": 1416
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 1.4746813931183209e-05,
+      "loss": 0.7161,
+      "step": 1417
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 1.473952038059855e-05,
+      "loss": 0.657,
+      "step": 1418
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 1.473222357691356e-05,
+      "loss": 0.8203,
+      "step": 1419
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 1.4724923525136597e-05,
+      "loss": 0.7454,
+      "step": 1420
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 1.4717620230278249e-05,
+      "loss": 0.812,
+      "step": 1421
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 1.4710313697351342e-05,
+      "loss": 0.6704,
+      "step": 1422
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 1.470300393137091e-05,
+      "loss": 0.7485,
+      "step": 1423
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 1.4695690937354212e-05,
+      "loss": 0.8604,
+      "step": 1424
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 1.4688374720320729e-05,
+      "loss": 0.7847,
+      "step": 1425
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 1.4681055285292138e-05,
+      "loss": 0.7573,
+      "step": 1426
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 1.4673732637292344e-05,
+      "loss": 0.8145,
+      "step": 1427
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 1.4666406781347443e-05,
+      "loss": 0.7566,
+      "step": 1428
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 1.4659077722485741e-05,
+      "loss": 0.7905,
+      "step": 1429
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 1.465174546573774e-05,
+      "loss": 0.8135,
+      "step": 1430
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 1.4644410016136132e-05,
+      "loss": 0.6963,
+      "step": 1431
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 1.4637071378715807e-05,
+      "loss": 0.6687,
+      "step": 1432
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 1.4629729558513841e-05,
+      "loss": 0.7412,
+      "step": 1433
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 1.4622384560569493e-05,
+      "loss": 0.7612,
+      "step": 1434
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 1.4615036389924209e-05,
+      "loss": 0.8069,
+      "step": 1435
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 1.4607685051621603e-05,
+      "loss": 0.7964,
+      "step": 1436
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 1.4600330550707473e-05,
+      "loss": 0.7998,
+      "step": 1437
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 1.4592972892229779e-05,
+      "loss": 0.7637,
+      "step": 1438
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 1.4585612081238656e-05,
+      "loss": 0.7979,
+      "step": 1439
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 1.4578248122786401e-05,
+      "loss": 0.6654,
+      "step": 1440
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 1.4570881021927461e-05,
+      "loss": 0.6729,
+      "step": 1441
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 1.4563510783718457e-05,
+      "loss": 0.8164,
+      "step": 1442
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 1.4556137413218153e-05,
+      "loss": 0.7349,
+      "step": 1443
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 1.4548760915487463e-05,
+      "loss": 0.8032,
+      "step": 1444
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 1.454138129558945e-05,
+      "loss": 0.7856,
+      "step": 1445
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 1.4533998558589319e-05,
+      "loss": 0.7827,
+      "step": 1446
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 1.452661270955442e-05,
+      "loss": 0.5737,
+      "step": 1447
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 1.4519223753554223e-05,
+      "loss": 0.6758,
+      "step": 1448
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 1.4511831695660354e-05,
+      "loss": 0.7183,
+      "step": 1449
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 1.4504436540946548e-05,
+      "loss": 0.7573,
+      "step": 1450
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 1.449703829448867e-05,
+      "loss": 0.7012,
+      "step": 1451
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 1.4489636961364719e-05,
+      "loss": 0.7139,
+      "step": 1452
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 1.4482232546654799e-05,
+      "loss": 0.7104,
+      "step": 1453
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 1.4474825055441135e-05,
+      "loss": 0.769,
+      "step": 1454
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 1.4467414492808061e-05,
+      "loss": 0.6653,
+      "step": 1455
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 1.4460000863842023e-05,
+      "loss": 0.6649,
+      "step": 1456
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 1.445258417363157e-05,
+      "loss": 0.769,
+      "step": 1457
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 1.4445164427267344e-05,
+      "loss": 0.6814,
+      "step": 1458
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 1.4437741629842103e-05,
+      "loss": 0.6948,
+      "step": 1459
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 1.4430315786450683e-05,
+      "loss": 0.7686,
+      "step": 1460
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 1.4422886902190014e-05,
+      "loss": 0.7896,
+      "step": 1461
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 1.4415454982159121e-05,
+      "loss": 0.7324,
+      "step": 1462
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 1.4408020031459103e-05,
+      "loss": 0.7227,
+      "step": 1463
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 1.4400582055193143e-05,
+      "loss": 0.6926,
+      "step": 1464
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.4393141058466502e-05,
+      "loss": 0.7046,
+      "step": 1465
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.4385697046386513e-05,
+      "loss": 0.697,
+      "step": 1466
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.4378250024062577e-05,
+      "loss": 0.7383,
+      "step": 1467
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.4370799996606169e-05,
+      "loss": 0.7803,
+      "step": 1468
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 1.4363346969130813e-05,
+      "loss": 0.7256,
+      "step": 1469
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 1.4355890946752102e-05,
+      "loss": 0.7837,
+      "step": 1470
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 1.4348431934587685e-05,
+      "loss": 0.7136,
+      "step": 1471
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 1.4340969937757256e-05,
+      "loss": 0.6628,
+      "step": 1472
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 1.4333504961382566e-05,
+      "loss": 0.6223,
+      "step": 1473
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 1.4326037010587406e-05,
+      "loss": 0.6682,
+      "step": 1474
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 1.4318566090497607e-05,
+      "loss": 0.7556,
+      "step": 1475
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 1.4311092206241046e-05,
+      "loss": 0.7417,
+      "step": 1476
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 1.4303615362947623e-05,
+      "loss": 0.7446,
+      "step": 1477
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 1.429613556574928e-05,
+      "loss": 0.7407,
+      "step": 1478
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 1.4288652819779982e-05,
+      "loss": 0.6567,
+      "step": 1479
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 1.4281167130175712e-05,
+      "loss": 0.6912,
+      "step": 1480
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 1.4273678502074487e-05,
+      "loss": 0.7041,
+      "step": 1481
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 1.426618694061633e-05,
+      "loss": 0.7051,
+      "step": 1482
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 1.4258692450943276e-05,
+      "loss": 0.6965,
+      "step": 1483
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 1.425119503819938e-05,
+      "loss": 0.7107,
+      "step": 1484
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 1.4243694707530696e-05,
+      "loss": 0.7214,
+      "step": 1485
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 1.4236191464085286e-05,
+      "loss": 0.658,
+      "step": 1486
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 1.4228685313013198e-05,
+      "loss": 0.6162,
+      "step": 1487
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 1.4221176259466493e-05,
+      "loss": 0.7896,
+      "step": 1488
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 1.421366430859922e-05,
+      "loss": 0.6829,
+      "step": 1489
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 1.4206149465567402e-05,
+      "loss": 0.833,
+      "step": 1490
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 1.419863173552907e-05,
+      "loss": 0.7014,
+      "step": 1491
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 1.419111112364422e-05,
+      "loss": 0.6421,
+      "step": 1492
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 1.4183587635074831e-05,
+      "loss": 0.7197,
+      "step": 1493
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 1.4176061274984858e-05,
+      "loss": 0.615,
+      "step": 1494
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 1.4168532048540225e-05,
+      "loss": 0.8335,
+      "step": 1495
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 1.4160999960908818e-05,
+      "loss": 0.8777,
+      "step": 1496
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 1.4153465017260504e-05,
+      "loss": 0.6807,
+      "step": 1497
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 1.414592722276709e-05,
+      "loss": 0.6212,
+      "step": 1498
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 1.4138386582602353e-05,
+      "loss": 0.7881,
+      "step": 1499
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 1.4130843101942017e-05,
+      "loss": 0.717,
+      "step": 1500
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 1.412329678596376e-05,
+      "loss": 0.6433,
+      "step": 1501
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 1.4115747639847204e-05,
+      "loss": 0.6875,
+      "step": 1502
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 1.410819566877391e-05,
+      "loss": 0.8037,
+      "step": 1503
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 1.4100640877927383e-05,
+      "loss": 0.6604,
+      "step": 1504
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 1.4093083272493061e-05,
+      "loss": 0.8123,
+      "step": 1505
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 1.4085522857658321e-05,
+      "loss": 0.6812,
+      "step": 1506
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 1.4077959638612451e-05,
+      "loss": 0.6414,
+      "step": 1507
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.4070393620546684e-05,
+      "loss": 0.7056,
+      "step": 1508
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.4062824808654158e-05,
+      "loss": 0.5588,
+      "step": 1509
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.405525320812994e-05,
+      "loss": 0.6592,
+      "step": 1510
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 1.4047678824171005e-05,
+      "loss": 0.5311,
+      "step": 1511
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.4040101661976239e-05,
+      "loss": 0.7866,
+      "step": 1512
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.4032521726746438e-05,
+      "loss": 0.6108,
+      "step": 1513
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.4024939023684298e-05,
+      "loss": 0.6807,
+      "step": 1514
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.401735355799442e-05,
+      "loss": 0.6873,
+      "step": 1515
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.4009765334883287e-05,
+      "loss": 0.7549,
+      "step": 1516
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.4002174359559298e-05,
+      "loss": 0.7068,
+      "step": 1517
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.3994580637232717e-05,
+      "loss": 0.7146,
+      "step": 1518
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 1.3986984173115709e-05,
+      "loss": 0.634,
+      "step": 1519
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.397938497242232e-05,
+      "loss": 0.7407,
+      "step": 1520
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.3971783040368465e-05,
+      "loss": 0.6768,
+      "step": 1521
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 1.3964178382171942e-05,
+      "loss": 0.6326,
+      "step": 1522
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.3956571003052416e-05,
+      "loss": 0.8008,
+      "step": 1523
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.3948960908231422e-05,
+      "loss": 0.8242,
+      "step": 1524
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.394134810293236e-05,
+      "loss": 0.7393,
+      "step": 1525
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 1.3933732592380485e-05,
+      "loss": 0.7529,
+      "step": 1526
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.3926114381802915e-05,
+      "loss": 0.7585,
+      "step": 1527
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.3918493476428618e-05,
+      "loss": 0.7734,
+      "step": 1528
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.391086988148841e-05,
+      "loss": 0.7214,
+      "step": 1529
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 1.390324360221496e-05,
+      "loss": 0.752,
+      "step": 1530
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.3895614643842773e-05,
+      "loss": 0.666,
+      "step": 1531
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.3887983011608192e-05,
+      "loss": 0.7136,
+      "step": 1532
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.3880348710749401e-05,
+      "loss": 0.7676,
+      "step": 1533
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 1.3872711746506413e-05,
+      "loss": 0.6753,
+      "step": 1534
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 1.3865072124121066e-05,
+      "loss": 0.7947,
+      "step": 1535
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 1.3857429848837022e-05,
+      "loss": 0.6772,
+      "step": 1536
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 1.3849784925899779e-05,
+      "loss": 0.814,
+      "step": 1537
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 1.3842137360556629e-05,
+      "loss": 0.571,
+      "step": 1538
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.3834487158056692e-05,
+      "loss": 0.6794,
+      "step": 1539
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.3826834323650899e-05,
+      "loss": 0.7275,
+      "step": 1540
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.3819178862591982e-05,
+      "loss": 0.7383,
+      "step": 1541
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.3811520780134471e-05,
+      "loss": 0.6936,
+      "step": 1542
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 1.380386008153471e-05,
+      "loss": 0.6128,
+      "step": 1543
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 1.3796196772050829e-05,
+      "loss": 0.5764,
+      "step": 1544
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 1.3788530856942747e-05,
+      "loss": 0.6895,
+      "step": 1545
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 1.3780862341472183e-05,
+      "loss": 0.6487,
+      "step": 1546
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 1.3773191230902629e-05,
+      "loss": 0.6201,
+      "step": 1547
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 1.3765517530499364e-05,
+      "loss": 0.6252,
+      "step": 1548
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 1.3757841245529442e-05,
+      "loss": 0.8218,
+      "step": 1549
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 1.3750162381261693e-05,
+      "loss": 0.8101,
+      "step": 1550
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 1.3742480942966721e-05,
+      "loss": 0.6121,
+      "step": 1551
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 1.3734796935916888e-05,
+      "loss": 0.7039,
+      "step": 1552
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 1.3727110365386325e-05,
+      "loss": 0.687,
+      "step": 1553
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 1.3719421236650923e-05,
+      "loss": 0.7622,
+      "step": 1554
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 1.3711729554988324e-05,
+      "loss": 0.5891,
+      "step": 1555
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 1.3704035325677931e-05,
+      "loss": 0.7998,
+      "step": 1556
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 1.3696338554000885e-05,
+      "loss": 0.7021,
+      "step": 1557
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 1.3688639245240078e-05,
+      "loss": 0.7842,
+      "step": 1558
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.3680937404680147e-05,
+      "loss": 0.6749,
+      "step": 1559
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.3673233037607458e-05,
+      "loss": 0.6709,
+      "step": 1560
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.3665526149310115e-05,
+      "loss": 0.6052,
+      "step": 1561
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 1.3657816745077957e-05,
+      "loss": 0.7354,
+      "step": 1562
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 1.3650104830202542e-05,
+      "loss": 0.7632,
+      "step": 1563
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 1.3642390409977156e-05,
+      "loss": 0.7544,
+      "step": 1564
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 1.3634673489696804e-05,
+      "loss": 0.6626,
+      "step": 1565
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 1.362695407465821e-05,
+      "loss": 0.6589,
+      "step": 1566
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 1.3619232170159803e-05,
+      "loss": 0.6982,
+      "step": 1567
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 1.3611507781501722e-05,
+      "loss": 0.7488,
+      "step": 1568
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 1.3603780913985825e-05,
+      "loss": 0.7974,
+      "step": 1569
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.359605157291565e-05,
+      "loss": 0.6653,
+      "step": 1570
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.3588319763596443e-05,
+      "loss": 0.603,
+      "step": 1571
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.3580585491335151e-05,
+      "loss": 0.6865,
+      "step": 1572
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 1.3572848761440403e-05,
+      "loss": 0.7739,
+      "step": 1573
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.3565109579222511e-05,
+      "loss": 0.7461,
+      "step": 1574
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.3557367949993483e-05,
+      "loss": 0.7378,
+      "step": 1575
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.3549623879066996e-05,
+      "loss": 0.6995,
+      "step": 1576
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 1.3541877371758405e-05,
+      "loss": 0.6643,
+      "step": 1577
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.3534128433384741e-05,
+      "loss": 0.7915,
+      "step": 1578
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.35263770692647e-05,
+      "loss": 0.7495,
+      "step": 1579
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.3518623284718644e-05,
+      "loss": 0.6165,
+      "step": 1580
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 1.3510867085068599e-05,
+      "loss": 0.7959,
+      "step": 1581
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.3503108475638244e-05,
+      "loss": 0.6819,
+      "step": 1582
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.3495347461752914e-05,
+      "loss": 0.7021,
+      "step": 1583
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.3487584048739598e-05,
+      "loss": 0.8359,
+      "step": 1584
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 1.3479818241926921e-05,
+      "loss": 0.6836,
+      "step": 1585
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.3472050046645167e-05,
+      "loss": 0.6768,
+      "step": 1586
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.3464279468226247e-05,
+      "loss": 0.6753,
+      "step": 1587
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.3456506512003707e-05,
+      "loss": 0.6426,
+      "step": 1588
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 1.3448731183312738e-05,
+      "loss": 0.7358,
+      "step": 1589
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.3440953487490145e-05,
+      "loss": 0.7725,
+      "step": 1590
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.3433173429874367e-05,
+      "loss": 0.7114,
+      "step": 1591
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.3425391015805458e-05,
+      "loss": 0.8379,
+      "step": 1592
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 1.3417606250625096e-05,
+      "loss": 0.676,
+      "step": 1593
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.3409819139676562e-05,
+      "loss": 0.7285,
+      "step": 1594
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.340202968830476e-05,
+      "loss": 0.752,
+      "step": 1595
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.3394237901856195e-05,
+      "loss": 0.751,
+      "step": 1596
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 1.3386443785678972e-05,
+      "loss": 0.7324,
+      "step": 1597
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.3378647345122795e-05,
+      "loss": 0.6755,
+      "step": 1598
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.337084858553897e-05,
+      "loss": 0.6721,
+      "step": 1599
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.3363047512280391e-05,
+      "loss": 0.7812,
+      "step": 1600
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 1.3355244130701532e-05,
+      "loss": 0.7036,
+      "step": 1601
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.3347438446158467e-05,
+      "loss": 0.6262,
+      "step": 1602
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.3339630464008838e-05,
+      "loss": 0.6958,
+      "step": 1603
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.3331820189611872e-05,
+      "loss": 0.6514,
+      "step": 1604
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 1.3324007628328363e-05,
+      "loss": 0.7446,
+      "step": 1605
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.331619278552068e-05,
+      "loss": 0.6685,
+      "step": 1606
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.3308375666552752e-05,
+      "loss": 0.6509,
+      "step": 1607
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 1.3300556276790078e-05,
+      "loss": 0.7302,
+      "step": 1608
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.3292734621599708e-05,
+      "loss": 0.6223,
+      "step": 1609
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.3284910706350249e-05,
+      "loss": 0.8291,
+      "step": 1610
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.3277084536411869e-05,
+      "loss": 0.6552,
+      "step": 1611
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 1.326925611715627e-05,
+      "loss": 0.7754,
+      "step": 1612
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.3261425453956698e-05,
+      "loss": 0.6899,
+      "step": 1613
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.325359255218795e-05,
+      "loss": 0.6699,
+      "step": 1614
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.3245757417226356e-05,
+      "loss": 0.7456,
+      "step": 1615
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 1.3237920054449775e-05,
+      "loss": 0.7084,
+      "step": 1616
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.3230080469237592e-05,
+      "loss": 0.6382,
+      "step": 1617
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.3222238666970728e-05,
+      "loss": 0.6672,
+      "step": 1618
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.3214394653031616e-05,
+      "loss": 0.6899,
+      "step": 1619
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 1.3206548432804213e-05,
+      "loss": 0.761,
+      "step": 1620
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.319870001167399e-05,
+      "loss": 0.7754,
+      "step": 1621
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.3190849395027926e-05,
+      "loss": 0.7324,
+      "step": 1622
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.3182996588254503e-05,
+      "loss": 0.7295,
+      "step": 1623
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 1.317514159674372e-05,
+      "loss": 0.751,
+      "step": 1624
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.316728442588706e-05,
+      "loss": 0.7061,
+      "step": 1625
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.315942508107751e-05,
+      "loss": 0.7061,
+      "step": 1626
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.3151563567709547e-05,
+      "loss": 0.7268,
+      "step": 1627
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 1.314369989117914e-05,
+      "loss": 0.6807,
+      "step": 1628
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.3135834056883738e-05,
+      "loss": 0.6423,
+      "step": 1629
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.3127966070222273e-05,
+      "loss": 0.7422,
+      "step": 1630
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.3120095936595156e-05,
+      "loss": 0.657,
+      "step": 1631
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 1.3112223661404268e-05,
+      "loss": 0.7842,
+      "step": 1632
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.310434925005296e-05,
+      "loss": 0.6626,
+      "step": 1633
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.3096472707946057e-05,
+      "loss": 0.6328,
+      "step": 1634
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.3088594040489837e-05,
+      "loss": 0.5903,
+      "step": 1635
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 1.3080713253092039e-05,
+      "loss": 0.7542,
+      "step": 1636
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.3072830351161862e-05,
+      "loss": 0.6519,
+      "step": 1637
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.306494534010995e-05,
+      "loss": 0.686,
+      "step": 1638
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.30570582253484e-05,
+      "loss": 0.665,
+      "step": 1639
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 1.3049169012290745e-05,
+      "loss": 0.8945,
+      "step": 1640
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.304127770635197e-05,
+      "loss": 0.8401,
+      "step": 1641
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3033384312948487e-05,
+      "loss": 0.7681,
+      "step": 1642
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3025488837498144e-05,
+      "loss": 0.7041,
+      "step": 1643
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 1.3017591285420219e-05,
+      "loss": 0.7104,
+      "step": 1644
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.3009691662135415e-05,
+      "loss": 0.6011,
+      "step": 1645
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.3001789973065852e-05,
+      "loss": 0.5583,
+      "step": 1646
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.2993886223635076e-05,
+      "loss": 0.7095,
+      "step": 1647
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 1.2985980419268045e-05,
+      "loss": 0.7314,
+      "step": 1648
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2978072565391118e-05,
+      "loss": 0.7209,
+      "step": 1649
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2970162667432075e-05,
+      "loss": 0.7544,
+      "step": 1650
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 1.2962250730820092e-05,
+      "loss": 0.7283,
+      "step": 1651
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.2954336760985737e-05,
+      "loss": 0.7688,
+      "step": 1652
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.294642076336099e-05,
+      "loss": 0.6667,
+      "step": 1653
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.2938502743379212e-05,
+      "loss": 0.6597,
+      "step": 1654
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 1.293058270647515e-05,
+      "loss": 0.6045,
+      "step": 1655
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2922660658084945e-05,
+      "loss": 0.6909,
+      "step": 1656
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2914736603646108e-05,
+      "loss": 0.7515,
+      "step": 1657
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.2906810548597533e-05,
+      "loss": 0.657,
+      "step": 1658
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 1.289888249837949e-05,
+      "loss": 0.8369,
+      "step": 1659
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.289095245843361e-05,
+      "loss": 0.6282,
+      "step": 1660
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.2883020434202897e-05,
+      "loss": 0.7441,
+      "step": 1661
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.2875086431131716e-05,
+      "loss": 0.6887,
+      "step": 1662
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 1.2867150454665783e-05,
+      "loss": 0.677,
+      "step": 1663
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.2859212510252176e-05,
+      "loss": 0.7056,
+      "step": 1664
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.2851272603339329e-05,
+      "loss": 0.7266,
+      "step": 1665
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.2843330739377003e-05,
+      "loss": 0.6804,
+      "step": 1666
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 1.2835386923816324e-05,
+      "loss": 0.7397,
+      "step": 1667
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.2827441162109749e-05,
+      "loss": 0.7742,
+      "step": 1668
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.2819493459711061e-05,
+      "loss": 0.668,
+      "step": 1669
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.2811543822075396e-05,
+      "loss": 0.676,
+      "step": 1670
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 1.2803592254659202e-05,
+      "loss": 0.6807,
+      "step": 1671
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.2795638762920254e-05,
+      "loss": 0.7207,
+      "step": 1672
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.2787683352317652e-05,
+      "loss": 0.7712,
+      "step": 1673
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.2779726028311812e-05,
+      "loss": 0.7026,
+      "step": 1674
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 1.2771766796364462e-05,
+      "loss": 0.585,
+      "step": 1675
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.276380566193864e-05,
+      "loss": 0.7402,
+      "step": 1676
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.2755842630498689e-05,
+      "loss": 0.5818,
+      "step": 1677
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.2747877707510252e-05,
+      "loss": 0.7969,
+      "step": 1678
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 1.2739910898440282e-05,
+      "loss": 0.6223,
+      "step": 1679
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.2731942208757012e-05,
+      "loss": 0.6431,
+      "step": 1680
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.272397164392997e-05,
+      "loss": 0.6504,
+      "step": 1681
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.2715999209429974e-05,
+      "loss": 0.6643,
+      "step": 1682
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 1.2708024910729127e-05,
+      "loss": 0.6472,
+      "step": 1683
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.2700048753300806e-05,
+      "loss": 0.6787,
+      "step": 1684
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.2692070742619665e-05,
+      "loss": 0.6714,
+      "step": 1685
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.2684090884161636e-05,
+      "loss": 0.6228,
+      "step": 1686
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 1.2676109183403909e-05,
+      "loss": 0.7043,
+      "step": 1687
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 1.2668125645824945e-05,
+      "loss": 0.657,
+      "step": 1688
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 1.2660140276904471e-05,
+      "loss": 0.6919,
+      "step": 1689
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 1.2652153082123458e-05,
+      "loss": 0.7666,
+      "step": 1690
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 1.2644164066964138e-05,
+      "loss": 0.7705,
+      "step": 1691
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 1.2636173236909995e-05,
+      "loss": 0.6836,
+      "step": 1692
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 1.2628180597445754e-05,
+      "loss": 0.8271,
+      "step": 1693
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 1.2620186154057383e-05,
+      "loss": 0.6641,
+      "step": 1694
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 1.2612189912232084e-05,
+      "loss": 0.6841,
+      "step": 1695
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 1.2604191877458307e-05,
+      "loss": 0.6626,
+      "step": 1696
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 1.2596192055225716e-05,
+      "loss": 0.6545,
+      "step": 1697
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 1.2588190451025209e-05,
+      "loss": 0.6975,
+      "step": 1698
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 1.2580187070348913e-05,
+      "loss": 0.5823,
+      "step": 1699
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 1.2572181918690162e-05,
+      "loss": 0.6772,
+      "step": 1700
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 1.2564175001543517e-05,
+      "loss": 0.7156,
+      "step": 1701
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 1.2556166324404747e-05,
+      "loss": 0.6973,
+      "step": 1702
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 1.2548155892770825e-05,
+      "loss": 0.7244,
+      "step": 1703
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 1.2540143712139935e-05,
+      "loss": 0.8359,
+      "step": 1704
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 1.2532129788011452e-05,
+      "loss": 0.7,
+      "step": 1705
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 1.2524114125885958e-05,
+      "loss": 0.6753,
+      "step": 1706
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 1.2516096731265222e-05,
+      "loss": 0.7104,
+      "step": 1707
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 1.2508077609652203e-05,
+      "loss": 0.7224,
+      "step": 1708
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 1.2500056766551047e-05,
+      "loss": 0.7861,
+      "step": 1709
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 1.2492034207467081e-05,
+      "loss": 0.6631,
+      "step": 1710
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 1.2484009937906807e-05,
+      "loss": 0.7222,
+      "step": 1711
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 1.2475983963377907e-05,
+      "loss": 0.6235,
+      "step": 1712
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 1.2467956289389228e-05,
+      "loss": 0.7383,
+      "step": 1713
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 1.245992692145078e-05,
+      "loss": 0.6799,
+      "step": 1714
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 1.2451895865073751e-05,
+      "loss": 0.6406,
+      "step": 1715
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 1.2443863125770473e-05,
+      "loss": 0.6516,
+      "step": 1716
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 1.2435828709054431e-05,
+      "loss": 0.6755,
+      "step": 1717
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 1.242779262044028e-05,
+      "loss": 0.6406,
+      "step": 1718
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 1.2419754865443803e-05,
+      "loss": 0.5977,
+      "step": 1719
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 1.2411715449581938e-05,
+      "loss": 0.6907,
+      "step": 1720
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 1.2403674378372755e-05,
+      "loss": 0.717,
+      "step": 1721
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 1.239563165733547e-05,
+      "loss": 0.634,
+      "step": 1722
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 1.2387587291990423e-05,
+      "loss": 0.6538,
+      "step": 1723
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 1.2379541287859085e-05,
+      "loss": 0.6597,
+      "step": 1724
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 1.2371493650464057e-05,
+      "loss": 0.6865,
+      "step": 1725
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 1.2363444385329052e-05,
+      "loss": 0.8081,
+      "step": 1726
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 1.2355393497978904e-05,
+      "loss": 0.6919,
+      "step": 1727
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 1.2347340993939567e-05,
+      "loss": 0.7969,
+      "step": 1728
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 1.2339286878738094e-05,
+      "loss": 0.6606,
+      "step": 1729
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 1.233123115790265e-05,
+      "loss": 0.6582,
+      "step": 1730
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 1.2323173836962502e-05,
+      "loss": 0.7178,
+      "step": 1731
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 1.2315114921448012e-05,
+      "loss": 0.7715,
+      "step": 1732
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 1.230705441689064e-05,
+      "loss": 0.6802,
+      "step": 1733
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 1.2298992328822937e-05,
+      "loss": 0.6304,
+      "step": 1734
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 1.2290928662778535e-05,
+      "loss": 0.5795,
+      "step": 1735
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 1.2282863424292155e-05,
+      "loss": 0.5634,
+      "step": 1736
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 1.2274796618899599e-05,
+      "loss": 0.6533,
+      "step": 1737
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 1.2266728252137735e-05,
+      "loss": 0.6628,
+      "step": 1738
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 1.2258658329544513e-05,
+      "loss": 0.6458,
+      "step": 1739
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 1.2250586856658946e-05,
+      "loss": 0.6721,
+      "step": 1740
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 1.2242513839021109e-05,
+      "loss": 0.5842,
+      "step": 1741
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 1.2234439282172144e-05,
+      "loss": 0.7222,
+      "step": 1742
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 1.2226363191654244e-05,
+      "loss": 0.6655,
+      "step": 1743
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 1.2218285573010654e-05,
+      "loss": 0.6497,
+      "step": 1744
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 1.2210206431785674e-05,
+      "loss": 0.667,
+      "step": 1745
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 1.220212577352464e-05,
+      "loss": 0.6938,
+      "step": 1746
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 1.2194043603773938e-05,
+      "loss": 0.614,
+      "step": 1747
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 1.2185959928080988e-05,
+      "loss": 0.7229,
+      "step": 1748
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 1.217787475199424e-05,
+      "loss": 0.6375,
+      "step": 1749
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 1.2169788081063181e-05,
+      "loss": 0.8057,
+      "step": 1750
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 1.216169992083832e-05,
+      "loss": 0.6841,
+      "step": 1751
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 1.2153610276871187e-05,
+      "loss": 0.6553,
+      "step": 1752
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 1.214551915471433e-05,
+      "loss": 0.6357,
+      "step": 1753
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 1.2137426559921317e-05,
+      "loss": 0.6284,
+      "step": 1754
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 1.2129332498046721e-05,
+      "loss": 0.6655,
+      "step": 1755
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 1.2121236974646127e-05,
+      "loss": 0.6973,
+      "step": 1756
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 1.2113139995276118e-05,
+      "loss": 0.6667,
+      "step": 1757
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 1.210504156549428e-05,
+      "loss": 0.7397,
+      "step": 1758
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 1.2096941690859194e-05,
+      "loss": 0.6572,
+      "step": 1759
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 1.2088840376930429e-05,
+      "loss": 0.7271,
+      "step": 1760
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 1.2080737629268546e-05,
+      "loss": 0.6506,
+      "step": 1761
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 1.2072633453435092e-05,
+      "loss": 0.6929,
+      "step": 1762
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 1.2064527854992587e-05,
+      "loss": 0.6787,
+      "step": 1763
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 1.2056420839504535e-05,
+      "loss": 0.6721,
+      "step": 1764
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 1.2048312412535407e-05,
+      "loss": 0.6855,
+      "step": 1765
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 1.2040202579650649e-05,
+      "loss": 0.7087,
+      "step": 1766
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 1.2032091346416663e-05,
+      "loss": 0.5825,
+      "step": 1767
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 1.202397871840082e-05,
+      "loss": 0.7324,
+      "step": 1768
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 1.2015864701171444e-05,
+      "loss": 0.7817,
+      "step": 1769
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 1.2007749300297818e-05,
+      "loss": 0.6689,
+      "step": 1770
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 1.1999632521350168e-05,
+      "loss": 0.7002,
+      "step": 1771
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 1.1991514369899668e-05,
+      "loss": 0.6417,
+      "step": 1772
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 1.1983394851518435e-05,
+      "loss": 0.6882,
+      "step": 1773
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 1.1975273971779528e-05,
+      "loss": 0.6029,
+      "step": 1774
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 1.1967151736256932e-05,
+      "loss": 0.7495,
+      "step": 1775
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 1.195902815052557e-05,
+      "loss": 0.7114,
+      "step": 1776
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 1.1950903220161286e-05,
+      "loss": 0.6318,
+      "step": 1777
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 1.1942776950740849e-05,
+      "loss": 0.6208,
+      "step": 1778
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 1.193464934784195e-05,
+      "loss": 0.6414,
+      "step": 1779
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 1.1926520417043195e-05,
+      "loss": 0.6975,
+      "step": 1780
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 1.1918390163924098e-05,
+      "loss": 0.7437,
+      "step": 1781
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 1.1910258594065079e-05,
+      "loss": 0.6111,
+      "step": 1782
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 1.1902125713047467e-05,
+      "loss": 0.6804,
+      "step": 1783
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 1.1893991526453493e-05,
+      "loss": 0.6035,
+      "step": 1784
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 1.1885856039866275e-05,
+      "loss": 0.624,
+      "step": 1785
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 1.1877719258869827e-05,
+      "loss": 0.6814,
+      "step": 1786
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 1.1869581189049057e-05,
+      "loss": 0.5999,
+      "step": 1787
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 1.1861441835989754e-05,
+      "loss": 0.6394,
+      "step": 1788
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 1.1853301205278579e-05,
+      "loss": 0.668,
+      "step": 1789
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 1.1845159302503086e-05,
+      "loss": 0.6292,
+      "step": 1790
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 1.1837016133251695e-05,
+      "loss": 0.6907,
+      "step": 1791
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 1.1828871703113686e-05,
+      "loss": 0.6804,
+      "step": 1792
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 1.1820726017679222e-05,
+      "loss": 0.5765,
+      "step": 1793
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 1.1812579082539317e-05,
+      "loss": 0.7715,
+      "step": 1794
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 1.1804430903285837e-05,
+      "loss": 0.7007,
+      "step": 1795
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 1.1796281485511519e-05,
+      "loss": 0.5284,
+      "step": 1796
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 1.1788130834809932e-05,
+      "loss": 0.6899,
+      "step": 1797
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 1.1779978956775507e-05,
+      "loss": 0.6119,
+      "step": 1798
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 1.17718258570035e-05,
+      "loss": 0.6714,
+      "step": 1799
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 1.1763671541090027e-05,
+      "loss": 0.6931,
+      "step": 1800
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 1.1755516014632023e-05,
+      "loss": 0.6333,
+      "step": 1801
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 1.1747359283227251e-05,
+      "loss": 0.5765,
+      "step": 1802
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 1.1739201352474322e-05,
+      "loss": 0.6646,
+      "step": 1803
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 1.1731042227972645e-05,
+      "loss": 0.7075,
+      "step": 1804
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 1.1722881915322462e-05,
+      "loss": 0.7302,
+      "step": 1805
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 1.1714720420124833e-05,
+      "loss": 0.7593,
+      "step": 1806
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 1.1706557747981622e-05,
+      "loss": 0.6594,
+      "step": 1807
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 1.16983939044955e-05,
+      "loss": 0.7163,
+      "step": 1808
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 1.1690228895269953e-05,
+      "loss": 0.6201,
+      "step": 1809
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 1.1682062725909257e-05,
+      "loss": 0.7002,
+      "step": 1810
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 1.1673895402018484e-05,
+      "loss": 0.7407,
+      "step": 1811
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 1.1665726929203509e-05,
+      "loss": 0.7102,
+      "step": 1812
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 1.165755731307098e-05,
+      "loss": 0.6345,
+      "step": 1813
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 1.1649386559228342e-05,
+      "loss": 0.7085,
+      "step": 1814
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 1.1641214673283818e-05,
+      "loss": 0.5945,
+      "step": 1815
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 1.1633041660846405e-05,
+      "loss": 0.584,
+      "step": 1816
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 1.1624867527525876e-05,
+      "loss": 0.6017,
+      "step": 1817
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 1.1616692278932772e-05,
+      "loss": 0.5801,
+      "step": 1818
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 1.1608515920678398e-05,
+      "loss": 0.6489,
+      "step": 1819
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 1.1600338458374829e-05,
+      "loss": 0.7993,
+      "step": 1820
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 1.1592159897634887e-05,
+      "loss": 0.7104,
+      "step": 1821
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 1.158398024407215e-05,
+      "loss": 0.6992,
+      "step": 1822
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 1.1575799503300952e-05,
+      "loss": 0.7285,
+      "step": 1823
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 1.1567617680936372e-05,
+      "loss": 0.707,
+      "step": 1824
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 1.1559434782594224e-05,
+      "loss": 0.7026,
+      "step": 1825
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 1.1551250813891066e-05,
+      "loss": 0.552,
+      "step": 1826
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 1.1543065780444196e-05,
+      "loss": 0.6692,
+      "step": 1827
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 1.1534879687871629e-05,
+      "loss": 0.729,
+      "step": 1828
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 1.1526692541792122e-05,
+      "loss": 0.5854,
+      "step": 1829
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 1.1518504347825146e-05,
+      "loss": 0.6399,
+      "step": 1830
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 1.151031511159089e-05,
+      "loss": 0.5911,
+      "step": 1831
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 1.1502124838710268e-05,
+      "loss": 0.7373,
+      "step": 1832
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 1.1493933534804892e-05,
+      "loss": 0.647,
+      "step": 1833
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 1.1485741205497094e-05,
+      "loss": 0.6431,
+      "step": 1834
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 1.1477547856409905e-05,
+      "loss": 0.6406,
+      "step": 1835
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 1.1469353493167056e-05,
+      "loss": 0.6055,
+      "step": 1836
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 1.1461158121392972e-05,
+      "loss": 0.7334,
+      "step": 1837
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 1.1452961746712769e-05,
+      "loss": 0.5376,
+      "step": 1838
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 1.1444764374752262e-05,
+      "loss": 0.6318,
+      "step": 1839
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.143656601113794e-05,
+      "loss": 0.6187,
+      "step": 1840
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.1428366661496971e-05,
+      "loss": 0.7173,
+      "step": 1841
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.1420166331457208e-05,
+      "loss": 0.7188,
+      "step": 1842
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 1.1411965026647177e-05,
+      "loss": 0.657,
+      "step": 1843
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.140376275269606e-05,
+      "loss": 0.572,
+      "step": 1844
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.1395559515233726e-05,
+      "loss": 0.6487,
+      "step": 1845
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.1387355319890685e-05,
+      "loss": 0.6919,
+      "step": 1846
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.137915017229811e-05,
+      "loss": 0.6895,
+      "step": 1847
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.1370944078087837e-05,
+      "loss": 0.5981,
+      "step": 1848
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.1362737042892342e-05,
+      "loss": 0.6631,
+      "step": 1849
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.1354529072344749e-05,
+      "loss": 0.6868,
+      "step": 1850
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 1.1346320172078822e-05,
+      "loss": 0.6599,
+      "step": 1851
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.1338110347728973e-05,
+      "loss": 0.7117,
+      "step": 1852
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.1329899604930235e-05,
+      "loss": 0.6426,
+      "step": 1853
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.1321687949318276e-05,
+      "loss": 0.6023,
+      "step": 1854
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 1.13134753865294e-05,
+      "loss": 0.5476,
+      "step": 1855
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.130526192220052e-05,
+      "loss": 0.7175,
+      "step": 1856
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.129704756196917e-05,
+      "loss": 0.6636,
+      "step": 1857
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.1288832311473507e-05,
+      "loss": 0.6855,
+      "step": 1858
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 1.1280616176352299e-05,
+      "loss": 0.6459,
+      "step": 1859
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.1272399162244907e-05,
+      "loss": 0.6399,
+      "step": 1860
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.1264181274791311e-05,
+      "loss": 0.6819,
+      "step": 1861
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.1255962519632082e-05,
+      "loss": 0.7441,
+      "step": 1862
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 1.1247742902408391e-05,
+      "loss": 0.536,
+      "step": 1863
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.1239522428761996e-05,
+      "loss": 0.6426,
+      "step": 1864
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.1231301104335248e-05,
+      "loss": 0.6494,
+      "step": 1865
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.1223078934771079e-05,
+      "loss": 0.6865,
+      "step": 1866
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 1.1214855925712998e-05,
+      "loss": 0.6394,
+      "step": 1867
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.1206632082805099e-05,
+      "loss": 0.645,
+      "step": 1868
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.1198407411692041e-05,
+      "loss": 0.5959,
+      "step": 1869
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 1.119018191801905e-05,
+      "loss": 0.7993,
+      "step": 1870
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.1181955607431925e-05,
+      "loss": 0.7285,
+      "step": 1871
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.117372848557702e-05,
+      "loss": 0.5854,
+      "step": 1872
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.116550055810124e-05,
+      "loss": 0.6458,
+      "step": 1873
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 1.1157271830652063e-05,
+      "loss": 0.6499,
+      "step": 1874
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 1.114904230887749e-05,
+      "loss": 0.7002,
+      "step": 1875
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 1.1140811998426089e-05,
+      "loss": 0.7263,
+      "step": 1876
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 1.1132580904946952e-05,
+      "loss": 0.6406,
+      "step": 1877
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 1.1124349034089724e-05,
+      "loss": 0.6323,
+      "step": 1878
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 1.1116116391504572e-05,
+      "loss": 0.6248,
+      "step": 1879
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 1.1107882982842195e-05,
+      "loss": 0.79,
+      "step": 1880
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 1.1099648813753827e-05,
+      "loss": 0.7324,
+      "step": 1881
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 1.1091413889891211e-05,
+      "loss": 0.6147,
+      "step": 1882
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 1.1083178216906614e-05,
+      "loss": 0.601,
+      "step": 1883
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 1.1074941800452818e-05,
+      "loss": 0.5955,
+      "step": 1884
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 1.1066704646183117e-05,
+      "loss": 0.6002,
+      "step": 1885
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 1.1058466759751302e-05,
+      "loss": 0.7224,
+      "step": 1886
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 1.1050228146811684e-05,
+      "loss": 0.752,
+      "step": 1887
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 1.1041988813019051e-05,
+      "loss": 0.5891,
+      "step": 1888
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 1.10337487640287e-05,
+      "loss": 0.5964,
+      "step": 1889
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 1.1025508005496417e-05,
+      "loss": 0.6819,
+      "step": 1890
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 1.1017266543078475e-05,
+      "loss": 0.6641,
+      "step": 1891
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 1.1009024382431623e-05,
+      "loss": 0.7114,
+      "step": 1892
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 1.1000781529213103e-05,
+      "loss": 0.7065,
+      "step": 1893
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 1.0992537989080618e-05,
+      "loss": 0.7256,
+      "step": 1894
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 1.0984293767692349e-05,
+      "loss": 0.6782,
+      "step": 1895
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 1.097604887070695e-05,
+      "loss": 0.5933,
+      "step": 1896
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 1.0967803303783523e-05,
+      "loss": 0.549,
+      "step": 1897
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 1.0959557072581652e-05,
+      "loss": 0.5486,
+      "step": 1898
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 1.0951310182761358e-05,
+      "loss": 0.6707,
+      "step": 1899
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 1.0943062639983119e-05,
+      "loss": 0.6775,
+      "step": 1900
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 1.093481444990787e-05,
+      "loss": 0.4968,
+      "step": 1901
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 1.0926565618196978e-05,
+      "loss": 0.6488,
+      "step": 1902
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 1.0918316150512256e-05,
+      "loss": 0.7129,
+      "step": 1903
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 1.0910066052515956e-05,
+      "loss": 0.6589,
+      "step": 1904
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 1.0901815329870758e-05,
+      "loss": 0.5464,
+      "step": 1905
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 1.0893563988239773e-05,
+      "loss": 0.7544,
+      "step": 1906
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 1.0885312033286537e-05,
+      "loss": 0.6851,
+      "step": 1907
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 1.0877059470675007e-05,
+      "loss": 0.6578,
+      "step": 1908
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 1.0868806306069555e-05,
+      "loss": 0.6829,
+      "step": 1909
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 1.086055254513497e-05,
+      "loss": 0.6404,
+      "step": 1910
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 1.0852298193536446e-05,
+      "loss": 0.6782,
+      "step": 1911
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 1.0844043256939585e-05,
+      "loss": 0.5823,
+      "step": 1912
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 1.083578774101039e-05,
+      "loss": 0.6611,
+      "step": 1913
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 1.0827531651415266e-05,
+      "loss": 0.6848,
+      "step": 1914
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 1.0819274993820998e-05,
+      "loss": 0.6179,
+      "step": 1915
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 1.0811017773894781e-05,
+      "loss": 0.6287,
+      "step": 1916
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 1.080275999730418e-05,
+      "loss": 0.603,
+      "step": 1917
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 1.0794501669717146e-05,
+      "loss": 0.5352,
+      "step": 1918
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 1.078624279680201e-05,
+      "loss": 0.6064,
+      "step": 1919
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 1.077798338422748e-05,
+      "loss": 0.6973,
+      "step": 1920
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 1.076972343766263e-05,
+      "loss": 0.6837,
+      "step": 1921
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.0761462962776896e-05,
+      "loss": 0.624,
+      "step": 1922
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.0753201965240095e-05,
+      "loss": 0.668,
+      "step": 1923
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.0744940450722378e-05,
+      "loss": 0.7417,
+      "step": 1924
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 1.0736678424894269e-05,
+      "loss": 0.6685,
+      "step": 1925
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.0728415893426636e-05,
+      "loss": 0.5142,
+      "step": 1926
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.0720152861990693e-05,
+      "loss": 0.7126,
+      "step": 1927
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.0711889336258004e-05,
+      "loss": 0.613,
+      "step": 1928
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 1.0703625321900462e-05,
+      "loss": 0.729,
+      "step": 1929
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.0695360824590304e-05,
+      "loss": 0.7141,
+      "step": 1930
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.0687095850000096e-05,
+      "loss": 0.748,
+      "step": 1931
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.0678830403802728e-05,
+      "loss": 0.679,
+      "step": 1932
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 1.0670564491671415e-05,
+      "loss": 0.626,
+      "step": 1933
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 1.0662298119279702e-05,
+      "loss": 0.6782,
+      "step": 1934
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 1.0654031292301432e-05,
+      "loss": 0.6704,
+      "step": 1935
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 1.0645764016410776e-05,
+      "loss": 0.7214,
+      "step": 1936
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 1.0637496297282202e-05,
+      "loss": 0.6504,
+      "step": 1937
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 1.0629228140590486e-05,
+      "loss": 0.7788,
+      "step": 1938
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 1.062095955201071e-05,
+      "loss": 0.5032,
+      "step": 1939
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 1.0612690537218246e-05,
+      "loss": 0.6799,
+      "step": 1940
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 1.0604421101888755e-05,
+      "loss": 0.6682,
+      "step": 1941
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 1.05961512516982e-05,
+      "loss": 0.5686,
+      "step": 1942
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 1.0587880992322814e-05,
+      "loss": 0.6726,
+      "step": 1943
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 1.0579610329439121e-05,
+      "loss": 0.7097,
+      "step": 1944
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 1.0571339268723915e-05,
+      "loss": 0.6177,
+      "step": 1945
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 1.0563067815854266e-05,
+      "loss": 0.6624,
+      "step": 1946
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 1.055479597650752e-05,
+      "loss": 0.6208,
+      "step": 1947
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 1.0546523756361271e-05,
+      "loss": 0.5876,
+      "step": 1948
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 1.0538251161093396e-05,
+      "loss": 0.6743,
+      "step": 1949
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.0529978196382011e-05,
+      "loss": 0.7305,
+      "step": 1950
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.0521704867905496e-05,
+      "loss": 0.6821,
+      "step": 1951
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.051343118134248e-05,
+      "loss": 0.6445,
+      "step": 1952
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 1.0505157142371832e-05,
+      "loss": 0.5571,
+      "step": 1953
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 1.0496882756672667e-05,
+      "loss": 0.6553,
+      "step": 1954
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 1.048860802992434e-05,
+      "loss": 0.599,
+      "step": 1955
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 1.0480332967806435e-05,
+      "loss": 0.6704,
+      "step": 1956
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 1.047205757599877e-05,
+      "loss": 0.7256,
+      "step": 1957
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 1.0463781860181385e-05,
+      "loss": 0.5873,
+      "step": 1958
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 1.045550582603455e-05,
+      "loss": 0.625,
+      "step": 1959
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 1.044722947923875e-05,
+      "loss": 0.6709,
+      "step": 1960
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 1.0438952825474678e-05,
+      "loss": 0.7065,
+      "step": 1961
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 1.0430675870423245e-05,
+      "loss": 0.7056,
+      "step": 1962
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 1.0422398619765571e-05,
+      "loss": 0.6536,
+      "step": 1963
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 1.0414121079182969e-05,
+      "loss": 0.6445,
+      "step": 1964
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.0405843254356959e-05,
+      "loss": 0.5923,
+      "step": 1965
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.039756515096926e-05,
+      "loss": 0.6392,
+      "step": 1966
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.0389286774701765e-05,
+      "loss": 0.6462,
+      "step": 1967
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 1.0381008131236576e-05,
+      "loss": 0.6191,
+      "step": 1968
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 1.0372729226255962e-05,
+      "loss": 0.6797,
+      "step": 1969
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 1.0364450065442377e-05,
+      "loss": 0.5686,
+      "step": 1970
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 1.0356170654478457e-05,
+      "loss": 0.5439,
+      "step": 1971
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 1.0347890999046998e-05,
+      "loss": 0.6344,
+      "step": 1972
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 1.033961110483097e-05,
+      "loss": 0.6841,
+      "step": 1973
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 1.033133097751351e-05,
+      "loss": 0.5625,
+      "step": 1974
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 1.0323050622777906e-05,
+      "loss": 0.5807,
+      "step": 1975
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 1.0314770046307611e-05,
+      "loss": 0.6465,
+      "step": 1976
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.030648925378623e-05,
+      "loss": 0.6418,
+      "step": 1977
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.0298208250897504e-05,
+      "loss": 0.5831,
+      "step": 1978
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.0289927043325336e-05,
+      "loss": 0.6323,
+      "step": 1979
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 1.0281645636753758e-05,
+      "loss": 0.5811,
+      "step": 1980
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 1.027336403686694e-05,
+      "loss": 0.6272,
+      "step": 1981
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 1.0265082249349188e-05,
+      "loss": 0.7214,
+      "step": 1982
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 1.0256800279884934e-05,
+      "loss": 0.5469,
+      "step": 1983
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 1.024851813415873e-05,
+      "loss": 0.6013,
+      "step": 1984
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 1.0240235817855269e-05,
+      "loss": 0.5516,
+      "step": 1985
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 1.0231953336659334e-05,
+      "loss": 0.4596,
+      "step": 1986
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 1.022367069625584e-05,
+      "loss": 0.6548,
+      "step": 1987
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 1.0215387902329806e-05,
+      "loss": 0.6398,
+      "step": 1988
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 1.020710496056635e-05,
+      "loss": 0.6616,
+      "step": 1989
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 1.0198821876650702e-05,
+      "loss": 0.646,
+      "step": 1990
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 1.0190538656268186e-05,
+      "loss": 0.6184,
+      "step": 1991
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 1.0182255305104215e-05,
+      "loss": 0.6533,
+      "step": 1992
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.017397182884429e-05,
+      "loss": 0.6135,
+      "step": 1993
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.0165688233174019e-05,
+      "loss": 0.6753,
+      "step": 1994
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.015740452377906e-05,
+      "loss": 0.6357,
+      "step": 1995
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 1.0149120706345175e-05,
+      "loss": 0.7358,
+      "step": 1996
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 1.0140836786558184e-05,
+      "loss": 0.5964,
+      "step": 1997
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 1.0132552770103987e-05,
+      "loss": 0.6753,
+      "step": 1998
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 1.0124268662668547e-05,
+      "loss": 0.6169,
+      "step": 1999
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 1.0115984469937883e-05,
+      "loss": 0.5857,
+      "step": 2000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 1.0107700197598089e-05,
+      "loss": 0.6453,
+      "step": 2001
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 1.00994158513353e-05,
+      "loss": 0.5758,
+      "step": 2002
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 1.0091131436835698e-05,
+      "loss": 0.6843,
+      "step": 2003
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.0082846959785528e-05,
+      "loss": 0.6643,
+      "step": 2004
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.0074562425871065e-05,
+      "loss": 0.6558,
+      "step": 2005
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.0066277840778626e-05,
+      "loss": 0.5471,
+      "step": 2006
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 1.005799321019457e-05,
+      "loss": 0.689,
+      "step": 2007
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 1.0049708539805274e-05,
+      "loss": 0.6243,
+      "step": 2008
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 1.0041423835297155e-05,
+      "loss": 0.6145,
+      "step": 2009
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 1.0033139102356641e-05,
+      "loss": 0.6333,
+      "step": 2010
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 1.0024854346670196e-05,
+      "loss": 0.6494,
+      "step": 2011
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 1.0016569573924287e-05,
+      "loss": 0.5791,
+      "step": 2012
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 1.0008284789805389e-05,
+      "loss": 0.6443,
+      "step": 2013
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 1e-05,
+      "loss": 0.6604,
+      "step": 2014
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 9.991715210194611e-06,
+      "loss": 0.6377,
+      "step": 2015
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.98343042607572e-06,
+      "loss": 0.6738,
+      "step": 2016
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.975145653329806e-06,
+      "loss": 0.7349,
+      "step": 2017
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.966860897643359e-06,
+      "loss": 0.6047,
+      "step": 2018
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 9.95857616470285e-06,
+      "loss": 0.647,
+      "step": 2019
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.950291460194729e-06,
+      "loss": 0.6216,
+      "step": 2020
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.942006789805432e-06,
+      "loss": 0.4255,
+      "step": 2021
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.933722159221375e-06,
+      "loss": 0.575,
+      "step": 2022
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 9.925437574128938e-06,
+      "loss": 0.665,
+      "step": 2023
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.917153040214473e-06,
+      "loss": 0.5845,
+      "step": 2024
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.908868563164305e-06,
+      "loss": 0.6682,
+      "step": 2025
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.900584148664705e-06,
+      "loss": 0.5823,
+      "step": 2026
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 9.892299802401911e-06,
+      "loss": 0.6462,
+      "step": 2027
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.884015530062119e-06,
+      "loss": 0.5491,
+      "step": 2028
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.875731337331458e-06,
+      "loss": 0.5635,
+      "step": 2029
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.86744722989602e-06,
+      "loss": 0.6411,
+      "step": 2030
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 9.85916321344182e-06,
+      "loss": 0.6934,
+      "step": 2031
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 9.850879293654829e-06,
+      "loss": 0.6445,
+      "step": 2032
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 9.842595476220945e-06,
+      "loss": 0.6372,
+      "step": 2033
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 9.834311766825986e-06,
+      "loss": 0.6812,
+      "step": 2034
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 9.82602817115571e-06,
+      "loss": 0.5808,
+      "step": 2035
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.817744694895792e-06,
+      "loss": 0.561,
+      "step": 2036
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.80946134373182e-06,
+      "loss": 0.6167,
+      "step": 2037
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.801178123349298e-06,
+      "loss": 0.5369,
+      "step": 2038
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 9.792895039433655e-06,
+      "loss": 0.7058,
+      "step": 2039
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 9.7846120976702e-06,
+      "loss": 0.6782,
+      "step": 2040
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 9.77632930374416e-06,
+      "loss": 0.7065,
+      "step": 2041
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 9.76804666334067e-06,
+      "loss": 0.6921,
+      "step": 2042
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.759764182144734e-06,
+      "loss": 0.5842,
+      "step": 2043
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.75148186584127e-06,
+      "loss": 0.5942,
+      "step": 2044
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.743199720115071e-06,
+      "loss": 0.5559,
+      "step": 2045
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 9.734917750650817e-06,
+      "loss": 0.6516,
+      "step": 2046
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.726635963133062e-06,
+      "loss": 0.6213,
+      "step": 2047
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.718354363246243e-06,
+      "loss": 0.7002,
+      "step": 2048
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.710072956674666e-06,
+      "loss": 0.6426,
+      "step": 2049
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 9.701791749102496e-06,
+      "loss": 0.6299,
+      "step": 2050
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.693510746213774e-06,
+      "loss": 0.6589,
+      "step": 2051
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.68522995369239e-06,
+      "loss": 0.635,
+      "step": 2052
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.676949377222096e-06,
+      "loss": 0.6387,
+      "step": 2053
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 9.668669022486495e-06,
+      "loss": 0.6841,
+      "step": 2054
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 9.660388895169034e-06,
+      "loss": 0.626,
+      "step": 2055
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 9.652109000953007e-06,
+      "loss": 0.6555,
+      "step": 2056
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 9.643829345521546e-06,
+      "loss": 0.7041,
+      "step": 2057
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 9.635549934557626e-06,
+      "loss": 0.7456,
+      "step": 2058
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.627270773744043e-06,
+      "loss": 0.6741,
+      "step": 2059
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.618991868763428e-06,
+      "loss": 0.6355,
+      "step": 2060
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.610713225298238e-06,
+      "loss": 0.6715,
+      "step": 2061
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 9.602434849030747e-06,
+      "loss": 0.6543,
+      "step": 2062
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.594156745643043e-06,
+      "loss": 0.677,
+      "step": 2063
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.585878920817032e-06,
+      "loss": 0.7017,
+      "step": 2064
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.577601380234434e-06,
+      "loss": 0.748,
+      "step": 2065
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 9.569324129576756e-06,
+      "loss": 0.4768,
+      "step": 2066
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.561047174525324e-06,
+      "loss": 0.5981,
+      "step": 2067
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.552770520761256e-06,
+      "loss": 0.5496,
+      "step": 2068
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.544494173965451e-06,
+      "loss": 0.5876,
+      "step": 2069
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 9.536218139818615e-06,
+      "loss": 0.7061,
+      "step": 2070
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.527942424001235e-06,
+      "loss": 0.6194,
+      "step": 2071
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.519667032193569e-06,
+      "loss": 0.7095,
+      "step": 2072
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.511391970075662e-06,
+      "loss": 0.5603,
+      "step": 2073
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 9.503117243327336e-06,
+      "loss": 0.6238,
+      "step": 2074
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.49484285762817e-06,
+      "loss": 0.709,
+      "step": 2075
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.48656881865752e-06,
+      "loss": 0.5378,
+      "step": 2076
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.478295132094505e-06,
+      "loss": 0.6262,
+      "step": 2077
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 9.47002180361799e-06,
+      "loss": 0.6364,
+      "step": 2078
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 9.461748838906606e-06,
+      "loss": 0.6733,
+      "step": 2079
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 9.45347624363873e-06,
+      "loss": 0.5892,
+      "step": 2080
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 9.445204023492484e-06,
+      "loss": 0.7046,
+      "step": 2081
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 9.436932184145737e-06,
+      "loss": 0.574,
+      "step": 2082
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 9.428660731276089e-06,
+      "loss": 0.58,
+      "step": 2083
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 9.42038967056088e-06,
+      "loss": 0.5615,
+      "step": 2084
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 9.41211900767719e-06,
+      "loss": 0.6729,
+      "step": 2085
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.403848748301802e-06,
+      "loss": 0.7227,
+      "step": 2086
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.395578898111243e-06,
+      "loss": 0.6635,
+      "step": 2087
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.387309462781759e-06,
+      "loss": 0.6692,
+      "step": 2088
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 9.379040447989292e-06,
+      "loss": 0.709,
+      "step": 2089
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.370771859409514e-06,
+      "loss": 0.6912,
+      "step": 2090
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.362503702717803e-06,
+      "loss": 0.635,
+      "step": 2091
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.354235983589229e-06,
+      "loss": 0.6593,
+      "step": 2092
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 9.34596870769857e-06,
+      "loss": 0.5569,
+      "step": 2093
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.337701880720303e-06,
+      "loss": 0.6418,
+      "step": 2094
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.329435508328587e-06,
+      "loss": 0.7192,
+      "step": 2095
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.321169596197274e-06,
+      "loss": 0.5942,
+      "step": 2096
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 9.312904149999909e-06,
+      "loss": 0.5146,
+      "step": 2097
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.304639175409699e-06,
+      "loss": 0.5671,
+      "step": 2098
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.29637467809954e-06,
+      "loss": 0.606,
+      "step": 2099
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.288110663742001e-06,
+      "loss": 0.6316,
+      "step": 2100
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 9.279847138009308e-06,
+      "loss": 0.688,
+      "step": 2101
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 9.271584106573364e-06,
+      "loss": 0.5637,
+      "step": 2102
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 9.263321575105734e-06,
+      "loss": 0.6963,
+      "step": 2103
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 9.255059549277624e-06,
+      "loss": 0.5469,
+      "step": 2104
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 9.246798034759907e-06,
+      "loss": 0.6238,
+      "step": 2105
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.238537037223106e-06,
+      "loss": 0.5819,
+      "step": 2106
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.230276562337374e-06,
+      "loss": 0.6868,
+      "step": 2107
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.222016615772525e-06,
+      "loss": 0.7422,
+      "step": 2108
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 9.213757203197994e-06,
+      "loss": 0.5967,
+      "step": 2109
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.205498330282857e-06,
+      "loss": 0.5693,
+      "step": 2110
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.197240002695826e-06,
+      "loss": 0.5993,
+      "step": 2111
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.188982226105222e-06,
+      "loss": 0.6392,
+      "step": 2112
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 9.180725006179002e-06,
+      "loss": 0.6069,
+      "step": 2113
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.17246834858474e-06,
+      "loss": 0.6885,
+      "step": 2114
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.164212258989611e-06,
+      "loss": 0.6592,
+      "step": 2115
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.155956743060417e-06,
+      "loss": 0.5874,
+      "step": 2116
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 9.147701806463559e-06,
+      "loss": 0.6519,
+      "step": 2117
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.139447454865034e-06,
+      "loss": 0.6577,
+      "step": 2118
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.131193693930447e-06,
+      "loss": 0.5435,
+      "step": 2119
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.122940529324997e-06,
+      "loss": 0.5638,
+      "step": 2120
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 9.114687966713465e-06,
+      "loss": 0.5452,
+      "step": 2121
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.106436011760229e-06,
+      "loss": 0.5444,
+      "step": 2122
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.098184670129244e-06,
+      "loss": 0.5057,
+      "step": 2123
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.089933947484048e-06,
+      "loss": 0.7407,
+      "step": 2124
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 9.081683849487745e-06,
+      "loss": 0.7,
+      "step": 2125
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 9.073434381803025e-06,
+      "loss": 0.6189,
+      "step": 2126
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 9.065185550092134e-06,
+      "loss": 0.6348,
+      "step": 2127
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 9.056937360016881e-06,
+      "loss": 0.6245,
+      "step": 2128
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.048689817238645e-06,
+      "loss": 0.5364,
+      "step": 2129
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.04044292741835e-06,
+      "loss": 0.5116,
+      "step": 2130
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.032196696216477e-06,
+      "loss": 0.5693,
+      "step": 2131
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 9.023951129293053e-06,
+      "loss": 0.6465,
+      "step": 2132
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 9.015706232307653e-06,
+      "loss": 0.5891,
+      "step": 2133
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 9.007462010919387e-06,
+      "loss": 0.5767,
+      "step": 2134
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 8.9992184707869e-06,
+      "loss": 0.5751,
+      "step": 2135
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 8.990975617568379e-06,
+      "loss": 0.644,
+      "step": 2136
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 8.98273345692153e-06,
+      "loss": 0.4606,
+      "step": 2137
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 8.974491994503585e-06,
+      "loss": 0.6189,
+      "step": 2138
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 8.966251235971302e-06,
+      "loss": 0.7036,
+      "step": 2139
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 8.958011186980954e-06,
+      "loss": 0.7319,
+      "step": 2140
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 8.94977185318832e-06,
+      "loss": 0.5808,
+      "step": 2141
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 8.9415332402487e-06,
+      "loss": 0.6401,
+      "step": 2142
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 8.933295353816888e-06,
+      "loss": 0.6123,
+      "step": 2143
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 8.925058199547183e-06,
+      "loss": 0.5078,
+      "step": 2144
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 8.916821783093386e-06,
+      "loss": 0.5715,
+      "step": 2145
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 8.908586110108794e-06,
+      "loss": 0.6418,
+      "step": 2146
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 8.900351186246177e-06,
+      "loss": 0.6068,
+      "step": 2147
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 8.892117017157805e-06,
+      "loss": 0.5995,
+      "step": 2148
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 8.883883608495433e-06,
+      "loss": 0.636,
+      "step": 2149
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 8.87565096591028e-06,
+      "loss": 0.5339,
+      "step": 2150
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 8.867419095053047e-06,
+      "loss": 0.6704,
+      "step": 2151
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 8.859188001573916e-06,
+      "loss": 0.6077,
+      "step": 2152
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 8.850957691122511e-06,
+      "loss": 0.4509,
+      "step": 2153
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 8.842728169347938e-06,
+      "loss": 0.6716,
+      "step": 2154
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 8.834499441898761e-06,
+      "loss": 0.6711,
+      "step": 2155
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 8.826271514422983e-06,
+      "loss": 0.603,
+      "step": 2156
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 8.818044392568076e-06,
+      "loss": 0.6279,
+      "step": 2157
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 8.809818081980954e-06,
+      "loss": 0.5465,
+      "step": 2158
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 8.801592588307964e-06,
+      "loss": 0.5898,
+      "step": 2159
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 8.793367917194906e-06,
+      "loss": 0.6182,
+      "step": 2160
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 8.785144074287006e-06,
+      "loss": 0.6365,
+      "step": 2161
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 8.776921065228924e-06,
+      "loss": 0.498,
+      "step": 2162
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 8.768698895664757e-06,
+      "loss": 0.655,
+      "step": 2163
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 8.760477571238006e-06,
+      "loss": 0.6479,
+      "step": 2164
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 8.75225709759161e-06,
+      "loss": 0.5272,
+      "step": 2165
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 8.744037480367922e-06,
+      "loss": 0.7202,
+      "step": 2166
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 8.735818725208692e-06,
+      "loss": 0.6384,
+      "step": 2167
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 8.727600837755095e-06,
+      "loss": 0.543,
+      "step": 2168
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 8.719383823647708e-06,
+      "loss": 0.6423,
+      "step": 2169
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 8.711167688526495e-06,
+      "loss": 0.5618,
+      "step": 2170
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 8.702952438030832e-06,
+      "loss": 0.5203,
+      "step": 2171
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 8.694738077799487e-06,
+      "loss": 0.6584,
+      "step": 2172
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 8.686524613470603e-06,
+      "loss": 0.6475,
+      "step": 2173
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 8.678312050681724e-06,
+      "loss": 0.5903,
+      "step": 2174
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 8.67010039506977e-06,
+      "loss": 0.6709,
+      "step": 2175
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 8.66188965227103e-06,
+      "loss": 0.6104,
+      "step": 2176
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 8.653679827921178e-06,
+      "loss": 0.5481,
+      "step": 2177
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 8.645470927655254e-06,
+      "loss": 0.5601,
+      "step": 2178
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 8.637262957107661e-06,
+      "loss": 0.5662,
+      "step": 2179
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 8.629055921912164e-06,
+      "loss": 0.5605,
+      "step": 2180
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 8.620849827701893e-06,
+      "loss": 0.6204,
+      "step": 2181
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 8.61264468010932e-06,
+      "loss": 0.5684,
+      "step": 2182
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 8.604440484766276e-06,
+      "loss": 0.6061,
+      "step": 2183
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 8.596237247303941e-06,
+      "loss": 0.6436,
+      "step": 2184
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 8.588034973352827e-06,
+      "loss": 0.6658,
+      "step": 2185
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 8.579833668542796e-06,
+      "loss": 0.5478,
+      "step": 2186
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 8.571633338503034e-06,
+      "loss": 0.5781,
+      "step": 2187
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 8.563433988862064e-06,
+      "loss": 0.7017,
+      "step": 2188
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 8.555235625247743e-06,
+      "loss": 0.6342,
+      "step": 2189
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 8.547038253287233e-06,
+      "loss": 0.6545,
+      "step": 2190
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 8.538841878607031e-06,
+      "loss": 0.5745,
+      "step": 2191
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 8.530646506832949e-06,
+      "loss": 0.6108,
+      "step": 2192
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 8.522452143590096e-06,
+      "loss": 0.5854,
+      "step": 2193
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 8.514258794502906e-06,
+      "loss": 0.6411,
+      "step": 2194
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 8.506066465195113e-06,
+      "loss": 0.6263,
+      "step": 2195
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 8.497875161289737e-06,
+      "loss": 0.5857,
+      "step": 2196
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 8.489684888409111e-06,
+      "loss": 0.5018,
+      "step": 2197
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 8.481495652174859e-06,
+      "loss": 0.5942,
+      "step": 2198
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 8.47330745820788e-06,
+      "loss": 0.6064,
+      "step": 2199
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 8.465120312128371e-06,
+      "loss": 0.5828,
+      "step": 2200
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 8.456934219555809e-06,
+      "loss": 0.5701,
+      "step": 2201
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 8.448749186108936e-06,
+      "loss": 0.5125,
+      "step": 2202
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 8.440565217405778e-06,
+      "loss": 0.6313,
+      "step": 2203
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 8.43238231906363e-06,
+      "loss": 0.5961,
+      "step": 2204
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 8.42420049669905e-06,
+      "loss": 0.6042,
+      "step": 2205
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 8.416019755927851e-06,
+      "loss": 0.525,
+      "step": 2206
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 8.407840102365116e-06,
+      "loss": 0.6768,
+      "step": 2207
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 8.399661541625174e-06,
+      "loss": 0.6096,
+      "step": 2208
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 8.391484079321602e-06,
+      "loss": 0.7085,
+      "step": 2209
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 8.383307721067231e-06,
+      "loss": 0.5601,
+      "step": 2210
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 8.375132472474127e-06,
+      "loss": 0.6814,
+      "step": 2211
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 8.3669583391536e-06,
+      "loss": 0.5906,
+      "step": 2212
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 8.358785326716184e-06,
+      "loss": 0.5051,
+      "step": 2213
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 8.350613440771661e-06,
+      "loss": 0.5544,
+      "step": 2214
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 8.342442686929024e-06,
+      "loss": 0.6589,
+      "step": 2215
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 8.334273070796495e-06,
+      "loss": 0.5164,
+      "step": 2216
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 8.326104597981517e-06,
+      "loss": 0.5247,
+      "step": 2217
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 8.317937274090748e-06,
+      "loss": 0.6338,
+      "step": 2218
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 8.309771104730049e-06,
+      "loss": 0.6807,
+      "step": 2219
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 8.301606095504501e-06,
+      "loss": 0.7598,
+      "step": 2220
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 8.293442252018384e-06,
+      "loss": 0.6226,
+      "step": 2221
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 8.28527957987517e-06,
+      "loss": 0.6497,
+      "step": 2222
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 8.277118084677538e-06,
+      "loss": 0.6118,
+      "step": 2223
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 8.26895777202736e-06,
+      "loss": 0.6855,
+      "step": 2224
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 8.260798647525684e-06,
+      "loss": 0.6208,
+      "step": 2225
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 8.252640716772748e-06,
+      "loss": 0.6323,
+      "step": 2226
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 8.244483985367984e-06,
+      "loss": 0.5468,
+      "step": 2227
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 8.236328458909975e-06,
+      "loss": 0.5347,
+      "step": 2228
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 8.2281741429965e-06,
+      "loss": 0.6006,
+      "step": 2229
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 8.2200210432245e-06,
+      "loss": 0.6821,
+      "step": 2230
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 8.21186916519007e-06,
+      "loss": 0.5491,
+      "step": 2231
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 8.203718514488483e-06,
+      "loss": 0.624,
+      "step": 2232
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 8.195569096714166e-06,
+      "loss": 0.575,
+      "step": 2233
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 8.187420917460686e-06,
+      "loss": 0.6428,
+      "step": 2234
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 8.179273982320778e-06,
+      "loss": 0.6348,
+      "step": 2235
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 8.171128296886316e-06,
+      "loss": 0.5163,
+      "step": 2236
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 8.162983866748308e-06,
+      "loss": 0.5436,
+      "step": 2237
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 8.154840697496917e-06,
+      "loss": 0.5981,
+      "step": 2238
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 8.146698794721423e-06,
+      "loss": 0.5774,
+      "step": 2239
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 8.138558164010251e-06,
+      "loss": 0.6243,
+      "step": 2240
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 8.130418810950948e-06,
+      "loss": 0.6528,
+      "step": 2241
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 8.122280741130177e-06,
+      "loss": 0.6934,
+      "step": 2242
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 8.114143960133728e-06,
+      "loss": 0.5542,
+      "step": 2243
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 8.106008473546512e-06,
+      "loss": 0.6221,
+      "step": 2244
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 8.097874286952534e-06,
+      "loss": 0.5803,
+      "step": 2245
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 8.089741405934923e-06,
+      "loss": 0.605,
+      "step": 2246
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 8.081609836075907e-06,
+      "loss": 0.6841,
+      "step": 2247
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 8.073479582956806e-06,
+      "loss": 0.4888,
+      "step": 2248
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 8.065350652158051e-06,
+      "loss": 0.6355,
+      "step": 2249
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 8.057223049259155e-06,
+      "loss": 0.5814,
+      "step": 2250
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 8.04909677983872e-06,
+      "loss": 0.5925,
+      "step": 2251
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 8.040971849474432e-06,
+      "loss": 0.6592,
+      "step": 2252
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 8.032848263743071e-06,
+      "loss": 0.5813,
+      "step": 2253
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 8.024726028220474e-06,
+      "loss": 0.5361,
+      "step": 2254
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 8.016605148481565e-06,
+      "loss": 0.6229,
+      "step": 2255
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 8.008485630100336e-06,
+      "loss": 0.5725,
+      "step": 2256
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 8.000367478649835e-06,
+      "loss": 0.6628,
+      "step": 2257
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 7.992250699702182e-06,
+      "loss": 0.553,
+      "step": 2258
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 7.98413529882856e-06,
+      "loss": 0.6748,
+      "step": 2259
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 7.976021281599182e-06,
+      "loss": 0.545,
+      "step": 2260
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 7.967908653583338e-06,
+      "loss": 0.6506,
+      "step": 2261
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 7.959797420349356e-06,
+      "loss": 0.5439,
+      "step": 2262
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 7.951687587464595e-06,
+      "loss": 0.5344,
+      "step": 2263
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 7.943579160495466e-06,
+      "loss": 0.6062,
+      "step": 2264
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 7.935472145007416e-06,
+      "loss": 0.6521,
+      "step": 2265
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 7.927366546564911e-06,
+      "loss": 0.5713,
+      "step": 2266
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 7.919262370731459e-06,
+      "loss": 0.7246,
+      "step": 2267
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 7.911159623069575e-06,
+      "loss": 0.4775,
+      "step": 2268
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 7.903058309140809e-06,
+      "loss": 0.6272,
+      "step": 2269
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.894958434505725e-06,
+      "loss": 0.5559,
+      "step": 2270
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.886860004723885e-06,
+      "loss": 0.5491,
+      "step": 2271
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.878763025353875e-06,
+      "loss": 0.656,
+      "step": 2272
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 7.870667501953284e-06,
+      "loss": 0.5693,
+      "step": 2273
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.862573440078686e-06,
+      "loss": 0.5632,
+      "step": 2274
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.854480845285672e-06,
+      "loss": 0.5432,
+      "step": 2275
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.846389723128819e-06,
+      "loss": 0.6185,
+      "step": 2276
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 7.838300079161684e-06,
+      "loss": 0.6577,
+      "step": 2277
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.83021191893682e-06,
+      "loss": 0.5945,
+      "step": 2278
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.822125248005764e-06,
+      "loss": 0.6147,
+      "step": 2279
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.814040071919016e-06,
+      "loss": 0.6467,
+      "step": 2280
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 7.805956396226064e-06,
+      "loss": 0.7183,
+      "step": 2281
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.797874226475362e-06,
+      "loss": 0.5953,
+      "step": 2282
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.78979356821433e-06,
+      "loss": 0.6602,
+      "step": 2283
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.781714426989346e-06,
+      "loss": 0.6599,
+      "step": 2284
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 7.773636808345757e-06,
+      "loss": 0.6272,
+      "step": 2285
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.765560717827857e-06,
+      "loss": 0.5642,
+      "step": 2286
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.757486160978891e-06,
+      "loss": 0.6475,
+      "step": 2287
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.749413143341056e-06,
+      "loss": 0.4841,
+      "step": 2288
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 7.741341670455489e-06,
+      "loss": 0.6523,
+      "step": 2289
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.733271747862265e-06,
+      "loss": 0.6392,
+      "step": 2290
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.725203381100403e-06,
+      "loss": 0.6313,
+      "step": 2291
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.717136575707846e-06,
+      "loss": 0.5349,
+      "step": 2292
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 7.709071337221468e-06,
+      "loss": 0.5587,
+      "step": 2293
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.701007671177066e-06,
+      "loss": 0.6479,
+      "step": 2294
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.692945583109362e-06,
+      "loss": 0.5918,
+      "step": 2295
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.684885078551991e-06,
+      "loss": 0.6553,
+      "step": 2296
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 7.6768261630375e-06,
+      "loss": 0.5383,
+      "step": 2297
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.668768842097353e-06,
+      "loss": 0.606,
+      "step": 2298
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.660713121261911e-06,
+      "loss": 0.5415,
+      "step": 2299
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.652659006060436e-06,
+      "loss": 0.541,
+      "step": 2300
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 7.644606502021096e-06,
+      "loss": 0.5774,
+      "step": 2301
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 7.636555614670953e-06,
+      "loss": 0.6421,
+      "step": 2302
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 7.6285063495359465e-06,
+      "loss": 0.5923,
+      "step": 2303
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 7.620458712140916e-06,
+      "loss": 0.6125,
+      "step": 2304
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.612412708009583e-06,
+      "loss": 0.5271,
+      "step": 2305
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.604368342664533e-06,
+      "loss": 0.6672,
+      "step": 2306
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.596325621627247e-06,
+      "loss": 0.6296,
+      "step": 2307
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 7.588284550418068e-06,
+      "loss": 0.5259,
+      "step": 2308
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.5802451345562e-06,
+      "loss": 0.6526,
+      "step": 2309
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.572207379559722e-06,
+      "loss": 0.5266,
+      "step": 2310
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.564171290945571e-06,
+      "loss": 0.5898,
+      "step": 2311
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 7.556136874229532e-06,
+      "loss": 0.5815,
+      "step": 2312
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 7.54810413492625e-06,
+      "loss": 0.5829,
+      "step": 2313
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 7.540073078549221e-06,
+      "loss": 0.6072,
+      "step": 2314
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 7.5320437106107756e-06,
+      "loss": 0.4783,
+      "step": 2315
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 7.524016036622094e-06,
+      "loss": 0.6921,
+      "step": 2316
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.515990062093196e-06,
+      "loss": 0.7051,
+      "step": 2317
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.507965792532922e-06,
+      "loss": 0.6936,
+      "step": 2318
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.499943233448958e-06,
+      "loss": 0.5818,
+      "step": 2319
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 7.491922390347801e-06,
+      "loss": 0.6213,
+      "step": 2320
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.48390326873478e-06,
+      "loss": 0.637,
+      "step": 2321
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.475885874114047e-06,
+      "loss": 0.4257,
+      "step": 2322
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.467870211988551e-06,
+      "loss": 0.6128,
+      "step": 2323
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 7.459856287860068e-06,
+      "loss": 0.5845,
+      "step": 2324
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.451844107229178e-06,
+      "loss": 0.5847,
+      "step": 2325
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.443833675595254e-06,
+      "loss": 0.5277,
+      "step": 2326
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.4358249984564824e-06,
+      "loss": 0.6396,
+      "step": 2327
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 7.427818081309841e-06,
+      "loss": 0.6138,
+      "step": 2328
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.419812929651091e-06,
+      "loss": 0.6672,
+      "step": 2329
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.411809548974792e-06,
+      "loss": 0.5369,
+      "step": 2330
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.403807944774291e-06,
+      "loss": 0.6675,
+      "step": 2331
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 7.395808122541697e-06,
+      "loss": 0.6221,
+      "step": 2332
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.387810087767916e-06,
+      "loss": 0.5916,
+      "step": 2333
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.3798138459426225e-06,
+      "loss": 0.605,
+      "step": 2334
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.371819402554248e-06,
+      "loss": 0.5686,
+      "step": 2335
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 7.363826763090005e-06,
+      "loss": 0.6172,
+      "step": 2336
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.355835933035864e-06,
+      "loss": 0.6238,
+      "step": 2337
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.3478469178765444e-06,
+      "loss": 0.5664,
+      "step": 2338
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.33985972309553e-06,
+      "loss": 0.4934,
+      "step": 2339
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 7.331874354175057e-06,
+      "loss": 0.5779,
+      "step": 2340
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.323890816596094e-06,
+      "loss": 0.6672,
+      "step": 2341
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.315909115838367e-06,
+      "loss": 0.571,
+      "step": 2342
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.307929257380337e-06,
+      "loss": 0.4821,
+      "step": 2343
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 7.299951246699198e-06,
+      "loss": 0.6162,
+      "step": 2344
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 7.291975089270877e-06,
+      "loss": 0.6853,
+      "step": 2345
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 7.284000790570029e-06,
+      "loss": 0.6038,
+      "step": 2346
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 7.276028356070033e-06,
+      "loss": 0.5499,
+      "step": 2347
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 7.268057791242994e-06,
+      "loss": 0.5767,
+      "step": 2348
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 7.2600891015597206e-06,
+      "loss": 0.696,
+      "step": 2349
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 7.252122292489747e-06,
+      "loss": 0.5571,
+      "step": 2350
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 7.244157369501317e-06,
+      "loss": 0.5095,
+      "step": 2351
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.236194338061364e-06,
+      "loss": 0.6481,
+      "step": 2352
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.22823320363554e-06,
+      "loss": 0.5256,
+      "step": 2353
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.220273971688192e-06,
+      "loss": 0.4335,
+      "step": 2354
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 7.212316647682351e-06,
+      "loss": 0.4873,
+      "step": 2355
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.204361237079746e-06,
+      "loss": 0.5273,
+      "step": 2356
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.1964077453408036e-06,
+      "loss": 0.5573,
+      "step": 2357
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.1884561779246055e-06,
+      "loss": 0.5513,
+      "step": 2358
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 7.180506540288939e-06,
+      "loss": 0.6694,
+      "step": 2359
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.172558837890256e-06,
+      "loss": 0.6262,
+      "step": 2360
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.164613076183678e-06,
+      "loss": 0.4849,
+      "step": 2361
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.156669260622997e-06,
+      "loss": 0.4659,
+      "step": 2362
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 7.148727396660675e-06,
+      "loss": 0.5588,
+      "step": 2363
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.140787489747824e-06,
+      "loss": 0.5745,
+      "step": 2364
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.132849545334218e-06,
+      "loss": 0.7861,
+      "step": 2365
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.124913568868288e-06,
+      "loss": 0.5283,
+      "step": 2366
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 7.116979565797104e-06,
+      "loss": 0.5706,
+      "step": 2367
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.109047541566392e-06,
+      "loss": 0.5537,
+      "step": 2368
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.101117501620514e-06,
+      "loss": 0.5886,
+      "step": 2369
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.0931894514024695e-06,
+      "loss": 0.7139,
+      "step": 2370
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 7.085263396353897e-06,
+      "loss": 0.571,
+      "step": 2371
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.077339341915059e-06,
+      "loss": 0.4788,
+      "step": 2372
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.069417293524854e-06,
+      "loss": 0.7007,
+      "step": 2373
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.061497256620793e-06,
+      "loss": 0.4895,
+      "step": 2374
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 7.053579236639012e-06,
+      "loss": 0.6384,
+      "step": 2375
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.045663239014265e-06,
+      "loss": 0.6052,
+      "step": 2376
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.037749269179914e-06,
+      "loss": 0.6206,
+      "step": 2377
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.029837332567928e-06,
+      "loss": 0.5474,
+      "step": 2378
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 7.0219274346088835e-06,
+      "loss": 0.6841,
+      "step": 2379
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 7.01401958073196e-06,
+      "loss": 0.5784,
+      "step": 2380
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 7.006113776364924e-06,
+      "loss": 0.5601,
+      "step": 2381
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 6.998210026934148e-06,
+      "loss": 0.6057,
+      "step": 2382
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 6.99030833786459e-06,
+      "loss": 0.6426,
+      "step": 2383
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 6.9824087145797825e-06,
+      "loss": 0.6572,
+      "step": 2384
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 6.974511162501857e-06,
+      "loss": 0.637,
+      "step": 2385
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 6.966615687051517e-06,
+      "loss": 0.5509,
+      "step": 2386
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 6.958722293648032e-06,
+      "loss": 0.5942,
+      "step": 2387
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 6.9508309877092566e-06,
+      "loss": 0.4984,
+      "step": 2388
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 6.942941774651606e-06,
+      "loss": 0.6138,
+      "step": 2389
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 6.935054659890053e-06,
+      "loss": 0.5869,
+      "step": 2390
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 6.927169648838139e-06,
+      "loss": 0.5283,
+      "step": 2391
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 6.9192867469079625e-06,
+      "loss": 0.585,
+      "step": 2392
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 6.911405959510164e-06,
+      "loss": 0.502,
+      "step": 2393
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 6.903527292053942e-06,
+      "loss": 0.6763,
+      "step": 2394
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 6.895650749947041e-06,
+      "loss": 0.7017,
+      "step": 2395
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 6.887776338595735e-06,
+      "loss": 0.5598,
+      "step": 2396
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 6.8799040634048496e-06,
+      "loss": 0.5471,
+      "step": 2397
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 6.872033929777731e-06,
+      "loss": 0.5034,
+      "step": 2398
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 6.864165943116265e-06,
+      "loss": 0.6045,
+      "step": 2399
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 6.856300108820865e-06,
+      "loss": 0.5543,
+      "step": 2400
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 6.8484364322904565e-06,
+      "loss": 0.7234,
+      "step": 2401
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 6.8405749189224935e-06,
+      "loss": 0.4517,
+      "step": 2402
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 6.832715574112946e-06,
+      "loss": 0.6157,
+      "step": 2403
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 6.824858403256284e-06,
+      "loss": 0.5994,
+      "step": 2404
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 6.8170034117454965e-06,
+      "loss": 0.6953,
+      "step": 2405
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 6.809150604972079e-06,
+      "loss": 0.5752,
+      "step": 2406
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 6.801299988326011e-06,
+      "loss": 0.6182,
+      "step": 2407
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 6.793451567195785e-06,
+      "loss": 0.5103,
+      "step": 2408
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 6.785605346968387e-06,
+      "loss": 0.5439,
+      "step": 2409
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 6.777761333029274e-06,
+      "loss": 0.4785,
+      "step": 2410
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.769919530762409e-06,
+      "loss": 0.5806,
+      "step": 2411
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.7620799455502305e-06,
+      "loss": 0.4839,
+      "step": 2412
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.754242582773646e-06,
+      "loss": 0.6499,
+      "step": 2413
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 6.746407447812049e-06,
+      "loss": 0.5027,
+      "step": 2414
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 6.738574546043306e-06,
+      "loss": 0.6025,
+      "step": 2415
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 6.730743882843736e-06,
+      "loss": 0.6221,
+      "step": 2416
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 6.7229154635881334e-06,
+      "loss": 0.5627,
+      "step": 2417
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 6.7150892936497525e-06,
+      "loss": 0.7578,
+      "step": 2418
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 6.7072653784002965e-06,
+      "loss": 0.4975,
+      "step": 2419
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 6.6994437232099245e-06,
+      "loss": 0.5876,
+      "step": 2420
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 6.691624333447252e-06,
+      "loss": 0.6636,
+      "step": 2421
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 6.683807214479323e-06,
+      "loss": 0.6775,
+      "step": 2422
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 6.675992371671641e-06,
+      "loss": 0.626,
+      "step": 2423
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 6.668179810388131e-06,
+      "loss": 0.5988,
+      "step": 2424
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 6.660369535991162e-06,
+      "loss": 0.5632,
+      "step": 2425
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 6.652561553841537e-06,
+      "loss": 0.5237,
+      "step": 2426
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 6.64475586929847e-06,
+      "loss": 0.6367,
+      "step": 2427
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 6.636952487719612e-06,
+      "loss": 0.6265,
+      "step": 2428
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 6.629151414461033e-06,
+      "loss": 0.5693,
+      "step": 2429
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 6.621352654877207e-06,
+      "loss": 0.7119,
+      "step": 2430
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 6.613556214321031e-06,
+      "loss": 0.5035,
+      "step": 2431
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 6.60576209814381e-06,
+      "loss": 0.522,
+      "step": 2432
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 6.597970311695242e-06,
+      "loss": 0.6487,
+      "step": 2433
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 6.59018086032344e-06,
+      "loss": 0.6221,
+      "step": 2434
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 6.582393749374911e-06,
+      "loss": 0.5273,
+      "step": 2435
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 6.574608984194546e-06,
+      "loss": 0.6331,
+      "step": 2436
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 6.566826570125635e-06,
+      "loss": 0.5186,
+      "step": 2437
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 6.559046512509859e-06,
+      "loss": 0.5367,
+      "step": 2438
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 6.551268816687265e-06,
+      "loss": 0.5104,
+      "step": 2439
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 6.543493487996293e-06,
+      "loss": 0.7012,
+      "step": 2440
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 6.535720531773757e-06,
+      "loss": 0.4891,
+      "step": 2441
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.527949953354835e-06,
+      "loss": 0.5876,
+      "step": 2442
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.5201817580730786e-06,
+      "loss": 0.5564,
+      "step": 2443
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.5124159512604055e-06,
+      "loss": 0.5806,
+      "step": 2444
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 6.504652538247088e-06,
+      "loss": 0.5591,
+      "step": 2445
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 6.496891524361757e-06,
+      "loss": 0.5144,
+      "step": 2446
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 6.489132914931403e-06,
+      "loss": 0.6924,
+      "step": 2447
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 6.481376715281358e-06,
+      "loss": 0.5664,
+      "step": 2448
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 6.4736229307353035e-06,
+      "loss": 0.5952,
+      "step": 2449
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 6.4658715666152625e-06,
+      "loss": 0.5566,
+      "step": 2450
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 6.458122628241599e-06,
+      "loss": 0.7073,
+      "step": 2451
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 6.450376120933008e-06,
+      "loss": 0.5753,
+      "step": 2452
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 6.4426320500065196e-06,
+      "loss": 0.6105,
+      "step": 2453
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.434890420777491e-06,
+      "loss": 0.5308,
+      "step": 2454
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.427151238559602e-06,
+      "loss": 0.5645,
+      "step": 2455
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.419414508664851e-06,
+      "loss": 0.6685,
+      "step": 2456
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 6.411680236403559e-06,
+      "loss": 0.4956,
+      "step": 2457
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 6.403948427084356e-06,
+      "loss": 0.4897,
+      "step": 2458
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 6.396219086014179e-06,
+      "loss": 0.5615,
+      "step": 2459
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 6.388492218498277e-06,
+      "loss": 0.5637,
+      "step": 2460
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 6.380767829840202e-06,
+      "loss": 0.5073,
+      "step": 2461
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 6.373045925341794e-06,
+      "loss": 0.6006,
+      "step": 2462
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 6.365326510303196e-06,
+      "loss": 0.6123,
+      "step": 2463
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 6.357609590022849e-06,
+      "loss": 0.5222,
+      "step": 2464
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 6.349895169797462e-06,
+      "loss": 0.5916,
+      "step": 2465
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 6.342183254922045e-06,
+      "loss": 0.5349,
+      "step": 2466
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 6.334473850689889e-06,
+      "loss": 0.577,
+      "step": 2467
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 6.326766962392544e-06,
+      "loss": 0.6077,
+      "step": 2468
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 6.319062595319853e-06,
+      "loss": 0.6404,
+      "step": 2469
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 6.311360754759923e-06,
+      "loss": 0.5872,
+      "step": 2470
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 6.303661445999116e-06,
+      "loss": 0.4237,
+      "step": 2471
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 6.295964674322069e-06,
+      "loss": 0.6597,
+      "step": 2472
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 6.288270445011677e-06,
+      "loss": 0.6614,
+      "step": 2473
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 6.280578763349079e-06,
+      "loss": 0.5771,
+      "step": 2474
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 6.272889634613679e-06,
+      "loss": 0.5627,
+      "step": 2475
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 6.265203064083117e-06,
+      "loss": 0.5632,
+      "step": 2476
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 6.257519057033281e-06,
+      "loss": 0.604,
+      "step": 2477
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 6.24983761873831e-06,
+      "loss": 0.6689,
+      "step": 2478
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 6.242158754470562e-06,
+      "loss": 0.4523,
+      "step": 2479
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 6.234482469500639e-06,
+      "loss": 0.6189,
+      "step": 2480
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 6.226808769097376e-06,
+      "loss": 0.4668,
+      "step": 2481
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 6.219137658527819e-06,
+      "loss": 0.5437,
+      "step": 2482
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 6.211469143057252e-06,
+      "loss": 0.5535,
+      "step": 2483
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 6.203803227949175e-06,
+      "loss": 0.6428,
+      "step": 2484
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 6.196139918465291e-06,
+      "loss": 0.6055,
+      "step": 2485
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 6.188479219865529e-06,
+      "loss": 0.5759,
+      "step": 2486
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 6.180821137408024e-06,
+      "loss": 0.7087,
+      "step": 2487
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 6.173165676349103e-06,
+      "loss": 0.4775,
+      "step": 2488
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 6.165512841943308e-06,
+      "loss": 0.6423,
+      "step": 2489
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 6.157862639443375e-06,
+      "loss": 0.583,
+      "step": 2490
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 6.1502150741002254e-06,
+      "loss": 0.6475,
+      "step": 2491
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 6.142570151162977e-06,
+      "loss": 0.6575,
+      "step": 2492
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 6.13492787587894e-06,
+      "loss": 0.5426,
+      "step": 2493
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 6.127288253493591e-06,
+      "loss": 0.5234,
+      "step": 2494
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 6.119651289250599e-06,
+      "loss": 0.5959,
+      "step": 2495
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 6.1120169883918115e-06,
+      "loss": 0.6753,
+      "step": 2496
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.10438535615723e-06,
+      "loss": 0.5344,
+      "step": 2497
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.096756397785039e-06,
+      "loss": 0.5916,
+      "step": 2498
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.089130118511593e-06,
+      "loss": 0.6523,
+      "step": 2499
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 6.081506523571384e-06,
+      "loss": 0.6704,
+      "step": 2500
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 6.0738856181970865e-06,
+      "loss": 0.5839,
+      "step": 2501
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 6.06626740761952e-06,
+      "loss": 0.6082,
+      "step": 2502
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 6.058651897067642e-06,
+      "loss": 0.6404,
+      "step": 2503
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 6.051039091768582e-06,
+      "loss": 0.6365,
+      "step": 2504
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 6.043428996947587e-06,
+      "loss": 0.6011,
+      "step": 2505
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 6.03582161782806e-06,
+      "loss": 0.4866,
+      "step": 2506
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 6.02821695963154e-06,
+      "loss": 0.6208,
+      "step": 2507
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 6.020615027577683e-06,
+      "loss": 0.6169,
+      "step": 2508
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 6.013015826884291e-06,
+      "loss": 0.6206,
+      "step": 2509
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 6.005419362767286e-06,
+      "loss": 0.575,
+      "step": 2510
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 5.997825640440706e-06,
+      "loss": 0.6257,
+      "step": 2511
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 5.990234665116713e-06,
+      "loss": 0.5667,
+      "step": 2512
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 5.982646442005586e-06,
+      "loss": 0.5889,
+      "step": 2513
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 5.975060976315703e-06,
+      "loss": 0.7297,
+      "step": 2514
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 5.967478273253562e-06,
+      "loss": 0.5906,
+      "step": 2515
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 5.959898338023764e-06,
+      "loss": 0.5446,
+      "step": 2516
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 5.952321175828998e-06,
+      "loss": 0.6599,
+      "step": 2517
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 5.944746791870062e-06,
+      "loss": 0.5688,
+      "step": 2518
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 5.937175191345844e-06,
+      "loss": 0.6592,
+      "step": 2519
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 5.92960637945332e-06,
+      "loss": 0.5127,
+      "step": 2520
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 5.92204036138755e-06,
+      "loss": 0.6133,
+      "step": 2521
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 5.914477142341682e-06,
+      "loss": 0.5813,
+      "step": 2522
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 5.906916727506939e-06,
+      "loss": 0.6008,
+      "step": 2523
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 5.899359122072617e-06,
+      "loss": 0.5349,
+      "step": 2524
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 5.891804331226092e-06,
+      "loss": 0.6367,
+      "step": 2525
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 5.8842523601528e-06,
+      "loss": 0.6125,
+      "step": 2526
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 5.876703214036241e-06,
+      "loss": 0.6243,
+      "step": 2527
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 5.869156898057983e-06,
+      "loss": 0.5448,
+      "step": 2528
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 5.86161341739765e-06,
+      "loss": 0.5322,
+      "step": 2529
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 5.854072777232914e-06,
+      "loss": 0.5254,
+      "step": 2530
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 5.8465349827395e-06,
+      "loss": 0.6345,
+      "step": 2531
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 5.839000039091185e-06,
+      "loss": 0.5055,
+      "step": 2532
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 5.831467951459784e-06,
+      "loss": 0.5886,
+      "step": 2533
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 5.823938725015148e-06,
+      "loss": 0.6577,
+      "step": 2534
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 5.816412364925173e-06,
+      "loss": 0.5242,
+      "step": 2535
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 5.808888876355785e-06,
+      "loss": 0.6389,
+      "step": 2536
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 5.8013682644709335e-06,
+      "loss": 0.4666,
+      "step": 2537
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 5.793850534432599e-06,
+      "loss": 0.4861,
+      "step": 2538
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 5.786335691400789e-06,
+      "loss": 0.5762,
+      "step": 2539
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 5.77882374053351e-06,
+      "loss": 0.5767,
+      "step": 2540
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 5.771314686986806e-06,
+      "loss": 0.5786,
+      "step": 2541
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 5.7638085359147235e-06,
+      "loss": 0.5679,
+      "step": 2542
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 5.756305292469308e-06,
+      "loss": 0.5746,
+      "step": 2543
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 5.748804961800623e-06,
+      "loss": 0.506,
+      "step": 2544
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 5.74130754905673e-06,
+      "loss": 0.5742,
+      "step": 2545
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 5.733813059383676e-06,
+      "loss": 0.5903,
+      "step": 2546
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 5.726321497925516e-06,
+      "loss": 0.5757,
+      "step": 2547
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 5.718832869824291e-06,
+      "loss": 0.4532,
+      "step": 2548
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 5.711347180220023e-06,
+      "loss": 0.6234,
+      "step": 2549
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 5.703864434250721e-06,
+      "loss": 0.5565,
+      "step": 2550
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 5.696384637052381e-06,
+      "loss": 0.5427,
+      "step": 2551
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 5.688907793758959e-06,
+      "loss": 0.658,
+      "step": 2552
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 5.681433909502394e-06,
+      "loss": 0.5642,
+      "step": 2553
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 5.673962989412599e-06,
+      "loss": 0.6697,
+      "step": 2554
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 5.666495038617437e-06,
+      "loss": 0.4456,
+      "step": 2555
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 5.659030062242746e-06,
+      "loss": 0.5298,
+      "step": 2556
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 5.65156806541232e-06,
+      "loss": 0.6648,
+      "step": 2557
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 5.6441090532479016e-06,
+      "loss": 0.6693,
+      "step": 2558
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 5.63665303086919e-06,
+      "loss": 0.5776,
+      "step": 2559
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 5.629200003393838e-06,
+      "loss": 0.561,
+      "step": 2560
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 5.621749975937425e-06,
+      "loss": 0.561,
+      "step": 2561
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 5.61430295361349e-06,
+      "loss": 0.5477,
+      "step": 2562
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 5.606858941533504e-06,
+      "loss": 0.644,
+      "step": 2563
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 5.5994179448068616e-06,
+      "loss": 0.645,
+      "step": 2564
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 5.591979968540902e-06,
+      "loss": 0.5251,
+      "step": 2565
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 5.584545017840886e-06,
+      "loss": 0.6101,
+      "step": 2566
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 5.5771130978099896e-06,
+      "loss": 0.5977,
+      "step": 2567
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 5.56968421354932e-06,
+      "loss": 0.6436,
+      "step": 2568
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 5.5622583701578984e-06,
+      "loss": 0.6233,
+      "step": 2569
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 5.554835572732658e-06,
+      "loss": 0.4753,
+      "step": 2570
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 5.547415826368434e-06,
+      "loss": 0.5011,
+      "step": 2571
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 5.539999136157977e-06,
+      "loss": 0.5457,
+      "step": 2572
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 5.532585507191941e-06,
+      "loss": 0.6001,
+      "step": 2573
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 5.525174944558866e-06,
+      "loss": 0.5442,
+      "step": 2574
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 5.517767453345199e-06,
+      "loss": 0.6331,
+      "step": 2575
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 5.510363038635282e-06,
+      "loss": 0.6047,
+      "step": 2576
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 5.502961705511329e-06,
+      "loss": 0.5691,
+      "step": 2577
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 5.495563459053455e-06,
+      "loss": 0.6289,
+      "step": 2578
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 5.48816830433965e-06,
+      "loss": 0.6035,
+      "step": 2579
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 5.480776246445776e-06,
+      "loss": 0.5215,
+      "step": 2580
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 5.473387290445582e-06,
+      "loss": 0.5764,
+      "step": 2581
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 5.4660014414106825e-06,
+      "loss": 0.5728,
+      "step": 2582
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 5.4586187044105524e-06,
+      "loss": 0.6606,
+      "step": 2583
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 5.451239084512537e-06,
+      "loss": 0.6079,
+      "step": 2584
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 5.44386258678185e-06,
+      "loss": 0.6719,
+      "step": 2585
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 5.436489216281544e-06,
+      "loss": 0.6453,
+      "step": 2586
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 5.4291189780725385e-06,
+      "loss": 0.6035,
+      "step": 2587
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 5.421751877213603e-06,
+      "loss": 0.5342,
+      "step": 2588
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 5.4143879187613435e-06,
+      "loss": 0.6704,
+      "step": 2589
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 5.40702710777022e-06,
+      "loss": 0.6011,
+      "step": 2590
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 5.399669449292529e-06,
+      "loss": 0.5648,
+      "step": 2591
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 5.392314948378396e-06,
+      "loss": 0.667,
+      "step": 2592
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 5.384963610075791e-06,
+      "loss": 0.6177,
+      "step": 2593
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 5.377615439430508e-06,
+      "loss": 0.4595,
+      "step": 2594
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 5.37027044148616e-06,
+      "loss": 0.6479,
+      "step": 2595
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 5.362928621284193e-06,
+      "loss": 0.5896,
+      "step": 2596
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 5.355589983863872e-06,
+      "loss": 0.5177,
+      "step": 2597
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 5.348254534262262e-06,
+      "loss": 0.5881,
+      "step": 2598
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 5.340922277514258e-06,
+      "loss": 0.5287,
+      "step": 2599
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 5.333593218652557e-06,
+      "loss": 0.6584,
+      "step": 2600
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 5.3262673627076555e-06,
+      "loss": 0.5242,
+      "step": 2601
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 5.318944714707861e-06,
+      "loss": 0.5282,
+      "step": 2602
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 5.3116252796792755e-06,
+      "loss": 0.5786,
+      "step": 2603
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 5.304309062645788e-06,
+      "loss": 0.5737,
+      "step": 2604
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 5.29699606862909e-06,
+      "loss": 0.5381,
+      "step": 2605
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 5.289686302648661e-06,
+      "loss": 0.5117,
+      "step": 2606
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 5.28237976972175e-06,
+      "loss": 0.5294,
+      "step": 2607
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 5.275076474863408e-06,
+      "loss": 0.5398,
+      "step": 2608
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 5.2677764230864434e-06,
+      "loss": 0.5952,
+      "step": 2609
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 5.260479619401451e-06,
+      "loss": 0.522,
+      "step": 2610
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 5.253186068816795e-06,
+      "loss": 0.5791,
+      "step": 2611
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 5.2458957763386005e-06,
+      "loss": 0.6079,
+      "step": 2612
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 5.2386087469707644e-06,
+      "loss": 0.6323,
+      "step": 2613
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 5.231324985714942e-06,
+      "loss": 0.5862,
+      "step": 2614
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 5.224044497570535e-06,
+      "loss": 0.5461,
+      "step": 2615
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 5.2167672875347135e-06,
+      "loss": 0.6548,
+      "step": 2616
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 5.209493360602392e-06,
+      "loss": 0.5605,
+      "step": 2617
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 5.202222721766226e-06,
+      "loss": 0.5867,
+      "step": 2618
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 5.19495537601662e-06,
+      "loss": 0.6394,
+      "step": 2619
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 5.1876913283417196e-06,
+      "loss": 0.6077,
+      "step": 2620
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 5.180430583727401e-06,
+      "loss": 0.4599,
+      "step": 2621
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 5.173173147157275e-06,
+      "loss": 0.604,
+      "step": 2622
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 5.165919023612691e-06,
+      "loss": 0.6151,
+      "step": 2623
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 5.158668218072706e-06,
+      "loss": 0.5206,
+      "step": 2624
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 5.151420735514114e-06,
+      "loss": 0.5498,
+      "step": 2625
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 5.144176580911431e-06,
+      "loss": 0.6462,
+      "step": 2626
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 5.13693575923687e-06,
+      "loss": 0.6177,
+      "step": 2627
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 5.129698275460375e-06,
+      "loss": 0.5576,
+      "step": 2628
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 5.122464134549596e-06,
+      "loss": 0.5574,
+      "step": 2629
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 5.1152333414698774e-06,
+      "loss": 0.4807,
+      "step": 2630
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 5.1080059011842764e-06,
+      "loss": 0.5447,
+      "step": 2631
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 5.100781818653549e-06,
+      "loss": 0.5173,
+      "step": 2632
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 5.093561098836135e-06,
+      "loss": 0.55,
+      "step": 2633
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 5.086343746688184e-06,
+      "loss": 0.5361,
+      "step": 2634
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 5.079129767163516e-06,
+      "loss": 0.5486,
+      "step": 2635
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 5.071919165213647e-06,
+      "loss": 0.6504,
+      "step": 2636
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 5.064711945787777e-06,
+      "loss": 0.584,
+      "step": 2637
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 5.057508113832772e-06,
+      "loss": 0.6519,
+      "step": 2638
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 5.050307674293185e-06,
+      "loss": 0.557,
+      "step": 2639
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 5.043110632111239e-06,
+      "loss": 0.4934,
+      "step": 2640
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 5.035916992226814e-06,
+      "loss": 0.5945,
+      "step": 2641
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 5.028726759577467e-06,
+      "loss": 0.6318,
+      "step": 2642
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 5.021539939098416e-06,
+      "loss": 0.6052,
+      "step": 2643
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 5.014356535722527e-06,
+      "loss": 0.5619,
+      "step": 2644
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 5.007176554380328e-06,
+      "loss": 0.6306,
+      "step": 2645
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 5.000000000000003e-06,
+      "loss": 0.479,
+      "step": 2646
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.9928268775073675e-06,
+      "loss": 0.5718,
+      "step": 2647
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.9856571918258976e-06,
+      "loss": 0.5063,
+      "step": 2648
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 4.978490947876709e-06,
+      "loss": 0.4893,
+      "step": 2649
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.9713281505785404e-06,
+      "loss": 0.6329,
+      "step": 2650
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.96416880484778e-06,
+      "loss": 0.6558,
+      "step": 2651
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 4.957012915598446e-06,
+      "loss": 0.6421,
+      "step": 2652
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.949860487742173e-06,
+      "loss": 0.6023,
+      "step": 2653
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.94271152618823e-06,
+      "loss": 0.5356,
+      "step": 2654
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.935566035843506e-06,
+      "loss": 0.4844,
+      "step": 2655
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 4.928424021612499e-06,
+      "loss": 0.5079,
+      "step": 2656
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.921285488397332e-06,
+      "loss": 0.6333,
+      "step": 2657
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.914150441097736e-06,
+      "loss": 0.6758,
+      "step": 2658
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.907018884611039e-06,
+      "loss": 0.5979,
+      "step": 2659
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 4.899890823832191e-06,
+      "loss": 0.5784,
+      "step": 2660
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.892766263653723e-06,
+      "loss": 0.5581,
+      "step": 2661
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.885645208965779e-06,
+      "loss": 0.6045,
+      "step": 2662
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.8785276646560935e-06,
+      "loss": 0.5364,
+      "step": 2663
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 4.871413635609981e-06,
+      "loss": 0.6387,
+      "step": 2664
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.8643031267103556e-06,
+      "loss": 0.6821,
+      "step": 2665
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.857196142837716e-06,
+      "loss": 0.6631,
+      "step": 2666
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.850092688870126e-06,
+      "loss": 0.4814,
+      "step": 2667
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 4.8429927696832434e-06,
+      "loss": 0.556,
+      "step": 2668
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.835896390150295e-06,
+      "loss": 0.5378,
+      "step": 2669
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.82880355514207e-06,
+      "loss": 0.5469,
+      "step": 2670
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.821714269526935e-06,
+      "loss": 0.6628,
+      "step": 2671
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 4.814628538170818e-06,
+      "loss": 0.573,
+      "step": 2672
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.807546365937201e-06,
+      "loss": 0.6038,
+      "step": 2673
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.800467757687131e-06,
+      "loss": 0.5488,
+      "step": 2674
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.793392718279207e-06,
+      "loss": 0.4623,
+      "step": 2675
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 4.786321252569574e-06,
+      "loss": 0.5542,
+      "step": 2676
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.779253365411927e-06,
+      "loss": 0.4998,
+      "step": 2677
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.772189061657511e-06,
+      "loss": 0.3364,
+      "step": 2678
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.765128346155098e-06,
+      "loss": 0.6309,
+      "step": 2679
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 4.758071223751006e-06,
+      "loss": 0.6855,
+      "step": 2680
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.751017699289092e-06,
+      "loss": 0.5291,
+      "step": 2681
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.743967777610727e-06,
+      "loss": 0.5996,
+      "step": 2682
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.736921463554824e-06,
+      "loss": 0.59,
+      "step": 2683
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 4.729878761957819e-06,
+      "loss": 0.4818,
+      "step": 2684
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.722839677653654e-06,
+      "loss": 0.6348,
+      "step": 2685
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.7158042154738094e-06,
+      "loss": 0.583,
+      "step": 2686
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.708772380247257e-06,
+      "loss": 0.6912,
+      "step": 2687
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 4.701744176800498e-06,
+      "loss": 0.6106,
+      "step": 2688
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.694719609957536e-06,
+      "loss": 0.4727,
+      "step": 2689
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.6876986845398665e-06,
+      "loss": 0.5515,
+      "step": 2690
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.6806814053665e-06,
+      "loss": 0.5357,
+      "step": 2691
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 4.673667777253944e-06,
+      "loss": 0.5643,
+      "step": 2692
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.666657805016187e-06,
+      "loss": 0.5308,
+      "step": 2693
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.659651493464721e-06,
+      "loss": 0.5857,
+      "step": 2694
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 4.652648847408523e-06,
+      "loss": 0.5894,
+      "step": 2695
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.645649871654045e-06,
+      "loss": 0.6061,
+      "step": 2696
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.638654571005231e-06,
+      "loss": 0.5535,
+      "step": 2697
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.631662950263502e-06,
+      "loss": 0.7366,
+      "step": 2698
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 4.624675014227743e-06,
+      "loss": 0.5366,
+      "step": 2699
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.617690767694319e-06,
+      "loss": 0.5994,
+      "step": 2700
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.610710215457062e-06,
+      "loss": 0.6096,
+      "step": 2701
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.603733362307261e-06,
+      "loss": 0.6428,
+      "step": 2702
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 4.596760213033674e-06,
+      "loss": 0.4724,
+      "step": 2703
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.589790772422519e-06,
+      "loss": 0.6379,
+      "step": 2704
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.582825045257455e-06,
+      "loss": 0.6685,
+      "step": 2705
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.575863036319604e-06,
+      "loss": 0.6882,
+      "step": 2706
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 4.568904750387538e-06,
+      "loss": 0.5659,
+      "step": 2707
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.561950192237259e-06,
+      "loss": 0.6465,
+      "step": 2708
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.554999366642225e-06,
+      "loss": 0.5977,
+      "step": 2709
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.548052278373327e-06,
+      "loss": 0.6057,
+      "step": 2710
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 4.541108932198887e-06,
+      "loss": 0.5388,
+      "step": 2711
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.534169332884662e-06,
+      "loss": 0.5779,
+      "step": 2712
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.527233485193843e-06,
+      "loss": 0.4216,
+      "step": 2713
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.520301393887032e-06,
+      "loss": 0.5977,
+      "step": 2714
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 4.513373063722266e-06,
+      "loss": 0.6108,
+      "step": 2715
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.506448499454996e-06,
+      "loss": 0.6053,
+      "step": 2716
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.499527705838081e-06,
+      "loss": 0.5981,
+      "step": 2717
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.492610687621805e-06,
+      "loss": 0.5737,
+      "step": 2718
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 4.485697449553853e-06,
+      "loss": 0.5671,
+      "step": 2719
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.478787996379311e-06,
+      "loss": 0.6226,
+      "step": 2720
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.471882332840677e-06,
+      "loss": 0.5791,
+      "step": 2721
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.464980463677846e-06,
+      "loss": 0.6191,
+      "step": 2722
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 4.458082393628098e-06,
+      "loss": 0.6177,
+      "step": 2723
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.451188127426119e-06,
+      "loss": 0.6562,
+      "step": 2724
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.444297669803981e-06,
+      "loss": 0.5635,
+      "step": 2725
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.437411025491131e-06,
+      "loss": 0.6469,
+      "step": 2726
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 4.4305281992144125e-06,
+      "loss": 0.4852,
+      "step": 2727
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.423649195698042e-06,
+      "loss": 0.4659,
+      "step": 2728
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.416774019663615e-06,
+      "loss": 0.6509,
+      "step": 2729
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.409902675830094e-06,
+      "loss": 0.378,
+      "step": 2730
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 4.403035168913817e-06,
+      "loss": 0.528,
+      "step": 2731
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.39617150362849e-06,
+      "loss": 0.6169,
+      "step": 2732
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.389311684685171e-06,
+      "loss": 0.5767,
+      "step": 2733
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.382455716792291e-06,
+      "loss": 0.6113,
+      "step": 2734
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 4.375603604655634e-06,
+      "loss": 0.5391,
+      "step": 2735
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.368755352978329e-06,
+      "loss": 0.5642,
+      "step": 2736
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.361910966460866e-06,
+      "loss": 0.5154,
+      "step": 2737
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 4.355070449801083e-06,
+      "loss": 0.5271,
+      "step": 2738
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.348233807694147e-06,
+      "loss": 0.5879,
+      "step": 2739
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.341401044832583e-06,
+      "loss": 0.6624,
+      "step": 2740
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.334572165906247e-06,
+      "loss": 0.5227,
+      "step": 2741
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 4.327747175602321e-06,
+      "loss": 0.4434,
+      "step": 2742
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.320926078605328e-06,
+      "loss": 0.5115,
+      "step": 2743
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.314108879597122e-06,
+      "loss": 0.6509,
+      "step": 2744
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.307295583256866e-06,
+      "loss": 0.6062,
+      "step": 2745
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 4.3004861942610575e-06,
+      "loss": 0.6384,
+      "step": 2746
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.293680717283512e-06,
+      "loss": 0.4541,
+      "step": 2747
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.286879156995349e-06,
+      "loss": 0.5569,
+      "step": 2748
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.28008151806501e-06,
+      "loss": 0.5581,
+      "step": 2749
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 4.273287805158245e-06,
+      "loss": 0.4563,
+      "step": 2750
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.2664980229381e-06,
+      "loss": 0.5472,
+      "step": 2751
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.25971217606493e-06,
+      "loss": 0.5219,
+      "step": 2752
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.252930269196395e-06,
+      "loss": 0.6018,
+      "step": 2753
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 4.246152306987434e-06,
+      "loss": 0.5457,
+      "step": 2754
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.239378294090292e-06,
+      "loss": 0.5718,
+      "step": 2755
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.2326082351545025e-06,
+      "loss": 0.6328,
+      "step": 2756
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.225842134826876e-06,
+      "loss": 0.587,
+      "step": 2757
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 4.219079997751515e-06,
+      "loss": 0.5261,
+      "step": 2758
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.212321828569801e-06,
+      "loss": 0.5923,
+      "step": 2759
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.2055676319203835e-06,
+      "loss": 0.4623,
+      "step": 2760
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.198817412439194e-06,
+      "loss": 0.4893,
+      "step": 2761
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 4.192071174759435e-06,
+      "loss": 0.585,
+      "step": 2762
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.185328923511568e-06,
+      "loss": 0.5133,
+      "step": 2763
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.178590663323323e-06,
+      "loss": 0.4915,
+      "step": 2764
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.171856398819696e-06,
+      "loss": 0.5226,
+      "step": 2765
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 4.165126134622927e-06,
+      "loss": 0.5679,
+      "step": 2766
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.158399875352525e-06,
+      "loss": 0.6064,
+      "step": 2767
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.151677625625237e-06,
+      "loss": 0.5094,
+      "step": 2768
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.1449593900550686e-06,
+      "loss": 0.5646,
+      "step": 2769
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 4.138245173253267e-06,
+      "loss": 0.6111,
+      "step": 2770
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.1315349798283145e-06,
+      "loss": 0.5255,
+      "step": 2771
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.124828814385939e-06,
+      "loss": 0.6107,
+      "step": 2772
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.118126681529108e-06,
+      "loss": 0.6038,
+      "step": 2773
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 4.111428585858005e-06,
+      "loss": 0.5986,
+      "step": 2774
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.104734531970058e-06,
+      "loss": 0.4851,
+      "step": 2775
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.0980445244599174e-06,
+      "loss": 0.5332,
+      "step": 2776
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.091358567919449e-06,
+      "loss": 0.6104,
+      "step": 2777
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 4.084676666937745e-06,
+      "loss": 0.564,
+      "step": 2778
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.077998826101115e-06,
+      "loss": 0.4876,
+      "step": 2779
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.071325049993073e-06,
+      "loss": 0.561,
+      "step": 2780
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 4.064655343194354e-06,
+      "loss": 0.532,
+      "step": 2781
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.057989710282897e-06,
+      "loss": 0.626,
+      "step": 2782
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.051328155833835e-06,
+      "loss": 0.5239,
+      "step": 2783
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.044670684419513e-06,
+      "loss": 0.6035,
+      "step": 2784
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 4.038017300609475e-06,
+      "loss": 0.5461,
+      "step": 2785
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.031368008970446e-06,
+      "loss": 0.5383,
+      "step": 2786
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.024722814066353e-06,
+      "loss": 0.5764,
+      "step": 2787
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.018081720458314e-06,
+      "loss": 0.5237,
+      "step": 2788
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 4.011444732704618e-06,
+      "loss": 0.6484,
+      "step": 2789
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 4.0048118553607485e-06,
+      "loss": 0.5523,
+      "step": 2790
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 3.998183092979368e-06,
+      "loss": 0.5771,
+      "step": 2791
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 3.9915584501103e-06,
+      "loss": 0.5039,
+      "step": 2792
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 3.9849379313005605e-06,
+      "loss": 0.5078,
+      "step": 2793
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 3.978321541094318e-06,
+      "loss": 0.625,
+      "step": 2794
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 3.9717092840329155e-06,
+      "loss": 0.6609,
+      "step": 2795
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 3.965101164654862e-06,
+      "loss": 0.668,
+      "step": 2796
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 3.958497187495815e-06,
+      "loss": 0.6044,
+      "step": 2797
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 3.951897357088602e-06,
+      "loss": 0.5833,
+      "step": 2798
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 3.9453016779631995e-06,
+      "loss": 0.5271,
+      "step": 2799
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 3.938710154646727e-06,
+      "loss": 0.5833,
+      "step": 2800
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 3.932122791663461e-06,
+      "loss": 0.5923,
+      "step": 2801
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 3.925539593534824e-06,
+      "loss": 0.5383,
+      "step": 2802
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 3.918960564779368e-06,
+      "loss": 0.5198,
+      "step": 2803
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 3.912385709912794e-06,
+      "loss": 0.5359,
+      "step": 2804
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 3.905815033447937e-06,
+      "loss": 0.5623,
+      "step": 2805
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 3.899248539894756e-06,
+      "loss": 0.5594,
+      "step": 2806
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 3.89268623376035e-06,
+      "loss": 0.6746,
+      "step": 2807
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 3.886128119548938e-06,
+      "loss": 0.6213,
+      "step": 2808
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 3.879574201761859e-06,
+      "loss": 0.5602,
+      "step": 2809
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 3.8730244848975766e-06,
+      "loss": 0.5073,
+      "step": 2810
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 3.866478973451673e-06,
+      "loss": 0.6836,
+      "step": 2811
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 3.859937671916833e-06,
+      "loss": 0.6067,
+      "step": 2812
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 3.853400584782862e-06,
+      "loss": 0.47,
+      "step": 2813
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 3.846867716536676e-06,
+      "loss": 0.4789,
+      "step": 2814
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 3.840339071662279e-06,
+      "loss": 0.6577,
+      "step": 2815
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 3.83381465464079e-06,
+      "loss": 0.5796,
+      "step": 2816
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 3.8272944699504285e-06,
+      "loss": 0.6772,
+      "step": 2817
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 3.8207785220664935e-06,
+      "loss": 0.636,
+      "step": 2818
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 3.8142668154613925e-06,
+      "loss": 0.6306,
+      "step": 2819
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 3.8077593546046087e-06,
+      "loss": 0.5515,
+      "step": 2820
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 3.801256143962719e-06,
+      "loss": 0.6304,
+      "step": 2821
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 3.794757187999386e-06,
+      "loss": 0.7383,
+      "step": 2822
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 3.7882624911753386e-06,
+      "loss": 0.4712,
+      "step": 2823
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 3.7817720579483962e-06,
+      "loss": 0.4912,
+      "step": 2824
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 3.7752858927734492e-06,
+      "loss": 0.6938,
+      "step": 2825
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 3.7688040001024473e-06,
+      "loss": 0.5084,
+      "step": 2826
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 3.7623263843844216e-06,
+      "loss": 0.5591,
+      "step": 2827
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 3.7558530500654643e-06,
+      "loss": 0.6311,
+      "step": 2828
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 3.7493840015887196e-06,
+      "loss": 0.5283,
+      "step": 2829
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 3.7429192433944016e-06,
+      "loss": 0.5413,
+      "step": 2830
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 3.736458779919777e-06,
+      "loss": 0.5955,
+      "step": 2831
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 3.7300026155991574e-06,
+      "loss": 0.5986,
+      "step": 2832
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 3.7235507548639127e-06,
+      "loss": 0.6243,
+      "step": 2833
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 3.717103202142457e-06,
+      "loss": 0.582,
+      "step": 2834
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 3.710659961860239e-06,
+      "loss": 0.6775,
+      "step": 2835
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 3.704221038439759e-06,
+      "loss": 0.7607,
+      "step": 2836
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 3.697786436300552e-06,
+      "loss": 0.6262,
+      "step": 2837
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 3.6913561598591775e-06,
+      "loss": 0.5328,
+      "step": 2838
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 3.6849302135292354e-06,
+      "loss": 0.4774,
+      "step": 2839
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 3.6785086017213555e-06,
+      "loss": 0.4949,
+      "step": 2840
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 3.6720913288431813e-06,
+      "loss": 0.6128,
+      "step": 2841
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 3.6656783992993885e-06,
+      "loss": 0.6156,
+      "step": 2842
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 3.6592698174916687e-06,
+      "loss": 0.6084,
+      "step": 2843
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 3.6528655878187247e-06,
+      "loss": 0.5798,
+      "step": 2844
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 3.6464657146762805e-06,
+      "loss": 0.5418,
+      "step": 2845
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 3.64007020245706e-06,
+      "loss": 0.5508,
+      "step": 2846
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 3.6336790555508018e-06,
+      "loss": 0.6018,
+      "step": 2847
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 3.62729227834425e-06,
+      "loss": 0.6472,
+      "step": 2848
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 3.6209098752211356e-06,
+      "loss": 0.5674,
+      "step": 2849
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 3.614531850562203e-06,
+      "loss": 0.5667,
+      "step": 2850
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 3.608158208745187e-06,
+      "loss": 0.5208,
+      "step": 2851
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 3.601788954144807e-06,
+      "loss": 0.6138,
+      "step": 2852
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 3.5954240911327777e-06,
+      "loss": 0.5271,
+      "step": 2853
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 3.589063624077802e-06,
+      "loss": 0.5585,
+      "step": 2854
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 3.582707557345555e-06,
+      "loss": 0.605,
+      "step": 2855
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 3.576355895298702e-06,
+      "loss": 0.561,
+      "step": 2856
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 3.570008642296885e-06,
+      "loss": 0.6023,
+      "step": 2857
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 3.5636658026967076e-06,
+      "loss": 0.5857,
+      "step": 2858
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 3.557327380851756e-06,
+      "loss": 0.5759,
+      "step": 2859
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 3.550993381112585e-06,
+      "loss": 0.6689,
+      "step": 2860
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 3.5446638078267027e-06,
+      "loss": 0.5942,
+      "step": 2861
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 3.5383386653385888e-06,
+      "loss": 0.7017,
+      "step": 2862
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 3.532017957989684e-06,
+      "loss": 0.5518,
+      "step": 2863
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 3.5257016901183707e-06,
+      "loss": 0.655,
+      "step": 2864
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 3.519389866059998e-06,
+      "loss": 0.5519,
+      "step": 2865
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 3.513082490146864e-06,
+      "loss": 0.5334,
+      "step": 2866
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 3.506779566708203e-06,
+      "loss": 0.5636,
+      "step": 2867
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 3.500481100070201e-06,
+      "loss": 0.5972,
+      "step": 2868
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 3.4941870945559906e-06,
+      "loss": 0.6443,
+      "step": 2869
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 3.4878975544856285e-06,
+      "loss": 0.5674,
+      "step": 2870
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 3.4816124841761157e-06,
+      "loss": 0.553,
+      "step": 2871
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 3.475331887941388e-06,
+      "loss": 0.6582,
+      "step": 2872
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 3.4690557700922976e-06,
+      "loss": 0.5453,
+      "step": 2873
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 3.4627841349366363e-06,
+      "loss": 0.5417,
+      "step": 2874
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 3.4565169867791147e-06,
+      "loss": 0.5721,
+      "step": 2875
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 3.4502543299213563e-06,
+      "loss": 0.559,
+      "step": 2876
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 3.443996168661912e-06,
+      "loss": 0.6162,
+      "step": 2877
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 3.4377425072962467e-06,
+      "loss": 0.496,
+      "step": 2878
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 3.4314933501167257e-06,
+      "loss": 0.5793,
+      "step": 2879
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 3.4252487014126336e-06,
+      "loss": 0.4272,
+      "step": 2880
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 3.419008565470161e-06,
+      "loss": 0.6577,
+      "step": 2881
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 3.412772946572389e-06,
+      "loss": 0.616,
+      "step": 2882
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 3.4065418489993118e-06,
+      "loss": 0.484,
+      "step": 2883
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 3.4003152770278124e-06,
+      "loss": 0.5491,
+      "step": 2884
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 3.3940932349316757e-06,
+      "loss": 0.5351,
+      "step": 2885
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 3.387875726981563e-06,
+      "loss": 0.6782,
+      "step": 2886
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 3.3816627574450367e-06,
+      "loss": 0.603,
+      "step": 2887
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 3.3754543305865418e-06,
+      "loss": 0.5776,
+      "step": 2888
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 3.3692504506673973e-06,
+      "loss": 0.5703,
+      "step": 2889
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 3.3630511219458093e-06,
+      "loss": 0.5381,
+      "step": 2890
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 3.3568563486768636e-06,
+      "loss": 0.5623,
+      "step": 2891
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 3.350666135112504e-06,
+      "loss": 0.4935,
+      "step": 2892
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 3.3444804855015576e-06,
+      "loss": 0.6155,
+      "step": 2893
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 3.3382994040897197e-06,
+      "loss": 0.5964,
+      "step": 2894
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 3.3321228951195396e-06,
+      "loss": 0.6177,
+      "step": 2895
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 3.3259509628304364e-06,
+      "loss": 0.5984,
+      "step": 2896
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 3.3197836114586903e-06,
+      "loss": 0.4606,
+      "step": 2897
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 3.313620845237425e-06,
+      "loss": 0.5874,
+      "step": 2898
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 3.307462668396628e-06,
+      "loss": 0.5366,
+      "step": 2899
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 3.3013090851631393e-06,
+      "loss": 0.5647,
+      "step": 2900
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 3.2951600997606305e-06,
+      "loss": 0.6052,
+      "step": 2901
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 3.2890157164096315e-06,
+      "loss": 0.4935,
+      "step": 2902
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 3.282875939327512e-06,
+      "loss": 0.5913,
+      "step": 2903
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 3.2767407727284716e-06,
+      "loss": 0.6052,
+      "step": 2904
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 3.2706102208235535e-06,
+      "loss": 0.564,
+      "step": 2905
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 3.264484287820634e-06,
+      "loss": 0.6724,
+      "step": 2906
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 3.2583629779244097e-06,
+      "loss": 0.5508,
+      "step": 2907
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 3.252246295336413e-06,
+      "loss": 0.5964,
+      "step": 2908
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 3.2461342442550017e-06,
+      "loss": 0.4131,
+      "step": 2909
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 3.2400268288753423e-06,
+      "loss": 0.6106,
+      "step": 2910
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 3.2339240533894323e-06,
+      "loss": 0.5157,
+      "step": 2911
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 3.227825921986082e-06,
+      "loss": 0.6084,
+      "step": 2912
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 3.2217324388509053e-06,
+      "loss": 0.6064,
+      "step": 2913
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 3.2156436081663356e-06,
+      "loss": 0.5914,
+      "step": 2914
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 3.2095594341116133e-06,
+      "loss": 0.5515,
+      "step": 2915
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 3.203479920862772e-06,
+      "loss": 0.6172,
+      "step": 2916
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 3.1974050725926552e-06,
+      "loss": 0.5303,
+      "step": 2917
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 3.1913348934709076e-06,
+      "loss": 0.5835,
+      "step": 2918
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 3.1852693876639563e-06,
+      "loss": 0.5747,
+      "step": 2919
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 3.179208559335031e-06,
+      "loss": 0.6292,
+      "step": 2920
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 3.1731524126441513e-06,
+      "loss": 0.5057,
+      "step": 2921
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 3.1671009517481155e-06,
+      "loss": 0.5028,
+      "step": 2922
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 3.161054180800516e-06,
+      "loss": 0.578,
+      "step": 2923
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 3.1550121039517125e-06,
+      "loss": 0.5731,
+      "step": 2924
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 3.1489747253488555e-06,
+      "loss": 0.5596,
+      "step": 2925
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 3.1429420491358696e-06,
+      "loss": 0.665,
+      "step": 2926
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 3.1369140794534412e-06,
+      "loss": 0.5049,
+      "step": 2927
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 3.1308908204390376e-06,
+      "loss": 0.6147,
+      "step": 2928
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 3.1248722762268902e-06,
+      "loss": 0.5853,
+      "step": 2929
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 3.1188584509479867e-06,
+      "loss": 0.6018,
+      "step": 2930
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 3.1128493487300847e-06,
+      "loss": 0.5134,
+      "step": 2931
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 3.1068449736977015e-06,
+      "loss": 0.647,
+      "step": 2932
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 3.1008453299720965e-06,
+      "loss": 0.6077,
+      "step": 2933
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 3.094850421671295e-06,
+      "loss": 0.5848,
+      "step": 2934
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 3.088860252910071e-06,
+      "loss": 0.4753,
+      "step": 2935
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 3.0828748277999354e-06,
+      "loss": 0.5005,
+      "step": 2936
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 3.0768941504491514e-06,
+      "loss": 0.5831,
+      "step": 2937
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 3.0709182249627255e-06,
+      "loss": 0.6489,
+      "step": 2938
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 3.0649470554423932e-06,
+      "loss": 0.5867,
+      "step": 2939
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 3.0589806459866336e-06,
+      "loss": 0.4681,
+      "step": 2940
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 3.053019000690659e-06,
+      "loss": 0.5034,
+      "step": 2941
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 3.047062123646404e-06,
+      "loss": 0.5447,
+      "step": 2942
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 3.0411100189425356e-06,
+      "loss": 0.562,
+      "step": 2943
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 3.0351626906644505e-06,
+      "loss": 0.5468,
+      "step": 2944
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 3.0292201428942535e-06,
+      "loss": 0.5088,
+      "step": 2945
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 3.0232823797107793e-06,
+      "loss": 0.5627,
+      "step": 2946
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 3.017349405189579e-06,
+      "loss": 0.4589,
+      "step": 2947
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 3.011421223402906e-06,
+      "loss": 0.6895,
+      "step": 2948
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 3.0054978384197386e-06,
+      "loss": 0.4534,
+      "step": 2949
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 2.999579254305748e-06,
+      "loss": 0.5664,
+      "step": 2950
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 2.9936654751233196e-06,
+      "loss": 0.4047,
+      "step": 2951
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 2.9877565049315425e-06,
+      "loss": 0.5854,
+      "step": 2952
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 2.981852347786196e-06,
+      "loss": 0.5065,
+      "step": 2953
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 2.9759530077397637e-06,
+      "loss": 0.5312,
+      "step": 2954
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 2.9700584888414217e-06,
+      "loss": 0.553,
+      "step": 2955
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 2.96416879513703e-06,
+      "loss": 0.606,
+      "step": 2956
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 2.9582839306691468e-06,
+      "loss": 0.5249,
+      "step": 2957
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 2.952403899477011e-06,
+      "loss": 0.5043,
+      "step": 2958
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 2.94652870559654e-06,
+      "loss": 0.5365,
+      "step": 2959
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 2.940658353060337e-06,
+      "loss": 0.6555,
+      "step": 2960
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 2.934792845897684e-06,
+      "loss": 0.5857,
+      "step": 2961
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 2.9289321881345257e-06,
+      "loss": 0.6274,
+      "step": 2962
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 2.9230763837934907e-06,
+      "loss": 0.5032,
+      "step": 2963
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 2.9172254368938736e-06,
+      "loss": 0.4509,
+      "step": 2964
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 2.9113793514516277e-06,
+      "loss": 0.4481,
+      "step": 2965
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 2.905538131479376e-06,
+      "loss": 0.6082,
+      "step": 2966
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 2.8997017809864068e-06,
+      "loss": 0.4905,
+      "step": 2967
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 2.8938703039786507e-06,
+      "loss": 0.5537,
+      "step": 2968
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 2.888043704458707e-06,
+      "loss": 0.5588,
+      "step": 2969
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 2.8822219864258272e-06,
+      "loss": 0.5264,
+      "step": 2970
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 2.8764051538758986e-06,
+      "loss": 0.5507,
+      "step": 2971
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 2.87059321080147e-06,
+      "loss": 0.5399,
+      "step": 2972
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 2.8647861611917303e-06,
+      "loss": 0.599,
+      "step": 2973
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 2.8589840090325028e-06,
+      "loss": 0.6045,
+      "step": 2974
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 2.8531867583062567e-06,
+      "loss": 0.5478,
+      "step": 2975
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 2.8473944129920985e-06,
+      "loss": 0.5974,
+      "step": 2976
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 2.841606977065757e-06,
+      "loss": 0.5291,
+      "step": 2977
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 2.8358244544996038e-06,
+      "loss": 0.5723,
+      "step": 2978
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 2.830046849262629e-06,
+      "loss": 0.6045,
+      "step": 2979
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 2.8242741653204508e-06,
+      "loss": 0.49,
+      "step": 2980
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 2.8185064066353153e-06,
+      "loss": 0.4649,
+      "step": 2981
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 2.812743577166075e-06,
+      "loss": 0.4911,
+      "step": 2982
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 2.8069856808682093e-06,
+      "loss": 0.5033,
+      "step": 2983
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 2.801232721693814e-06,
+      "loss": 0.6067,
+      "step": 2984
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 2.7954847035915846e-06,
+      "loss": 0.4817,
+      "step": 2985
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 2.7897416305068325e-06,
+      "loss": 0.5393,
+      "step": 2986
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 2.7840035063814784e-06,
+      "loss": 0.4937,
+      "step": 2987
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 2.778270335154036e-06,
+      "loss": 0.5009,
+      "step": 2988
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 2.772542120759628e-06,
+      "loss": 0.4285,
+      "step": 2989
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 2.766818867129976e-06,
+      "loss": 0.5167,
+      "step": 2990
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 2.761100578193385e-06,
+      "loss": 0.5654,
+      "step": 2991
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 2.755387257874764e-06,
+      "loss": 0.5864,
+      "step": 2992
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 2.749678910095611e-06,
+      "loss": 0.6169,
+      "step": 2993
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 2.743975538774002e-06,
+      "loss": 0.5798,
+      "step": 2994
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 2.738277147824605e-06,
+      "loss": 0.5903,
+      "step": 2995
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 2.732583741158671e-06,
+      "loss": 0.6216,
+      "step": 2996
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 2.7268953226840213e-06,
+      "loss": 0.6094,
+      "step": 2997
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 2.721211896305059e-06,
+      "loss": 0.4773,
+      "step": 2998
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 2.715533465922766e-06,
+      "loss": 0.5315,
+      "step": 2999
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 2.7098600354346826e-06,
+      "loss": 0.6372,
+      "step": 3000
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 2.7041916087349264e-06,
+      "loss": 0.6013,
+      "step": 3001
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 2.6985281897141815e-06,
+      "loss": 0.614,
+      "step": 3002
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 2.6928697822596863e-06,
+      "loss": 0.5073,
+      "step": 3003
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 2.6872163902552494e-06,
+      "loss": 0.5649,
+      "step": 3004
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 2.6815680175812266e-06,
+      "loss": 0.486,
+      "step": 3005
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 2.675924668114537e-06,
+      "loss": 0.5349,
+      "step": 3006
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 2.670286345728652e-06,
+      "loss": 0.5155,
+      "step": 3007
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 2.664653054293583e-06,
+      "loss": 0.4755,
+      "step": 3008
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 2.6590247976758988e-06,
+      "loss": 0.5852,
+      "step": 3009
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 2.6534015797387103e-06,
+      "loss": 0.5427,
+      "step": 3010
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 2.647783404341663e-06,
+      "loss": 0.6265,
+      "step": 3011
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 2.6421702753409474e-06,
+      "loss": 0.5625,
+      "step": 3012
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 2.636562196589294e-06,
+      "loss": 0.5374,
+      "step": 3013
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 2.6309591719359563e-06,
+      "loss": 0.598,
+      "step": 3014
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 2.6253612052267254e-06,
+      "loss": 0.5302,
+      "step": 3015
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 2.6197683003039253e-06,
+      "loss": 0.53,
+      "step": 3016
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 2.6141804610063926e-06,
+      "loss": 0.6411,
+      "step": 3017
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 2.6085976911694986e-06,
+      "loss": 0.4822,
+      "step": 3018
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 2.6030199946251334e-06,
+      "loss": 0.5972,
+      "step": 3019
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 2.5974473752016984e-06,
+      "loss": 0.5461,
+      "step": 3020
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 2.5918798367241147e-06,
+      "loss": 0.5522,
+      "step": 3021
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 2.5863173830138212e-06,
+      "loss": 0.5359,
+      "step": 3022
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 2.5807600178887536e-06,
+      "loss": 0.5262,
+      "step": 3023
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 2.575207745163366e-06,
+      "loss": 0.4326,
+      "step": 3024
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 2.5696605686486166e-06,
+      "loss": 0.5303,
+      "step": 3025
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 2.5641184921519568e-06,
+      "loss": 0.4374,
+      "step": 3026
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 2.5585815194773466e-06,
+      "loss": 0.5825,
+      "step": 3027
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 2.5530496544252426e-06,
+      "loss": 0.6484,
+      "step": 3028
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 2.5475229007925863e-06,
+      "loss": 0.5303,
+      "step": 3029
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 2.542001262372821e-06,
+      "loss": 0.5471,
+      "step": 3030
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 2.536484742955878e-06,
+      "loss": 0.5676,
+      "step": 3031
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 2.530973346328166e-06,
+      "loss": 0.5325,
+      "step": 3032
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 2.5254670762725864e-06,
+      "loss": 0.5869,
+      "step": 3033
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 2.5199659365685235e-06,
+      "loss": 0.4054,
+      "step": 3034
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 2.5144699309918296e-06,
+      "loss": 0.5417,
+      "step": 3035
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 2.5089790633148426e-06,
+      "loss": 0.5337,
+      "step": 3036
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 2.503493337306373e-06,
+      "loss": 0.6729,
+      "step": 3037
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 2.4980127567316948e-06,
+      "loss": 0.5135,
+      "step": 3038
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 2.4925373253525584e-06,
+      "loss": 0.484,
+      "step": 3039
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 2.4870670469271783e-06,
+      "loss": 0.6016,
+      "step": 3040
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 2.4816019252102274e-06,
+      "loss": 0.4188,
+      "step": 3041
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 2.476141963952844e-06,
+      "loss": 0.5557,
+      "step": 3042
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 2.470687166902622e-06,
+      "loss": 0.5908,
+      "step": 3043
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 2.4652375378036164e-06,
+      "loss": 0.6636,
+      "step": 3044
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 2.459793080396322e-06,
+      "loss": 0.5013,
+      "step": 3045
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 2.454353798417698e-06,
+      "loss": 0.6042,
+      "step": 3046
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 2.4489196956011464e-06,
+      "loss": 0.4836,
+      "step": 3047
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 2.443490775676507e-06,
+      "loss": 0.5189,
+      "step": 3048
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 2.4380670423700726e-06,
+      "loss": 0.6321,
+      "step": 3049
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 2.432648499404575e-06,
+      "loss": 0.4728,
+      "step": 3050
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 2.4272351504991744e-06,
+      "loss": 0.6914,
+      "step": 3051
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 2.4218269993694733e-06,
+      "loss": 0.5854,
+      "step": 3052
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 2.416424049727509e-06,
+      "loss": 0.5996,
+      "step": 3053
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 2.4110263052817396e-06,
+      "loss": 0.4982,
+      "step": 3054
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 2.405633769737059e-06,
+      "loss": 0.5974,
+      "step": 3055
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 2.400246446794783e-06,
+      "loss": 0.5806,
+      "step": 3056
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 2.3948643401526463e-06,
+      "loss": 0.4695,
+      "step": 3057
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 2.389487453504806e-06,
+      "loss": 0.6133,
+      "step": 3058
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 2.384115790541841e-06,
+      "loss": 0.5596,
+      "step": 3059
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 2.3787493549507346e-06,
+      "loss": 0.54,
+      "step": 3060
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 2.373388150414889e-06,
+      "loss": 0.5559,
+      "step": 3061
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 2.3680321806141182e-06,
+      "loss": 0.4848,
+      "step": 3062
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 2.362681449224634e-06,
+      "loss": 0.6313,
+      "step": 3063
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 2.3573359599190616e-06,
+      "loss": 0.6045,
+      "step": 3064
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 2.3519957163664263e-06,
+      "loss": 0.5933,
+      "step": 3065
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 2.346660722232148e-06,
+      "loss": 0.5791,
+      "step": 3066
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 2.341330981178046e-06,
+      "loss": 0.6064,
+      "step": 3067
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 2.336006496862342e-06,
+      "loss": 0.5652,
+      "step": 3068
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 2.3306872729396323e-06,
+      "loss": 0.5602,
+      "step": 3069
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 2.325373313060919e-06,
+      "loss": 0.459,
+      "step": 3070
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 2.3200646208735866e-06,
+      "loss": 0.6353,
+      "step": 3071
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 2.3147612000213958e-06,
+      "loss": 0.6011,
+      "step": 3072
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 2.3094630541444997e-06,
+      "loss": 0.5593,
+      "step": 3073
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 2.3041701868794285e-06,
+      "loss": 0.432,
+      "step": 3074
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 2.2988826018590816e-06,
+      "loss": 0.6279,
+      "step": 3075
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 2.2936003027127418e-06,
+      "loss": 0.5366,
+      "step": 3076
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 2.2883232930660638e-06,
+      "loss": 0.6121,
+      "step": 3077
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 2.283051576541062e-06,
+      "loss": 0.5471,
+      "step": 3078
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 2.2777851567561272e-06,
+      "loss": 0.4512,
+      "step": 3079
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 2.272524037326015e-06,
+      "loss": 0.5635,
+      "step": 3080
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 2.2672682218618337e-06,
+      "loss": 0.5934,
+      "step": 3081
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 2.262017713971063e-06,
+      "loss": 0.4488,
+      "step": 3082
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 2.256772517257527e-06,
+      "loss": 0.5454,
+      "step": 3083
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 2.251532635321415e-06,
+      "loss": 0.6455,
+      "step": 3084
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 2.2462980717592662e-06,
+      "loss": 0.6665,
+      "step": 3085
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 2.2410688301639617e-06,
+      "loss": 0.6233,
+      "step": 3086
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 2.2358449141247395e-06,
+      "loss": 0.5812,
+      "step": 3087
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 2.2306263272271787e-06,
+      "loss": 0.4873,
+      "step": 3088
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 2.2254130730531977e-06,
+      "loss": 0.6433,
+      "step": 3089
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 2.2202051551810568e-06,
+      "loss": 0.6436,
+      "step": 3090
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 2.215002577185359e-06,
+      "loss": 0.5586,
+      "step": 3091
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 2.209805342637029e-06,
+      "loss": 0.5197,
+      "step": 3092
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 2.204613455103337e-06,
+      "loss": 0.5637,
+      "step": 3093
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 2.19942691814788e-06,
+      "loss": 0.54,
+      "step": 3094
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 2.1942457353305746e-06,
+      "loss": 0.5884,
+      "step": 3095
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 2.1890699102076707e-06,
+      "loss": 0.5591,
+      "step": 3096
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 2.183899446331742e-06,
+      "loss": 0.5808,
+      "step": 3097
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 2.178734347251673e-06,
+      "loss": 0.5312,
+      "step": 3098
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 2.173574616512675e-06,
+      "loss": 0.4269,
+      "step": 3099
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 2.1684202576562718e-06,
+      "loss": 0.5417,
+      "step": 3100
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 2.1632712742202966e-06,
+      "loss": 0.386,
+      "step": 3101
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 2.1581276697388974e-06,
+      "loss": 0.5358,
+      "step": 3102
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 2.1529894477425327e-06,
+      "loss": 0.5955,
+      "step": 3103
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 2.147856611757957e-06,
+      "loss": 0.521,
+      "step": 3104
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 2.1427291653082337e-06,
+      "loss": 0.5593,
+      "step": 3105
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 2.137607111912734e-06,
+      "loss": 0.5837,
+      "step": 3106
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 2.1324904550871107e-06,
+      "loss": 0.5396,
+      "step": 3107
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 2.12737919834333e-06,
+      "loss": 0.5139,
+      "step": 3108
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 2.122273345189638e-06,
+      "loss": 0.5078,
+      "step": 3109
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 2.1171728991305797e-06,
+      "loss": 0.5085,
+      "step": 3110
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 2.1120778636669905e-06,
+      "loss": 0.5784,
+      "step": 3111
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 2.106988242295981e-06,
+      "loss": 0.5703,
+      "step": 3112
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 2.101904038510958e-06,
+      "loss": 0.5408,
+      "step": 3113
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 2.0968252558016055e-06,
+      "loss": 0.5662,
+      "step": 3114
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 2.091751897653881e-06,
+      "loss": 0.6414,
+      "step": 3115
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 2.086683967550027e-06,
+      "loss": 0.5385,
+      "step": 3116
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 2.0816214689685598e-06,
+      "loss": 0.5205,
+      "step": 3117
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 2.0765644053842583e-06,
+      "loss": 0.5623,
+      "step": 3118
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 2.07151278026818e-06,
+      "loss": 0.4583,
+      "step": 3119
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 2.0664665970876496e-06,
+      "loss": 0.7588,
+      "step": 3120
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 2.0614258593062496e-06,
+      "loss": 0.6438,
+      "step": 3121
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 2.0563905703838315e-06,
+      "loss": 0.4549,
+      "step": 3122
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 2.0513607337765074e-06,
+      "loss": 0.5195,
+      "step": 3123
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 2.0463363529366375e-06,
+      "loss": 0.4646,
+      "step": 3124
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 2.0413174313128483e-06,
+      "loss": 0.5601,
+      "step": 3125
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 2.0363039723500155e-06,
+      "loss": 0.5088,
+      "step": 3126
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 2.0312959794892618e-06,
+      "loss": 0.6831,
+      "step": 3127
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 2.0262934561679615e-06,
+      "loss": 0.5554,
+      "step": 3128
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 2.021296405819738e-06,
+      "loss": 0.6226,
+      "step": 3129
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 2.016304831874449e-06,
+      "loss": 0.5156,
+      "step": 3130
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 2.0113187377582023e-06,
+      "loss": 0.5049,
+      "step": 3131
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 2.0063381268933426e-06,
+      "loss": 0.5446,
+      "step": 3132
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 2.0013630026984433e-06,
+      "loss": 0.4443,
+      "step": 3133
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.996393368588325e-06,
+      "loss": 0.5491,
+      "step": 3134
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.9914292279740256e-06,
+      "loss": 0.5737,
+      "step": 3135
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.986470584262824e-06,
+      "loss": 0.6592,
+      "step": 3136
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 1.9815174408582237e-06,
+      "loss": 0.5017,
+      "step": 3137
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.9765698011599464e-06,
+      "loss": 0.6018,
+      "step": 3138
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.9716276685639426e-06,
+      "loss": 0.5503,
+      "step": 3139
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.9666910464623844e-06,
+      "loss": 0.5721,
+      "step": 3140
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 1.9617599382436526e-06,
+      "loss": 0.5564,
+      "step": 3141
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.9568343472923524e-06,
+      "loss": 0.6082,
+      "step": 3142
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.9519142769893006e-06,
+      "loss": 0.626,
+      "step": 3143
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.9469997307115185e-06,
+      "loss": 0.5598,
+      "step": 3144
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 1.9420907118322428e-06,
+      "loss": 0.5227,
+      "step": 3145
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.9371872237209167e-06,
+      "loss": 0.4705,
+      "step": 3146
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.9322892697431783e-06,
+      "loss": 0.5044,
+      "step": 3147
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.9273968532608754e-06,
+      "loss": 0.517,
+      "step": 3148
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 1.9225099776320567e-06,
+      "loss": 0.5383,
+      "step": 3149
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.9176286462109574e-06,
+      "loss": 0.4364,
+      "step": 3150
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.9127528623480175e-06,
+      "loss": 0.6741,
+      "step": 3151
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.9078826293898676e-06,
+      "loss": 0.5698,
+      "step": 3152
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 1.9030179506793201e-06,
+      "loss": 0.4875,
+      "step": 3153
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.8981588295553855e-06,
+      "loss": 0.4993,
+      "step": 3154
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.893305269353257e-06,
+      "loss": 0.4958,
+      "step": 3155
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.888457273404305e-06,
+      "loss": 0.5762,
+      "step": 3156
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 1.883614845036087e-06,
+      "loss": 0.5771,
+      "step": 3157
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.8787779875723389e-06,
+      "loss": 0.5974,
+      "step": 3158
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.8739467043329663e-06,
+      "loss": 0.3894,
+      "step": 3159
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.8691209986340598e-06,
+      "loss": 0.5309,
+      "step": 3160
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 1.8643008737878687e-06,
+      "loss": 0.4956,
+      "step": 3161
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.8594863331028224e-06,
+      "loss": 0.5713,
+      "step": 3162
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.8546773798835149e-06,
+      "loss": 0.5911,
+      "step": 3163
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.8498740174306984e-06,
+      "loss": 0.4919,
+      "step": 3164
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 1.8450762490412965e-06,
+      "loss": 0.4252,
+      "step": 3165
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.840284078008393e-06,
+      "loss": 0.5886,
+      "step": 3166
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.8354975076212188e-06,
+      "loss": 0.5764,
+      "step": 3167
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.8307165411651728e-06,
+      "loss": 0.5764,
+      "step": 3168
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 1.8259411819218054e-06,
+      "loss": 0.6265,
+      "step": 3169
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.821171433168809e-06,
+      "loss": 0.4866,
+      "step": 3170
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.8164072981800374e-06,
+      "loss": 0.4764,
+      "step": 3171
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 1.8116487802254868e-06,
+      "loss": 0.4115,
+      "step": 3172
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.8068958825712912e-06,
+      "loss": 0.5613,
+      "step": 3173
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.8021486084797368e-06,
+      "loss": 0.4578,
+      "step": 3174
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.7974069612092481e-06,
+      "loss": 0.5725,
+      "step": 3175
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 1.7926709440143797e-06,
+      "loss": 0.4621,
+      "step": 3176
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.7879405601458311e-06,
+      "loss": 0.5952,
+      "step": 3177
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.7832158128504328e-06,
+      "loss": 0.6128,
+      "step": 3178
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.77849670537114e-06,
+      "loss": 0.6326,
+      "step": 3179
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 1.7737832409470458e-06,
+      "loss": 0.5504,
+      "step": 3180
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.7690754228133688e-06,
+      "loss": 0.505,
+      "step": 3181
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.7643732542014437e-06,
+      "loss": 0.5586,
+      "step": 3182
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.7596767383387359e-06,
+      "loss": 0.5869,
+      "step": 3183
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 1.7549858784488317e-06,
+      "loss": 0.5327,
+      "step": 3184
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.7503006777514254e-06,
+      "loss": 0.5076,
+      "step": 3185
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.7456211394623379e-06,
+      "loss": 0.5527,
+      "step": 3186
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.7409472667935012e-06,
+      "loss": 0.461,
+      "step": 3187
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 1.7362790629529513e-06,
+      "loss": 0.6125,
+      "step": 3188
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.7316165311448407e-06,
+      "loss": 0.5113,
+      "step": 3189
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.7269596745694295e-06,
+      "loss": 0.5906,
+      "step": 3190
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.722308496423074e-06,
+      "loss": 0.644,
+      "step": 3191
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 1.7176629998982408e-06,
+      "loss": 0.637,
+      "step": 3192
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.713023188183498e-06,
+      "loss": 0.6187,
+      "step": 3193
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.7083890644635015e-06,
+      "loss": 0.5994,
+      "step": 3194
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.7037606319190137e-06,
+      "loss": 0.5664,
+      "step": 3195
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 1.6991378937268888e-06,
+      "loss": 0.4792,
+      "step": 3196
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.6945208530600644e-06,
+      "loss": 0.5554,
+      "step": 3197
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.6899095130875775e-06,
+      "loss": 0.4822,
+      "step": 3198
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.6853038769745466e-06,
+      "loss": 0.593,
+      "step": 3199
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 1.680703947882182e-06,
+      "loss": 0.5383,
+      "step": 3200
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.6761097289677642e-06,
+      "loss": 0.4519,
+      "step": 3201
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.6715212233846656e-06,
+      "loss": 0.5693,
+      "step": 3202
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.6669384342823347e-06,
+      "loss": 0.5496,
+      "step": 3203
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 1.662361364806292e-06,
+      "loss": 0.609,
+      "step": 3204
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.6577900180981366e-06,
+      "loss": 0.4481,
+      "step": 3205
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.6532243972955397e-06,
+      "loss": 0.6025,
+      "step": 3206
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.6486645055322381e-06,
+      "loss": 0.5496,
+      "step": 3207
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 1.6441103459380392e-06,
+      "loss": 0.6282,
+      "step": 3208
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.6395619216388204e-06,
+      "loss": 0.4741,
+      "step": 3209
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.635019235756511e-06,
+      "loss": 0.5253,
+      "step": 3210
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.6304822914091134e-06,
+      "loss": 0.4591,
+      "step": 3211
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 1.6259510917106847e-06,
+      "loss": 0.4802,
+      "step": 3212
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.6214256397713347e-06,
+      "loss": 0.5129,
+      "step": 3213
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.6169059386972342e-06,
+      "loss": 0.464,
+      "step": 3214
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 1.612391991590606e-06,
+      "loss": 0.578,
+      "step": 3215
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.607883801549719e-06,
+      "loss": 0.5918,
+      "step": 3216
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.603381371668895e-06,
+      "loss": 0.5879,
+      "step": 3217
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.5988847050385037e-06,
+      "loss": 0.4954,
+      "step": 3218
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 1.5943938047449493e-06,
+      "loss": 0.5554,
+      "step": 3219
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.5899086738706904e-06,
+      "loss": 0.6843,
+      "step": 3220
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.5854293154942203e-06,
+      "loss": 0.6274,
+      "step": 3221
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.580955732690065e-06,
+      "loss": 0.5433,
+      "step": 3222
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 1.576487928528795e-06,
+      "loss": 0.5305,
+      "step": 3223
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.5720259060770128e-06,
+      "loss": 0.4854,
+      "step": 3224
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.5675696683973452e-06,
+      "loss": 0.563,
+      "step": 3225
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.5631192185484557e-06,
+      "loss": 0.5139,
+      "step": 3226
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 1.558674559585036e-06,
+      "loss": 0.6265,
+      "step": 3227
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.5542356945577964e-06,
+      "loss": 0.5757,
+      "step": 3228
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.5498026265134747e-06,
+      "loss": 0.6074,
+      "step": 3229
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.5453753584948329e-06,
+      "loss": 0.4529,
+      "step": 3230
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 1.5409538935406422e-06,
+      "loss": 0.5763,
+      "step": 3231
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.5365382346857005e-06,
+      "loss": 0.514,
+      "step": 3232
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.5321283849608193e-06,
+      "loss": 0.4784,
+      "step": 3233
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.527724347392815e-06,
+      "loss": 0.6182,
+      "step": 3234
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 1.5233261250045218e-06,
+      "loss": 0.6204,
+      "step": 3235
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5189337208147848e-06,
+      "loss": 0.624,
+      "step": 3236
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5145471378384458e-06,
+      "loss": 0.6064,
+      "step": 3237
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5101663790863597e-06,
+      "loss": 0.6245,
+      "step": 3238
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 1.5057914475653835e-06,
+      "loss": 0.541,
+      "step": 3239
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.501422346278365e-06,
+      "loss": 0.4565,
+      "step": 3240
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.4970590782241645e-06,
+      "loss": 0.5288,
+      "step": 3241
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.4927016463976262e-06,
+      "loss": 0.583,
+      "step": 3242
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 1.4883500537895946e-06,
+      "loss": 0.5356,
+      "step": 3243
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.4840043033869079e-06,
+      "loss": 0.4797,
+      "step": 3244
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.479664398172387e-06,
+      "loss": 0.5121,
+      "step": 3245
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.4753303411248477e-06,
+      "loss": 0.6145,
+      "step": 3246
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 1.4710021352190918e-06,
+      "loss": 0.6042,
+      "step": 3247
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.4666797834258984e-06,
+      "loss": 0.5172,
+      "step": 3248
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.4623632887120343e-06,
+      "loss": 0.6709,
+      "step": 3249
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.4580526540402462e-06,
+      "loss": 0.636,
+      "step": 3250
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 1.453747882369254e-06,
+      "loss": 0.5038,
+      "step": 3251
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.4494489766537579e-06,
+      "loss": 0.614,
+      "step": 3252
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.4451559398444314e-06,
+      "loss": 0.522,
+      "step": 3253
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.4408687748879157e-06,
+      "loss": 0.5566,
+      "step": 3254
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 1.4365874847268257e-06,
+      "loss": 0.6469,
+      "step": 3255
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.4323120722997463e-06,
+      "loss": 0.4883,
+      "step": 3256
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.428042540541218e-06,
+      "loss": 0.4314,
+      "step": 3257
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 1.4237788923817553e-06,
+      "loss": 0.478,
+      "step": 3258
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.419521130747833e-06,
+      "loss": 0.6462,
+      "step": 3259
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.4152692585618767e-06,
+      "loss": 0.6343,
+      "step": 3260
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.4110232787422807e-06,
+      "loss": 0.5358,
+      "step": 3261
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 1.4067831942033904e-06,
+      "loss": 0.7314,
+      "step": 3262
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.4025490078555015e-06,
+      "loss": 0.5308,
+      "step": 3263
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.3983207226048656e-06,
+      "loss": 0.5302,
+      "step": 3264
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.3940983413536847e-06,
+      "loss": 0.5962,
+      "step": 3265
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 1.3898818670001035e-06,
+      "loss": 0.4824,
+      "step": 3266
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.3856713024382195e-06,
+      "loss": 0.4952,
+      "step": 3267
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.3814666505580632e-06,
+      "loss": 0.4783,
+      "step": 3268
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.377267914245618e-06,
+      "loss": 0.4937,
+      "step": 3269
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 1.3730750963828033e-06,
+      "loss": 0.5454,
+      "step": 3270
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.36888819984747e-06,
+      "loss": 0.4902,
+      "step": 3271
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.3647072275134122e-06,
+      "loss": 0.54,
+      "step": 3272
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.3605321822503592e-06,
+      "loss": 0.4814,
+      "step": 3273
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 1.3563630669239625e-06,
+      "loss": 0.6167,
+      "step": 3274
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.3521998843958117e-06,
+      "loss": 0.6022,
+      "step": 3275
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.3480426375234234e-06,
+      "loss": 0.5723,
+      "step": 3276
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.343891329160235e-06,
+      "loss": 0.468,
+      "step": 3277
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 1.339745962155613e-06,
+      "loss": 0.442,
+      "step": 3278
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.3356065393548467e-06,
+      "loss": 0.4227,
+      "step": 3279
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.3314730635991391e-06,
+      "loss": 0.5359,
+      "step": 3280
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.3273455377256162e-06,
+      "loss": 0.5413,
+      "step": 3281
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 1.3232239645673217e-06,
+      "loss": 0.4606,
+      "step": 3282
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.3191083469532062e-06,
+      "loss": 0.5746,
+      "step": 3283
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.31499868770814e-06,
+      "loss": 0.5156,
+      "step": 3284
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.3108949896529023e-06,
+      "loss": 0.4377,
+      "step": 3285
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 1.3067972556041753e-06,
+      "loss": 0.4318,
+      "step": 3286
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.3027054883745538e-06,
+      "loss": 0.5513,
+      "step": 3287
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.2986196907725368e-06,
+      "loss": 0.5891,
+      "step": 3288
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.294539865602521e-06,
+      "loss": 0.6038,
+      "step": 3289
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 1.2904660156648074e-06,
+      "loss": 0.5513,
+      "step": 3290
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.2863981437555972e-06,
+      "loss": 0.5193,
+      "step": 3291
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.2823362526669825e-06,
+      "loss": 0.4542,
+      "step": 3292
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.2782803451869597e-06,
+      "loss": 0.5229,
+      "step": 3293
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 1.2742304240994053e-06,
+      "loss": 0.4629,
+      "step": 3294
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.2701864921840989e-06,
+      "loss": 0.6121,
+      "step": 3295
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.2661485522167062e-06,
+      "loss": 0.4894,
+      "step": 3296
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.262116606968773e-06,
+      "loss": 0.5625,
+      "step": 3297
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 1.2580906592077403e-06,
+      "loss": 0.5427,
+      "step": 3298
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.2540707116969274e-06,
+      "loss": 0.6125,
+      "step": 3299
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.2500567671955345e-06,
+      "loss": 0.572,
+      "step": 3300
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 1.2460488284586437e-06,
+      "loss": 0.5977,
+      "step": 3301
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.2420468982372158e-06,
+      "loss": 0.5979,
+      "step": 3302
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.2380509792780826e-06,
+      "loss": 0.5168,
+      "step": 3303
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.2340610743239545e-06,
+      "loss": 0.5833,
+      "step": 3304
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 1.2300771861134153e-06,
+      "loss": 0.5566,
+      "step": 3305
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.2260993173809122e-06,
+      "loss": 0.36,
+      "step": 3306
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.2221274708567664e-06,
+      "loss": 0.5776,
+      "step": 3307
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.2181616492671677e-06,
+      "loss": 0.5107,
+      "step": 3308
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 1.2142018553341607e-06,
+      "loss": 0.5607,
+      "step": 3309
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.2102480917756632e-06,
+      "loss": 0.5286,
+      "step": 3310
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.2063003613054491e-06,
+      "loss": 0.6248,
+      "step": 3311
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.202358666633151e-06,
+      "loss": 0.5571,
+      "step": 3312
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 1.198423010464259e-06,
+      "loss": 0.6487,
+      "step": 3313
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.1944933955001226e-06,
+      "loss": 0.6323,
+      "step": 3314
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.190569824437937e-06,
+      "loss": 0.5347,
+      "step": 3315
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.1866522999707553e-06,
+      "loss": 0.502,
+      "step": 3316
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 1.1827408247874795e-06,
+      "loss": 0.5496,
+      "step": 3317
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.1788354015728543e-06,
+      "loss": 0.5706,
+      "step": 3318
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.1749360330074799e-06,
+      "loss": 0.5532,
+      "step": 3319
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.171042721767791e-06,
+      "loss": 0.5532,
+      "step": 3320
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 1.167155470526069e-06,
+      "loss": 0.4429,
+      "step": 3321
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.1632742819504406e-06,
+      "loss": 0.4766,
+      "step": 3322
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.1593991587048604e-06,
+      "loss": 0.4429,
+      "step": 3323
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.1555301034491284e-06,
+      "loss": 0.5652,
+      "step": 3324
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 1.1516671188388805e-06,
+      "loss": 0.5576,
+      "step": 3325
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.147810207525577e-06,
+      "loss": 0.6008,
+      "step": 3326
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.143959372156518e-06,
+      "loss": 0.4504,
+      "step": 3327
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.140114615374831e-06,
+      "loss": 0.4512,
+      "step": 3328
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 1.136275939819469e-06,
+      "loss": 0.6096,
+      "step": 3329
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.132443348125214e-06,
+      "loss": 0.4823,
+      "step": 3330
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.128616842922672e-06,
+      "loss": 0.564,
+      "step": 3331
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.1247964268382661e-06,
+      "loss": 0.5977,
+      "step": 3332
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 1.120982102494247e-06,
+      "loss": 0.6333,
+      "step": 3333
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.1171738725086833e-06,
+      "loss": 0.6466,
+      "step": 3334
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.1133717394954536e-06,
+      "loss": 0.4181,
+      "step": 3335
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.1095757060642588e-06,
+      "loss": 0.5714,
+      "step": 3336
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 1.1057857748206148e-06,
+      "loss": 0.5479,
+      "step": 3337
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.1020019483658385e-06,
+      "loss": 0.4695,
+      "step": 3338
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.0982242292970668e-06,
+      "loss": 0.5967,
+      "step": 3339
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.0944526202072424e-06,
+      "loss": 0.5283,
+      "step": 3340
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 1.0906871236851114e-06,
+      "loss": 0.5444,
+      "step": 3341
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.0869277423152248e-06,
+      "loss": 0.5823,
+      "step": 3342
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.083174478677942e-06,
+      "loss": 0.5356,
+      "step": 3343
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.0794273353494144e-06,
+      "loss": 0.6654,
+      "step": 3344
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 1.0756863149015972e-06,
+      "loss": 0.4235,
+      "step": 3345
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.0719514199022473e-06,
+      "loss": 0.6277,
+      "step": 3346
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.0682226529149076e-06,
+      "loss": 0.625,
+      "step": 3347
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 1.0645000164989217e-06,
+      "loss": 0.5189,
+      "step": 3348
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.0607835132094258e-06,
+      "loss": 0.4294,
+      "step": 3349
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.0570731455973415e-06,
+      "loss": 0.5021,
+      "step": 3350
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.053368916209382e-06,
+      "loss": 0.5828,
+      "step": 3351
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 1.0496708275880497e-06,
+      "loss": 0.5454,
+      "step": 3352
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.0459788822716255e-06,
+      "loss": 0.5999,
+      "step": 3353
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.0422930827941803e-06,
+      "loss": 0.5333,
+      "step": 3354
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.0386134316855667e-06,
+      "loss": 0.5593,
+      "step": 3355
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 1.034939931471408e-06,
+      "loss": 0.571,
+      "step": 3356
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.0312725846731174e-06,
+      "loss": 0.6003,
+      "step": 3357
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.0276113938078768e-06,
+      "loss": 0.5392,
+      "step": 3358
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.0239563613886495e-06,
+      "loss": 0.6123,
+      "step": 3359
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 1.0203074899241616e-06,
+      "loss": 0.4927,
+      "step": 3360
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.0166647819189202e-06,
+      "loss": 0.5774,
+      "step": 3361
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.0130282398731984e-06,
+      "loss": 0.4829,
+      "step": 3362
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.0093978662830351e-06,
+      "loss": 0.4934,
+      "step": 3363
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 1.0057736636402382e-06,
+      "loss": 0.5171,
+      "step": 3364
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 1.0021556344323812e-06,
+      "loss": 0.5826,
+      "step": 3365
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 9.985437811427934e-07,
+      "loss": 0.6008,
+      "step": 3366
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 9.949381062505726e-07,
+      "loss": 0.5679,
+      "step": 3367
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 9.913386122305747e-07,
+      "loss": 0.4064,
+      "step": 3368
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 9.877453015534066e-07,
+      "loss": 0.5334,
+      "step": 3369
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 9.841581766854402e-07,
+      "loss": 0.2993,
+      "step": 3370
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 9.80577240088798e-07,
+      "loss": 0.5707,
+      "step": 3371
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 9.770024942213507e-07,
+      "loss": 0.6353,
+      "step": 3372
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 9.734339415367254e-07,
+      "loss": 0.6003,
+      "step": 3373
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 9.698715844842988e-07,
+      "loss": 0.6152,
+      "step": 3374
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 9.663154255091889e-07,
+      "loss": 0.5417,
+      "step": 3375
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 9.627654670522646e-07,
+      "loss": 0.5962,
+      "step": 3376
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 9.5922171155014e-07,
+      "loss": 0.5168,
+      "step": 3377
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 9.556841614351664e-07,
+      "loss": 0.5217,
+      "step": 3378
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 9.52152819135439e-07,
+      "loss": 0.5249,
+      "step": 3379
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 9.486276870747946e-07,
+      "loss": 0.5919,
+      "step": 3380
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 9.45108767672801e-07,
+      "loss": 0.6017,
+      "step": 3381
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 9.415960633447674e-07,
+      "loss": 0.4632,
+      "step": 3382
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 9.380895765017361e-07,
+      "loss": 0.5105,
+      "step": 3383
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 9.345893095504777e-07,
+      "loss": 0.5164,
+      "step": 3384
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 9.310952648935001e-07,
+      "loss": 0.5387,
+      "step": 3385
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 9.276074449290362e-07,
+      "loss": 0.4668,
+      "step": 3386
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 9.241258520510454e-07,
+      "loss": 0.4548,
+      "step": 3387
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 9.206504886492162e-07,
+      "loss": 0.6584,
+      "step": 3388
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 9.171813571089628e-07,
+      "loss": 0.729,
+      "step": 3389
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 9.137184598114135e-07,
+      "loss": 0.478,
+      "step": 3390
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 9.102617991334273e-07,
+      "loss": 0.5809,
+      "step": 3391
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 9.068113774475784e-07,
+      "loss": 0.6008,
+      "step": 3392
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 9.033671971221558e-07,
+      "loss": 0.4246,
+      "step": 3393
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 8.999292605211695e-07,
+      "loss": 0.5908,
+      "step": 3394
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 8.964975700043433e-07,
+      "loss": 0.6755,
+      "step": 3395
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 8.930721279271093e-07,
+      "loss": 0.6125,
+      "step": 3396
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 8.896529366406181e-07,
+      "loss": 0.5776,
+      "step": 3397
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 8.862399984917214e-07,
+      "loss": 0.48,
+      "step": 3398
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 8.828333158229851e-07,
+      "loss": 0.6201,
+      "step": 3399
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 8.794328909726823e-07,
+      "loss": 0.6426,
+      "step": 3400
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 8.760387262747838e-07,
+      "loss": 0.5879,
+      "step": 3401
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 8.726508240589693e-07,
+      "loss": 0.5596,
+      "step": 3402
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 8.69269186650622e-07,
+      "loss": 0.595,
+      "step": 3403
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 8.658938163708175e-07,
+      "loss": 0.3941,
+      "step": 3404
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 8.625247155363358e-07,
+      "loss": 0.5383,
+      "step": 3405
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 8.591618864596541e-07,
+      "loss": 0.6079,
+      "step": 3406
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 8.5580533144894e-07,
+      "loss": 0.6423,
+      "step": 3407
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 8.524550528080578e-07,
+      "loss": 0.4342,
+      "step": 3408
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 8.491110528365653e-07,
+      "loss": 0.5305,
+      "step": 3409
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 8.457733338297069e-07,
+      "loss": 0.4489,
+      "step": 3410
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 8.424418980784177e-07,
+      "loss": 0.5088,
+      "step": 3411
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 8.391167478693241e-07,
+      "loss": 0.5863,
+      "step": 3412
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 8.357978854847304e-07,
+      "loss": 0.4772,
+      "step": 3413
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 8.3248531320263e-07,
+      "loss": 0.5647,
+      "step": 3414
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 8.291790332967009e-07,
+      "loss": 0.4695,
+      "step": 3415
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 8.258790480362955e-07,
+      "loss": 0.5091,
+      "step": 3416
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 8.225853596864519e-07,
+      "loss": 0.4998,
+      "step": 3417
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 8.192979705078852e-07,
+      "loss": 0.6104,
+      "step": 3418
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 8.160168827569804e-07,
+      "loss": 0.4866,
+      "step": 3419
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 8.12742098685807e-07,
+      "loss": 0.4553,
+      "step": 3420
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 8.094736205421027e-07,
+      "loss": 0.4873,
+      "step": 3421
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 8.062114505692742e-07,
+      "loss": 0.5842,
+      "step": 3422
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 8.029555910064069e-07,
+      "loss": 0.5857,
+      "step": 3423
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 7.997060440882453e-07,
+      "loss": 0.576,
+      "step": 3424
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 7.964628120452067e-07,
+      "loss": 0.5691,
+      "step": 3425
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 7.932258971033746e-07,
+      "loss": 0.5166,
+      "step": 3426
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 7.899953014844919e-07,
+      "loss": 0.5203,
+      "step": 3427
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 7.86771027405967e-07,
+      "loss": 0.5742,
+      "step": 3428
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 7.835530770808731e-07,
+      "loss": 0.5244,
+      "step": 3429
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 7.803414527179343e-07,
+      "loss": 0.58,
+      "step": 3430
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 7.771361565215385e-07,
+      "loss": 0.5071,
+      "step": 3431
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 7.739371906917326e-07,
+      "loss": 0.5071,
+      "step": 3432
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 7.7074455742421e-07,
+      "loss": 0.6213,
+      "step": 3433
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 7.675582589103248e-07,
+      "loss": 0.5283,
+      "step": 3434
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 7.643782973370817e-07,
+      "loss": 0.5825,
+      "step": 3435
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 7.612046748871327e-07,
+      "loss": 0.6101,
+      "step": 3436
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 7.580373937387819e-07,
+      "loss": 0.5691,
+      "step": 3437
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 7.548764560659815e-07,
+      "loss": 0.496,
+      "step": 3438
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 7.517218640383262e-07,
+      "loss": 0.613,
+      "step": 3439
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 7.485736198210559e-07,
+      "loss": 0.5642,
+      "step": 3440
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 7.454317255750587e-07,
+      "loss": 0.4353,
+      "step": 3441
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 7.422961834568565e-07,
+      "loss": 0.5088,
+      "step": 3442
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 7.391669956186165e-07,
+      "loss": 0.7021,
+      "step": 3443
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 7.36044164208145e-07,
+      "loss": 0.4816,
+      "step": 3444
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 7.329276913688788e-07,
+      "loss": 0.4854,
+      "step": 3445
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 7.298175792398976e-07,
+      "loss": 0.571,
+      "step": 3446
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 7.267138299559129e-07,
+      "loss": 0.522,
+      "step": 3447
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 7.236164456472672e-07,
+      "loss": 0.6392,
+      "step": 3448
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 7.205254284399366e-07,
+      "loss": 0.4292,
+      "step": 3449
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 7.174407804555261e-07,
+      "loss": 0.5819,
+      "step": 3450
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 7.143625038112667e-07,
+      "loss": 0.6189,
+      "step": 3451
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 7.112906006200215e-07,
+      "loss": 0.6519,
+      "step": 3452
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 7.082250729902718e-07,
+      "loss": 0.4625,
+      "step": 3453
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 7.051659230261299e-07,
+      "loss": 0.5507,
+      "step": 3454
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 7.021131528273284e-07,
+      "loss": 0.5161,
+      "step": 3455
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.990667644892168e-07,
+      "loss": 0.5577,
+      "step": 3456
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.960267601027693e-07,
+      "loss": 0.498,
+      "step": 3457
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 6.929931417545787e-07,
+      "loss": 0.5791,
+      "step": 3458
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.899659115268498e-07,
+      "loss": 0.6128,
+      "step": 3459
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.869450714974058e-07,
+      "loss": 0.6182,
+      "step": 3460
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.839306237396859e-07,
+      "loss": 0.6243,
+      "step": 3461
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 6.809225703227352e-07,
+      "loss": 0.5366,
+      "step": 3462
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.779209133112164e-07,
+      "loss": 0.5833,
+      "step": 3463
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.749256547654004e-07,
+      "loss": 0.5417,
+      "step": 3464
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.719367967411638e-07,
+      "loss": 0.557,
+      "step": 3465
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 6.689543412899913e-07,
+      "loss": 0.5043,
+      "step": 3466
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.659782904589752e-07,
+      "loss": 0.5964,
+      "step": 3467
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.630086462908081e-07,
+      "loss": 0.5328,
+      "step": 3468
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.600454108237875e-07,
+      "loss": 0.4108,
+      "step": 3469
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 6.57088586091813e-07,
+      "loss": 0.5017,
+      "step": 3470
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.541381741243802e-07,
+      "loss": 0.4865,
+      "step": 3471
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.51194176946588e-07,
+      "loss": 0.4702,
+      "step": 3472
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.482565965791299e-07,
+      "loss": 0.5757,
+      "step": 3473
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 6.453254350382921e-07,
+      "loss": 0.4761,
+      "step": 3474
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.424006943359606e-07,
+      "loss": 0.5376,
+      "step": 3475
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.394823764796121e-07,
+      "loss": 0.5018,
+      "step": 3476
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 6.36570483472313e-07,
+      "loss": 0.4624,
+      "step": 3477
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.336650173127224e-07,
+      "loss": 0.5248,
+      "step": 3478
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.30765979995086e-07,
+      "loss": 0.4602,
+      "step": 3479
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.278733735092379e-07,
+      "loss": 0.5674,
+      "step": 3480
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 6.249871998405998e-07,
+      "loss": 0.5952,
+      "step": 3481
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.221074609701739e-07,
+      "loss": 0.4302,
+      "step": 3482
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.192341588745498e-07,
+      "loss": 0.5068,
+      "step": 3483
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.163672955258982e-07,
+      "loss": 0.6135,
+      "step": 3484
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 6.135068728919691e-07,
+      "loss": 0.5535,
+      "step": 3485
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.106528929360911e-07,
+      "loss": 0.5811,
+      "step": 3486
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.078053576171738e-07,
+      "loss": 0.5808,
+      "step": 3487
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.049642688897007e-07,
+      "loss": 0.5835,
+      "step": 3488
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 6.021296287037304e-07,
+      "loss": 0.3251,
+      "step": 3489
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.99301439004899e-07,
+      "loss": 0.6472,
+      "step": 3490
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.964797017344081e-07,
+      "loss": 0.454,
+      "step": 3491
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.936644188290386e-07,
+      "loss": 0.4725,
+      "step": 3492
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.908555922211368e-07,
+      "loss": 0.5974,
+      "step": 3493
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.880532238386161e-07,
+      "loss": 0.5623,
+      "step": 3494
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.852573156049601e-07,
+      "loss": 0.5549,
+      "step": 3495
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.824678694392194e-07,
+      "loss": 0.5292,
+      "step": 3496
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.796848872560045e-07,
+      "loss": 0.5505,
+      "step": 3497
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.769083709654933e-07,
+      "loss": 0.5337,
+      "step": 3498
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.741383224734254e-07,
+      "loss": 0.5216,
+      "step": 3499
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.713747436810968e-07,
+      "loss": 0.4438,
+      "step": 3500
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.68617636485368e-07,
+      "loss": 0.6157,
+      "step": 3501
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.658670027786561e-07,
+      "loss": 0.5713,
+      "step": 3502
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.631228444489312e-07,
+      "loss": 0.5447,
+      "step": 3503
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.603851633797242e-07,
+      "loss": 0.5388,
+      "step": 3504
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.576539614501197e-07,
+      "loss": 0.5298,
+      "step": 3505
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.549292405347495e-07,
+      "loss": 0.552,
+      "step": 3506
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.522110025038019e-07,
+      "loss": 0.6206,
+      "step": 3507
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.494992492230167e-07,
+      "loss": 0.4111,
+      "step": 3508
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.467939825536783e-07,
+      "loss": 0.4997,
+      "step": 3509
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.440952043526215e-07,
+      "loss": 0.592,
+      "step": 3510
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.414029164722279e-07,
+      "loss": 0.5466,
+      "step": 3511
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.387171207604214e-07,
+      "loss": 0.5741,
+      "step": 3512
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.36037819060673e-07,
+      "loss": 0.5413,
+      "step": 3513
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.333650132119972e-07,
+      "loss": 0.5815,
+      "step": 3514
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.306987050489442e-07,
+      "loss": 0.533,
+      "step": 3515
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.280388964016093e-07,
+      "loss": 0.5669,
+      "step": 3516
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.253855890956272e-07,
+      "loss": 0.4937,
+      "step": 3517
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.22738784952167e-07,
+      "loss": 0.4569,
+      "step": 3518
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.200984857879355e-07,
+      "loss": 0.4291,
+      "step": 3519
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.174646934151751e-07,
+      "loss": 0.4935,
+      "step": 3520
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.14837409641663e-07,
+      "loss": 0.5433,
+      "step": 3521
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.122166362707048e-07,
+      "loss": 0.5405,
+      "step": 3522
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.096023751011414e-07,
+      "loss": 0.5833,
+      "step": 3523
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 5.069946279273464e-07,
+      "loss": 0.5001,
+      "step": 3524
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 5.043933965392134e-07,
+      "loss": 0.4509,
+      "step": 3525
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 5.017986827221733e-07,
+      "loss": 0.4362,
+      "step": 3526
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.99210488257178e-07,
+      "loss": 0.4448,
+      "step": 3527
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.966288149207055e-07,
+      "loss": 0.4034,
+      "step": 3528
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.940536644847593e-07,
+      "loss": 0.5774,
+      "step": 3529
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.914850387168657e-07,
+      "loss": 0.5366,
+      "step": 3530
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.889229393800687e-07,
+      "loss": 0.5132,
+      "step": 3531
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.863673682329373e-07,
+      "loss": 0.5325,
+      "step": 3532
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.838183270295604e-07,
+      "loss": 0.5491,
+      "step": 3533
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.812758175195397e-07,
+      "loss": 0.6001,
+      "step": 3534
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.78739841447996e-07,
+      "loss": 0.5,
+      "step": 3535
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.762104005555701e-07,
+      "loss": 0.5321,
+      "step": 3536
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.7368749657840997e-07,
+      "loss": 0.605,
+      "step": 3537
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.711711312481815e-07,
+      "loss": 0.622,
+      "step": 3538
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.68661306292062e-07,
+      "loss": 0.5239,
+      "step": 3539
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.6615802343273787e-07,
+      "loss": 0.424,
+      "step": 3540
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.6366128438840586e-07,
+      "loss": 0.6567,
+      "step": 3541
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.61171090872774e-07,
+      "loss": 0.5538,
+      "step": 3542
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.5868744459505174e-07,
+      "loss": 0.5085,
+      "step": 3543
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.5621034725995994e-07,
+      "loss": 0.554,
+      "step": 3544
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.5373980056772405e-07,
+      "loss": 0.6125,
+      "step": 3545
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.5127580621406877e-07,
+      "loss": 0.6924,
+      "step": 3546
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.4881836589022566e-07,
+      "loss": 0.6147,
+      "step": 3547
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.463674812829277e-07,
+      "loss": 0.5592,
+      "step": 3548
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.43923154074406e-07,
+      "loss": 0.4747,
+      "step": 3549
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.4148538594239176e-07,
+      "loss": 0.5349,
+      "step": 3550
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.3905417856011543e-07,
+      "loss": 0.4734,
+      "step": 3551
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.366295335963011e-07,
+      "loss": 0.5597,
+      "step": 3552
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.3421145271517193e-07,
+      "loss": 0.5303,
+      "step": 3553
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.317999375764459e-07,
+      "loss": 0.5198,
+      "step": 3554
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.293949898353289e-07,
+      "loss": 0.5063,
+      "step": 3555
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.269966111425272e-07,
+      "loss": 0.55,
+      "step": 3556
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.246048031442307e-07,
+      "loss": 0.5718,
+      "step": 3557
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.222195674821239e-07,
+      "loss": 0.5838,
+      "step": 3558
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.1984090579338054e-07,
+      "loss": 0.6345,
+      "step": 3559
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.17468819710658e-07,
+      "loss": 0.6328,
+      "step": 3560
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.15103310862105e-07,
+      "loss": 0.5566,
+      "step": 3561
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.1274438087135273e-07,
+      "loss": 0.6184,
+      "step": 3562
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.1039203135751717e-07,
+      "loss": 0.5803,
+      "step": 3563
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.0804626393519897e-07,
+      "loss": 0.5814,
+      "step": 3564
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.0570708021448134e-07,
+      "loss": 0.4961,
+      "step": 3565
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.0337448180092444e-07,
+      "loss": 0.4648,
+      "step": 3566
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.010484702955741e-07,
+      "loss": 0.5942,
+      "step": 3567
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 3.9872904729495123e-07,
+      "loss": 0.6051,
+      "step": 3568
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 3.964162143910555e-07,
+      "loss": 0.5601,
+      "step": 3569
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 3.941099731713638e-07,
+      "loss": 0.5037,
+      "step": 3570
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 3.918103252188299e-07,
+      "loss": 0.3874,
+      "step": 3571
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.895172721118778e-07,
+      "loss": 0.4976,
+      "step": 3572
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.8723081542440866e-07,
+      "loss": 0.5325,
+      "step": 3573
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.8495095672579584e-07,
+      "loss": 0.4418,
+      "step": 3574
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 3.826776975808821e-07,
+      "loss": 0.5267,
+      "step": 3575
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 3.8041103954998046e-07,
+      "loss": 0.5184,
+      "step": 3576
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 3.781509841888775e-07,
+      "loss": 0.5283,
+      "step": 3577
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 3.7589753304882125e-07,
+      "loss": 0.5151,
+      "step": 3578
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 3.736506876765311e-07,
+      "loss": 0.5448,
+      "step": 3579
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 3.714104496141924e-07,
+      "loss": 0.4598,
+      "step": 3580
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 3.6917682039945167e-07,
+      "loss": 0.4488,
+      "step": 3581
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 3.6694980156542494e-07,
+      "loss": 0.5894,
+      "step": 3582
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 3.647293946406849e-07,
+      "loss": 0.6074,
+      "step": 3583
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 3.625156011492703e-07,
+      "loss": 0.51,
+      "step": 3584
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 3.603084226106801e-07,
+      "loss": 0.4961,
+      "step": 3585
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 3.5810786053987025e-07,
+      "loss": 0.4855,
+      "step": 3586
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 3.5591391644725805e-07,
+      "loss": 0.5554,
+      "step": 3587
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.5372659183871895e-07,
+      "loss": 0.5391,
+      "step": 3588
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.515458882155798e-07,
+      "loss": 0.5238,
+      "step": 3589
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.493718070746299e-07,
+      "loss": 0.5823,
+      "step": 3590
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 3.4720434990810904e-07,
+      "loss": 0.4266,
+      "step": 3591
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 3.450435182037104e-07,
+      "loss": 0.3876,
+      "step": 3592
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 3.4288931344457987e-07,
+      "loss": 0.6011,
+      "step": 3593
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 3.4074173710931804e-07,
+      "loss": 0.5952,
+      "step": 3594
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 3.3860079067196906e-07,
+      "loss": 0.5637,
+      "step": 3595
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 3.3646647560203303e-07,
+      "loss": 0.5178,
+      "step": 3596
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 3.3433879336445597e-07,
+      "loss": 0.5,
+      "step": 3597
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 3.322177454196285e-07,
+      "loss": 0.6133,
+      "step": 3598
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 3.301033332233938e-07,
+      "loss": 0.6152,
+      "step": 3599
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.2799555822703554e-07,
+      "loss": 0.5413,
+      "step": 3600
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.2589442187728193e-07,
+      "loss": 0.4758,
+      "step": 3601
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.237999256163071e-07,
+      "loss": 0.5083,
+      "step": 3602
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.2171207088172676e-07,
+      "loss": 0.4847,
+      "step": 3603
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.196308591065966e-07,
+      "loss": 0.6748,
+      "step": 3604
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.1755629171941196e-07,
+      "loss": 0.4634,
+      "step": 3605
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.154883701441136e-07,
+      "loss": 0.4473,
+      "step": 3606
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.134270958000718e-07,
+      "loss": 0.6491,
+      "step": 3607
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.113724701021026e-07,
+      "loss": 0.5453,
+      "step": 3608
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.093244944604523e-07,
+      "loss": 0.625,
+      "step": 3609
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.0728317028080656e-07,
+      "loss": 0.5583,
+      "step": 3610
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 3.0524849896428567e-07,
+      "loss": 0.665,
+      "step": 3611
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 3.0322048190744025e-07,
+      "loss": 0.5044,
+      "step": 3612
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 3.0119912050225576e-07,
+      "loss": 0.4989,
+      "step": 3613
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 2.991844161361512e-07,
+      "loss": 0.5076,
+      "step": 3614
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 2.9717637019197163e-07,
+      "loss": 0.5818,
+      "step": 3615
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 2.9517498404799674e-07,
+      "loss": 0.5366,
+      "step": 3616
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 2.9318025907793314e-07,
+      "loss": 0.5708,
+      "step": 3617
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 2.9119219665091345e-07,
+      "loss": 0.5801,
+      "step": 3618
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 2.8921079813150066e-07,
+      "loss": 0.5214,
+      "step": 3619
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 2.8723606487968235e-07,
+      "loss": 0.5315,
+      "step": 3620
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 2.8526799825086993e-07,
+      "loss": 0.5107,
+      "step": 3621
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 2.8330659959589944e-07,
+      "loss": 0.6023,
+      "step": 3622
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 2.8135187026103403e-07,
+      "loss": 0.4827,
+      "step": 3623
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 2.794038115879538e-07,
+      "loss": 0.4895,
+      "step": 3624
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 2.774624249137614e-07,
+      "loss": 0.4245,
+      "step": 3625
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 2.755277115709842e-07,
+      "loss": 0.4358,
+      "step": 3626
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 2.735996728875623e-07,
+      "loss": 0.4791,
+      "step": 3627
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 2.716783101868614e-07,
+      "loss": 0.4863,
+      "step": 3628
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 2.6976362478766003e-07,
+      "loss": 0.5298,
+      "step": 3629
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 2.678556180041547e-07,
+      "loss": 0.5278,
+      "step": 3630
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 2.6595429114595894e-07,
+      "loss": 0.5885,
+      "step": 3631
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 2.6405964551810325e-07,
+      "loss": 0.5935,
+      "step": 3632
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 2.621716824210263e-07,
+      "loss": 0.5391,
+      "step": 3633
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 2.6029040315058487e-07,
+      "loss": 0.4845,
+      "step": 3634
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 2.5841580899804706e-07,
+      "loss": 0.474,
+      "step": 3635
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 2.565479012500915e-07,
+      "loss": 0.5753,
+      "step": 3636
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 2.5468668118880935e-07,
+      "loss": 0.5132,
+      "step": 3637
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 2.528321500916986e-07,
+      "loss": 0.4875,
+      "step": 3638
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 2.509843092316688e-07,
+      "loss": 0.489,
+      "step": 3639
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 2.4914315987703664e-07,
+      "loss": 0.6641,
+      "step": 3640
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 2.4730870329152557e-07,
+      "loss": 0.4641,
+      "step": 3641
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 2.45480940734264e-07,
+      "loss": 0.4497,
+      "step": 3642
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 2.4365987345978946e-07,
+      "loss": 0.519,
+      "step": 3643
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 2.418455027180389e-07,
+      "loss": 0.6079,
+      "step": 3644
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 2.400378297543582e-07,
+      "loss": 0.5185,
+      "step": 3645
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 2.3823685580949273e-07,
+      "loss": 0.5333,
+      "step": 3646
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 2.3644258211958926e-07,
+      "loss": 0.6313,
+      "step": 3647
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 2.3465500991619815e-07,
+      "loss": 0.5859,
+      "step": 3648
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 2.328741404262691e-07,
+      "loss": 0.5513,
+      "step": 3649
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 2.3109997487214987e-07,
+      "loss": 0.5898,
+      "step": 3650
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 2.2933251447158744e-07,
+      "loss": 0.5396,
+      "step": 3651
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 2.275717604377292e-07,
+      "loss": 0.5096,
+      "step": 3652
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 2.2581771397911512e-07,
+      "loss": 0.5403,
+      "step": 3653
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 2.2407037629968431e-07,
+      "loss": 0.4802,
+      "step": 3654
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 2.2232974859877076e-07,
+      "loss": 0.5901,
+      "step": 3655
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 2.2059583207109992e-07,
+      "loss": 0.6406,
+      "step": 3656
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 2.188686279067942e-07,
+      "loss": 0.5845,
+      "step": 3657
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 2.1714813729136975e-07,
+      "loss": 0.593,
+      "step": 3658
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 2.1543436140572972e-07,
+      "loss": 0.5151,
+      "step": 3659
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 2.1372730142617205e-07,
+      "loss": 0.4554,
+      "step": 3660
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 2.1202695852438725e-07,
+      "loss": 0.5405,
+      "step": 3661
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 2.1033333386744848e-07,
+      "loss": 0.5319,
+      "step": 3662
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 2.0864642861782246e-07,
+      "loss": 0.4668,
+      "step": 3663
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 2.069662439333664e-07,
+      "loss": 0.5684,
+      "step": 3664
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 2.0529278096731664e-07,
+      "loss": 0.4915,
+      "step": 3665
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 2.0362604086830328e-07,
+      "loss": 0.4886,
+      "step": 3666
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 2.0196602478034012e-07,
+      "loss": 0.459,
+      "step": 3667
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 2.0031273384282347e-07,
+      "loss": 0.5886,
+      "step": 3668
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 1.986661691905345e-07,
+      "loss": 0.5292,
+      "step": 3669
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.9702633195363918e-07,
+      "loss": 0.5,
+      "step": 3670
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.9539322325768494e-07,
+      "loss": 0.4742,
+      "step": 3671
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.937668442235996e-07,
+      "loss": 0.5364,
+      "step": 3672
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 1.921471959676957e-07,
+      "loss": 0.5323,
+      "step": 3673
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.9053427960166183e-07,
+      "loss": 0.6123,
+      "step": 3674
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.889280962325668e-07,
+      "loss": 0.5469,
+      "step": 3675
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.87328646962861e-07,
+      "loss": 0.4966,
+      "step": 3676
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 1.857359328903685e-07,
+      "loss": 0.5916,
+      "step": 3677
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.841499551082937e-07,
+      "loss": 0.5564,
+      "step": 3678
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.825707147052147e-07,
+      "loss": 0.5026,
+      "step": 3679
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.8099821276508888e-07,
+      "loss": 0.552,
+      "step": 3680
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 1.7943245036724512e-07,
+      "loss": 0.4823,
+      "step": 3681
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.7787342858638589e-07,
+      "loss": 0.5085,
+      "step": 3682
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.7632114849259197e-07,
+      "loss": 0.4843,
+      "step": 3683
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.747756111513099e-07,
+      "loss": 0.5168,
+      "step": 3684
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 1.7323681762336453e-07,
+      "loss": 0.5906,
+      "step": 3685
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.717047689649487e-07,
+      "loss": 0.4578,
+      "step": 3686
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.701794662276257e-07,
+      "loss": 0.5342,
+      "step": 3687
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.6866091045832923e-07,
+      "loss": 0.5493,
+      "step": 3688
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 1.6714910269936213e-07,
+      "loss": 0.5999,
+      "step": 3689
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.656440439883944e-07,
+      "loss": 0.5312,
+      "step": 3690
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.6414573535846523e-07,
+      "loss": 0.5669,
+      "step": 3691
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 1.6265417783798087e-07,
+      "loss": 0.7036,
+      "step": 3692
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.6116937245071018e-07,
+      "loss": 0.539,
+      "step": 3693
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.5969132021579347e-07,
+      "loss": 0.6401,
+      "step": 3694
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.5822002214773035e-07,
+      "loss": 0.574,
+      "step": 3695
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 1.5675547925638856e-07,
+      "loss": 0.4796,
+      "step": 3696
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.5529769254699512e-07,
+      "loss": 0.5559,
+      "step": 3697
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.5384666302014407e-07,
+      "loss": 0.541,
+      "step": 3698
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.5240239167178872e-07,
+      "loss": 0.5414,
+      "step": 3699
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 1.50964879493245e-07,
+      "loss": 0.5635,
+      "step": 3700
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.4953412747118813e-07,
+      "loss": 0.6963,
+      "step": 3701
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.4811013658765471e-07,
+      "loss": 0.5129,
+      "step": 3702
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.4669290782003963e-07,
+      "loss": 0.5764,
+      "step": 3703
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 1.452824421410981e-07,
+      "loss": 0.4956,
+      "step": 3704
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.4387874051894013e-07,
+      "loss": 0.6299,
+      "step": 3705
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.4248180391703614e-07,
+      "loss": 0.4708,
+      "step": 3706
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.410916332942125e-07,
+      "loss": 0.6455,
+      "step": 3707
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 1.3970822960465035e-07,
+      "loss": 0.6455,
+      "step": 3708
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.3833159379788686e-07,
+      "loss": 0.6331,
+      "step": 3709
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.3696172681881504e-07,
+      "loss": 0.5015,
+      "step": 3710
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.3559862960767944e-07,
+      "loss": 0.4909,
+      "step": 3711
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 1.3424230310007946e-07,
+      "loss": 0.511,
+      "step": 3712
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.3289274822696818e-07,
+      "loss": 0.5583,
+      "step": 3713
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.315499659146491e-07,
+      "loss": 0.5193,
+      "step": 3714
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.3021395708477824e-07,
+      "loss": 0.6309,
+      "step": 3715
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 1.2888472265436104e-07,
+      "loss": 0.5876,
+      "step": 3716
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.2756226353575428e-07,
+      "loss": 0.5024,
+      "step": 3717
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.262465806366664e-07,
+      "loss": 0.4034,
+      "step": 3718
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.249376748601494e-07,
+      "loss": 0.5979,
+      "step": 3719
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 1.2363554710460913e-07,
+      "loss": 0.6135,
+      "step": 3720
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.2234019826379618e-07,
+      "loss": 0.5442,
+      "step": 3721
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.2105162922680823e-07,
+      "loss": 0.4657,
+      "step": 3722
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.1976984087809007e-07,
+      "loss": 0.5264,
+      "step": 3723
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 1.1849483409743457e-07,
+      "loss": 0.5825,
+      "step": 3724
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.1722660975997501e-07,
+      "loss": 0.6658,
+      "step": 3725
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.1596516873619401e-07,
+      "loss": 0.5474,
+      "step": 3726
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.1471051189191562e-07,
+      "loss": 0.4574,
+      "step": 3727
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 1.1346264008830765e-07,
+      "loss": 0.5154,
+      "step": 3728
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.1222155418188163e-07,
+      "loss": 0.4559,
+      "step": 3729
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.109872550244917e-07,
+      "loss": 0.4839,
+      "step": 3730
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.097597434633324e-07,
+      "loss": 0.6108,
+      "step": 3731
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 1.0853902034093977e-07,
+      "loss": 0.5449,
+      "step": 3732
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0732508649519247e-07,
+      "loss": 0.646,
+      "step": 3733
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0611794275930398e-07,
+      "loss": 0.5069,
+      "step": 3734
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0491758996183487e-07,
+      "loss": 0.6191,
+      "step": 3735
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 1.0372402892667721e-07,
+      "loss": 0.5947,
+      "step": 3736
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0253726047306567e-07,
+      "loss": 0.5427,
+      "step": 3737
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0135728541557199e-07,
+      "loss": 0.5266,
+      "step": 3738
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 1.0018410456410388e-07,
+      "loss": 0.5713,
+      "step": 3739
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 9.901771872390497e-08,
+      "loss": 0.5165,
+      "step": 3740
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 9.785812869555933e-08,
+      "loss": 0.5493,
+      "step": 3741
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 9.670533527498139e-08,
+      "loss": 0.5507,
+      "step": 3742
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 9.555933925342265e-08,
+      "loss": 0.5322,
+      "step": 3743
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 9.442014141747058e-08,
+      "loss": 0.5464,
+      "step": 3744
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 9.328774254904305e-08,
+      "loss": 0.6077,
+      "step": 3745
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 9.216214342539387e-08,
+      "loss": 0.4544,
+      "step": 3746
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 9.104334481911059e-08,
+      "loss": 0.4967,
+      "step": 3747
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 8.993134749810895e-08,
+      "loss": 0.4736,
+      "step": 3748
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 8.88261522256395e-08,
+      "loss": 0.5255,
+      "step": 3749
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 8.772775976028547e-08,
+      "loss": 0.625,
+      "step": 3750
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 8.66361708559571e-08,
+      "loss": 0.4745,
+      "step": 3751
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 8.555138626189619e-08,
+      "loss": 0.6423,
+      "step": 3752
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 8.44734067226749e-08,
+      "loss": 0.5852,
+      "step": 3753
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 8.340223297819472e-08,
+      "loss": 0.5789,
+      "step": 3754
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 8.233786576368418e-08,
+      "loss": 0.4646,
+      "step": 3755
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 8.128030580970336e-08,
+      "loss": 0.6335,
+      "step": 3756
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 8.022955384213493e-08,
+      "loss": 0.4252,
+      "step": 3757
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 7.918561058219198e-08,
+      "loss": 0.3809,
+      "step": 3758
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 7.814847674641468e-08,
+      "loss": 0.5996,
+      "step": 3759
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 7.711815304666804e-08,
+      "loss": 0.6394,
+      "step": 3760
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 7.609464019014301e-08,
+      "loss": 0.5608,
+      "step": 3761
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 7.507793887935655e-08,
+      "loss": 0.496,
+      "step": 3762
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 7.406804981214822e-08,
+      "loss": 0.6619,
+      "step": 3763
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 7.306497368168352e-08,
+      "loss": 0.5829,
+      "step": 3764
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 7.206871117645286e-08,
+      "loss": 0.5144,
+      "step": 3765
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 7.10792629802659e-08,
+      "loss": 0.5635,
+      "step": 3766
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 7.00966297722605e-08,
+      "loss": 0.5129,
+      "step": 3767
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 6.91208122268916e-08,
+      "loss": 0.5853,
+      "step": 3768
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 6.815181101393898e-08,
+      "loss": 0.4511,
+      "step": 3769
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 6.718962679850504e-08,
+      "loss": 0.3652,
+      "step": 3770
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 6.623426024100931e-08,
+      "loss": 0.4648,
+      "step": 3771
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 6.528571199719502e-08,
+      "loss": 0.4696,
+      "step": 3772
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 6.43439827181247e-08,
+      "loss": 0.5208,
+      "step": 3773
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 6.340907305017907e-08,
+      "loss": 0.5461,
+      "step": 3774
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 6.248098363505928e-08,
+      "loss": 0.5542,
+      "step": 3775
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 6.155971510978576e-08,
+      "loss": 0.5468,
+      "step": 3776
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 6.064526810669602e-08,
+      "loss": 0.5156,
+      "step": 3777
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 5.973764325344689e-08,
+      "loss": 0.5828,
+      "step": 3778
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 5.8836841173011136e-08,
+      "loss": 0.5208,
+      "step": 3779
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 5.794286248367753e-08,
+      "loss": 0.4275,
+      "step": 3780
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 5.705570779905523e-08,
+      "loss": 0.5054,
+      "step": 3781
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 5.617537772806603e-08,
+      "loss": 0.4125,
+      "step": 3782
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.530187287494882e-08,
+      "loss": 0.5144,
+      "step": 3783
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.443519383925844e-08,
+      "loss": 0.5588,
+      "step": 3784
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.3575341215863496e-08,
+      "loss": 0.6001,
+      "step": 3785
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 5.272231559494634e-08,
+      "loss": 0.4402,
+      "step": 3786
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 5.187611756200639e-08,
+      "loss": 0.594,
+      "step": 3787
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 5.103674769785461e-08,
+      "loss": 0.5524,
+      "step": 3788
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 5.020420657861569e-08,
+      "loss": 0.6577,
+      "step": 3789
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 4.937849477572587e-08,
+      "loss": 0.5938,
+      "step": 3790
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.8559612855936244e-08,
+      "loss": 0.5295,
+      "step": 3791
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.774756138130831e-08,
+      "loss": 0.4608,
+      "step": 3792
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.694234090921734e-08,
+      "loss": 0.4446,
+      "step": 3793
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 4.6143951992346783e-08,
+      "loss": 0.5867,
+      "step": 3794
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.535239517869494e-08,
+      "loss": 0.5656,
+      "step": 3795
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.4567671011567224e-08,
+      "loss": 0.5352,
+      "step": 3796
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.3789780029579456e-08,
+      "loss": 0.5215,
+      "step": 3797
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 4.3018722766661194e-08,
+      "loss": 0.4575,
+      "step": 3798
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.225449975204798e-08,
+      "loss": 0.4249,
+      "step": 3799
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.1497111510283574e-08,
+      "loss": 0.5925,
+      "step": 3800
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.0746558561224336e-08,
+      "loss": 0.5028,
+      "step": 3801
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 4.000284142003264e-08,
+      "loss": 0.3465,
+      "step": 3802
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.926596059717791e-08,
+      "loss": 0.4948,
+      "step": 3803
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.853591659844003e-08,
+      "loss": 0.5491,
+      "step": 3804
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.781270992490482e-08,
+      "loss": 0.575,
+      "step": 3805
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 3.7096341072964116e-08,
+      "loss": 0.6101,
+      "step": 3806
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 3.638681053431792e-08,
+      "loss": 0.5701,
+      "step": 3807
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 3.568411879597111e-08,
+      "loss": 0.5378,
+      "step": 3808
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 3.498826634023789e-08,
+      "loss": 0.5627,
+      "step": 3809
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 3.429925364473285e-08,
+      "loss": 0.4832,
+      "step": 3810
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 3.361708118238105e-08,
+      "loss": 0.6338,
+      "step": 3811
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 3.294174942140904e-08,
+      "loss": 0.6064,
+      "step": 3812
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 3.227325882535048e-08,
+      "loss": 0.5962,
+      "step": 3813
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 3.161160985304168e-08,
+      "loss": 0.5253,
+      "step": 3814
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.09568029586238e-08,
+      "loss": 0.5513,
+      "step": 3815
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 3.030883859154066e-08,
+      "loss": 0.5947,
+      "step": 3816
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 2.9667717196542045e-08,
+      "loss": 0.4725,
+      "step": 3817
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 2.903343921367707e-08,
+      "loss": 0.5999,
+      "step": 3818
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 2.8406005078301934e-08,
+      "loss": 0.5178,
+      "step": 3819
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 2.7785415221073254e-08,
+      "loss": 0.6047,
+      "step": 3820
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 2.7171670067949186e-08,
+      "loss": 0.5718,
+      "step": 3821
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 2.6564770040190535e-08,
+      "loss": 0.5479,
+      "step": 3822
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 2.5964715554360753e-08,
+      "loss": 0.6296,
+      "step": 3823
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 2.5371507022322606e-08,
+      "loss": 0.585,
+      "step": 3824
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 2.4785144851241506e-08,
+      "loss": 0.6001,
+      "step": 3825
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 2.420562944358329e-08,
+      "loss": 0.3743,
+      "step": 3826
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 2.3632961197115335e-08,
+      "loss": 0.5715,
+      "step": 3827
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 2.3067140504902108e-08,
+      "loss": 0.4631,
+      "step": 3828
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 2.250816775531295e-08,
+      "loss": 0.4866,
+      "step": 3829
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 2.1956043332010957e-08,
+      "loss": 0.5038,
+      "step": 3830
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 2.1410767613965212e-08,
+      "loss": 0.5615,
+      "step": 3831
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 2.0872340975438555e-08,
+      "loss": 0.5012,
+      "step": 3832
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 2.0340763785995366e-08,
+      "loss": 0.5667,
+      "step": 3833
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.981603641049934e-08,
+      "loss": 0.6267,
+      "step": 3834
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.9298159209110157e-08,
+      "loss": 0.4756,
+      "step": 3835
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.878713253728792e-08,
+      "loss": 0.5698,
+      "step": 3836
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 1.8282956745789837e-08,
+      "loss": 0.5144,
+      "step": 3837
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 1.77856321806702e-08,
+      "loss": 0.5305,
+      "step": 3838
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 1.7295159183281507e-08,
+      "loss": 0.5482,
+      "step": 3839
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 1.6811538090274472e-08,
+      "loss": 0.4468,
+      "step": 3840
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 1.6334769233594673e-08,
+      "loss": 0.5253,
+      "step": 3841
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 1.5864852940485896e-08,
+      "loss": 0.6155,
+      "step": 3842
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 1.540178953348792e-08,
+      "loss": 0.5936,
+      "step": 3843
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 1.494557933043761e-08,
+      "loss": 0.4492,
+      "step": 3844
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 1.4496222644467816e-08,
+      "loss": 0.5275,
+      "step": 3845
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.405371978400516e-08,
+      "loss": 0.491,
+      "step": 3846
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.3618071052774462e-08,
+      "loss": 0.5319,
+      "step": 3847
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.3189276749795421e-08,
+      "loss": 0.582,
+      "step": 3848
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 1.276733716938372e-08,
+      "loss": 0.5334,
+      "step": 3849
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 1.2352252601147696e-08,
+      "loss": 0.6414,
+      "step": 3850
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 1.1944023329992782e-08,
+      "loss": 0.4254,
+      "step": 3851
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 1.1542649636118175e-08,
+      "loss": 0.6124,
+      "step": 3852
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 1.1148131795017946e-08,
+      "loss": 0.5479,
+      "step": 3853
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 1.0760470077479934e-08,
+      "loss": 0.4006,
+      "step": 3854
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 1.0379664749586849e-08,
+      "loss": 0.5176,
+      "step": 3855
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 1.0005716072714055e-08,
+      "loss": 0.6396,
+      "step": 3856
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 9.638624303532907e-09,
+      "loss": 0.413,
+      "step": 3857
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 9.278389694006296e-09,
+      "loss": 0.532,
+      "step": 3858
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.925012491390882e-09,
+      "loss": 0.5486,
+      "step": 3859
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.578492938238203e-09,
+      "loss": 0.6406,
+      "step": 3860
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 8.238831272389114e-09,
+      "loss": 0.5137,
+      "step": 3861
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 7.906027726981568e-09,
+      "loss": 0.4819,
+      "step": 3862
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 7.580082530443955e-09,
+      "loss": 0.5322,
+      "step": 3863
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 7.260995906497314e-09,
+      "loss": 0.3779,
+      "step": 3864
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 6.9487680741553435e-09,
+      "loss": 0.5659,
+      "step": 3865
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 6.643399247725502e-09,
+      "loss": 0.552,
+      "step": 3866
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 6.3448896368056845e-09,
+      "loss": 0.5435,
+      "step": 3867
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 6.05323944628533e-09,
+      "loss": 0.5436,
+      "step": 3868
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.768448876347643e-09,
+      "loss": 0.6685,
+      "step": 3869
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.49051812246515e-09,
+      "loss": 0.5969,
+      "step": 3870
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.219447375405251e-09,
+      "loss": 0.4907,
+      "step": 3871
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 4.955236821222453e-09,
+      "loss": 0.4987,
+      "step": 3872
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.697886641266136e-09,
+      "loss": 0.5386,
+      "step": 3873
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.447397012177224e-09,
+      "loss": 0.4445,
+      "step": 3874
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.203768105882633e-09,
+      "loss": 0.5592,
+      "step": 3875
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 3.9670000896074865e-09,
+      "loss": 0.4521,
+      "step": 3876
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.73709312586068e-09,
+      "loss": 0.6167,
+      "step": 3877
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.514047372448204e-09,
+      "loss": 0.5205,
+      "step": 3878
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.297862982460931e-09,
+      "loss": 0.6355,
+      "step": 3879
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 3.0885401042857197e-09,
+      "loss": 0.5632,
+      "step": 3880
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 2.886078881595422e-09,
+      "loss": 0.5234,
+      "step": 3881
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 2.690479453354433e-09,
+      "loss": 0.5352,
+      "step": 3882
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 2.501741953818693e-09,
+      "loss": 0.5496,
+      "step": 3883
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 2.319866512533464e-09,
+      "loss": 0.516,
+      "step": 3884
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 2.1448532543333344e-09,
+      "loss": 0.4863,
+      "step": 3885
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 1.9767022993444353e-09,
+      "loss": 0.603,
+      "step": 3886
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 1.8154137629811108e-09,
+      "loss": 0.6238,
+      "step": 3887
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 1.6609877559481402e-09,
+      "loss": 0.5454,
+      "step": 3888
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.5134243842396257e-09,
+      "loss": 0.5415,
+      "step": 3889
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.3727237491412138e-09,
+      "loss": 0.5153,
+      "step": 3890
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.2388859472256542e-09,
+      "loss": 0.5896,
+      "step": 3891
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 1.111911070356131e-09,
+      "loss": 0.5339,
+      "step": 3892
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 9.917992056851511e-10,
+      "loss": 0.5242,
+      "step": 3893
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 8.785504356556562e-10,
+      "loss": 0.5278,
+      "step": 3894
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 7.721648379988011e-10,
+      "loss": 0.5022,
+      "step": 3895
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 6.726424857339542e-10,
+      "loss": 0.3636,
+      "step": 3896
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 5.79983447173138e-10,
+      "loss": 0.5707,
+      "step": 3897
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.941877859143684e-10,
+      "loss": 0.522,
+      "step": 3898
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 4.1525556084609507e-10,
+      "loss": 0.4795,
+      "step": 3899
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 3.4318682614498113e-10,
+      "loss": 0.5857,
+      "step": 3900
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 2.7798163127812363e-10,
+      "loss": 0.4921,
+      "step": 3901
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 2.1964002100083314e-10,
+      "loss": 0.4469,
+      "step": 3902
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 1.6816203535663379e-10,
+      "loss": 0.4678,
+      "step": 3903
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 1.235477096805937e-10,
+      "loss": 0.667,
+      "step": 3904
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 8.579707459266396e-11,
+      "loss": 0.5101,
+      "step": 3905
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 5.491015600656013e-11,
+      "loss": 0.4891,
+      "step": 3906
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 3.088697512088068e-11,
+      "loss": 0.4841,
+      "step": 3907
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 1.3727548423547732e-11,
+      "loss": 0.5796,
+      "step": 3908
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 3.4318876951378034e-12,
+      "loss": 0.5481,
+      "step": 3909
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 0.0,
+      "loss": 0.4433,
+      "step": 3910
+    },
+    {
+      "epoch": 10.0,
+      "step": 3910,
+      "total_flos": 950066681610240.0,
+      "train_loss": 0.7636732799012947,
+      "train_runtime": 48931.4526,
+      "train_samples_per_second": 10.218,
+      "train_steps_per_second": 0.08
+    }
+  ],
+  "max_steps": 3910,
+  "num_train_epochs": 10,
+  "total_flos": 950066681610240.0,
+  "trial_name": null,
+  "trial_params": null
+}