Upload trainer_state.json
Browse files- trainer_state.json +220 -0
trainer_state.json
ADDED
|
@@ -0,0 +1,220 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 9.025093793163277,
|
| 3 |
+
"best_model_checkpoint": "./envi_checkpoints/checkpoint-31250",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"global_step": 31250,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"learning_rate": 3.1808000000000005e-06,
|
| 13 |
+
"loss": 7.9834,
|
| 14 |
+
"step": 1000
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.13,
|
| 18 |
+
"learning_rate": 6.380800000000001e-06,
|
| 19 |
+
"loss": 5.5297,
|
| 20 |
+
"step": 2000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.19,
|
| 24 |
+
"learning_rate": 9.580800000000002e-06,
|
| 25 |
+
"loss": 4.7262,
|
| 26 |
+
"step": 3000
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.26,
|
| 30 |
+
"learning_rate": 1.2780800000000001e-05,
|
| 31 |
+
"loss": 4.4162,
|
| 32 |
+
"step": 4000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.32,
|
| 36 |
+
"learning_rate": 1.59808e-05,
|
| 37 |
+
"loss": 4.2072,
|
| 38 |
+
"step": 5000
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.38,
|
| 42 |
+
"learning_rate": 1.91808e-05,
|
| 43 |
+
"loss": 4.0103,
|
| 44 |
+
"step": 6000
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.45,
|
| 48 |
+
"learning_rate": 2.23808e-05,
|
| 49 |
+
"loss": 3.7644,
|
| 50 |
+
"step": 7000
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.51,
|
| 54 |
+
"learning_rate": 2.55808e-05,
|
| 55 |
+
"loss": 3.3987,
|
| 56 |
+
"step": 8000
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.58,
|
| 60 |
+
"learning_rate": 2.87808e-05,
|
| 61 |
+
"loss": 2.8234,
|
| 62 |
+
"step": 9000
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.64,
|
| 66 |
+
"learning_rate": 3.19808e-05,
|
| 67 |
+
"loss": 2.0638,
|
| 68 |
+
"step": 10000
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.7,
|
| 72 |
+
"learning_rate": 3.51808e-05,
|
| 73 |
+
"loss": 1.5503,
|
| 74 |
+
"step": 11000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.77,
|
| 78 |
+
"learning_rate": 3.83776e-05,
|
| 79 |
+
"loss": 1.2126,
|
| 80 |
+
"step": 12000
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.83,
|
| 84 |
+
"learning_rate": 4.15776e-05,
|
| 85 |
+
"loss": 0.9281,
|
| 86 |
+
"step": 13000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.9,
|
| 90 |
+
"learning_rate": 4.47776e-05,
|
| 91 |
+
"loss": 0.713,
|
| 92 |
+
"step": 14000
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.96,
|
| 96 |
+
"learning_rate": 4.79776e-05,
|
| 97 |
+
"loss": 0.6009,
|
| 98 |
+
"step": 15000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.0,
|
| 102 |
+
"eval_bleu": 8.6159300527749,
|
| 103 |
+
"eval_loss": 0.45428285002708435,
|
| 104 |
+
"eval_runtime": 93.2143,
|
| 105 |
+
"eval_samples_per_second": 26.82,
|
| 106 |
+
"eval_steps_per_second": 0.848,
|
| 107 |
+
"step": 15625
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.02,
|
| 111 |
+
"learning_rate": 4.88224e-05,
|
| 112 |
+
"loss": 0.54,
|
| 113 |
+
"step": 16000
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.09,
|
| 117 |
+
"learning_rate": 4.56256e-05,
|
| 118 |
+
"loss": 0.5017,
|
| 119 |
+
"step": 17000
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.15,
|
| 123 |
+
"learning_rate": 4.24288e-05,
|
| 124 |
+
"loss": 0.4777,
|
| 125 |
+
"step": 18000
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 1.22,
|
| 129 |
+
"learning_rate": 3.9228800000000006e-05,
|
| 130 |
+
"loss": 0.4661,
|
| 131 |
+
"step": 19000
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 1.28,
|
| 135 |
+
"learning_rate": 3.60288e-05,
|
| 136 |
+
"loss": 0.4526,
|
| 137 |
+
"step": 20000
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 1.34,
|
| 141 |
+
"learning_rate": 3.2828800000000007e-05,
|
| 142 |
+
"loss": 0.4457,
|
| 143 |
+
"step": 21000
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.41,
|
| 147 |
+
"learning_rate": 2.96288e-05,
|
| 148 |
+
"loss": 0.4386,
|
| 149 |
+
"step": 22000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 1.47,
|
| 153 |
+
"learning_rate": 2.6428800000000004e-05,
|
| 154 |
+
"loss": 0.4365,
|
| 155 |
+
"step": 23000
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 1.54,
|
| 159 |
+
"learning_rate": 2.32288e-05,
|
| 160 |
+
"loss": 0.4305,
|
| 161 |
+
"step": 24000
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.6,
|
| 165 |
+
"learning_rate": 2.0032e-05,
|
| 166 |
+
"loss": 0.4216,
|
| 167 |
+
"step": 25000
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 1.66,
|
| 171 |
+
"learning_rate": 1.6832e-05,
|
| 172 |
+
"loss": 0.4096,
|
| 173 |
+
"step": 26000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 1.73,
|
| 177 |
+
"learning_rate": 1.3632000000000001e-05,
|
| 178 |
+
"loss": 0.4066,
|
| 179 |
+
"step": 27000
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 1.79,
|
| 183 |
+
"learning_rate": 1.0432e-05,
|
| 184 |
+
"loss": 0.4071,
|
| 185 |
+
"step": 28000
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 1.86,
|
| 189 |
+
"learning_rate": 7.2319999999999995e-06,
|
| 190 |
+
"loss": 0.4048,
|
| 191 |
+
"step": 29000
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 1.92,
|
| 195 |
+
"learning_rate": 4.0320000000000005e-06,
|
| 196 |
+
"loss": 0.3997,
|
| 197 |
+
"step": 30000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 1.98,
|
| 201 |
+
"learning_rate": 8.352000000000001e-07,
|
| 202 |
+
"loss": 0.4015,
|
| 203 |
+
"step": 31000
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.0,
|
| 207 |
+
"eval_bleu": 9.025093793163277,
|
| 208 |
+
"eval_loss": 0.36966755986213684,
|
| 209 |
+
"eval_runtime": 94.07,
|
| 210 |
+
"eval_samples_per_second": 26.576,
|
| 211 |
+
"eval_steps_per_second": 0.84,
|
| 212 |
+
"step": 31250
|
| 213 |
+
}
|
| 214 |
+
],
|
| 215 |
+
"max_steps": 31250,
|
| 216 |
+
"num_train_epochs": 2,
|
| 217 |
+
"total_flos": 1.505048664e+17,
|
| 218 |
+
"trial_name": null,
|
| 219 |
+
"trial_params": null
|
| 220 |
+
}
|