Upload checkpoint-7245/trainer_state.json with huggingface_hub
Browse files
checkpoint-7245/trainer_state.json
ADDED
|
@@ -0,0 +1,480 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 7245,
|
| 3 |
+
"best_metric": 0.8153345585743379,
|
| 4 |
+
"best_model_checkpoint": "outputs/runs/bartpho/checkpoint-7245",
|
| 5 |
+
"epoch": 23.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 7245,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.9968253968253968,
|
| 14 |
+
"grad_norm": 5.682558536529541,
|
| 15 |
+
"learning_rate": 4.999432333543028e-06,
|
| 16 |
+
"loss": 0.6382,
|
| 17 |
+
"step": 314
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 1.0,
|
| 21 |
+
"eval_accuracy": 0.6618786257685761,
|
| 22 |
+
"eval_f1": 0.5275329196417486,
|
| 23 |
+
"eval_loss": 0.623261570930481,
|
| 24 |
+
"eval_precision": 0.7377902962250189,
|
| 25 |
+
"eval_recall": 0.5717004519029796,
|
| 26 |
+
"eval_runtime": 1.4418,
|
| 27 |
+
"eval_samples_per_second": 902.363,
|
| 28 |
+
"eval_steps_per_second": 28.437,
|
| 29 |
+
"step": 315
|
| 30 |
+
},
|
| 31 |
+
{
|
| 32 |
+
"epoch": 1.9936507936507937,
|
| 33 |
+
"grad_norm": 1.4926543235778809,
|
| 34 |
+
"learning_rate": 4.996525669931999e-06,
|
| 35 |
+
"loss": 0.5683,
|
| 36 |
+
"step": 628
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 2.0,
|
| 40 |
+
"eval_accuracy": 0.6871157119533122,
|
| 41 |
+
"eval_f1": 0.6270341888525142,
|
| 42 |
+
"eval_loss": 0.5687424540519714,
|
| 43 |
+
"eval_precision": 0.6851236426126874,
|
| 44 |
+
"eval_recall": 0.6290258325510966,
|
| 45 |
+
"eval_runtime": 1.4493,
|
| 46 |
+
"eval_samples_per_second": 897.649,
|
| 47 |
+
"eval_steps_per_second": 28.289,
|
| 48 |
+
"step": 630
|
| 49 |
+
},
|
| 50 |
+
{
|
| 51 |
+
"epoch": 2.9904761904761905,
|
| 52 |
+
"grad_norm": 6.10944128036499,
|
| 53 |
+
"learning_rate": 4.991155236893945e-06,
|
| 54 |
+
"loss": 0.5534,
|
| 55 |
+
"step": 942
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"epoch": 3.0,
|
| 59 |
+
"eval_accuracy": 0.6981102581026157,
|
| 60 |
+
"eval_f1": 0.6686886646196348,
|
| 61 |
+
"eval_loss": 0.5649789571762085,
|
| 62 |
+
"eval_precision": 0.6820856346718416,
|
| 63 |
+
"eval_recall": 0.6647145253749493,
|
| 64 |
+
"eval_runtime": 1.4482,
|
| 65 |
+
"eval_samples_per_second": 898.381,
|
| 66 |
+
"eval_steps_per_second": 28.312,
|
| 67 |
+
"step": 945
|
| 68 |
+
},
|
| 69 |
+
{
|
| 70 |
+
"epoch": 3.9873015873015873,
|
| 71 |
+
"grad_norm": 3.255537509918213,
|
| 72 |
+
"learning_rate": 4.983326334397891e-06,
|
| 73 |
+
"loss": 0.5436,
|
| 74 |
+
"step": 1256
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 4.0,
|
| 78 |
+
"eval_accuracy": 0.697224441588217,
|
| 79 |
+
"eval_f1": 0.6146617747343325,
|
| 80 |
+
"eval_loss": 0.5817087888717651,
|
| 81 |
+
"eval_precision": 0.7360948991758735,
|
| 82 |
+
"eval_recall": 0.625299233298471,
|
| 83 |
+
"eval_runtime": 1.4497,
|
| 84 |
+
"eval_samples_per_second": 897.453,
|
| 85 |
+
"eval_steps_per_second": 28.283,
|
| 86 |
+
"step": 1260
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.984126984126984,
|
| 90 |
+
"grad_norm": 4.936647891998291,
|
| 91 |
+
"learning_rate": 4.97304668862541e-06,
|
| 92 |
+
"loss": 0.5387,
|
| 93 |
+
"step": 1570
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 5.0,
|
| 97 |
+
"eval_accuracy": 0.6994824052523708,
|
| 98 |
+
"eval_f1": 0.6780463785641409,
|
| 99 |
+
"eval_loss": 0.5464158654212952,
|
| 100 |
+
"eval_precision": 0.6826491609942078,
|
| 101 |
+
"eval_recall": 0.6754857450135558,
|
| 102 |
+
"eval_runtime": 1.4501,
|
| 103 |
+
"eval_samples_per_second": 897.181,
|
| 104 |
+
"eval_steps_per_second": 28.274,
|
| 105 |
+
"step": 1575
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 5.980952380952381,
|
| 109 |
+
"grad_norm": 5.54253625869751,
|
| 110 |
+
"learning_rate": 4.9603264443458e-06,
|
| 111 |
+
"loss": 0.5347,
|
| 112 |
+
"step": 1884
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 6.0,
|
| 116 |
+
"eval_accuracy": 0.703234098725119,
|
| 117 |
+
"eval_f1": 0.658548053122289,
|
| 118 |
+
"eval_loss": 0.5504359602928162,
|
| 119 |
+
"eval_precision": 0.6971966581407725,
|
| 120 |
+
"eval_recall": 0.6552219050544602,
|
| 121 |
+
"eval_runtime": 1.4502,
|
| 122 |
+
"eval_samples_per_second": 897.134,
|
| 123 |
+
"eval_steps_per_second": 28.272,
|
| 124 |
+
"step": 1890
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 6.977777777777778,
|
| 128 |
+
"grad_norm": 2.980400800704956,
|
| 129 |
+
"learning_rate": 4.945178154904432e-06,
|
| 130 |
+
"loss": 0.5295,
|
| 131 |
+
"step": 2198
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 7.0,
|
| 135 |
+
"eval_accuracy": 0.7049188869976031,
|
| 136 |
+
"eval_f1": 0.6712744945089447,
|
| 137 |
+
"eval_loss": 0.5498743057250977,
|
| 138 |
+
"eval_precision": 0.6921020382053092,
|
| 139 |
+
"eval_recall": 0.6667597786262732,
|
| 140 |
+
"eval_runtime": 1.451,
|
| 141 |
+
"eval_samples_per_second": 896.602,
|
| 142 |
+
"eval_steps_per_second": 28.256,
|
| 143 |
+
"step": 2205
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 7.974603174603175,
|
| 147 |
+
"grad_norm": 1.8561100959777832,
|
| 148 |
+
"learning_rate": 4.92761676983411e-06,
|
| 149 |
+
"loss": 0.526,
|
| 150 |
+
"step": 2512
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 8.0,
|
| 154 |
+
"eval_accuracy": 0.7058568103657901,
|
| 155 |
+
"eval_f1": 0.662676465048434,
|
| 156 |
+
"eval_loss": 0.5571920871734619,
|
| 157 |
+
"eval_precision": 0.6997671412894451,
|
| 158 |
+
"eval_recall": 0.6589087948228982,
|
| 159 |
+
"eval_runtime": 1.4516,
|
| 160 |
+
"eval_samples_per_second": 896.228,
|
| 161 |
+
"eval_steps_per_second": 28.244,
|
| 162 |
+
"step": 2520
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 8.971428571428572,
|
| 166 |
+
"grad_norm": 3.5451998710632324,
|
| 167 |
+
"learning_rate": 4.9076596201016856e-06,
|
| 168 |
+
"loss": 0.523,
|
| 169 |
+
"step": 2826
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 9.0,
|
| 173 |
+
"eval_accuracy": 0.7018619515753639,
|
| 174 |
+
"eval_f1": 0.625374057862767,
|
| 175 |
+
"eval_loss": 0.5914136171340942,
|
| 176 |
+
"eval_precision": 0.7358493036678011,
|
| 177 |
+
"eval_recall": 0.6327511530635204,
|
| 178 |
+
"eval_runtime": 1.4534,
|
| 179 |
+
"eval_samples_per_second": 895.149,
|
| 180 |
+
"eval_steps_per_second": 28.21,
|
| 181 |
+
"step": 2835
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 9.968253968253968,
|
| 185 |
+
"grad_norm": 5.0408406257629395,
|
| 186 |
+
"learning_rate": 4.88532640100449e-06,
|
| 187 |
+
"loss": 0.5171,
|
| 188 |
+
"step": 3140
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 10.0,
|
| 192 |
+
"eval_accuracy": 0.7077847639559524,
|
| 193 |
+
"eval_f1": 0.6786121509955985,
|
| 194 |
+
"eval_loss": 0.5660568475723267,
|
| 195 |
+
"eval_precision": 0.6936028147526183,
|
| 196 |
+
"eval_recall": 0.6741239127178265,
|
| 197 |
+
"eval_runtime": 1.4513,
|
| 198 |
+
"eval_samples_per_second": 896.413,
|
| 199 |
+
"eval_steps_per_second": 28.25,
|
| 200 |
+
"step": 3150
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 10.965079365079365,
|
| 204 |
+
"grad_norm": 9.36005973815918,
|
| 205 |
+
"learning_rate": 4.860639152733449e-06,
|
| 206 |
+
"loss": 0.5147,
|
| 207 |
+
"step": 3454
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 11.0,
|
| 211 |
+
"eval_accuracy": 0.701914058429152,
|
| 212 |
+
"eval_f1": 0.6364582078339008,
|
| 213 |
+
"eval_loss": 0.5687407851219177,
|
| 214 |
+
"eval_precision": 0.7180110728497148,
|
| 215 |
+
"eval_recall": 0.6392313775049545,
|
| 216 |
+
"eval_runtime": 1.4526,
|
| 217 |
+
"eval_samples_per_second": 895.624,
|
| 218 |
+
"eval_steps_per_second": 28.225,
|
| 219 |
+
"step": 3465
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 11.961904761904762,
|
| 223 |
+
"grad_norm": 3.487974166870117,
|
| 224 |
+
"learning_rate": 4.833622238622079e-06,
|
| 225 |
+
"loss": 0.5095,
|
| 226 |
+
"step": 3768
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 12.0,
|
| 230 |
+
"eval_accuracy": 0.7085316288602495,
|
| 231 |
+
"eval_f1": 0.6626885625948004,
|
| 232 |
+
"eval_loss": 0.5603615045547485,
|
| 233 |
+
"eval_precision": 0.7060931620291221,
|
| 234 |
+
"eval_recall": 0.6591844125186699,
|
| 235 |
+
"eval_runtime": 1.4523,
|
| 236 |
+
"eval_samples_per_second": 895.797,
|
| 237 |
+
"eval_steps_per_second": 28.23,
|
| 238 |
+
"step": 3780
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 12.958730158730159,
|
| 242 |
+
"grad_norm": 5.526001930236816,
|
| 243 |
+
"learning_rate": 4.804302321102816e-06,
|
| 244 |
+
"loss": 0.508,
|
| 245 |
+
"step": 4082
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 13.0,
|
| 249 |
+
"eval_accuracy": 0.7037377983117379,
|
| 250 |
+
"eval_f1": 0.6584786752528541,
|
| 251 |
+
"eval_loss": 0.5470253229141235,
|
| 252 |
+
"eval_precision": 0.6984109009927764,
|
| 253 |
+
"eval_recall": 0.6552206326603638,
|
| 254 |
+
"eval_runtime": 1.4537,
|
| 255 |
+
"eval_samples_per_second": 894.952,
|
| 256 |
+
"eval_steps_per_second": 28.204,
|
| 257 |
+
"step": 4095
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 13.955555555555556,
|
| 261 |
+
"grad_norm": 7.560145378112793,
|
| 262 |
+
"learning_rate": 4.772708335394416e-06,
|
| 263 |
+
"loss": 0.5029,
|
| 264 |
+
"step": 4396
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 14.0,
|
| 268 |
+
"eval_accuracy": 0.713568624726439,
|
| 269 |
+
"eval_f1": 0.6795572312733509,
|
| 270 |
+
"eval_loss": 0.5456140637397766,
|
| 271 |
+
"eval_precision": 0.7034572312241041,
|
| 272 |
+
"eval_recall": 0.6744872926065051,
|
| 273 |
+
"eval_runtime": 1.4534,
|
| 274 |
+
"eval_samples_per_second": 895.17,
|
| 275 |
+
"eval_steps_per_second": 28.211,
|
| 276 |
+
"step": 4410
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 14.952380952380953,
|
| 280 |
+
"grad_norm": 5.061767578125,
|
| 281 |
+
"learning_rate": 4.738871460946384e-06,
|
| 282 |
+
"loss": 0.4971,
|
| 283 |
+
"step": 4710
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 15.0,
|
| 287 |
+
"eval_accuracy": 0.7147323444610414,
|
| 288 |
+
"eval_f1": 0.6712151153220013,
|
| 289 |
+
"eval_loss": 0.5506519675254822,
|
| 290 |
+
"eval_precision": 0.7131289966118773,
|
| 291 |
+
"eval_recall": 0.66683731470523,
|
| 292 |
+
"eval_runtime": 1.4555,
|
| 293 |
+
"eval_samples_per_second": 893.846,
|
| 294 |
+
"eval_steps_per_second": 28.169,
|
| 295 |
+
"step": 4725
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 15.94920634920635,
|
| 299 |
+
"grad_norm": 3.6310696601867676,
|
| 300 |
+
"learning_rate": 4.702825090668624e-06,
|
| 301 |
+
"loss": 0.4883,
|
| 302 |
+
"step": 5024
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 16.0,
|
| 306 |
+
"eval_accuracy": 0.7275853683954563,
|
| 307 |
+
"eval_f1": 0.6800593771748749,
|
| 308 |
+
"eval_loss": 0.5249310731887817,
|
| 309 |
+
"eval_precision": 0.7385652261130415,
|
| 310 |
+
"eval_recall": 0.6753129875296177,
|
| 311 |
+
"eval_runtime": 1.4568,
|
| 312 |
+
"eval_samples_per_second": 893.074,
|
| 313 |
+
"eval_steps_per_second": 28.145,
|
| 314 |
+
"step": 5040
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 16.946031746031746,
|
| 318 |
+
"grad_norm": 3.5576395988464355,
|
| 319 |
+
"learning_rate": 4.664604797976672e-06,
|
| 320 |
+
"loss": 0.4785,
|
| 321 |
+
"step": 5338
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 17.0,
|
| 325 |
+
"eval_accuracy": 0.7470733317122312,
|
| 326 |
+
"eval_f1": 0.7220592428721022,
|
| 327 |
+
"eval_loss": 0.5220229625701904,
|
| 328 |
+
"eval_precision": 0.7393366406970809,
|
| 329 |
+
"eval_recall": 0.7157473225861406,
|
| 330 |
+
"eval_runtime": 1.4557,
|
| 331 |
+
"eval_samples_per_second": 893.714,
|
| 332 |
+
"eval_steps_per_second": 28.165,
|
| 333 |
+
"step": 5355
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 17.942857142857143,
|
| 337 |
+
"grad_norm": 3.7962605953216553,
|
| 338 |
+
"learning_rate": 4.6242483016850204e-06,
|
| 339 |
+
"loss": 0.4601,
|
| 340 |
+
"step": 5652
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 18.0,
|
| 344 |
+
"eval_accuracy": 0.7800743391114044,
|
| 345 |
+
"eval_f1": 0.7542114810572249,
|
| 346 |
+
"eval_loss": 0.47945165634155273,
|
| 347 |
+
"eval_precision": 0.7845629818690574,
|
| 348 |
+
"eval_recall": 0.7447658710776668,
|
| 349 |
+
"eval_runtime": 1.4542,
|
| 350 |
+
"eval_samples_per_second": 894.647,
|
| 351 |
+
"eval_steps_per_second": 28.194,
|
| 352 |
+
"step": 5670
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 18.93968253968254,
|
| 356 |
+
"grad_norm": 6.58827543258667,
|
| 357 |
+
"learning_rate": 4.581795428783211e-06,
|
| 358 |
+
"loss": 0.438,
|
| 359 |
+
"step": 5966
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 19.0,
|
| 363 |
+
"eval_accuracy": 0.8036266370236566,
|
| 364 |
+
"eval_f1": 0.7794844647965122,
|
| 365 |
+
"eval_loss": 0.45785731077194214,
|
| 366 |
+
"eval_precision": 0.8159505749196594,
|
| 367 |
+
"eval_recall": 0.7680218463402948,
|
| 368 |
+
"eval_runtime": 1.4553,
|
| 369 |
+
"eval_samples_per_second": 893.958,
|
| 370 |
+
"eval_steps_per_second": 28.172,
|
| 371 |
+
"step": 5985
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 19.936507936507937,
|
| 375 |
+
"grad_norm": 5.3035383224487305,
|
| 376 |
+
"learning_rate": 4.537288075131401e-06,
|
| 377 |
+
"loss": 0.4133,
|
| 378 |
+
"step": 6280
|
| 379 |
+
},
|
| 380 |
+
{
|
| 381 |
+
"epoch": 20.0,
|
| 382 |
+
"eval_accuracy": 0.8041129676590127,
|
| 383 |
+
"eval_f1": 0.797301642675663,
|
| 384 |
+
"eval_loss": 0.4323329031467438,
|
| 385 |
+
"eval_precision": 0.7942179843107702,
|
| 386 |
+
"eval_recall": 0.8029752318176708,
|
| 387 |
+
"eval_runtime": 1.452,
|
| 388 |
+
"eval_samples_per_second": 895.983,
|
| 389 |
+
"eval_steps_per_second": 28.236,
|
| 390 |
+
"step": 6300
|
| 391 |
+
},
|
| 392 |
+
{
|
| 393 |
+
"epoch": 20.933333333333334,
|
| 394 |
+
"grad_norm": 5.589924335479736,
|
| 395 |
+
"learning_rate": 4.490770164114206e-06,
|
| 396 |
+
"loss": 0.3947,
|
| 397 |
+
"step": 6594
|
| 398 |
+
},
|
| 399 |
+
{
|
| 400 |
+
"epoch": 21.0,
|
| 401 |
+
"eval_accuracy": 0.8210650640914302,
|
| 402 |
+
"eval_f1": 0.8050070699973129,
|
| 403 |
+
"eval_loss": 0.41711243987083435,
|
| 404 |
+
"eval_precision": 0.821898032840692,
|
| 405 |
+
"eval_recall": 0.7966116925782902,
|
| 406 |
+
"eval_runtime": 1.4546,
|
| 407 |
+
"eval_samples_per_second": 894.434,
|
| 408 |
+
"eval_steps_per_second": 28.187,
|
| 409 |
+
"step": 6615
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 21.93015873015873,
|
| 413 |
+
"grad_norm": 4.1770501136779785,
|
| 414 |
+
"learning_rate": 4.442287603293625e-06,
|
| 415 |
+
"loss": 0.3787,
|
| 416 |
+
"step": 6908
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 22.0,
|
| 420 |
+
"eval_accuracy": 0.8151943585646299,
|
| 421 |
+
"eval_f1": 0.807236518842082,
|
| 422 |
+
"eval_loss": 0.4175125062465668,
|
| 423 |
+
"eval_precision": 0.8051094915340469,
|
| 424 |
+
"eval_recall": 0.8099929034678388,
|
| 425 |
+
"eval_runtime": 1.4549,
|
| 426 |
+
"eval_samples_per_second": 894.205,
|
| 427 |
+
"eval_steps_per_second": 28.18,
|
| 428 |
+
"step": 6930
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 22.926984126984127,
|
| 432 |
+
"grad_norm": 6.486811637878418,
|
| 433 |
+
"learning_rate": 4.391888239103818e-06,
|
| 434 |
+
"loss": 0.367,
|
| 435 |
+
"step": 7222
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 23.0,
|
| 439 |
+
"eval_accuracy": 0.8269878764720187,
|
| 440 |
+
"eval_f1": 0.8153345585743379,
|
| 441 |
+
"eval_loss": 0.4084073305130005,
|
| 442 |
+
"eval_precision": 0.8211330177764262,
|
| 443 |
+
"eval_recall": 0.8111882260202181,
|
| 444 |
+
"eval_runtime": 1.4534,
|
| 445 |
+
"eval_samples_per_second": 895.129,
|
| 446 |
+
"eval_steps_per_second": 28.209,
|
| 447 |
+
"step": 7245
|
| 448 |
+
}
|
| 449 |
+
],
|
| 450 |
+
"logging_steps": 314,
|
| 451 |
+
"max_steps": 31500,
|
| 452 |
+
"num_input_tokens_seen": 0,
|
| 453 |
+
"num_train_epochs": 100,
|
| 454 |
+
"save_steps": 500,
|
| 455 |
+
"stateful_callbacks": {
|
| 456 |
+
"EarlyStoppingCallback": {
|
| 457 |
+
"args": {
|
| 458 |
+
"early_stopping_patience": 5,
|
| 459 |
+
"early_stopping_threshold": 0.0
|
| 460 |
+
},
|
| 461 |
+
"attributes": {
|
| 462 |
+
"early_stopping_patience_counter": 0
|
| 463 |
+
}
|
| 464 |
+
},
|
| 465 |
+
"TrainerControl": {
|
| 466 |
+
"args": {
|
| 467 |
+
"should_epoch_stop": false,
|
| 468 |
+
"should_evaluate": false,
|
| 469 |
+
"should_log": false,
|
| 470 |
+
"should_save": true,
|
| 471 |
+
"should_training_stop": false
|
| 472 |
+
},
|
| 473 |
+
"attributes": {}
|
| 474 |
+
}
|
| 475 |
+
},
|
| 476 |
+
"total_flos": 0.0,
|
| 477 |
+
"train_batch_size": 32,
|
| 478 |
+
"trial_name": null,
|
| 479 |
+
"trial_params": null
|
| 480 |
+
}
|