Commit
·
c9f81bd
1
Parent(s):
ca69592
yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor
Browse files- README.md +10 -15
- all_results.json +11 -11
- eval_results.json +6 -6
- pytorch_model.bin +1 -1
- train_results.json +6 -6
- trainer_state.json +167 -524
- training_args.bin +1 -1
README.md
CHANGED
|
@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor](https://huggingface.co/yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
-
- Loss: 0.
|
| 21 |
-
- Accuracy: 0.
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -39,27 +39,22 @@ More information needed
|
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
- learning_rate: 5e-05
|
| 41 |
- train_batch_size: 16
|
| 42 |
-
- eval_batch_size:
|
| 43 |
- seed: 42
|
| 44 |
-
- gradient_accumulation_steps:
|
| 45 |
-
- total_train_batch_size:
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
-
- num_epochs:
|
| 49 |
|
| 50 |
### Training results
|
| 51 |
|
| 52 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 53 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 54 |
-
| 0.
|
| 55 |
-
| 0.
|
| 56 |
-
| 0.
|
| 57 |
-
| 0.
|
| 58 |
-
| 0.0141 | 52.77 | 5000 | 0.2184 | 0.9439 |
|
| 59 |
-
| 0.0106 | 63.32 | 6000 | 0.2401 | 0.9348 |
|
| 60 |
-
| 0.0112 | 73.88 | 7000 | 0.2206 | 0.9493 |
|
| 61 |
-
| 0.0085 | 84.43 | 8000 | 0.1907 | 0.9526 |
|
| 62 |
-
| 0.0079 | 94.99 | 9000 | 0.2052 | 0.9484 |
|
| 63 |
|
| 64 |
|
| 65 |
### Framework versions
|
|
|
|
| 17 |
|
| 18 |
This model is a fine-tuned version of [yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor](https://huggingface.co/yashcode00/wav2vec2-large-xlsr-indian-language-classification-featureExtractor) on an unknown dataset.
|
| 19 |
It achieves the following results on the evaluation set:
|
| 20 |
+
- Loss: 0.1719
|
| 21 |
+
- Accuracy: 0.9554
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
|
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
- learning_rate: 5e-05
|
| 41 |
- train_batch_size: 16
|
| 42 |
+
- eval_batch_size: 16
|
| 43 |
- seed: 42
|
| 44 |
+
- gradient_accumulation_steps: 16
|
| 45 |
+
- total_train_batch_size: 256
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
+
- num_epochs: 90
|
| 49 |
|
| 50 |
### Training results
|
| 51 |
|
| 52 |
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|
| 53 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 54 |
+
| 0.0103 | 21.11 | 1000 | 0.1802 | 0.9501 |
|
| 55 |
+
| 0.009 | 42.22 | 2000 | 0.1717 | 0.9497 |
|
| 56 |
+
| 0.0086 | 63.32 | 3000 | 0.1675 | 0.9546 |
|
| 57 |
+
| 0.0073 | 84.43 | 4000 | 0.1686 | 0.9538 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
|
| 59 |
|
| 60 |
### Framework versions
|
all_results.json
CHANGED
|
@@ -1,15 +1,15 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 2424,
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
-
"total_flos": 3.
|
| 10 |
-
"train_loss": 0.
|
| 11 |
-
"train_runtime":
|
| 12 |
"train_samples": 12120,
|
| 13 |
-
"train_samples_per_second":
|
| 14 |
-
"train_steps_per_second": 0.
|
| 15 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 89.29,
|
| 3 |
+
"eval_accuracy": 0.9554455280303955,
|
| 4 |
+
"eval_loss": 0.1719195693731308,
|
| 5 |
+
"eval_runtime": 50.8715,
|
| 6 |
"eval_samples": 2424,
|
| 7 |
+
"eval_samples_per_second": 47.65,
|
| 8 |
+
"eval_steps_per_second": 2.988,
|
| 9 |
+
"total_flos": 3.2880550437308154e+19,
|
| 10 |
+
"train_loss": 0.00991302564845863,
|
| 11 |
+
"train_runtime": 33902.9866,
|
| 12 |
"train_samples": 12120,
|
| 13 |
+
"train_samples_per_second": 32.174,
|
| 14 |
+
"train_steps_per_second": 0.125
|
| 15 |
}
|
eval_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"eval_accuracy": 0.
|
| 4 |
-
"eval_loss": 0.
|
| 5 |
-
"eval_runtime":
|
| 6 |
"eval_samples": 2424,
|
| 7 |
-
"eval_samples_per_second":
|
| 8 |
-
"eval_steps_per_second":
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 89.29,
|
| 3 |
+
"eval_accuracy": 0.9554455280303955,
|
| 4 |
+
"eval_loss": 0.1719195693731308,
|
| 5 |
+
"eval_runtime": 50.8715,
|
| 6 |
"eval_samples": 2424,
|
| 7 |
+
"eval_samples_per_second": 47.65,
|
| 8 |
+
"eval_steps_per_second": 2.988
|
| 9 |
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 1266146037
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7ceaacefa2250a52d0d79eeb85b6a0da21680d9b9b79e1d64c35a9ab6bd911c1
|
| 3 |
size 1266146037
|
train_results.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
| 1 |
{
|
| 2 |
-
"epoch":
|
| 3 |
-
"total_flos": 3.
|
| 4 |
-
"train_loss": 0.
|
| 5 |
-
"train_runtime":
|
| 6 |
"train_samples": 12120,
|
| 7 |
-
"train_samples_per_second":
|
| 8 |
-
"train_steps_per_second": 0.
|
| 9 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"epoch": 89.29,
|
| 3 |
+
"total_flos": 3.2880550437308154e+19,
|
| 4 |
+
"train_loss": 0.00991302564845863,
|
| 5 |
+
"train_runtime": 33902.9866,
|
| 6 |
"train_samples": 12120,
|
| 7 |
+
"train_samples_per_second": 32.174,
|
| 8 |
+
"train_steps_per_second": 0.125
|
| 9 |
}
|
trainer_state.json
CHANGED
|
@@ -1,682 +1,325 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
-
"epoch":
|
| 5 |
"eval_steps": 1000,
|
| 6 |
-
"global_step":
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
| 11 |
-
{
|
| 12 |
-
"epoch": 1.06,
|
| 13 |
-
"learning_rate": 4.946808510638298e-05,
|
| 14 |
-
"loss": 0.0232,
|
| 15 |
-
"step": 100
|
| 16 |
-
},
|
| 17 |
{
|
| 18 |
"epoch": 2.11,
|
| 19 |
-
"learning_rate": 4.
|
| 20 |
-
"loss": 0.
|
| 21 |
-
"step":
|
| 22 |
-
},
|
| 23 |
-
{
|
| 24 |
-
"epoch": 3.17,
|
| 25 |
-
"learning_rate": 4.840425531914894e-05,
|
| 26 |
-
"loss": 0.0271,
|
| 27 |
-
"step": 300
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 4.22,
|
| 31 |
-
"learning_rate": 4.
|
| 32 |
-
"loss": 0.
|
| 33 |
-
"step":
|
| 34 |
-
},
|
| 35 |
-
{
|
| 36 |
-
"epoch": 5.28,
|
| 37 |
-
"learning_rate": 4.734042553191489e-05,
|
| 38 |
-
"loss": 0.025,
|
| 39 |
-
"step": 500
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 6.33,
|
| 43 |
-
"learning_rate": 4.
|
| 44 |
-
"loss": 0.
|
| 45 |
-
"step":
|
| 46 |
-
},
|
| 47 |
-
{
|
| 48 |
-
"epoch": 7.39,
|
| 49 |
-
"learning_rate": 4.627659574468085e-05,
|
| 50 |
-
"loss": 0.0216,
|
| 51 |
-
"step": 700
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 8.44,
|
| 55 |
-
"learning_rate": 4.
|
| 56 |
-
"loss": 0.
|
| 57 |
-
"step":
|
| 58 |
-
},
|
| 59 |
-
{
|
| 60 |
-
"epoch": 9.5,
|
| 61 |
-
"learning_rate": 4.521808510638298e-05,
|
| 62 |
-
"loss": 0.025,
|
| 63 |
-
"step": 900
|
| 64 |
-
},
|
| 65 |
-
{
|
| 66 |
-
"epoch": 10.55,
|
| 67 |
-
"learning_rate": 4.468617021276596e-05,
|
| 68 |
-
"loss": 0.0213,
|
| 69 |
-
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 10.55,
|
| 73 |
-
"
|
| 74 |
-
"
|
| 75 |
-
"
|
| 76 |
-
"eval_samples_per_second": 49.113,
|
| 77 |
-
"eval_steps_per_second": 6.139,
|
| 78 |
-
"step": 1000
|
| 79 |
-
},
|
| 80 |
-
{
|
| 81 |
-
"epoch": 11.61,
|
| 82 |
-
"learning_rate": 4.415425531914894e-05,
|
| 83 |
-
"loss": 0.0226,
|
| 84 |
-
"step": 1100
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 12.66,
|
| 88 |
-
"learning_rate": 4.
|
| 89 |
-
"loss": 0.
|
| 90 |
-
"step":
|
| 91 |
-
},
|
| 92 |
-
{
|
| 93 |
-
"epoch": 13.72,
|
| 94 |
-
"learning_rate": 4.30904255319149e-05,
|
| 95 |
-
"loss": 0.0196,
|
| 96 |
-
"step": 1300
|
| 97 |
},
|
| 98 |
{
|
| 99 |
"epoch": 14.78,
|
| 100 |
-
"learning_rate": 4.
|
| 101 |
-
"loss": 0.
|
| 102 |
-
"step":
|
| 103 |
-
},
|
| 104 |
-
{
|
| 105 |
-
"epoch": 15.83,
|
| 106 |
-
"learning_rate": 4.2026595744680855e-05,
|
| 107 |
-
"loss": 0.0209,
|
| 108 |
-
"step": 1500
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"epoch": 16.89,
|
| 112 |
-
"learning_rate": 4.
|
| 113 |
-
"loss": 0.
|
| 114 |
-
"step":
|
| 115 |
-
},
|
| 116 |
-
{
|
| 117 |
-
"epoch": 17.94,
|
| 118 |
-
"learning_rate": 4.096276595744681e-05,
|
| 119 |
-
"loss": 0.0193,
|
| 120 |
-
"step": 1700
|
| 121 |
},
|
| 122 |
{
|
| 123 |
"epoch": 19.0,
|
| 124 |
-
"learning_rate":
|
| 125 |
-
"loss": 0.
|
| 126 |
-
"step":
|
| 127 |
-
},
|
| 128 |
-
{
|
| 129 |
-
"epoch": 20.05,
|
| 130 |
-
"learning_rate": 3.989893617021277e-05,
|
| 131 |
-
"loss": 0.0206,
|
| 132 |
-
"step": 1900
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 21.11,
|
| 136 |
-
"learning_rate": 3.
|
| 137 |
-
"loss": 0.
|
| 138 |
-
"step":
|
| 139 |
},
|
| 140 |
{
|
| 141 |
"epoch": 21.11,
|
| 142 |
-
"eval_accuracy": 0.
|
| 143 |
-
"eval_loss": 0.
|
| 144 |
-
"eval_runtime":
|
| 145 |
-
"eval_samples_per_second":
|
| 146 |
-
"eval_steps_per_second":
|
| 147 |
-
"step":
|
| 148 |
-
},
|
| 149 |
-
{
|
| 150 |
-
"epoch": 22.16,
|
| 151 |
-
"learning_rate": 3.8835106382978724e-05,
|
| 152 |
-
"loss": 0.015,
|
| 153 |
-
"step": 2100
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 23.22,
|
| 157 |
-
"learning_rate": 3.
|
| 158 |
-
"loss": 0.
|
| 159 |
-
"step":
|
| 160 |
-
},
|
| 161 |
-
{
|
| 162 |
-
"epoch": 24.27,
|
| 163 |
-
"learning_rate": 3.777127659574468e-05,
|
| 164 |
-
"loss": 0.0159,
|
| 165 |
-
"step": 2300
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 25.33,
|
| 169 |
-
"learning_rate": 3.
|
| 170 |
-
"loss": 0.
|
| 171 |
-
"step":
|
| 172 |
-
},
|
| 173 |
-
{
|
| 174 |
-
"epoch": 26.39,
|
| 175 |
-
"learning_rate": 3.670744680851064e-05,
|
| 176 |
-
"loss": 0.0145,
|
| 177 |
-
"step": 2500
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 27.44,
|
| 181 |
-
"learning_rate": 3.
|
| 182 |
-
"loss": 0.
|
| 183 |
-
"step":
|
| 184 |
-
},
|
| 185 |
-
{
|
| 186 |
-
"epoch": 28.5,
|
| 187 |
-
"learning_rate": 3.56436170212766e-05,
|
| 188 |
-
"loss": 0.0214,
|
| 189 |
-
"step": 2700
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 29.55,
|
| 193 |
-
"learning_rate": 3.
|
| 194 |
-
"loss": 0.
|
| 195 |
-
"step":
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 30.61,
|
| 199 |
-
"learning_rate": 3.457978723404256e-05,
|
| 200 |
-
"loss": 0.0147,
|
| 201 |
-
"step": 2900
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 31.66,
|
| 205 |
-
"learning_rate": 3.4047872340425536e-05,
|
| 206 |
-
"loss": 0.0196,
|
| 207 |
-
"step": 3000
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 31.66,
|
| 211 |
-
"
|
| 212 |
-
"
|
| 213 |
-
"
|
| 214 |
-
"eval_samples_per_second": 48.902,
|
| 215 |
-
"eval_steps_per_second": 6.113,
|
| 216 |
-
"step": 3000
|
| 217 |
-
},
|
| 218 |
-
{
|
| 219 |
-
"epoch": 32.72,
|
| 220 |
-
"learning_rate": 3.351595744680851e-05,
|
| 221 |
-
"loss": 0.0196,
|
| 222 |
-
"step": 3100
|
| 223 |
},
|
| 224 |
{
|
| 225 |
"epoch": 33.77,
|
| 226 |
-
"learning_rate": 3.
|
| 227 |
-
"loss": 0.
|
| 228 |
-
"step":
|
| 229 |
-
},
|
| 230 |
-
{
|
| 231 |
-
"epoch": 34.83,
|
| 232 |
-
"learning_rate": 3.2452127659574474e-05,
|
| 233 |
-
"loss": 0.015,
|
| 234 |
-
"step": 3300
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 35.88,
|
| 238 |
-
"learning_rate":
|
| 239 |
-
"loss": 0.
|
| 240 |
-
"step":
|
| 241 |
-
},
|
| 242 |
-
{
|
| 243 |
-
"epoch": 36.94,
|
| 244 |
-
"learning_rate": 3.1388297872340426e-05,
|
| 245 |
-
"loss": 0.0155,
|
| 246 |
-
"step": 3500
|
| 247 |
},
|
| 248 |
{
|
| 249 |
"epoch": 37.99,
|
| 250 |
-
"learning_rate":
|
| 251 |
-
"loss": 0.
|
| 252 |
-
"step":
|
| 253 |
-
},
|
| 254 |
-
{
|
| 255 |
-
"epoch": 39.05,
|
| 256 |
-
"learning_rate": 3.0329787234042556e-05,
|
| 257 |
-
"loss": 0.0166,
|
| 258 |
-
"step": 3700
|
| 259 |
},
|
| 260 |
{
|
| 261 |
"epoch": 40.11,
|
| 262 |
-
"learning_rate": 2.
|
| 263 |
-
"loss": 0.
|
| 264 |
-
"step":
|
| 265 |
-
},
|
| 266 |
-
{
|
| 267 |
-
"epoch": 41.16,
|
| 268 |
-
"learning_rate": 2.926595744680851e-05,
|
| 269 |
-
"loss": 0.0176,
|
| 270 |
-
"step": 3900
|
| 271 |
},
|
| 272 |
{
|
| 273 |
"epoch": 42.22,
|
| 274 |
-
"learning_rate": 2.
|
| 275 |
-
"loss": 0.
|
| 276 |
-
"step":
|
| 277 |
},
|
| 278 |
{
|
| 279 |
"epoch": 42.22,
|
| 280 |
-
"eval_accuracy": 0.
|
| 281 |
-
"eval_loss": 0.
|
| 282 |
-
"eval_runtime":
|
| 283 |
-
"eval_samples_per_second":
|
| 284 |
-
"eval_steps_per_second":
|
| 285 |
-
"step":
|
| 286 |
-
},
|
| 287 |
-
{
|
| 288 |
-
"epoch": 43.27,
|
| 289 |
-
"learning_rate": 2.820212765957447e-05,
|
| 290 |
-
"loss": 0.0154,
|
| 291 |
-
"step": 4100
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 44.33,
|
| 295 |
-
"learning_rate": 2.
|
| 296 |
-
"loss": 0.
|
| 297 |
-
"step":
|
| 298 |
-
},
|
| 299 |
-
{
|
| 300 |
-
"epoch": 45.38,
|
| 301 |
-
"learning_rate": 2.713829787234043e-05,
|
| 302 |
-
"loss": 0.0121,
|
| 303 |
-
"step": 4300
|
| 304 |
},
|
| 305 |
{
|
| 306 |
"epoch": 46.44,
|
| 307 |
-
"learning_rate": 2.
|
| 308 |
-
"loss": 0.
|
| 309 |
-
"step":
|
| 310 |
-
},
|
| 311 |
-
{
|
| 312 |
-
"epoch": 47.49,
|
| 313 |
-
"learning_rate": 2.607978723404255e-05,
|
| 314 |
-
"loss": 0.0169,
|
| 315 |
-
"step": 4500
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"epoch": 48.55,
|
| 319 |
-
"learning_rate": 2.
|
| 320 |
-
"loss": 0.
|
| 321 |
-
"step":
|
| 322 |
-
},
|
| 323 |
-
{
|
| 324 |
-
"epoch": 49.6,
|
| 325 |
-
"learning_rate": 2.5015957446808512e-05,
|
| 326 |
-
"loss": 0.0147,
|
| 327 |
-
"step": 4700
|
| 328 |
},
|
| 329 |
{
|
| 330 |
"epoch": 50.66,
|
| 331 |
-
"learning_rate": 2.
|
| 332 |
-
"loss": 0.
|
| 333 |
-
"step":
|
| 334 |
-
},
|
| 335 |
-
{
|
| 336 |
-
"epoch": 51.72,
|
| 337 |
-
"learning_rate": 2.395212765957447e-05,
|
| 338 |
-
"loss": 0.0113,
|
| 339 |
-
"step": 4900
|
| 340 |
-
},
|
| 341 |
-
{
|
| 342 |
-
"epoch": 52.77,
|
| 343 |
-
"learning_rate": 2.3420212765957447e-05,
|
| 344 |
-
"loss": 0.0141,
|
| 345 |
-
"step": 5000
|
| 346 |
},
|
| 347 |
{
|
| 348 |
"epoch": 52.77,
|
| 349 |
-
"
|
| 350 |
-
"
|
| 351 |
-
"
|
| 352 |
-
"eval_samples_per_second": 48.424,
|
| 353 |
-
"eval_steps_per_second": 6.053,
|
| 354 |
-
"step": 5000
|
| 355 |
-
},
|
| 356 |
-
{
|
| 357 |
-
"epoch": 53.83,
|
| 358 |
-
"learning_rate": 2.288829787234043e-05,
|
| 359 |
-
"loss": 0.011,
|
| 360 |
-
"step": 5100
|
| 361 |
},
|
| 362 |
{
|
| 363 |
"epoch": 54.88,
|
| 364 |
-
"learning_rate":
|
| 365 |
-
"loss": 0.
|
| 366 |
-
"step":
|
| 367 |
-
},
|
| 368 |
-
{
|
| 369 |
-
"epoch": 55.94,
|
| 370 |
-
"learning_rate": 2.1824468085106384e-05,
|
| 371 |
-
"loss": 0.0101,
|
| 372 |
-
"step": 5300
|
| 373 |
},
|
| 374 |
{
|
| 375 |
"epoch": 56.99,
|
| 376 |
-
"learning_rate":
|
| 377 |
-
"loss": 0.
|
| 378 |
-
"step":
|
| 379 |
-
},
|
| 380 |
-
{
|
| 381 |
-
"epoch": 58.05,
|
| 382 |
-
"learning_rate": 2.0760638297872343e-05,
|
| 383 |
-
"loss": 0.0105,
|
| 384 |
-
"step": 5500
|
| 385 |
},
|
| 386 |
{
|
| 387 |
"epoch": 59.1,
|
| 388 |
-
"learning_rate":
|
| 389 |
-
"loss": 0.
|
| 390 |
-
"step":
|
| 391 |
-
},
|
| 392 |
-
{
|
| 393 |
-
"epoch": 60.16,
|
| 394 |
-
"learning_rate": 1.9696808510638298e-05,
|
| 395 |
-
"loss": 0.0089,
|
| 396 |
-
"step": 5700
|
| 397 |
},
|
| 398 |
{
|
| 399 |
"epoch": 61.21,
|
| 400 |
-
"learning_rate": 1.
|
| 401 |
-
"loss": 0.
|
| 402 |
-
"step":
|
| 403 |
-
},
|
| 404 |
-
{
|
| 405 |
-
"epoch": 62.27,
|
| 406 |
-
"learning_rate": 1.8632978723404256e-05,
|
| 407 |
-
"loss": 0.0118,
|
| 408 |
-
"step": 5900
|
| 409 |
},
|
| 410 |
{
|
| 411 |
"epoch": 63.32,
|
| 412 |
-
"learning_rate": 1.
|
| 413 |
-
"loss": 0.
|
| 414 |
-
"step":
|
| 415 |
},
|
| 416 |
{
|
| 417 |
"epoch": 63.32,
|
| 418 |
-
"eval_accuracy": 0.
|
| 419 |
-
"eval_loss": 0.
|
| 420 |
-
"eval_runtime":
|
| 421 |
-
"eval_samples_per_second":
|
| 422 |
-
"eval_steps_per_second":
|
| 423 |
-
"step":
|
| 424 |
-
},
|
| 425 |
-
{
|
| 426 |
-
"epoch": 64.38,
|
| 427 |
-
"learning_rate": 1.7569148936170214e-05,
|
| 428 |
-
"loss": 0.0104,
|
| 429 |
-
"step": 6100
|
| 430 |
},
|
| 431 |
{
|
| 432 |
"epoch": 65.44,
|
| 433 |
-
"learning_rate": 1.
|
| 434 |
-
"loss": 0.
|
| 435 |
-
"step":
|
| 436 |
-
},
|
| 437 |
-
{
|
| 438 |
-
"epoch": 66.49,
|
| 439 |
-
"learning_rate": 1.6505319148936173e-05,
|
| 440 |
-
"loss": 0.0105,
|
| 441 |
-
"step": 6300
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"epoch": 67.55,
|
| 445 |
-
"learning_rate": 1.
|
| 446 |
-
"loss": 0.
|
| 447 |
-
"step":
|
| 448 |
-
},
|
| 449 |
-
{
|
| 450 |
-
"epoch": 68.6,
|
| 451 |
-
"learning_rate": 1.5441489361702128e-05,
|
| 452 |
-
"loss": 0.0088,
|
| 453 |
-
"step": 6500
|
| 454 |
},
|
| 455 |
{
|
| 456 |
"epoch": 69.66,
|
| 457 |
-
"learning_rate": 1.
|
| 458 |
-
"loss": 0.
|
| 459 |
-
"step":
|
| 460 |
-
},
|
| 461 |
-
{
|
| 462 |
-
"epoch": 70.71,
|
| 463 |
-
"learning_rate": 1.4377659574468086e-05,
|
| 464 |
-
"loss": 0.0081,
|
| 465 |
-
"step": 6700
|
| 466 |
},
|
| 467 |
{
|
| 468 |
"epoch": 71.77,
|
| 469 |
-
"learning_rate":
|
| 470 |
-
"loss": 0.0096,
|
| 471 |
-
"step": 6800
|
| 472 |
-
},
|
| 473 |
-
{
|
| 474 |
-
"epoch": 72.82,
|
| 475 |
-
"learning_rate": 1.3313829787234045e-05,
|
| 476 |
"loss": 0.0074,
|
| 477 |
-
"step":
|
| 478 |
-
},
|
| 479 |
-
{
|
| 480 |
-
"epoch": 73.88,
|
| 481 |
-
"learning_rate": 1.2781914893617022e-05,
|
| 482 |
-
"loss": 0.0112,
|
| 483 |
-
"step": 7000
|
| 484 |
},
|
| 485 |
{
|
| 486 |
"epoch": 73.88,
|
| 487 |
-
"
|
| 488 |
-
"
|
| 489 |
-
"
|
| 490 |
-
"eval_samples_per_second": 48.555,
|
| 491 |
-
"eval_steps_per_second": 6.069,
|
| 492 |
-
"step": 7000
|
| 493 |
-
},
|
| 494 |
-
{
|
| 495 |
-
"epoch": 74.93,
|
| 496 |
-
"learning_rate": 1.225e-05,
|
| 497 |
-
"loss": 0.0097,
|
| 498 |
-
"step": 7100
|
| 499 |
},
|
| 500 |
{
|
| 501 |
"epoch": 75.99,
|
| 502 |
-
"learning_rate":
|
| 503 |
-
"loss": 0.
|
| 504 |
-
"step":
|
| 505 |
-
},
|
| 506 |
-
{
|
| 507 |
-
"epoch": 77.04,
|
| 508 |
-
"learning_rate": 1.1186170212765958e-05,
|
| 509 |
-
"loss": 0.0103,
|
| 510 |
-
"step": 7300
|
| 511 |
},
|
| 512 |
{
|
| 513 |
"epoch": 78.1,
|
| 514 |
-
"learning_rate":
|
| 515 |
-
"loss": 0.
|
| 516 |
-
"step":
|
| 517 |
-
},
|
| 518 |
-
{
|
| 519 |
-
"epoch": 79.16,
|
| 520 |
-
"learning_rate": 1.0122340425531915e-05,
|
| 521 |
-
"loss": 0.0085,
|
| 522 |
-
"step": 7500
|
| 523 |
},
|
| 524 |
{
|
| 525 |
"epoch": 80.21,
|
| 526 |
-
"learning_rate":
|
| 527 |
-
"loss": 0.
|
| 528 |
-
"step":
|
| 529 |
-
},
|
| 530 |
-
{
|
| 531 |
-
"epoch": 81.27,
|
| 532 |
-
"learning_rate": 9.058510638297871e-06,
|
| 533 |
-
"loss": 0.0102,
|
| 534 |
-
"step": 7700
|
| 535 |
},
|
| 536 |
{
|
| 537 |
"epoch": 82.32,
|
| 538 |
-
"learning_rate":
|
| 539 |
"loss": 0.0072,
|
| 540 |
-
"step":
|
| 541 |
-
},
|
| 542 |
-
{
|
| 543 |
-
"epoch": 83.38,
|
| 544 |
-
"learning_rate": 7.994680851063832e-06,
|
| 545 |
-
"loss": 0.0076,
|
| 546 |
-
"step": 7900
|
| 547 |
},
|
| 548 |
{
|
| 549 |
"epoch": 84.43,
|
| 550 |
-
"learning_rate":
|
| 551 |
-
"loss": 0.
|
| 552 |
-
"step":
|
| 553 |
},
|
| 554 |
{
|
| 555 |
"epoch": 84.43,
|
| 556 |
-
"eval_accuracy": 0.
|
| 557 |
-
"eval_loss": 0.
|
| 558 |
-
"eval_runtime":
|
| 559 |
-
"eval_samples_per_second":
|
| 560 |
-
"eval_steps_per_second":
|
| 561 |
-
"step":
|
| 562 |
-
},
|
| 563 |
-
{
|
| 564 |
-
"epoch": 85.49,
|
| 565 |
-
"learning_rate": 6.930851063829788e-06,
|
| 566 |
-
"loss": 0.0082,
|
| 567 |
-
"step": 8100
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"epoch": 86.54,
|
| 571 |
-
"learning_rate":
|
| 572 |
-
"loss": 0.
|
| 573 |
-
"step":
|
| 574 |
-
},
|
| 575 |
-
{
|
| 576 |
-
"epoch": 87.6,
|
| 577 |
-
"learning_rate": 5.867021276595745e-06,
|
| 578 |
-
"loss": 0.0093,
|
| 579 |
-
"step": 8300
|
| 580 |
},
|
| 581 |
{
|
| 582 |
"epoch": 88.65,
|
| 583 |
-
"learning_rate":
|
| 584 |
-
"loss": 0.
|
| 585 |
-
"step":
|
| 586 |
-
},
|
| 587 |
-
{
|
| 588 |
-
"epoch": 89.71,
|
| 589 |
-
"learning_rate": 4.8031914893617025e-06,
|
| 590 |
-
"loss": 0.0066,
|
| 591 |
-
"step": 8500
|
| 592 |
-
},
|
| 593 |
-
{
|
| 594 |
-
"epoch": 90.77,
|
| 595 |
-
"learning_rate": 4.271276595744681e-06,
|
| 596 |
-
"loss": 0.007,
|
| 597 |
-
"step": 8600
|
| 598 |
-
},
|
| 599 |
-
{
|
| 600 |
-
"epoch": 91.82,
|
| 601 |
-
"learning_rate": 3.7393617021276596e-06,
|
| 602 |
-
"loss": 0.0079,
|
| 603 |
-
"step": 8700
|
| 604 |
-
},
|
| 605 |
-
{
|
| 606 |
-
"epoch": 92.88,
|
| 607 |
-
"learning_rate": 3.2074468085106384e-06,
|
| 608 |
-
"loss": 0.0075,
|
| 609 |
-
"step": 8800
|
| 610 |
-
},
|
| 611 |
-
{
|
| 612 |
-
"epoch": 93.93,
|
| 613 |
-
"learning_rate": 2.6808510638297874e-06,
|
| 614 |
-
"loss": 0.008,
|
| 615 |
-
"step": 8900
|
| 616 |
-
},
|
| 617 |
-
{
|
| 618 |
-
"epoch": 94.99,
|
| 619 |
-
"learning_rate": 2.148936170212766e-06,
|
| 620 |
-
"loss": 0.0079,
|
| 621 |
-
"step": 9000
|
| 622 |
-
},
|
| 623 |
-
{
|
| 624 |
-
"epoch": 94.99,
|
| 625 |
-
"eval_accuracy": 0.9484323263168335,
|
| 626 |
-
"eval_loss": 0.20523911714553833,
|
| 627 |
-
"eval_runtime": 49.881,
|
| 628 |
-
"eval_samples_per_second": 48.596,
|
| 629 |
-
"eval_steps_per_second": 6.074,
|
| 630 |
-
"step": 9000
|
| 631 |
-
},
|
| 632 |
-
{
|
| 633 |
-
"epoch": 96.04,
|
| 634 |
-
"learning_rate": 1.6170212765957448e-06,
|
| 635 |
-
"loss": 0.0061,
|
| 636 |
-
"step": 9100
|
| 637 |
-
},
|
| 638 |
-
{
|
| 639 |
-
"epoch": 97.1,
|
| 640 |
-
"learning_rate": 1.0851063829787236e-06,
|
| 641 |
-
"loss": 0.0088,
|
| 642 |
-
"step": 9200
|
| 643 |
-
},
|
| 644 |
-
{
|
| 645 |
-
"epoch": 98.15,
|
| 646 |
-
"learning_rate": 5.531914893617021e-07,
|
| 647 |
-
"loss": 0.0077,
|
| 648 |
-
"step": 9300
|
| 649 |
-
},
|
| 650 |
-
{
|
| 651 |
-
"epoch": 99.21,
|
| 652 |
-
"learning_rate": 2.1276595744680853e-08,
|
| 653 |
-
"loss": 0.0077,
|
| 654 |
-
"step": 9400
|
| 655 |
},
|
| 656 |
{
|
| 657 |
-
"epoch":
|
| 658 |
-
"step":
|
| 659 |
-
"total_flos": 3.
|
| 660 |
-
"train_loss": 0.
|
| 661 |
-
"train_runtime":
|
| 662 |
-
"train_samples_per_second":
|
| 663 |
-
"train_steps_per_second": 0.
|
| 664 |
},
|
| 665 |
{
|
| 666 |
-
"epoch":
|
| 667 |
-
"eval_accuracy": 0.
|
| 668 |
-
"eval_loss": 0.
|
| 669 |
-
"eval_runtime":
|
| 670 |
-
"eval_samples_per_second":
|
| 671 |
-
"eval_steps_per_second":
|
| 672 |
-
"step":
|
| 673 |
}
|
| 674 |
],
|
| 675 |
"logging_steps": 100,
|
| 676 |
-
"max_steps":
|
| 677 |
-
"num_train_epochs":
|
| 678 |
"save_steps": 2000,
|
| 679 |
-
"total_flos": 3.
|
| 680 |
"trial_name": null,
|
| 681 |
"trial_params": null
|
| 682 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 89.28759894459102,
|
| 5 |
"eval_steps": 1000,
|
| 6 |
+
"global_step": 4230,
|
| 7 |
"is_hyper_param_search": false,
|
| 8 |
"is_local_process_zero": true,
|
| 9 |
"is_world_process_zero": true,
|
| 10 |
"log_history": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
{
|
| 12 |
"epoch": 2.11,
|
| 13 |
+
"learning_rate": 4.8817966903073283e-05,
|
| 14 |
+
"loss": 0.0118,
|
| 15 |
+
"step": 100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
},
|
| 17 |
{
|
| 18 |
"epoch": 4.22,
|
| 19 |
+
"learning_rate": 4.763593380614658e-05,
|
| 20 |
+
"loss": 0.0156,
|
| 21 |
+
"step": 200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
},
|
| 23 |
{
|
| 24 |
"epoch": 6.33,
|
| 25 |
+
"learning_rate": 4.645390070921986e-05,
|
| 26 |
+
"loss": 0.0122,
|
| 27 |
+
"step": 300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"epoch": 8.44,
|
| 31 |
+
"learning_rate": 4.527186761229315e-05,
|
| 32 |
+
"loss": 0.0148,
|
| 33 |
+
"step": 400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
},
|
| 35 |
{
|
| 36 |
"epoch": 10.55,
|
| 37 |
+
"learning_rate": 4.4089834515366435e-05,
|
| 38 |
+
"loss": 0.0114,
|
| 39 |
+
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 40 |
},
|
| 41 |
{
|
| 42 |
"epoch": 12.66,
|
| 43 |
+
"learning_rate": 4.2907801418439716e-05,
|
| 44 |
+
"loss": 0.0143,
|
| 45 |
+
"step": 600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"epoch": 14.78,
|
| 49 |
+
"learning_rate": 4.1725768321513004e-05,
|
| 50 |
+
"loss": 0.0149,
|
| 51 |
+
"step": 700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
},
|
| 53 |
{
|
| 54 |
"epoch": 16.89,
|
| 55 |
+
"learning_rate": 4.0543735224586285e-05,
|
| 56 |
+
"loss": 0.0139,
|
| 57 |
+
"step": 800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
},
|
| 59 |
{
|
| 60 |
"epoch": 19.0,
|
| 61 |
+
"learning_rate": 3.936170212765958e-05,
|
| 62 |
+
"loss": 0.0125,
|
| 63 |
+
"step": 900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"epoch": 21.11,
|
| 67 |
+
"learning_rate": 3.817966903073286e-05,
|
| 68 |
+
"loss": 0.0103,
|
| 69 |
+
"step": 1000
|
| 70 |
},
|
| 71 |
{
|
| 72 |
"epoch": 21.11,
|
| 73 |
+
"eval_accuracy": 0.9500824809074402,
|
| 74 |
+
"eval_loss": 0.18024244904518127,
|
| 75 |
+
"eval_runtime": 47.0038,
|
| 76 |
+
"eval_samples_per_second": 51.57,
|
| 77 |
+
"eval_steps_per_second": 3.234,
|
| 78 |
+
"step": 1000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
},
|
| 80 |
{
|
| 81 |
"epoch": 23.22,
|
| 82 |
+
"learning_rate": 3.699763593380615e-05,
|
| 83 |
+
"loss": 0.0111,
|
| 84 |
+
"step": 1100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 85 |
},
|
| 86 |
{
|
| 87 |
"epoch": 25.33,
|
| 88 |
+
"learning_rate": 3.5815602836879437e-05,
|
| 89 |
+
"loss": 0.0093,
|
| 90 |
+
"step": 1200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"epoch": 27.44,
|
| 94 |
+
"learning_rate": 3.463356973995272e-05,
|
| 95 |
+
"loss": 0.0109,
|
| 96 |
+
"step": 1300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 97 |
},
|
| 98 |
{
|
| 99 |
"epoch": 29.55,
|
| 100 |
+
"learning_rate": 3.3451536643026005e-05,
|
| 101 |
+
"loss": 0.0102,
|
| 102 |
+
"step": 1400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 103 |
},
|
| 104 |
{
|
| 105 |
"epoch": 31.66,
|
| 106 |
+
"learning_rate": 3.226950354609929e-05,
|
| 107 |
+
"loss": 0.012,
|
| 108 |
+
"step": 1500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"epoch": 33.77,
|
| 112 |
+
"learning_rate": 3.108747044917258e-05,
|
| 113 |
+
"loss": 0.0116,
|
| 114 |
+
"step": 1600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 115 |
},
|
| 116 |
{
|
| 117 |
"epoch": 35.88,
|
| 118 |
+
"learning_rate": 2.9905437352245862e-05,
|
| 119 |
+
"loss": 0.0145,
|
| 120 |
+
"step": 1700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 121 |
},
|
| 122 |
{
|
| 123 |
"epoch": 37.99,
|
| 124 |
+
"learning_rate": 2.8723404255319154e-05,
|
| 125 |
+
"loss": 0.011,
|
| 126 |
+
"step": 1800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 127 |
},
|
| 128 |
{
|
| 129 |
"epoch": 40.11,
|
| 130 |
+
"learning_rate": 2.7541371158392438e-05,
|
| 131 |
+
"loss": 0.0108,
|
| 132 |
+
"step": 1900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 133 |
},
|
| 134 |
{
|
| 135 |
"epoch": 42.22,
|
| 136 |
+
"learning_rate": 2.6359338061465723e-05,
|
| 137 |
+
"loss": 0.009,
|
| 138 |
+
"step": 2000
|
| 139 |
},
|
| 140 |
{
|
| 141 |
"epoch": 42.22,
|
| 142 |
+
"eval_accuracy": 0.9496699571609497,
|
| 143 |
+
"eval_loss": 0.1716560274362564,
|
| 144 |
+
"eval_runtime": 51.12,
|
| 145 |
+
"eval_samples_per_second": 47.418,
|
| 146 |
+
"eval_steps_per_second": 2.973,
|
| 147 |
+
"step": 2000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 148 |
},
|
| 149 |
{
|
| 150 |
"epoch": 44.33,
|
| 151 |
+
"learning_rate": 2.5177304964539007e-05,
|
| 152 |
+
"loss": 0.0101,
|
| 153 |
+
"step": 2100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"epoch": 46.44,
|
| 157 |
+
"learning_rate": 2.3995271867612295e-05,
|
| 158 |
+
"loss": 0.0087,
|
| 159 |
+
"step": 2200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 160 |
},
|
| 161 |
{
|
| 162 |
"epoch": 48.55,
|
| 163 |
+
"learning_rate": 2.281323877068558e-05,
|
| 164 |
+
"loss": 0.0114,
|
| 165 |
+
"step": 2300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 166 |
},
|
| 167 |
{
|
| 168 |
"epoch": 50.66,
|
| 169 |
+
"learning_rate": 2.1631205673758867e-05,
|
| 170 |
+
"loss": 0.0076,
|
| 171 |
+
"step": 2400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"epoch": 52.77,
|
| 175 |
+
"learning_rate": 2.0449172576832152e-05,
|
| 176 |
+
"loss": 0.0088,
|
| 177 |
+
"step": 2500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 178 |
},
|
| 179 |
{
|
| 180 |
"epoch": 54.88,
|
| 181 |
+
"learning_rate": 1.926713947990544e-05,
|
| 182 |
+
"loss": 0.0084,
|
| 183 |
+
"step": 2600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 184 |
},
|
| 185 |
{
|
| 186 |
"epoch": 56.99,
|
| 187 |
+
"learning_rate": 1.8085106382978724e-05,
|
| 188 |
+
"loss": 0.0095,
|
| 189 |
+
"step": 2700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"epoch": 59.1,
|
| 193 |
+
"learning_rate": 1.690307328605201e-05,
|
| 194 |
+
"loss": 0.0075,
|
| 195 |
+
"step": 2800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 196 |
},
|
| 197 |
{
|
| 198 |
"epoch": 61.21,
|
| 199 |
+
"learning_rate": 1.5721040189125296e-05,
|
| 200 |
+
"loss": 0.0097,
|
| 201 |
+
"step": 2900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 202 |
},
|
| 203 |
{
|
| 204 |
"epoch": 63.32,
|
| 205 |
+
"learning_rate": 1.4539007092198581e-05,
|
| 206 |
+
"loss": 0.0086,
|
| 207 |
+
"step": 3000
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"epoch": 63.32,
|
| 211 |
+
"eval_accuracy": 0.9546204805374146,
|
| 212 |
+
"eval_loss": 0.16754871606826782,
|
| 213 |
+
"eval_runtime": 52.1005,
|
| 214 |
+
"eval_samples_per_second": 46.525,
|
| 215 |
+
"eval_steps_per_second": 2.917,
|
| 216 |
+
"step": 3000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"epoch": 65.44,
|
| 220 |
+
"learning_rate": 1.3356973995271869e-05,
|
| 221 |
+
"loss": 0.0079,
|
| 222 |
+
"step": 3100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 223 |
},
|
| 224 |
{
|
| 225 |
"epoch": 67.55,
|
| 226 |
+
"learning_rate": 1.2174940898345153e-05,
|
| 227 |
+
"loss": 0.0076,
|
| 228 |
+
"step": 3200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 229 |
},
|
| 230 |
{
|
| 231 |
"epoch": 69.66,
|
| 232 |
+
"learning_rate": 1.0992907801418441e-05,
|
| 233 |
+
"loss": 0.0072,
|
| 234 |
+
"step": 3300
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"epoch": 71.77,
|
| 238 |
+
"learning_rate": 9.810874704491727e-06,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 239 |
"loss": 0.0074,
|
| 240 |
+
"step": 3400
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 241 |
},
|
| 242 |
{
|
| 243 |
"epoch": 73.88,
|
| 244 |
+
"learning_rate": 8.628841607565012e-06,
|
| 245 |
+
"loss": 0.0076,
|
| 246 |
+
"step": 3500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 247 |
},
|
| 248 |
{
|
| 249 |
"epoch": 75.99,
|
| 250 |
+
"learning_rate": 7.446808510638298e-06,
|
| 251 |
+
"loss": 0.0069,
|
| 252 |
+
"step": 3600
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 253 |
},
|
| 254 |
{
|
| 255 |
"epoch": 78.1,
|
| 256 |
+
"learning_rate": 6.264775413711583e-06,
|
| 257 |
+
"loss": 0.0068,
|
| 258 |
+
"step": 3700
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 259 |
},
|
| 260 |
{
|
| 261 |
"epoch": 80.21,
|
| 262 |
+
"learning_rate": 5.08274231678487e-06,
|
| 263 |
+
"loss": 0.007,
|
| 264 |
+
"step": 3800
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 265 |
},
|
| 266 |
{
|
| 267 |
"epoch": 82.32,
|
| 268 |
+
"learning_rate": 3.9007092198581565e-06,
|
| 269 |
"loss": 0.0072,
|
| 270 |
+
"step": 3900
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 271 |
},
|
| 272 |
{
|
| 273 |
"epoch": 84.43,
|
| 274 |
+
"learning_rate": 2.7186761229314422e-06,
|
| 275 |
+
"loss": 0.0073,
|
| 276 |
+
"step": 4000
|
| 277 |
},
|
| 278 |
{
|
| 279 |
"epoch": 84.43,
|
| 280 |
+
"eval_accuracy": 0.9537953734397888,
|
| 281 |
+
"eval_loss": 0.16863200068473816,
|
| 282 |
+
"eval_runtime": 51.419,
|
| 283 |
+
"eval_samples_per_second": 47.142,
|
| 284 |
+
"eval_steps_per_second": 2.956,
|
| 285 |
+
"step": 4000
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 286 |
},
|
| 287 |
{
|
| 288 |
"epoch": 86.54,
|
| 289 |
+
"learning_rate": 1.5484633569739953e-06,
|
| 290 |
+
"loss": 0.0065,
|
| 291 |
+
"step": 4100
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 292 |
},
|
| 293 |
{
|
| 294 |
"epoch": 88.65,
|
| 295 |
+
"learning_rate": 3.6643026004728135e-07,
|
| 296 |
+
"loss": 0.0058,
|
| 297 |
+
"step": 4200
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 298 |
},
|
| 299 |
{
|
| 300 |
+
"epoch": 89.29,
|
| 301 |
+
"step": 4230,
|
| 302 |
+
"total_flos": 3.2880550437308154e+19,
|
| 303 |
+
"train_loss": 0.00991302564845863,
|
| 304 |
+
"train_runtime": 33902.9866,
|
| 305 |
+
"train_samples_per_second": 32.174,
|
| 306 |
+
"train_steps_per_second": 0.125
|
| 307 |
},
|
| 308 |
{
|
| 309 |
+
"epoch": 89.29,
|
| 310 |
+
"eval_accuracy": 0.9554455280303955,
|
| 311 |
+
"eval_loss": 0.1719195693731308,
|
| 312 |
+
"eval_runtime": 50.8715,
|
| 313 |
+
"eval_samples_per_second": 47.65,
|
| 314 |
+
"eval_steps_per_second": 2.988,
|
| 315 |
+
"step": 4230
|
| 316 |
}
|
| 317 |
],
|
| 318 |
"logging_steps": 100,
|
| 319 |
+
"max_steps": 4230,
|
| 320 |
+
"num_train_epochs": 90,
|
| 321 |
"save_steps": 2000,
|
| 322 |
+
"total_flos": 3.2880550437308154e+19,
|
| 323 |
"trial_name": null,
|
| 324 |
"trial_params": null
|
| 325 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 4155
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:afafbb8d751fc33fbf51d1298497b2fbbc858aa6e7af5c8ee9fc1310c74fcc53
|
| 3 |
size 4155
|