Upload with huggingface_hub
Browse files- optimizer.pt +1 -1
- pytorch_model.bin +1 -1
- rng_state.pth +1 -1
- scheduler.pt +1 -1
- trainer_state.json +321 -21
- training_args.bin +1 -1
optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 535706209
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab1197d12a90025fd234d66a5889a9f66fc50238fad434021bcce3243ff6c60b
|
| 3 |
size 535706209
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 267857393
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e0245e6591a24422073905214bcbe73098b2f73a6d8ca6a40b32c136ffaf21e
|
| 3 |
size 267857393
|
rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14503
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f95227265e995dfc259b58c87461d534728d8071eae99155ac7f7b6ee69e0c1d
|
| 3 |
size 14503
|
scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 623
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65580af7c9779932b2720e89a7f2f04027fc4dde0aec900b7eb7585247a6d196
|
| 3 |
size 623
|
trainer_state.json
CHANGED
|
@@ -1,46 +1,346 @@
|
|
| 1 |
{
|
| 2 |
-
"best_metric": 0.
|
| 3 |
-
"best_model_checkpoint": "./results/checkpoint-
|
| 4 |
-
"epoch":
|
| 5 |
-
"global_step":
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 1.0,
|
| 12 |
-
"learning_rate": 1.
|
| 13 |
-
"loss":
|
| 14 |
"step": 200
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 1.25,
|
| 18 |
-
"eval_accuracy": 0.
|
| 19 |
-
"eval_loss": 0.
|
| 20 |
-
"eval_runtime":
|
| 21 |
-
"eval_samples_per_second":
|
| 22 |
-
"eval_steps_per_second": 3.
|
| 23 |
"step": 250
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 2.0,
|
| 27 |
-
"learning_rate": 1.
|
| 28 |
-
"loss": 0.
|
| 29 |
"step": 400
|
| 30 |
},
|
| 31 |
{
|
| 32 |
"epoch": 2.5,
|
| 33 |
-
"eval_accuracy": 0.
|
| 34 |
-
"eval_loss": 0.
|
| 35 |
-
"eval_runtime":
|
| 36 |
-
"eval_samples_per_second":
|
| 37 |
-
"eval_steps_per_second": 3.
|
| 38 |
"step": 500
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
}
|
| 40 |
],
|
| 41 |
-
"max_steps":
|
| 42 |
-
"num_train_epochs":
|
| 43 |
-
"total_flos":
|
| 44 |
"trial_name": null,
|
| 45 |
"trial_params": null
|
| 46 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"best_metric": 0.64125,
|
| 3 |
+
"best_model_checkpoint": "./results/checkpoint-5000",
|
| 4 |
+
"epoch": 25.0,
|
| 5 |
+
"global_step": 5000,
|
| 6 |
"is_hyper_param_search": false,
|
| 7 |
"is_local_process_zero": true,
|
| 8 |
"is_world_process_zero": true,
|
| 9 |
"log_history": [
|
| 10 |
{
|
| 11 |
"epoch": 1.0,
|
| 12 |
+
"learning_rate": 1.9333333333333333e-05,
|
| 13 |
+
"loss": 1.0006,
|
| 14 |
"step": 200
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 1.25,
|
| 18 |
+
"eval_accuracy": 0.58125,
|
| 19 |
+
"eval_loss": 0.8775196075439453,
|
| 20 |
+
"eval_runtime": 16.4637,
|
| 21 |
+
"eval_samples_per_second": 48.592,
|
| 22 |
+
"eval_steps_per_second": 3.037,
|
| 23 |
"step": 250
|
| 24 |
},
|
| 25 |
{
|
| 26 |
"epoch": 2.0,
|
| 27 |
+
"learning_rate": 1.866666666666667e-05,
|
| 28 |
+
"loss": 0.8654,
|
| 29 |
"step": 400
|
| 30 |
},
|
| 31 |
{
|
| 32 |
"epoch": 2.5,
|
| 33 |
+
"eval_accuracy": 0.6275,
|
| 34 |
+
"eval_loss": 0.8210939168930054,
|
| 35 |
+
"eval_runtime": 15.3938,
|
| 36 |
+
"eval_samples_per_second": 51.969,
|
| 37 |
+
"eval_steps_per_second": 3.248,
|
| 38 |
"step": 500
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 3.0,
|
| 42 |
+
"learning_rate": 1.8e-05,
|
| 43 |
+
"loss": 0.7041,
|
| 44 |
+
"step": 600
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 3.75,
|
| 48 |
+
"eval_accuracy": 0.62125,
|
| 49 |
+
"eval_loss": 0.9076805710792542,
|
| 50 |
+
"eval_runtime": 15.4386,
|
| 51 |
+
"eval_samples_per_second": 51.818,
|
| 52 |
+
"eval_steps_per_second": 3.239,
|
| 53 |
+
"step": 750
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 4.0,
|
| 57 |
+
"learning_rate": 1.7333333333333336e-05,
|
| 58 |
+
"loss": 0.5119,
|
| 59 |
+
"step": 800
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 5.0,
|
| 63 |
+
"learning_rate": 1.6666666666666667e-05,
|
| 64 |
+
"loss": 0.3314,
|
| 65 |
+
"step": 1000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 5.0,
|
| 69 |
+
"eval_accuracy": 0.63125,
|
| 70 |
+
"eval_loss": 1.0315085649490356,
|
| 71 |
+
"eval_runtime": 15.309,
|
| 72 |
+
"eval_samples_per_second": 52.257,
|
| 73 |
+
"eval_steps_per_second": 3.266,
|
| 74 |
+
"step": 1000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 6.0,
|
| 78 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 79 |
+
"loss": 0.2029,
|
| 80 |
+
"step": 1200
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 6.25,
|
| 84 |
+
"eval_accuracy": 0.64,
|
| 85 |
+
"eval_loss": 1.349707841873169,
|
| 86 |
+
"eval_runtime": 15.3657,
|
| 87 |
+
"eval_samples_per_second": 52.064,
|
| 88 |
+
"eval_steps_per_second": 3.254,
|
| 89 |
+
"step": 1250
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 7.0,
|
| 93 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 94 |
+
"loss": 0.1332,
|
| 95 |
+
"step": 1400
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 7.5,
|
| 99 |
+
"eval_accuracy": 0.56375,
|
| 100 |
+
"eval_loss": 2.2334282398223877,
|
| 101 |
+
"eval_runtime": 15.4014,
|
| 102 |
+
"eval_samples_per_second": 51.943,
|
| 103 |
+
"eval_steps_per_second": 3.246,
|
| 104 |
+
"step": 1500
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 8.0,
|
| 108 |
+
"learning_rate": 1.4666666666666666e-05,
|
| 109 |
+
"loss": 0.0748,
|
| 110 |
+
"step": 1600
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 8.75,
|
| 114 |
+
"eval_accuracy": 0.605,
|
| 115 |
+
"eval_loss": 2.0466909408569336,
|
| 116 |
+
"eval_runtime": 15.38,
|
| 117 |
+
"eval_samples_per_second": 52.016,
|
| 118 |
+
"eval_steps_per_second": 3.251,
|
| 119 |
+
"step": 1750
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 9.0,
|
| 123 |
+
"learning_rate": 1.4e-05,
|
| 124 |
+
"loss": 0.0476,
|
| 125 |
+
"step": 1800
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 10.0,
|
| 129 |
+
"learning_rate": 1.3333333333333333e-05,
|
| 130 |
+
"loss": 0.0282,
|
| 131 |
+
"step": 2000
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 10.0,
|
| 135 |
+
"eval_accuracy": 0.6175,
|
| 136 |
+
"eval_loss": 2.3013718128204346,
|
| 137 |
+
"eval_runtime": 15.3412,
|
| 138 |
+
"eval_samples_per_second": 52.147,
|
| 139 |
+
"eval_steps_per_second": 3.259,
|
| 140 |
+
"step": 2000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 11.0,
|
| 144 |
+
"learning_rate": 1.2666666666666667e-05,
|
| 145 |
+
"loss": 0.038,
|
| 146 |
+
"step": 2200
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 11.25,
|
| 150 |
+
"eval_accuracy": 0.63,
|
| 151 |
+
"eval_loss": 2.488283157348633,
|
| 152 |
+
"eval_runtime": 15.3496,
|
| 153 |
+
"eval_samples_per_second": 52.119,
|
| 154 |
+
"eval_steps_per_second": 3.257,
|
| 155 |
+
"step": 2250
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 12.0,
|
| 159 |
+
"learning_rate": 1.2e-05,
|
| 160 |
+
"loss": 0.0218,
|
| 161 |
+
"step": 2400
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 12.5,
|
| 165 |
+
"eval_accuracy": 0.62875,
|
| 166 |
+
"eval_loss": 2.5122828483581543,
|
| 167 |
+
"eval_runtime": 15.4178,
|
| 168 |
+
"eval_samples_per_second": 51.888,
|
| 169 |
+
"eval_steps_per_second": 3.243,
|
| 170 |
+
"step": 2500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 13.0,
|
| 174 |
+
"learning_rate": 1.1333333333333334e-05,
|
| 175 |
+
"loss": 0.0174,
|
| 176 |
+
"step": 2600
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 13.75,
|
| 180 |
+
"eval_accuracy": 0.62125,
|
| 181 |
+
"eval_loss": 2.5520098209381104,
|
| 182 |
+
"eval_runtime": 15.4539,
|
| 183 |
+
"eval_samples_per_second": 51.767,
|
| 184 |
+
"eval_steps_per_second": 3.235,
|
| 185 |
+
"step": 2750
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 14.0,
|
| 189 |
+
"learning_rate": 1.0666666666666667e-05,
|
| 190 |
+
"loss": 0.0114,
|
| 191 |
+
"step": 2800
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 15.0,
|
| 195 |
+
"learning_rate": 1e-05,
|
| 196 |
+
"loss": 0.0134,
|
| 197 |
+
"step": 3000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 15.0,
|
| 201 |
+
"eval_accuracy": 0.61,
|
| 202 |
+
"eval_loss": 2.702610969543457,
|
| 203 |
+
"eval_runtime": 15.3837,
|
| 204 |
+
"eval_samples_per_second": 52.003,
|
| 205 |
+
"eval_steps_per_second": 3.25,
|
| 206 |
+
"step": 3000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 16.0,
|
| 210 |
+
"learning_rate": 9.333333333333334e-06,
|
| 211 |
+
"loss": 0.0107,
|
| 212 |
+
"step": 3200
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 16.25,
|
| 216 |
+
"eval_accuracy": 0.605,
|
| 217 |
+
"eval_loss": 2.7754039764404297,
|
| 218 |
+
"eval_runtime": 15.3762,
|
| 219 |
+
"eval_samples_per_second": 52.029,
|
| 220 |
+
"eval_steps_per_second": 3.252,
|
| 221 |
+
"step": 3250
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 17.0,
|
| 225 |
+
"learning_rate": 8.666666666666668e-06,
|
| 226 |
+
"loss": 0.0088,
|
| 227 |
+
"step": 3400
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 17.5,
|
| 231 |
+
"eval_accuracy": 0.59875,
|
| 232 |
+
"eval_loss": 3.0348358154296875,
|
| 233 |
+
"eval_runtime": 15.3746,
|
| 234 |
+
"eval_samples_per_second": 52.034,
|
| 235 |
+
"eval_steps_per_second": 3.252,
|
| 236 |
+
"step": 3500
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 18.0,
|
| 240 |
+
"learning_rate": 8.000000000000001e-06,
|
| 241 |
+
"loss": 0.0084,
|
| 242 |
+
"step": 3600
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 18.75,
|
| 246 |
+
"eval_accuracy": 0.6375,
|
| 247 |
+
"eval_loss": 2.828749656677246,
|
| 248 |
+
"eval_runtime": 15.4159,
|
| 249 |
+
"eval_samples_per_second": 51.894,
|
| 250 |
+
"eval_steps_per_second": 3.243,
|
| 251 |
+
"step": 3750
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 19.0,
|
| 255 |
+
"learning_rate": 7.333333333333333e-06,
|
| 256 |
+
"loss": 0.0085,
|
| 257 |
+
"step": 3800
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 20.0,
|
| 261 |
+
"learning_rate": 6.666666666666667e-06,
|
| 262 |
+
"loss": 0.0044,
|
| 263 |
+
"step": 4000
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 20.0,
|
| 267 |
+
"eval_accuracy": 0.62875,
|
| 268 |
+
"eval_loss": 2.9390206336975098,
|
| 269 |
+
"eval_runtime": 15.4137,
|
| 270 |
+
"eval_samples_per_second": 51.902,
|
| 271 |
+
"eval_steps_per_second": 3.244,
|
| 272 |
+
"step": 4000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 21.0,
|
| 276 |
+
"learning_rate": 6e-06,
|
| 277 |
+
"loss": 0.0093,
|
| 278 |
+
"step": 4200
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 21.25,
|
| 282 |
+
"eval_accuracy": 0.60375,
|
| 283 |
+
"eval_loss": 3.133685350418091,
|
| 284 |
+
"eval_runtime": 15.4418,
|
| 285 |
+
"eval_samples_per_second": 51.808,
|
| 286 |
+
"eval_steps_per_second": 3.238,
|
| 287 |
+
"step": 4250
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 22.0,
|
| 291 |
+
"learning_rate": 5.333333333333334e-06,
|
| 292 |
+
"loss": 0.0071,
|
| 293 |
+
"step": 4400
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 22.5,
|
| 297 |
+
"eval_accuracy": 0.6175,
|
| 298 |
+
"eval_loss": 2.979722261428833,
|
| 299 |
+
"eval_runtime": 15.4954,
|
| 300 |
+
"eval_samples_per_second": 51.628,
|
| 301 |
+
"eval_steps_per_second": 3.227,
|
| 302 |
+
"step": 4500
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 23.0,
|
| 306 |
+
"learning_rate": 4.666666666666667e-06,
|
| 307 |
+
"loss": 0.0087,
|
| 308 |
+
"step": 4600
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 23.75,
|
| 312 |
+
"eval_accuracy": 0.61625,
|
| 313 |
+
"eval_loss": 3.0937910079956055,
|
| 314 |
+
"eval_runtime": 15.473,
|
| 315 |
+
"eval_samples_per_second": 51.703,
|
| 316 |
+
"eval_steps_per_second": 3.231,
|
| 317 |
+
"step": 4750
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 24.0,
|
| 321 |
+
"learning_rate": 4.000000000000001e-06,
|
| 322 |
+
"loss": 0.0048,
|
| 323 |
+
"step": 4800
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 25.0,
|
| 327 |
+
"learning_rate": 3.3333333333333333e-06,
|
| 328 |
+
"loss": 0.0033,
|
| 329 |
+
"step": 5000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 25.0,
|
| 333 |
+
"eval_accuracy": 0.64125,
|
| 334 |
+
"eval_loss": 2.959096908569336,
|
| 335 |
+
"eval_runtime": 15.3539,
|
| 336 |
+
"eval_samples_per_second": 52.104,
|
| 337 |
+
"eval_steps_per_second": 3.257,
|
| 338 |
+
"step": 5000
|
| 339 |
}
|
| 340 |
],
|
| 341 |
+
"max_steps": 6000,
|
| 342 |
+
"num_train_epochs": 30,
|
| 343 |
+
"total_flos": 1.059758088192e+16,
|
| 344 |
"trial_name": null,
|
| 345 |
"trial_params": null
|
| 346 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3375
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:baebf3b382853feb0420e7a2f05d5defe1e8998e21926fd410842e611c022574
|
| 3 |
size 3375
|