Commit
·
50c347b
1
Parent(s):
2001a24
Training in progress, epoch 1
Browse files- last-checkpoint/optimizer.pt +2 -2
- last-checkpoint/pytorch_model.bin +1 -1
- last-checkpoint/rng_state.pth +1 -1
- last-checkpoint/scaler.pt +1 -1
- last-checkpoint/scheduler.pt +1 -1
- last-checkpoint/trainer_state.json +9 -247
- last-checkpoint/training_args.bin +1 -1
- pytorch_model.bin +1 -1
- training_args.bin +1 -1
last-checkpoint/optimizer.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8676f9764c07e6b933f4fa352af9b98b3b044de2eff759af6bf5baa43998241a
|
| 3 |
+
size 721649669
|
last-checkpoint/pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 377640289
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3321050bfb777e92d4a9535bacf26233d6d93c177f40afb72e9a4edb621469ad
|
| 3 |
size 377640289
|
last-checkpoint/rng_state.pth
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 14639
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02a382c9fbd249cde2e9b0ca8a6981276f1d798e47ccb2d4445b039ddbd74cae
|
| 3 |
size 14639
|
last-checkpoint/scaler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 557
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:af1f41381fb3909e7aa415aa1d1f2d9db8344ae4f5d0ae7b0cdf50527d59f2cb
|
| 3 |
size 557
|
last-checkpoint/scheduler.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 627
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:00e56ff1f9aeef202866d18c52f49cf630af4e847aa9aca70191fe768f73377a
|
| 3 |
size 627
|
last-checkpoint/trainer_state.json
CHANGED
|
@@ -1,271 +1,33 @@
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 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": 9.
|
| 13 |
-
"loss":
|
| 14 |
"step": 322
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 1.0,
|
| 18 |
"eval_cer": 1.0,
|
| 19 |
-
"eval_loss":
|
| 20 |
"eval_new_wer": 1.0,
|
| 21 |
"eval_old_wer": 1.0,
|
| 22 |
-
"eval_runtime": 7.
|
| 23 |
-
"eval_samples_per_second":
|
| 24 |
-
"eval_steps_per_second": 3.
|
| 25 |
"step": 322
|
| 26 |
-
},
|
| 27 |
-
{
|
| 28 |
-
"epoch": 2.0,
|
| 29 |
-
"learning_rate": 9.139246023098715e-05,
|
| 30 |
-
"loss": 5.7752,
|
| 31 |
-
"step": 644
|
| 32 |
-
},
|
| 33 |
-
{
|
| 34 |
-
"epoch": 2.0,
|
| 35 |
-
"eval_cer": 1.0,
|
| 36 |
-
"eval_loss": 3.7167434692382812,
|
| 37 |
-
"eval_new_wer": 1.0,
|
| 38 |
-
"eval_old_wer": 1.0,
|
| 39 |
-
"eval_runtime": 7.7609,
|
| 40 |
-
"eval_samples_per_second": 27.059,
|
| 41 |
-
"eval_steps_per_second": 3.479,
|
| 42 |
-
"step": 644
|
| 43 |
-
},
|
| 44 |
-
{
|
| 45 |
-
"epoch": 3.0,
|
| 46 |
-
"learning_rate": 8.437568097624755e-05,
|
| 47 |
-
"loss": 3.4632,
|
| 48 |
-
"step": 966
|
| 49 |
-
},
|
| 50 |
-
{
|
| 51 |
-
"epoch": 3.0,
|
| 52 |
-
"eval_cer": 1.0,
|
| 53 |
-
"eval_loss": 3.6892776489257812,
|
| 54 |
-
"eval_new_wer": 1.0,
|
| 55 |
-
"eval_old_wer": 1.0,
|
| 56 |
-
"eval_runtime": 8.762,
|
| 57 |
-
"eval_samples_per_second": 23.967,
|
| 58 |
-
"eval_steps_per_second": 3.081,
|
| 59 |
-
"step": 966
|
| 60 |
-
},
|
| 61 |
-
{
|
| 62 |
-
"epoch": 4.0,
|
| 63 |
-
"learning_rate": 7.735890172150796e-05,
|
| 64 |
-
"loss": 3.3883,
|
| 65 |
-
"step": 1288
|
| 66 |
-
},
|
| 67 |
-
{
|
| 68 |
-
"epoch": 4.0,
|
| 69 |
-
"eval_cer": 1.0,
|
| 70 |
-
"eval_loss": 3.6128814220428467,
|
| 71 |
-
"eval_new_wer": 1.0,
|
| 72 |
-
"eval_old_wer": 1.0,
|
| 73 |
-
"eval_runtime": 8.0803,
|
| 74 |
-
"eval_samples_per_second": 25.989,
|
| 75 |
-
"eval_steps_per_second": 3.341,
|
| 76 |
-
"step": 1288
|
| 77 |
-
},
|
| 78 |
-
{
|
| 79 |
-
"epoch": 5.0,
|
| 80 |
-
"learning_rate": 7.034212246676837e-05,
|
| 81 |
-
"loss": 3.3461,
|
| 82 |
-
"step": 1610
|
| 83 |
-
},
|
| 84 |
-
{
|
| 85 |
-
"epoch": 5.0,
|
| 86 |
-
"eval_cer": 1.0,
|
| 87 |
-
"eval_loss": 3.5834929943084717,
|
| 88 |
-
"eval_new_wer": 1.0,
|
| 89 |
-
"eval_old_wer": 1.0,
|
| 90 |
-
"eval_runtime": 8.5699,
|
| 91 |
-
"eval_samples_per_second": 24.504,
|
| 92 |
-
"eval_steps_per_second": 3.151,
|
| 93 |
-
"step": 1610
|
| 94 |
-
},
|
| 95 |
-
{
|
| 96 |
-
"epoch": 6.0,
|
| 97 |
-
"learning_rate": 6.332534321202876e-05,
|
| 98 |
-
"loss": 3.3228,
|
| 99 |
-
"step": 1932
|
| 100 |
-
},
|
| 101 |
-
{
|
| 102 |
-
"epoch": 6.0,
|
| 103 |
-
"eval_cer": 1.0,
|
| 104 |
-
"eval_loss": 3.564229726791382,
|
| 105 |
-
"eval_new_wer": 1.0,
|
| 106 |
-
"eval_old_wer": 1.0,
|
| 107 |
-
"eval_runtime": 7.6036,
|
| 108 |
-
"eval_samples_per_second": 27.619,
|
| 109 |
-
"eval_steps_per_second": 3.551,
|
| 110 |
-
"step": 1932
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 7.0,
|
| 114 |
-
"learning_rate": 5.6308563957289177e-05,
|
| 115 |
-
"loss": 3.3009,
|
| 116 |
-
"step": 2254
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 7.0,
|
| 120 |
-
"eval_cer": 1.0,
|
| 121 |
-
"eval_loss": 3.5358943939208984,
|
| 122 |
-
"eval_new_wer": 1.0,
|
| 123 |
-
"eval_old_wer": 1.0,
|
| 124 |
-
"eval_runtime": 7.829,
|
| 125 |
-
"eval_samples_per_second": 26.824,
|
| 126 |
-
"eval_steps_per_second": 3.449,
|
| 127 |
-
"step": 2254
|
| 128 |
-
},
|
| 129 |
-
{
|
| 130 |
-
"epoch": 8.0,
|
| 131 |
-
"learning_rate": 4.9291784702549575e-05,
|
| 132 |
-
"loss": 3.2842,
|
| 133 |
-
"step": 2576
|
| 134 |
-
},
|
| 135 |
-
{
|
| 136 |
-
"epoch": 8.0,
|
| 137 |
-
"eval_cer": 1.0,
|
| 138 |
-
"eval_loss": 3.5163114070892334,
|
| 139 |
-
"eval_new_wer": 1.0,
|
| 140 |
-
"eval_old_wer": 1.0,
|
| 141 |
-
"eval_runtime": 7.665,
|
| 142 |
-
"eval_samples_per_second": 27.397,
|
| 143 |
-
"eval_steps_per_second": 3.523,
|
| 144 |
-
"step": 2576
|
| 145 |
-
},
|
| 146 |
-
{
|
| 147 |
-
"epoch": 9.0,
|
| 148 |
-
"learning_rate": 4.2275005447809986e-05,
|
| 149 |
-
"loss": 3.2793,
|
| 150 |
-
"step": 2898
|
| 151 |
-
},
|
| 152 |
-
{
|
| 153 |
-
"epoch": 9.0,
|
| 154 |
-
"eval_cer": 1.0,
|
| 155 |
-
"eval_loss": 3.4834606647491455,
|
| 156 |
-
"eval_new_wer": 1.0,
|
| 157 |
-
"eval_old_wer": 1.0,
|
| 158 |
-
"eval_runtime": 8.9806,
|
| 159 |
-
"eval_samples_per_second": 23.384,
|
| 160 |
-
"eval_steps_per_second": 3.006,
|
| 161 |
-
"step": 2898
|
| 162 |
-
},
|
| 163 |
-
{
|
| 164 |
-
"epoch": 10.0,
|
| 165 |
-
"learning_rate": 3.5258226193070384e-05,
|
| 166 |
-
"loss": 3.2629,
|
| 167 |
-
"step": 3220
|
| 168 |
-
},
|
| 169 |
-
{
|
| 170 |
-
"epoch": 10.0,
|
| 171 |
-
"eval_cer": 1.0,
|
| 172 |
-
"eval_loss": 3.4674203395843506,
|
| 173 |
-
"eval_new_wer": 1.0,
|
| 174 |
-
"eval_old_wer": 1.0,
|
| 175 |
-
"eval_runtime": 9.6156,
|
| 176 |
-
"eval_samples_per_second": 21.839,
|
| 177 |
-
"eval_steps_per_second": 2.808,
|
| 178 |
-
"step": 3220
|
| 179 |
-
},
|
| 180 |
-
{
|
| 181 |
-
"epoch": 11.0,
|
| 182 |
-
"learning_rate": 2.824144693833079e-05,
|
| 183 |
-
"loss": 3.2617,
|
| 184 |
-
"step": 3542
|
| 185 |
-
},
|
| 186 |
-
{
|
| 187 |
-
"epoch": 11.0,
|
| 188 |
-
"eval_cer": 1.0,
|
| 189 |
-
"eval_loss": 3.4477264881134033,
|
| 190 |
-
"eval_new_wer": 1.0,
|
| 191 |
-
"eval_old_wer": 1.0,
|
| 192 |
-
"eval_runtime": 8.5536,
|
| 193 |
-
"eval_samples_per_second": 24.551,
|
| 194 |
-
"eval_steps_per_second": 3.157,
|
| 195 |
-
"step": 3542
|
| 196 |
-
},
|
| 197 |
-
{
|
| 198 |
-
"epoch": 12.0,
|
| 199 |
-
"learning_rate": 2.1224667683591198e-05,
|
| 200 |
-
"loss": 3.251,
|
| 201 |
-
"step": 3864
|
| 202 |
-
},
|
| 203 |
-
{
|
| 204 |
-
"epoch": 12.0,
|
| 205 |
-
"eval_cer": 1.0,
|
| 206 |
-
"eval_loss": 3.451367139816284,
|
| 207 |
-
"eval_new_wer": 1.0,
|
| 208 |
-
"eval_old_wer": 1.0,
|
| 209 |
-
"eval_runtime": 7.3215,
|
| 210 |
-
"eval_samples_per_second": 28.683,
|
| 211 |
-
"eval_steps_per_second": 3.688,
|
| 212 |
-
"step": 3864
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 13.0,
|
| 216 |
-
"learning_rate": 1.4207888428851601e-05,
|
| 217 |
-
"loss": 3.2499,
|
| 218 |
-
"step": 4186
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 13.0,
|
| 222 |
-
"eval_cer": 1.0,
|
| 223 |
-
"eval_loss": 3.4498093128204346,
|
| 224 |
-
"eval_new_wer": 1.0,
|
| 225 |
-
"eval_old_wer": 1.0,
|
| 226 |
-
"eval_runtime": 8.0836,
|
| 227 |
-
"eval_samples_per_second": 25.979,
|
| 228 |
-
"eval_steps_per_second": 3.34,
|
| 229 |
-
"step": 4186
|
| 230 |
-
},
|
| 231 |
-
{
|
| 232 |
-
"epoch": 14.0,
|
| 233 |
-
"learning_rate": 7.191109174112007e-06,
|
| 234 |
-
"loss": 3.2462,
|
| 235 |
-
"step": 4508
|
| 236 |
-
},
|
| 237 |
-
{
|
| 238 |
-
"epoch": 14.0,
|
| 239 |
-
"eval_cer": 1.0,
|
| 240 |
-
"eval_loss": 3.4502668380737305,
|
| 241 |
-
"eval_new_wer": 1.0,
|
| 242 |
-
"eval_old_wer": 1.0,
|
| 243 |
-
"eval_runtime": 7.7485,
|
| 244 |
-
"eval_samples_per_second": 27.102,
|
| 245 |
-
"eval_steps_per_second": 3.485,
|
| 246 |
-
"step": 4508
|
| 247 |
-
},
|
| 248 |
-
{
|
| 249 |
-
"epoch": 15.0,
|
| 250 |
-
"learning_rate": 1.743299193724123e-07,
|
| 251 |
-
"loss": 3.2407,
|
| 252 |
-
"step": 4830
|
| 253 |
-
},
|
| 254 |
-
{
|
| 255 |
-
"epoch": 15.0,
|
| 256 |
-
"eval_cer": 1.0,
|
| 257 |
-
"eval_loss": 3.4475154876708984,
|
| 258 |
-
"eval_new_wer": 1.0,
|
| 259 |
-
"eval_old_wer": 1.0,
|
| 260 |
-
"eval_runtime": 7.7552,
|
| 261 |
-
"eval_samples_per_second": 27.079,
|
| 262 |
-
"eval_steps_per_second": 3.482,
|
| 263 |
-
"step": 4830
|
| 264 |
}
|
| 265 |
],
|
| 266 |
"max_steps": 4830,
|
| 267 |
"num_train_epochs": 15,
|
| 268 |
-
"total_flos":
|
| 269 |
"trial_name": null,
|
| 270 |
"trial_params": null
|
| 271 |
}
|
|
|
|
| 1 |
{
|
| 2 |
"best_metric": null,
|
| 3 |
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 322,
|
| 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": 9.83874482458052e-05,
|
| 13 |
+
"loss": 20.722,
|
| 14 |
"step": 322
|
| 15 |
},
|
| 16 |
{
|
| 17 |
"epoch": 1.0,
|
| 18 |
"eval_cer": 1.0,
|
| 19 |
+
"eval_loss": 3.2618589401245117,
|
| 20 |
"eval_new_wer": 1.0,
|
| 21 |
"eval_old_wer": 1.0,
|
| 22 |
+
"eval_runtime": 7.6687,
|
| 23 |
+
"eval_samples_per_second": 27.384,
|
| 24 |
+
"eval_steps_per_second": 3.521,
|
| 25 |
"step": 322
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
}
|
| 27 |
],
|
| 28 |
"max_steps": 4830,
|
| 29 |
"num_train_epochs": 15,
|
| 30 |
+
"total_flos": 2.331097571376e+17,
|
| 31 |
"trial_name": null,
|
| 32 |
"trial_params": null
|
| 33 |
}
|
last-checkpoint/training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3387
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec767e1aae092c1f11b99196ba7b3d812c3d6f0dd8fc08a3236cbdaed665c876
|
| 3 |
size 3387
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 377640289
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3321050bfb777e92d4a9535bacf26233d6d93c177f40afb72e9a4edb621469ad
|
| 3 |
size 377640289
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3387
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ec767e1aae092c1f11b99196ba7b3d812c3d6f0dd8fc08a3236cbdaed665c876
|
| 3 |
size 3387
|