Delete checkpoint-2000
Browse files- checkpoint-2000/config.json +0 -42
- checkpoint-2000/optimizer.pt +0 -3
- checkpoint-2000/preprocessor_config.json +0 -0
- checkpoint-2000/pytorch_model.bin +0 -3
- checkpoint-2000/rng_state.pth +0 -3
- checkpoint-2000/scaler.pt +0 -3
- checkpoint-2000/scheduler.pt +0 -3
- checkpoint-2000/trainer_state.json +0 -514
- checkpoint-2000/training_args.bin +0 -3
checkpoint-2000/config.json
DELETED
|
@@ -1,42 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"_name_or_path": "./lab2_id2223",
|
| 3 |
-
"activation_dropout": 0.0,
|
| 4 |
-
"activation_function": "gelu",
|
| 5 |
-
"architectures": [
|
| 6 |
-
"WhisperForConditionalGeneration"
|
| 7 |
-
],
|
| 8 |
-
"attention_dropout": 0.0,
|
| 9 |
-
"begin_suppress_tokens": [
|
| 10 |
-
220,
|
| 11 |
-
50257
|
| 12 |
-
],
|
| 13 |
-
"bos_token_id": 50257,
|
| 14 |
-
"d_model": 768,
|
| 15 |
-
"decoder_attention_heads": 12,
|
| 16 |
-
"decoder_ffn_dim": 3072,
|
| 17 |
-
"decoder_layerdrop": 0.0,
|
| 18 |
-
"decoder_layers": 12,
|
| 19 |
-
"decoder_start_token_id": 50258,
|
| 20 |
-
"dropout": 0.0,
|
| 21 |
-
"encoder_attention_heads": 12,
|
| 22 |
-
"encoder_ffn_dim": 3072,
|
| 23 |
-
"encoder_layerdrop": 0.0,
|
| 24 |
-
"encoder_layers": 12,
|
| 25 |
-
"eos_token_id": 50257,
|
| 26 |
-
"forced_decoder_ids": null,
|
| 27 |
-
"init_std": 0.02,
|
| 28 |
-
"is_encoder_decoder": true,
|
| 29 |
-
"max_length": 448,
|
| 30 |
-
"max_source_positions": 1500,
|
| 31 |
-
"max_target_positions": 448,
|
| 32 |
-
"model_type": "whisper",
|
| 33 |
-
"num_hidden_layers": 12,
|
| 34 |
-
"num_mel_bins": 80,
|
| 35 |
-
"pad_token_id": 50257,
|
| 36 |
-
"scale_embedding": false,
|
| 37 |
-
"suppress_tokens": [],
|
| 38 |
-
"torch_dtype": "float32",
|
| 39 |
-
"transformers_version": "4.26.0.dev0",
|
| 40 |
-
"use_cache": false,
|
| 41 |
-
"vocab_size": 51865
|
| 42 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/optimizer.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:1280b664e2299772ee558c8b50a070bb2ecb8aeeff2becf9dacd5f8eeaa96fae
|
| 3 |
-
size 1934161093
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/preprocessor_config.json
DELETED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-2000/pytorch_model.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:f5bc6e356b2f66f6b694c2bd274a577c6d9586eac3215b17afee61b5217f3579
|
| 3 |
-
size 967102601
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/rng_state.pth
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:fc2370b290bbb9b0161d6bb54c04e80e95148a66774a7fc043af0ab9a03d7ecd
|
| 3 |
-
size 14575
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/scaler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:738953d7bfb3f612f66f32f7e2019c92256bd40f51a66f25241750d74e38d0b1
|
| 3 |
-
size 557
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/scheduler.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0fe7ca1c4fdccf931b9bdbd5399f5fc0f8d8eee4c47f56dd3b27eacf280f86a3
|
| 3 |
-
size 627
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/trainer_state.json
DELETED
|
@@ -1,514 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"best_metric": 19.680970568411592,
|
| 3 |
-
"best_model_checkpoint": "./lab2_id2223/checkpoint-2000",
|
| 4 |
-
"epoch": 2.5873221216041395,
|
| 5 |
-
"global_step": 2000,
|
| 6 |
-
"is_hyper_param_search": false,
|
| 7 |
-
"is_local_process_zero": true,
|
| 8 |
-
"is_world_process_zero": true,
|
| 9 |
-
"log_history": [
|
| 10 |
-
{
|
| 11 |
-
"epoch": 0.03,
|
| 12 |
-
"learning_rate": 4.800000000000001e-07,
|
| 13 |
-
"loss": 0.2225,
|
| 14 |
-
"step": 25
|
| 15 |
-
},
|
| 16 |
-
{
|
| 17 |
-
"epoch": 0.06,
|
| 18 |
-
"learning_rate": 9.800000000000001e-07,
|
| 19 |
-
"loss": 0.2081,
|
| 20 |
-
"step": 50
|
| 21 |
-
},
|
| 22 |
-
{
|
| 23 |
-
"epoch": 0.1,
|
| 24 |
-
"learning_rate": 1.48e-06,
|
| 25 |
-
"loss": 0.1992,
|
| 26 |
-
"step": 75
|
| 27 |
-
},
|
| 28 |
-
{
|
| 29 |
-
"epoch": 0.13,
|
| 30 |
-
"learning_rate": 1.98e-06,
|
| 31 |
-
"loss": 0.2014,
|
| 32 |
-
"step": 100
|
| 33 |
-
},
|
| 34 |
-
{
|
| 35 |
-
"epoch": 0.16,
|
| 36 |
-
"learning_rate": 2.4800000000000004e-06,
|
| 37 |
-
"loss": 0.1914,
|
| 38 |
-
"step": 125
|
| 39 |
-
},
|
| 40 |
-
{
|
| 41 |
-
"epoch": 0.19,
|
| 42 |
-
"learning_rate": 2.9800000000000003e-06,
|
| 43 |
-
"loss": 0.1978,
|
| 44 |
-
"step": 150
|
| 45 |
-
},
|
| 46 |
-
{
|
| 47 |
-
"epoch": 0.23,
|
| 48 |
-
"learning_rate": 3.48e-06,
|
| 49 |
-
"loss": 0.1988,
|
| 50 |
-
"step": 175
|
| 51 |
-
},
|
| 52 |
-
{
|
| 53 |
-
"epoch": 0.26,
|
| 54 |
-
"learning_rate": 3.980000000000001e-06,
|
| 55 |
-
"loss": 0.2153,
|
| 56 |
-
"step": 200
|
| 57 |
-
},
|
| 58 |
-
{
|
| 59 |
-
"epoch": 0.29,
|
| 60 |
-
"learning_rate": 4.48e-06,
|
| 61 |
-
"loss": 0.1907,
|
| 62 |
-
"step": 225
|
| 63 |
-
},
|
| 64 |
-
{
|
| 65 |
-
"epoch": 0.32,
|
| 66 |
-
"learning_rate": 4.980000000000001e-06,
|
| 67 |
-
"loss": 0.1976,
|
| 68 |
-
"step": 250
|
| 69 |
-
},
|
| 70 |
-
{
|
| 71 |
-
"epoch": 0.36,
|
| 72 |
-
"learning_rate": 5.480000000000001e-06,
|
| 73 |
-
"loss": 0.2256,
|
| 74 |
-
"step": 275
|
| 75 |
-
},
|
| 76 |
-
{
|
| 77 |
-
"epoch": 0.39,
|
| 78 |
-
"learning_rate": 5.98e-06,
|
| 79 |
-
"loss": 0.1805,
|
| 80 |
-
"step": 300
|
| 81 |
-
},
|
| 82 |
-
{
|
| 83 |
-
"epoch": 0.42,
|
| 84 |
-
"learning_rate": 6.480000000000001e-06,
|
| 85 |
-
"loss": 0.2144,
|
| 86 |
-
"step": 325
|
| 87 |
-
},
|
| 88 |
-
{
|
| 89 |
-
"epoch": 0.45,
|
| 90 |
-
"learning_rate": 6.98e-06,
|
| 91 |
-
"loss": 0.195,
|
| 92 |
-
"step": 350
|
| 93 |
-
},
|
| 94 |
-
{
|
| 95 |
-
"epoch": 0.49,
|
| 96 |
-
"learning_rate": 7.48e-06,
|
| 97 |
-
"loss": 0.2053,
|
| 98 |
-
"step": 375
|
| 99 |
-
},
|
| 100 |
-
{
|
| 101 |
-
"epoch": 0.52,
|
| 102 |
-
"learning_rate": 7.980000000000002e-06,
|
| 103 |
-
"loss": 0.2157,
|
| 104 |
-
"step": 400
|
| 105 |
-
},
|
| 106 |
-
{
|
| 107 |
-
"epoch": 0.55,
|
| 108 |
-
"learning_rate": 8.48e-06,
|
| 109 |
-
"loss": 0.2062,
|
| 110 |
-
"step": 425
|
| 111 |
-
},
|
| 112 |
-
{
|
| 113 |
-
"epoch": 0.58,
|
| 114 |
-
"learning_rate": 8.96e-06,
|
| 115 |
-
"loss": 0.1979,
|
| 116 |
-
"step": 450
|
| 117 |
-
},
|
| 118 |
-
{
|
| 119 |
-
"epoch": 0.61,
|
| 120 |
-
"learning_rate": 9.460000000000001e-06,
|
| 121 |
-
"loss": 0.1868,
|
| 122 |
-
"step": 475
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"epoch": 0.65,
|
| 126 |
-
"learning_rate": 9.960000000000001e-06,
|
| 127 |
-
"loss": 0.2211,
|
| 128 |
-
"step": 500
|
| 129 |
-
},
|
| 130 |
-
{
|
| 131 |
-
"epoch": 0.68,
|
| 132 |
-
"learning_rate": 9.934285714285715e-06,
|
| 133 |
-
"loss": 0.2049,
|
| 134 |
-
"step": 525
|
| 135 |
-
},
|
| 136 |
-
{
|
| 137 |
-
"epoch": 0.71,
|
| 138 |
-
"learning_rate": 9.862857142857144e-06,
|
| 139 |
-
"loss": 0.1989,
|
| 140 |
-
"step": 550
|
| 141 |
-
},
|
| 142 |
-
{
|
| 143 |
-
"epoch": 0.74,
|
| 144 |
-
"learning_rate": 9.791428571428571e-06,
|
| 145 |
-
"loss": 0.2127,
|
| 146 |
-
"step": 575
|
| 147 |
-
},
|
| 148 |
-
{
|
| 149 |
-
"epoch": 0.78,
|
| 150 |
-
"learning_rate": 9.72e-06,
|
| 151 |
-
"loss": 0.2261,
|
| 152 |
-
"step": 600
|
| 153 |
-
},
|
| 154 |
-
{
|
| 155 |
-
"epoch": 0.81,
|
| 156 |
-
"learning_rate": 9.648571428571429e-06,
|
| 157 |
-
"loss": 0.2081,
|
| 158 |
-
"step": 625
|
| 159 |
-
},
|
| 160 |
-
{
|
| 161 |
-
"epoch": 0.84,
|
| 162 |
-
"learning_rate": 9.577142857142858e-06,
|
| 163 |
-
"loss": 0.2281,
|
| 164 |
-
"step": 650
|
| 165 |
-
},
|
| 166 |
-
{
|
| 167 |
-
"epoch": 0.87,
|
| 168 |
-
"learning_rate": 9.505714285714287e-06,
|
| 169 |
-
"loss": 0.204,
|
| 170 |
-
"step": 675
|
| 171 |
-
},
|
| 172 |
-
{
|
| 173 |
-
"epoch": 0.91,
|
| 174 |
-
"learning_rate": 9.434285714285714e-06,
|
| 175 |
-
"loss": 0.2053,
|
| 176 |
-
"step": 700
|
| 177 |
-
},
|
| 178 |
-
{
|
| 179 |
-
"epoch": 0.94,
|
| 180 |
-
"learning_rate": 9.362857142857143e-06,
|
| 181 |
-
"loss": 0.2208,
|
| 182 |
-
"step": 725
|
| 183 |
-
},
|
| 184 |
-
{
|
| 185 |
-
"epoch": 0.97,
|
| 186 |
-
"learning_rate": 9.291428571428572e-06,
|
| 187 |
-
"loss": 0.2076,
|
| 188 |
-
"step": 750
|
| 189 |
-
},
|
| 190 |
-
{
|
| 191 |
-
"epoch": 1.0,
|
| 192 |
-
"learning_rate": 9.220000000000002e-06,
|
| 193 |
-
"loss": 0.1941,
|
| 194 |
-
"step": 775
|
| 195 |
-
},
|
| 196 |
-
{
|
| 197 |
-
"epoch": 1.03,
|
| 198 |
-
"learning_rate": 9.148571428571429e-06,
|
| 199 |
-
"loss": 0.0949,
|
| 200 |
-
"step": 800
|
| 201 |
-
},
|
| 202 |
-
{
|
| 203 |
-
"epoch": 1.07,
|
| 204 |
-
"learning_rate": 9.077142857142858e-06,
|
| 205 |
-
"loss": 0.1011,
|
| 206 |
-
"step": 825
|
| 207 |
-
},
|
| 208 |
-
{
|
| 209 |
-
"epoch": 1.1,
|
| 210 |
-
"learning_rate": 9.005714285714287e-06,
|
| 211 |
-
"loss": 0.1078,
|
| 212 |
-
"step": 850
|
| 213 |
-
},
|
| 214 |
-
{
|
| 215 |
-
"epoch": 1.13,
|
| 216 |
-
"learning_rate": 8.934285714285716e-06,
|
| 217 |
-
"loss": 0.0956,
|
| 218 |
-
"step": 875
|
| 219 |
-
},
|
| 220 |
-
{
|
| 221 |
-
"epoch": 1.16,
|
| 222 |
-
"learning_rate": 8.862857142857143e-06,
|
| 223 |
-
"loss": 0.1093,
|
| 224 |
-
"step": 900
|
| 225 |
-
},
|
| 226 |
-
{
|
| 227 |
-
"epoch": 1.2,
|
| 228 |
-
"learning_rate": 8.791428571428572e-06,
|
| 229 |
-
"loss": 0.1133,
|
| 230 |
-
"step": 925
|
| 231 |
-
},
|
| 232 |
-
{
|
| 233 |
-
"epoch": 1.23,
|
| 234 |
-
"learning_rate": 8.720000000000001e-06,
|
| 235 |
-
"loss": 0.1128,
|
| 236 |
-
"step": 950
|
| 237 |
-
},
|
| 238 |
-
{
|
| 239 |
-
"epoch": 1.26,
|
| 240 |
-
"learning_rate": 8.64857142857143e-06,
|
| 241 |
-
"loss": 0.1068,
|
| 242 |
-
"step": 975
|
| 243 |
-
},
|
| 244 |
-
{
|
| 245 |
-
"epoch": 1.29,
|
| 246 |
-
"learning_rate": 8.577142857142858e-06,
|
| 247 |
-
"loss": 0.1015,
|
| 248 |
-
"step": 1000
|
| 249 |
-
},
|
| 250 |
-
{
|
| 251 |
-
"epoch": 1.29,
|
| 252 |
-
"eval_loss": 0.2880099415779114,
|
| 253 |
-
"eval_runtime": 2234.5268,
|
| 254 |
-
"eval_samples_per_second": 1.343,
|
| 255 |
-
"eval_steps_per_second": 0.168,
|
| 256 |
-
"eval_wer": 20.41339024938216,
|
| 257 |
-
"step": 1000
|
| 258 |
-
},
|
| 259 |
-
{
|
| 260 |
-
"epoch": 1.33,
|
| 261 |
-
"learning_rate": 8.505714285714287e-06,
|
| 262 |
-
"loss": 0.1154,
|
| 263 |
-
"step": 1025
|
| 264 |
-
},
|
| 265 |
-
{
|
| 266 |
-
"epoch": 1.36,
|
| 267 |
-
"learning_rate": 8.434285714285716e-06,
|
| 268 |
-
"loss": 0.1035,
|
| 269 |
-
"step": 1050
|
| 270 |
-
},
|
| 271 |
-
{
|
| 272 |
-
"epoch": 1.39,
|
| 273 |
-
"learning_rate": 8.362857142857143e-06,
|
| 274 |
-
"loss": 0.103,
|
| 275 |
-
"step": 1075
|
| 276 |
-
},
|
| 277 |
-
{
|
| 278 |
-
"epoch": 1.42,
|
| 279 |
-
"learning_rate": 8.291428571428572e-06,
|
| 280 |
-
"loss": 0.1089,
|
| 281 |
-
"step": 1100
|
| 282 |
-
},
|
| 283 |
-
{
|
| 284 |
-
"epoch": 1.46,
|
| 285 |
-
"learning_rate": 8.220000000000001e-06,
|
| 286 |
-
"loss": 0.0986,
|
| 287 |
-
"step": 1125
|
| 288 |
-
},
|
| 289 |
-
{
|
| 290 |
-
"epoch": 1.49,
|
| 291 |
-
"learning_rate": 8.148571428571428e-06,
|
| 292 |
-
"loss": 0.0999,
|
| 293 |
-
"step": 1150
|
| 294 |
-
},
|
| 295 |
-
{
|
| 296 |
-
"epoch": 1.52,
|
| 297 |
-
"learning_rate": 8.077142857142857e-06,
|
| 298 |
-
"loss": 0.1075,
|
| 299 |
-
"step": 1175
|
| 300 |
-
},
|
| 301 |
-
{
|
| 302 |
-
"epoch": 1.55,
|
| 303 |
-
"learning_rate": 8.005714285714286e-06,
|
| 304 |
-
"loss": 0.1155,
|
| 305 |
-
"step": 1200
|
| 306 |
-
},
|
| 307 |
-
{
|
| 308 |
-
"epoch": 1.58,
|
| 309 |
-
"learning_rate": 7.934285714285715e-06,
|
| 310 |
-
"loss": 0.0978,
|
| 311 |
-
"step": 1225
|
| 312 |
-
},
|
| 313 |
-
{
|
| 314 |
-
"epoch": 1.62,
|
| 315 |
-
"learning_rate": 7.862857142857143e-06,
|
| 316 |
-
"loss": 0.1035,
|
| 317 |
-
"step": 1250
|
| 318 |
-
},
|
| 319 |
-
{
|
| 320 |
-
"epoch": 1.65,
|
| 321 |
-
"learning_rate": 7.791428571428572e-06,
|
| 322 |
-
"loss": 0.1085,
|
| 323 |
-
"step": 1275
|
| 324 |
-
},
|
| 325 |
-
{
|
| 326 |
-
"epoch": 1.68,
|
| 327 |
-
"learning_rate": 7.72e-06,
|
| 328 |
-
"loss": 0.1008,
|
| 329 |
-
"step": 1300
|
| 330 |
-
},
|
| 331 |
-
{
|
| 332 |
-
"epoch": 1.71,
|
| 333 |
-
"learning_rate": 7.64857142857143e-06,
|
| 334 |
-
"loss": 0.1083,
|
| 335 |
-
"step": 1325
|
| 336 |
-
},
|
| 337 |
-
{
|
| 338 |
-
"epoch": 1.75,
|
| 339 |
-
"learning_rate": 7.577142857142857e-06,
|
| 340 |
-
"loss": 0.0918,
|
| 341 |
-
"step": 1350
|
| 342 |
-
},
|
| 343 |
-
{
|
| 344 |
-
"epoch": 1.78,
|
| 345 |
-
"learning_rate": 7.505714285714286e-06,
|
| 346 |
-
"loss": 0.1124,
|
| 347 |
-
"step": 1375
|
| 348 |
-
},
|
| 349 |
-
{
|
| 350 |
-
"epoch": 1.81,
|
| 351 |
-
"learning_rate": 7.434285714285715e-06,
|
| 352 |
-
"loss": 0.0997,
|
| 353 |
-
"step": 1400
|
| 354 |
-
},
|
| 355 |
-
{
|
| 356 |
-
"epoch": 1.84,
|
| 357 |
-
"learning_rate": 7.362857142857144e-06,
|
| 358 |
-
"loss": 0.0954,
|
| 359 |
-
"step": 1425
|
| 360 |
-
},
|
| 361 |
-
{
|
| 362 |
-
"epoch": 1.88,
|
| 363 |
-
"learning_rate": 7.291428571428571e-06,
|
| 364 |
-
"loss": 0.0946,
|
| 365 |
-
"step": 1450
|
| 366 |
-
},
|
| 367 |
-
{
|
| 368 |
-
"epoch": 1.91,
|
| 369 |
-
"learning_rate": 7.22e-06,
|
| 370 |
-
"loss": 0.1,
|
| 371 |
-
"step": 1475
|
| 372 |
-
},
|
| 373 |
-
{
|
| 374 |
-
"epoch": 1.94,
|
| 375 |
-
"learning_rate": 7.148571428571429e-06,
|
| 376 |
-
"loss": 0.0961,
|
| 377 |
-
"step": 1500
|
| 378 |
-
},
|
| 379 |
-
{
|
| 380 |
-
"epoch": 1.97,
|
| 381 |
-
"learning_rate": 7.077142857142858e-06,
|
| 382 |
-
"loss": 0.1074,
|
| 383 |
-
"step": 1525
|
| 384 |
-
},
|
| 385 |
-
{
|
| 386 |
-
"epoch": 2.01,
|
| 387 |
-
"learning_rate": 7.0057142857142865e-06,
|
| 388 |
-
"loss": 0.0821,
|
| 389 |
-
"step": 1550
|
| 390 |
-
},
|
| 391 |
-
{
|
| 392 |
-
"epoch": 2.04,
|
| 393 |
-
"learning_rate": 6.934285714285715e-06,
|
| 394 |
-
"loss": 0.0392,
|
| 395 |
-
"step": 1575
|
| 396 |
-
},
|
| 397 |
-
{
|
| 398 |
-
"epoch": 2.07,
|
| 399 |
-
"learning_rate": 6.862857142857144e-06,
|
| 400 |
-
"loss": 0.0371,
|
| 401 |
-
"step": 1600
|
| 402 |
-
},
|
| 403 |
-
{
|
| 404 |
-
"epoch": 2.1,
|
| 405 |
-
"learning_rate": 6.791428571428572e-06,
|
| 406 |
-
"loss": 0.0441,
|
| 407 |
-
"step": 1625
|
| 408 |
-
},
|
| 409 |
-
{
|
| 410 |
-
"epoch": 2.13,
|
| 411 |
-
"learning_rate": 6.720000000000001e-06,
|
| 412 |
-
"loss": 0.0416,
|
| 413 |
-
"step": 1650
|
| 414 |
-
},
|
| 415 |
-
{
|
| 416 |
-
"epoch": 2.17,
|
| 417 |
-
"learning_rate": 6.648571428571429e-06,
|
| 418 |
-
"loss": 0.0434,
|
| 419 |
-
"step": 1675
|
| 420 |
-
},
|
| 421 |
-
{
|
| 422 |
-
"epoch": 2.2,
|
| 423 |
-
"learning_rate": 6.577142857142857e-06,
|
| 424 |
-
"loss": 0.0457,
|
| 425 |
-
"step": 1700
|
| 426 |
-
},
|
| 427 |
-
{
|
| 428 |
-
"epoch": 2.23,
|
| 429 |
-
"learning_rate": 6.505714285714286e-06,
|
| 430 |
-
"loss": 0.0376,
|
| 431 |
-
"step": 1725
|
| 432 |
-
},
|
| 433 |
-
{
|
| 434 |
-
"epoch": 2.26,
|
| 435 |
-
"learning_rate": 6.434285714285715e-06,
|
| 436 |
-
"loss": 0.039,
|
| 437 |
-
"step": 1750
|
| 438 |
-
},
|
| 439 |
-
{
|
| 440 |
-
"epoch": 2.3,
|
| 441 |
-
"learning_rate": 6.3628571428571426e-06,
|
| 442 |
-
"loss": 0.0405,
|
| 443 |
-
"step": 1775
|
| 444 |
-
},
|
| 445 |
-
{
|
| 446 |
-
"epoch": 2.33,
|
| 447 |
-
"learning_rate": 6.2914285714285716e-06,
|
| 448 |
-
"loss": 0.0352,
|
| 449 |
-
"step": 1800
|
| 450 |
-
},
|
| 451 |
-
{
|
| 452 |
-
"epoch": 2.36,
|
| 453 |
-
"learning_rate": 6.220000000000001e-06,
|
| 454 |
-
"loss": 0.0397,
|
| 455 |
-
"step": 1825
|
| 456 |
-
},
|
| 457 |
-
{
|
| 458 |
-
"epoch": 2.39,
|
| 459 |
-
"learning_rate": 6.14857142857143e-06,
|
| 460 |
-
"loss": 0.0396,
|
| 461 |
-
"step": 1850
|
| 462 |
-
},
|
| 463 |
-
{
|
| 464 |
-
"epoch": 2.43,
|
| 465 |
-
"learning_rate": 6.077142857142858e-06,
|
| 466 |
-
"loss": 0.0388,
|
| 467 |
-
"step": 1875
|
| 468 |
-
},
|
| 469 |
-
{
|
| 470 |
-
"epoch": 2.46,
|
| 471 |
-
"learning_rate": 6.005714285714286e-06,
|
| 472 |
-
"loss": 0.038,
|
| 473 |
-
"step": 1900
|
| 474 |
-
},
|
| 475 |
-
{
|
| 476 |
-
"epoch": 2.49,
|
| 477 |
-
"learning_rate": 5.934285714285715e-06,
|
| 478 |
-
"loss": 0.0391,
|
| 479 |
-
"step": 1925
|
| 480 |
-
},
|
| 481 |
-
{
|
| 482 |
-
"epoch": 2.52,
|
| 483 |
-
"learning_rate": 5.862857142857143e-06,
|
| 484 |
-
"loss": 0.0358,
|
| 485 |
-
"step": 1950
|
| 486 |
-
},
|
| 487 |
-
{
|
| 488 |
-
"epoch": 2.55,
|
| 489 |
-
"learning_rate": 5.791428571428572e-06,
|
| 490 |
-
"loss": 0.0399,
|
| 491 |
-
"step": 1975
|
| 492 |
-
},
|
| 493 |
-
{
|
| 494 |
-
"epoch": 2.59,
|
| 495 |
-
"learning_rate": 5.72e-06,
|
| 496 |
-
"loss": 0.0387,
|
| 497 |
-
"step": 2000
|
| 498 |
-
},
|
| 499 |
-
{
|
| 500 |
-
"epoch": 2.59,
|
| 501 |
-
"eval_loss": 0.29592692852020264,
|
| 502 |
-
"eval_runtime": 2465.3193,
|
| 503 |
-
"eval_samples_per_second": 1.217,
|
| 504 |
-
"eval_steps_per_second": 0.152,
|
| 505 |
-
"eval_wer": 19.680970568411592,
|
| 506 |
-
"step": 2000
|
| 507 |
-
}
|
| 508 |
-
],
|
| 509 |
-
"max_steps": 4000,
|
| 510 |
-
"num_train_epochs": 6,
|
| 511 |
-
"total_flos": 9.23011544383488e+18,
|
| 512 |
-
"trial_name": null,
|
| 513 |
-
"trial_params": null
|
| 514 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
checkpoint-2000/training_args.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:e622acaf222233592b6831832007ec4d3cdb12856c8cb7d9c6a33ac3ba0638d9
|
| 3 |
-
size 3579
|
|
|
|
|
|
|
|
|
|
|
|