End of training
Browse files- all_results.json +7 -0
- train_results.json +7 -0
- trainer_state.json +355 -0
all_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.18131806907653808,
|
| 4 |
+
"train_runtime": 89475.4174,
|
| 5 |
+
"train_samples_per_second": 3.576,
|
| 6 |
+
"train_steps_per_second": 0.112
|
| 7 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1.0,
|
| 3 |
+
"train_loss": 0.18131806907653808,
|
| 4 |
+
"train_runtime": 89475.4174,
|
| 5 |
+
"train_samples_per_second": 3.576,
|
| 6 |
+
"train_steps_per_second": 0.112
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,355 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 10.718358038768528,
|
| 3 |
+
"best_model_checkpoint": "../whisper-NST-cons2e5/checkpoint-6000",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"global_step": 10000,
|
| 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": 9.88e-06,
|
| 13 |
+
"loss": 1.2287,
|
| 14 |
+
"step": 250
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 1.9880000000000003e-05,
|
| 19 |
+
"loss": 0.2979,
|
| 20 |
+
"step": 500
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 2e-05,
|
| 25 |
+
"loss": 0.2728,
|
| 26 |
+
"step": 750
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 2e-05,
|
| 31 |
+
"loss": 0.2517,
|
| 32 |
+
"step": 1000
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.1,
|
| 36 |
+
"eval_loss": 0.4131389260292053,
|
| 37 |
+
"eval_runtime": 51.3583,
|
| 38 |
+
"eval_samples_per_second": 1.947,
|
| 39 |
+
"eval_steps_per_second": 0.136,
|
| 40 |
+
"eval_wer": 18.47206385404789,
|
| 41 |
+
"step": 1000
|
| 42 |
+
},
|
| 43 |
+
{
|
| 44 |
+
"epoch": 0.12,
|
| 45 |
+
"learning_rate": 2e-05,
|
| 46 |
+
"loss": 0.2254,
|
| 47 |
+
"step": 1250
|
| 48 |
+
},
|
| 49 |
+
{
|
| 50 |
+
"epoch": 0.15,
|
| 51 |
+
"learning_rate": 2e-05,
|
| 52 |
+
"loss": 0.2115,
|
| 53 |
+
"step": 1500
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"epoch": 0.17,
|
| 57 |
+
"learning_rate": 2e-05,
|
| 58 |
+
"loss": 0.2018,
|
| 59 |
+
"step": 1750
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.2,
|
| 63 |
+
"learning_rate": 2e-05,
|
| 64 |
+
"loss": 0.1931,
|
| 65 |
+
"step": 2000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.2,
|
| 69 |
+
"eval_loss": 0.3531236946582794,
|
| 70 |
+
"eval_runtime": 74.7257,
|
| 71 |
+
"eval_samples_per_second": 1.338,
|
| 72 |
+
"eval_steps_per_second": 0.094,
|
| 73 |
+
"eval_wer": 19.04218928164196,
|
| 74 |
+
"step": 2000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.23,
|
| 78 |
+
"learning_rate": 2e-05,
|
| 79 |
+
"loss": 0.1786,
|
| 80 |
+
"step": 2250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.25,
|
| 84 |
+
"learning_rate": 2e-05,
|
| 85 |
+
"loss": 0.1746,
|
| 86 |
+
"step": 2500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.28,
|
| 90 |
+
"learning_rate": 2e-05,
|
| 91 |
+
"loss": 0.1726,
|
| 92 |
+
"step": 2750
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.3,
|
| 96 |
+
"learning_rate": 2e-05,
|
| 97 |
+
"loss": 0.1598,
|
| 98 |
+
"step": 3000
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.3,
|
| 102 |
+
"eval_loss": 0.3605496883392334,
|
| 103 |
+
"eval_runtime": 59.9526,
|
| 104 |
+
"eval_samples_per_second": 1.668,
|
| 105 |
+
"eval_steps_per_second": 0.117,
|
| 106 |
+
"eval_wer": 16.875712656784494,
|
| 107 |
+
"step": 3000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.33,
|
| 111 |
+
"learning_rate": 2e-05,
|
| 112 |
+
"loss": 0.1636,
|
| 113 |
+
"step": 3250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 0.35,
|
| 117 |
+
"learning_rate": 2e-05,
|
| 118 |
+
"loss": 0.158,
|
| 119 |
+
"step": 3500
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 0.38,
|
| 123 |
+
"learning_rate": 2e-05,
|
| 124 |
+
"loss": 0.1494,
|
| 125 |
+
"step": 3750
|
| 126 |
+
},
|
| 127 |
+
{
|
| 128 |
+
"epoch": 0.4,
|
| 129 |
+
"learning_rate": 2e-05,
|
| 130 |
+
"loss": 0.1541,
|
| 131 |
+
"step": 4000
|
| 132 |
+
},
|
| 133 |
+
{
|
| 134 |
+
"epoch": 0.4,
|
| 135 |
+
"eval_loss": 0.33670979738235474,
|
| 136 |
+
"eval_runtime": 50.4438,
|
| 137 |
+
"eval_samples_per_second": 1.982,
|
| 138 |
+
"eval_steps_per_second": 0.139,
|
| 139 |
+
"eval_wer": 14.481185860889395,
|
| 140 |
+
"step": 4000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.42,
|
| 144 |
+
"learning_rate": 2e-05,
|
| 145 |
+
"loss": 0.1498,
|
| 146 |
+
"step": 4250
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.45,
|
| 150 |
+
"learning_rate": 2e-05,
|
| 151 |
+
"loss": 0.1486,
|
| 152 |
+
"step": 4500
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.47,
|
| 156 |
+
"learning_rate": 2e-05,
|
| 157 |
+
"loss": 0.1471,
|
| 158 |
+
"step": 4750
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.5,
|
| 162 |
+
"learning_rate": 2e-05,
|
| 163 |
+
"loss": 0.1443,
|
| 164 |
+
"step": 5000
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.5,
|
| 168 |
+
"eval_loss": 0.3274496793746948,
|
| 169 |
+
"eval_runtime": 55.4869,
|
| 170 |
+
"eval_samples_per_second": 1.802,
|
| 171 |
+
"eval_steps_per_second": 0.126,
|
| 172 |
+
"eval_wer": 13.340935005701255,
|
| 173 |
+
"step": 5000
|
| 174 |
+
},
|
| 175 |
+
{
|
| 176 |
+
"epoch": 0.53,
|
| 177 |
+
"learning_rate": 2e-05,
|
| 178 |
+
"loss": 0.1337,
|
| 179 |
+
"step": 5250
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 0.55,
|
| 183 |
+
"learning_rate": 2e-05,
|
| 184 |
+
"loss": 0.1358,
|
| 185 |
+
"step": 5500
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.57,
|
| 189 |
+
"learning_rate": 2e-05,
|
| 190 |
+
"loss": 0.1232,
|
| 191 |
+
"step": 5750
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.6,
|
| 195 |
+
"learning_rate": 2e-05,
|
| 196 |
+
"loss": 0.1301,
|
| 197 |
+
"step": 6000
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 0.6,
|
| 201 |
+
"eval_loss": 0.34810879826545715,
|
| 202 |
+
"eval_runtime": 52.0858,
|
| 203 |
+
"eval_samples_per_second": 1.92,
|
| 204 |
+
"eval_steps_per_second": 0.134,
|
| 205 |
+
"eval_wer": 10.718358038768528,
|
| 206 |
+
"step": 6000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.62,
|
| 210 |
+
"learning_rate": 2e-05,
|
| 211 |
+
"loss": 0.1308,
|
| 212 |
+
"step": 6250
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.65,
|
| 216 |
+
"learning_rate": 2e-05,
|
| 217 |
+
"loss": 0.121,
|
| 218 |
+
"step": 6500
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.68,
|
| 222 |
+
"learning_rate": 2e-05,
|
| 223 |
+
"loss": 0.1343,
|
| 224 |
+
"step": 6750
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.7,
|
| 228 |
+
"learning_rate": 2e-05,
|
| 229 |
+
"loss": 0.1266,
|
| 230 |
+
"step": 7000
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.7,
|
| 234 |
+
"eval_loss": 0.3451653718948364,
|
| 235 |
+
"eval_runtime": 52.2228,
|
| 236 |
+
"eval_samples_per_second": 1.915,
|
| 237 |
+
"eval_steps_per_second": 0.134,
|
| 238 |
+
"eval_wer": 12.998859749144811,
|
| 239 |
+
"step": 7000
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 0.72,
|
| 243 |
+
"learning_rate": 2e-05,
|
| 244 |
+
"loss": 0.1223,
|
| 245 |
+
"step": 7250
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 0.75,
|
| 249 |
+
"learning_rate": 2e-05,
|
| 250 |
+
"loss": 0.1264,
|
| 251 |
+
"step": 7500
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"epoch": 0.78,
|
| 255 |
+
"learning_rate": 2e-05,
|
| 256 |
+
"loss": 0.119,
|
| 257 |
+
"step": 7750
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 0.8,
|
| 261 |
+
"learning_rate": 2e-05,
|
| 262 |
+
"loss": 0.1216,
|
| 263 |
+
"step": 8000
|
| 264 |
+
},
|
| 265 |
+
{
|
| 266 |
+
"epoch": 0.8,
|
| 267 |
+
"eval_loss": 0.32154810428619385,
|
| 268 |
+
"eval_runtime": 82.8333,
|
| 269 |
+
"eval_samples_per_second": 1.207,
|
| 270 |
+
"eval_steps_per_second": 0.085,
|
| 271 |
+
"eval_wer": 10.832383124287343,
|
| 272 |
+
"step": 8000
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.82,
|
| 276 |
+
"learning_rate": 2e-05,
|
| 277 |
+
"loss": 0.1242,
|
| 278 |
+
"step": 8250
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.85,
|
| 282 |
+
"learning_rate": 2e-05,
|
| 283 |
+
"loss": 0.1217,
|
| 284 |
+
"step": 8500
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.88,
|
| 288 |
+
"learning_rate": 2e-05,
|
| 289 |
+
"loss": 0.1245,
|
| 290 |
+
"step": 8750
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.9,
|
| 294 |
+
"learning_rate": 2e-05,
|
| 295 |
+
"loss": 0.1121,
|
| 296 |
+
"step": 9000
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.9,
|
| 300 |
+
"eval_loss": 0.31604325771331787,
|
| 301 |
+
"eval_runtime": 50.7948,
|
| 302 |
+
"eval_samples_per_second": 1.969,
|
| 303 |
+
"eval_steps_per_second": 0.138,
|
| 304 |
+
"eval_wer": 11.516533637400228,
|
| 305 |
+
"step": 9000
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 0.93,
|
| 309 |
+
"learning_rate": 2e-05,
|
| 310 |
+
"loss": 0.1161,
|
| 311 |
+
"step": 9250
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.95,
|
| 315 |
+
"learning_rate": 2e-05,
|
| 316 |
+
"loss": 0.112,
|
| 317 |
+
"step": 9500
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.97,
|
| 321 |
+
"learning_rate": 2e-05,
|
| 322 |
+
"loss": 0.1167,
|
| 323 |
+
"step": 9750
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 1.0,
|
| 327 |
+
"learning_rate": 2e-05,
|
| 328 |
+
"loss": 0.1171,
|
| 329 |
+
"step": 10000
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 1.0,
|
| 333 |
+
"eval_loss": 0.3520647883415222,
|
| 334 |
+
"eval_runtime": 84.9296,
|
| 335 |
+
"eval_samples_per_second": 1.177,
|
| 336 |
+
"eval_steps_per_second": 0.082,
|
| 337 |
+
"eval_wer": 11.85860889395667,
|
| 338 |
+
"step": 10000
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.0,
|
| 342 |
+
"step": 10000,
|
| 343 |
+
"total_flos": 4.61736640512e+19,
|
| 344 |
+
"train_loss": 0.18131806907653808,
|
| 345 |
+
"train_runtime": 89475.4174,
|
| 346 |
+
"train_samples_per_second": 3.576,
|
| 347 |
+
"train_steps_per_second": 0.112
|
| 348 |
+
}
|
| 349 |
+
],
|
| 350 |
+
"max_steps": 10000,
|
| 351 |
+
"num_train_epochs": 9223372036854775807,
|
| 352 |
+
"total_flos": 4.61736640512e+19,
|
| 353 |
+
"trial_name": null,
|
| 354 |
+
"trial_params": null
|
| 355 |
+
}
|