End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.01,
|
| 3 |
+
"eval_loss": 0.25427135825157166,
|
| 4 |
+
"eval_runtime": 1952.0893,
|
| 5 |
+
"eval_samples_per_second": 2.358,
|
| 6 |
+
"eval_steps_per_second": 0.074,
|
| 7 |
+
"eval_wer": 13.768684731417652,
|
| 8 |
+
"train_loss": 0.18922350925207138,
|
| 9 |
+
"train_runtime": 17278.1753,
|
| 10 |
+
"train_samples_per_second": 3.704,
|
| 11 |
+
"train_steps_per_second": 0.058
|
| 12 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.01,
|
| 3 |
+
"eval_loss": 0.25427135825157166,
|
| 4 |
+
"eval_runtime": 1952.0893,
|
| 5 |
+
"eval_samples_per_second": 2.358,
|
| 6 |
+
"eval_steps_per_second": 0.074,
|
| 7 |
+
"eval_wer": 13.768684731417652
|
| 8 |
+
}
|
runs/Dec12_21-54-54_129-213-131-105/events.out.tfevents.1670901386.129-213-131-105.1405782.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b285ff5666eaf1ed4c3b9a29ef0299e01d50439db8306d7f0751fee6d0601a3
|
| 3 |
+
size 358
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 7.01,
|
| 3 |
+
"train_loss": 0.18922350925207138,
|
| 4 |
+
"train_runtime": 17278.1753,
|
| 5 |
+
"train_samples_per_second": 3.704,
|
| 6 |
+
"train_steps_per_second": 0.058
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,310 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 13.768684731417652,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-600",
|
| 4 |
+
"epoch": 7.013,
|
| 5 |
+
"global_step": 1000,
|
| 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": 4.6000000000000004e-07,
|
| 13 |
+
"loss": 0.8523,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.05,
|
| 18 |
+
"learning_rate": 9.600000000000001e-07,
|
| 19 |
+
"loss": 0.5983,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.07,
|
| 24 |
+
"learning_rate": 1.46e-06,
|
| 25 |
+
"loss": 0.4674,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.1,
|
| 30 |
+
"learning_rate": 1.9600000000000003e-06,
|
| 31 |
+
"loss": 0.5018,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.12,
|
| 36 |
+
"learning_rate": 2.46e-06,
|
| 37 |
+
"loss": 0.4059,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 1.01,
|
| 42 |
+
"learning_rate": 2.96e-06,
|
| 43 |
+
"loss": 0.3495,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 1.03,
|
| 48 |
+
"learning_rate": 3.46e-06,
|
| 49 |
+
"loss": 0.2695,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 1.06,
|
| 54 |
+
"learning_rate": 3.96e-06,
|
| 55 |
+
"loss": 0.2515,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.06,
|
| 60 |
+
"eval_loss": 0.28808125853538513,
|
| 61 |
+
"eval_runtime": 1946.1012,
|
| 62 |
+
"eval_samples_per_second": 2.366,
|
| 63 |
+
"eval_steps_per_second": 0.074,
|
| 64 |
+
"eval_wer": 16.944235888335268,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 1.08,
|
| 69 |
+
"learning_rate": 4.4600000000000005e-06,
|
| 70 |
+
"loss": 0.3219,
|
| 71 |
+
"step": 225
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 1.11,
|
| 75 |
+
"learning_rate": 4.960000000000001e-06,
|
| 76 |
+
"loss": 0.3448,
|
| 77 |
+
"step": 250
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.13,
|
| 81 |
+
"learning_rate": 5.460000000000001e-06,
|
| 82 |
+
"loss": 0.2704,
|
| 83 |
+
"step": 275
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 2.02,
|
| 87 |
+
"learning_rate": 5.9600000000000005e-06,
|
| 88 |
+
"loss": 0.2282,
|
| 89 |
+
"step": 300
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 2.04,
|
| 93 |
+
"learning_rate": 6.460000000000001e-06,
|
| 94 |
+
"loss": 0.1848,
|
| 95 |
+
"step": 325
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 2.07,
|
| 99 |
+
"learning_rate": 6.96e-06,
|
| 100 |
+
"loss": 0.2006,
|
| 101 |
+
"step": 350
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 2.09,
|
| 105 |
+
"learning_rate": 7.4600000000000006e-06,
|
| 106 |
+
"loss": 0.2534,
|
| 107 |
+
"step": 375
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 2.12,
|
| 111 |
+
"learning_rate": 7.960000000000002e-06,
|
| 112 |
+
"loss": 0.2212,
|
| 113 |
+
"step": 400
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 2.12,
|
| 117 |
+
"eval_loss": 0.2615717649459839,
|
| 118 |
+
"eval_runtime": 1940.3687,
|
| 119 |
+
"eval_samples_per_second": 2.373,
|
| 120 |
+
"eval_steps_per_second": 0.074,
|
| 121 |
+
"eval_wer": 14.688417172889224,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 3.0,
|
| 126 |
+
"learning_rate": 8.46e-06,
|
| 127 |
+
"loss": 0.1822,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 3.03,
|
| 132 |
+
"learning_rate": 8.96e-06,
|
| 133 |
+
"loss": 0.1378,
|
| 134 |
+
"step": 450
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 3.05,
|
| 138 |
+
"learning_rate": 9.460000000000001e-06,
|
| 139 |
+
"loss": 0.1197,
|
| 140 |
+
"step": 475
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 3.08,
|
| 144 |
+
"learning_rate": 9.960000000000001e-06,
|
| 145 |
+
"loss": 0.1561,
|
| 146 |
+
"step": 500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 3.1,
|
| 150 |
+
"learning_rate": 9.54e-06,
|
| 151 |
+
"loss": 0.1683,
|
| 152 |
+
"step": 525
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 3.13,
|
| 156 |
+
"learning_rate": 9.040000000000002e-06,
|
| 157 |
+
"loss": 0.1323,
|
| 158 |
+
"step": 550
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 4.01,
|
| 162 |
+
"learning_rate": 8.540000000000001e-06,
|
| 163 |
+
"loss": 0.1061,
|
| 164 |
+
"step": 575
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 4.04,
|
| 168 |
+
"learning_rate": 8.040000000000001e-06,
|
| 169 |
+
"loss": 0.0774,
|
| 170 |
+
"step": 600
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 4.04,
|
| 174 |
+
"eval_loss": 0.25427135825157166,
|
| 175 |
+
"eval_runtime": 1983.6254,
|
| 176 |
+
"eval_samples_per_second": 2.321,
|
| 177 |
+
"eval_steps_per_second": 0.073,
|
| 178 |
+
"eval_wer": 13.768684731417652,
|
| 179 |
+
"step": 600
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 4.06,
|
| 183 |
+
"learning_rate": 7.540000000000001e-06,
|
| 184 |
+
"loss": 0.0755,
|
| 185 |
+
"step": 625
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 4.09,
|
| 189 |
+
"learning_rate": 7.04e-06,
|
| 190 |
+
"loss": 0.0925,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 4.11,
|
| 195 |
+
"learning_rate": 6.540000000000001e-06,
|
| 196 |
+
"loss": 0.0908,
|
| 197 |
+
"step": 675
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 4.14,
|
| 201 |
+
"learning_rate": 6.040000000000001e-06,
|
| 202 |
+
"loss": 0.0684,
|
| 203 |
+
"step": 700
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 5.02,
|
| 207 |
+
"learning_rate": 5.540000000000001e-06,
|
| 208 |
+
"loss": 0.0529,
|
| 209 |
+
"step": 725
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 5.04,
|
| 213 |
+
"learning_rate": 5.04e-06,
|
| 214 |
+
"loss": 0.0401,
|
| 215 |
+
"step": 750
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 5.07,
|
| 219 |
+
"learning_rate": 4.540000000000001e-06,
|
| 220 |
+
"loss": 0.0469,
|
| 221 |
+
"step": 775
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 5.09,
|
| 225 |
+
"learning_rate": 4.04e-06,
|
| 226 |
+
"loss": 0.0564,
|
| 227 |
+
"step": 800
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 5.09,
|
| 231 |
+
"eval_loss": 0.27310237288475037,
|
| 232 |
+
"eval_runtime": 1976.319,
|
| 233 |
+
"eval_samples_per_second": 2.33,
|
| 234 |
+
"eval_steps_per_second": 0.073,
|
| 235 |
+
"eval_wer": 13.9768616476691,
|
| 236 |
+
"step": 800
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 5.12,
|
| 240 |
+
"learning_rate": 3.54e-06,
|
| 241 |
+
"loss": 0.0452,
|
| 242 |
+
"step": 825
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 6.0,
|
| 246 |
+
"learning_rate": 3.04e-06,
|
| 247 |
+
"loss": 0.0372,
|
| 248 |
+
"step": 850
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 6.03,
|
| 252 |
+
"learning_rate": 2.5400000000000002e-06,
|
| 253 |
+
"loss": 0.026,
|
| 254 |
+
"step": 875
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 6.05,
|
| 258 |
+
"learning_rate": 2.04e-06,
|
| 259 |
+
"loss": 0.0216,
|
| 260 |
+
"step": 900
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 6.08,
|
| 264 |
+
"learning_rate": 1.54e-06,
|
| 265 |
+
"loss": 0.031,
|
| 266 |
+
"step": 925
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 6.1,
|
| 270 |
+
"learning_rate": 1.04e-06,
|
| 271 |
+
"loss": 0.0347,
|
| 272 |
+
"step": 950
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 6.13,
|
| 276 |
+
"learning_rate": 5.4e-07,
|
| 277 |
+
"loss": 0.0264,
|
| 278 |
+
"step": 975
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 7.01,
|
| 282 |
+
"learning_rate": 4e-08,
|
| 283 |
+
"loss": 0.0221,
|
| 284 |
+
"step": 1000
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 7.01,
|
| 288 |
+
"eval_loss": 0.28137078881263733,
|
| 289 |
+
"eval_runtime": 2000.1135,
|
| 290 |
+
"eval_samples_per_second": 2.302,
|
| 291 |
+
"eval_steps_per_second": 0.072,
|
| 292 |
+
"eval_wer": 13.970036175005118,
|
| 293 |
+
"step": 1000
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 7.01,
|
| 297 |
+
"step": 1000,
|
| 298 |
+
"total_flos": 1.836038033915904e+19,
|
| 299 |
+
"train_loss": 0.18922350925207138,
|
| 300 |
+
"train_runtime": 17278.1753,
|
| 301 |
+
"train_samples_per_second": 3.704,
|
| 302 |
+
"train_steps_per_second": 0.058
|
| 303 |
+
}
|
| 304 |
+
],
|
| 305 |
+
"max_steps": 1000,
|
| 306 |
+
"num_train_epochs": 9223372036854775807,
|
| 307 |
+
"total_flos": 1.836038033915904e+19,
|
| 308 |
+
"trial_name": null,
|
| 309 |
+
"trial_params": null
|
| 310 |
+
}
|