End of training
Browse files
all_results.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.09,
|
| 3 |
+
"eval_cer": 5.2581777082703605,
|
| 4 |
+
"eval_loss": 0.2789727747440338,
|
| 5 |
+
"eval_runtime": 4205.1455,
|
| 6 |
+
"eval_samples_per_second": 1.095,
|
| 7 |
+
"eval_steps_per_second": 0.068,
|
| 8 |
+
"eval_wer": 8.39864864864865,
|
| 9 |
+
"train_loss": 0.04369454556107521,
|
| 10 |
+
"train_runtime": 84974.4249,
|
| 11 |
+
"train_samples_per_second": 3.766,
|
| 12 |
+
"train_steps_per_second": 0.118
|
| 13 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.09,
|
| 3 |
+
"eval_cer": 5.2581777082703605,
|
| 4 |
+
"eval_loss": 0.2789727747440338,
|
| 5 |
+
"eval_runtime": 4205.1455,
|
| 6 |
+
"eval_samples_per_second": 1.095,
|
| 7 |
+
"eval_steps_per_second": 0.068,
|
| 8 |
+
"eval_wer": 8.39864864864865
|
| 9 |
+
}
|
runs/Dec22_02-42-31_129-146-162-127/events.out.tfevents.1671769773.129-146-162-127.4079130.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38708817151242b78a77bfaca7415d2f83c95569983960219456d92b45d81a9d
|
| 3 |
+
size 405
|
train_results.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.09,
|
| 3 |
+
"train_loss": 0.04369454556107521,
|
| 4 |
+
"train_runtime": 84974.4249,
|
| 5 |
+
"train_samples_per_second": 3.766,
|
| 6 |
+
"train_steps_per_second": 0.118
|
| 7 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,245 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 8.39864864864865,
|
| 3 |
+
"best_model_checkpoint": "./checkpoint-10000",
|
| 4 |
+
"epoch": 10.090817356205852,
|
| 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": 0.5,
|
| 12 |
+
"learning_rate": 9.940000000000001e-06,
|
| 13 |
+
"loss": 0.3462,
|
| 14 |
+
"step": 500
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 1.01,
|
| 18 |
+
"learning_rate": 9.476842105263158e-06,
|
| 19 |
+
"loss": 0.1691,
|
| 20 |
+
"step": 1000
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 1.01,
|
| 24 |
+
"eval_cer": 6.350948206752441,
|
| 25 |
+
"eval_loss": 0.1870781034231186,
|
| 26 |
+
"eval_runtime": 4246.4995,
|
| 27 |
+
"eval_samples_per_second": 1.084,
|
| 28 |
+
"eval_steps_per_second": 0.068,
|
| 29 |
+
"eval_wer": 10.173986486486486,
|
| 30 |
+
"step": 1000
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 1.51,
|
| 34 |
+
"learning_rate": 8.950526315789474e-06,
|
| 35 |
+
"loss": 0.097,
|
| 36 |
+
"step": 1500
|
| 37 |
+
},
|
| 38 |
+
{
|
| 39 |
+
"epoch": 2.02,
|
| 40 |
+
"learning_rate": 8.42421052631579e-06,
|
| 41 |
+
"loss": 0.0916,
|
| 42 |
+
"step": 2000
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 2.02,
|
| 46 |
+
"eval_cer": 5.5499336320984085,
|
| 47 |
+
"eval_loss": 0.16914600133895874,
|
| 48 |
+
"eval_runtime": 4299.9346,
|
| 49 |
+
"eval_samples_per_second": 1.071,
|
| 50 |
+
"eval_steps_per_second": 0.067,
|
| 51 |
+
"eval_wer": 8.97972972972973,
|
| 52 |
+
"step": 2000
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 2.52,
|
| 56 |
+
"learning_rate": 7.897894736842106e-06,
|
| 57 |
+
"loss": 0.0453,
|
| 58 |
+
"step": 2500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 3.03,
|
| 62 |
+
"learning_rate": 7.371578947368422e-06,
|
| 63 |
+
"loss": 0.0452,
|
| 64 |
+
"step": 3000
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"epoch": 3.03,
|
| 68 |
+
"eval_cer": 5.58673167654519,
|
| 69 |
+
"eval_loss": 0.19017226994037628,
|
| 70 |
+
"eval_runtime": 4203.2916,
|
| 71 |
+
"eval_samples_per_second": 1.095,
|
| 72 |
+
"eval_steps_per_second": 0.069,
|
| 73 |
+
"eval_wer": 8.98141891891892,
|
| 74 |
+
"step": 3000
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 3.53,
|
| 78 |
+
"learning_rate": 6.845263157894737e-06,
|
| 79 |
+
"loss": 0.0214,
|
| 80 |
+
"step": 3500
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 4.04,
|
| 84 |
+
"learning_rate": 6.318947368421054e-06,
|
| 85 |
+
"loss": 0.0213,
|
| 86 |
+
"step": 4000
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 4.04,
|
| 90 |
+
"eval_cer": 5.653099578136705,
|
| 91 |
+
"eval_loss": 0.20616890490055084,
|
| 92 |
+
"eval_runtime": 4260.0725,
|
| 93 |
+
"eval_samples_per_second": 1.081,
|
| 94 |
+
"eval_steps_per_second": 0.068,
|
| 95 |
+
"eval_wer": 8.9375,
|
| 96 |
+
"step": 4000
|
| 97 |
+
},
|
| 98 |
+
{
|
| 99 |
+
"epoch": 4.54,
|
| 100 |
+
"learning_rate": 5.7926315789473685e-06,
|
| 101 |
+
"loss": 0.0096,
|
| 102 |
+
"step": 4500
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 5.05,
|
| 106 |
+
"learning_rate": 5.266315789473685e-06,
|
| 107 |
+
"loss": 0.0096,
|
| 108 |
+
"step": 5000
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 5.05,
|
| 112 |
+
"eval_cer": 5.620244181309221,
|
| 113 |
+
"eval_loss": 0.22839371860027313,
|
| 114 |
+
"eval_runtime": 4320.8285,
|
| 115 |
+
"eval_samples_per_second": 1.066,
|
| 116 |
+
"eval_steps_per_second": 0.067,
|
| 117 |
+
"eval_wer": 8.733108108108107,
|
| 118 |
+
"step": 5000
|
| 119 |
+
},
|
| 120 |
+
{
|
| 121 |
+
"epoch": 5.55,
|
| 122 |
+
"learning_rate": 4.74e-06,
|
| 123 |
+
"loss": 0.0045,
|
| 124 |
+
"step": 5500
|
| 125 |
+
},
|
| 126 |
+
{
|
| 127 |
+
"epoch": 6.05,
|
| 128 |
+
"learning_rate": 4.213684210526316e-06,
|
| 129 |
+
"loss": 0.0041,
|
| 130 |
+
"step": 6000
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 6.05,
|
| 134 |
+
"eval_cer": 5.300889724146089,
|
| 135 |
+
"eval_loss": 0.2394615262746811,
|
| 136 |
+
"eval_runtime": 4207.8926,
|
| 137 |
+
"eval_samples_per_second": 1.094,
|
| 138 |
+
"eval_steps_per_second": 0.068,
|
| 139 |
+
"eval_wer": 8.505067567567567,
|
| 140 |
+
"step": 6000
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 6.56,
|
| 144 |
+
"learning_rate": 3.687368421052632e-06,
|
| 145 |
+
"loss": 0.0024,
|
| 146 |
+
"step": 6500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 7.06,
|
| 150 |
+
"learning_rate": 3.1610526315789474e-06,
|
| 151 |
+
"loss": 0.0022,
|
| 152 |
+
"step": 7000
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 7.06,
|
| 156 |
+
"eval_cer": 5.363972086054855,
|
| 157 |
+
"eval_loss": 0.2534642517566681,
|
| 158 |
+
"eval_runtime": 4214.0551,
|
| 159 |
+
"eval_samples_per_second": 1.093,
|
| 160 |
+
"eval_steps_per_second": 0.068,
|
| 161 |
+
"eval_wer": 8.550675675675675,
|
| 162 |
+
"step": 7000
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 7.57,
|
| 166 |
+
"learning_rate": 2.6347368421052633e-06,
|
| 167 |
+
"loss": 0.0013,
|
| 168 |
+
"step": 7500
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 8.07,
|
| 172 |
+
"learning_rate": 2.1084210526315792e-06,
|
| 173 |
+
"loss": 0.001,
|
| 174 |
+
"step": 8000
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 8.07,
|
| 178 |
+
"eval_cer": 5.379085568595497,
|
| 179 |
+
"eval_loss": 0.2655620574951172,
|
| 180 |
+
"eval_runtime": 4212.0826,
|
| 181 |
+
"eval_samples_per_second": 1.093,
|
| 182 |
+
"eval_steps_per_second": 0.068,
|
| 183 |
+
"eval_wer": 8.555743243243242,
|
| 184 |
+
"step": 8000
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 8.58,
|
| 188 |
+
"learning_rate": 1.5821052631578948e-06,
|
| 189 |
+
"loss": 0.0007,
|
| 190 |
+
"step": 8500
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 9.08,
|
| 194 |
+
"learning_rate": 1.0557894736842105e-06,
|
| 195 |
+
"loss": 0.0006,
|
| 196 |
+
"step": 9000
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 9.08,
|
| 200 |
+
"eval_cer": 5.273948298747552,
|
| 201 |
+
"eval_loss": 0.2720721364021301,
|
| 202 |
+
"eval_runtime": 4204.0827,
|
| 203 |
+
"eval_samples_per_second": 1.095,
|
| 204 |
+
"eval_steps_per_second": 0.069,
|
| 205 |
+
"eval_wer": 8.403716216216216,
|
| 206 |
+
"step": 9000
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 9.59,
|
| 210 |
+
"learning_rate": 5.294736842105263e-07,
|
| 211 |
+
"loss": 0.0005,
|
| 212 |
+
"step": 9500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 10.09,
|
| 216 |
+
"learning_rate": 3.1578947368421054e-09,
|
| 217 |
+
"loss": 0.0004,
|
| 218 |
+
"step": 10000
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 10.09,
|
| 222 |
+
"eval_cer": 5.2581777082703605,
|
| 223 |
+
"eval_loss": 0.2789727747440338,
|
| 224 |
+
"eval_runtime": 4206.3116,
|
| 225 |
+
"eval_samples_per_second": 1.095,
|
| 226 |
+
"eval_steps_per_second": 0.068,
|
| 227 |
+
"eval_wer": 8.39864864864865,
|
| 228 |
+
"step": 10000
|
| 229 |
+
},
|
| 230 |
+
{
|
| 231 |
+
"epoch": 10.09,
|
| 232 |
+
"step": 10000,
|
| 233 |
+
"total_flos": 3.265119220137984e+20,
|
| 234 |
+
"train_loss": 0.04369454556107521,
|
| 235 |
+
"train_runtime": 84974.4249,
|
| 236 |
+
"train_samples_per_second": 3.766,
|
| 237 |
+
"train_steps_per_second": 0.118
|
| 238 |
+
}
|
| 239 |
+
],
|
| 240 |
+
"max_steps": 10000,
|
| 241 |
+
"num_train_epochs": 11,
|
| 242 |
+
"total_flos": 3.265119220137984e+20,
|
| 243 |
+
"trial_name": null,
|
| 244 |
+
"trial_params": null
|
| 245 |
+
}
|