Upload 9 files
Browse files- pytorch_model.bin +1 -1
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- trainer_state.json +244 -0
- training_args.bin +1 -1
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 967102601
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6b96916efb04821b8f34e0eb2cb9b774612ddee6a64a2456a60010bf3c2267ff
|
| 3 |
size 967102601
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc7429d9d6abaf0f9a7f0884ac5f19724496065526eee29c4e31250843167a65
|
| 3 |
+
size 14511
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3d15232b452895e36dc91e0446f6e8ef9af073f8677f732bb005840a10ca7266
|
| 3 |
+
size 557
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2478370688e920d6f6653d668ddacc94f12fd60a8da1b22d5ff9a3df70437e30
|
| 3 |
+
size 627
|
trainer_state.json
ADDED
|
@@ -0,0 +1,244 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 12.92352371732817,
|
| 3 |
+
"best_model_checkpoint": "./whisper-small-ja/checkpoint-800",
|
| 4 |
+
"epoch": 2.9962546816479403,
|
| 5 |
+
"global_step": 800,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.09,
|
| 12 |
+
"learning_rate": 2.1000000000000002e-06,
|
| 13 |
+
"loss": 0.7741,
|
| 14 |
+
"step": 25
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.19,
|
| 18 |
+
"learning_rate": 4.600000000000001e-06,
|
| 19 |
+
"loss": 0.6088,
|
| 20 |
+
"step": 50
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.28,
|
| 24 |
+
"learning_rate": 7.100000000000001e-06,
|
| 25 |
+
"loss": 0.642,
|
| 26 |
+
"step": 75
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.37,
|
| 30 |
+
"learning_rate": 9.600000000000001e-06,
|
| 31 |
+
"loss": 0.4286,
|
| 32 |
+
"step": 100
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.47,
|
| 36 |
+
"learning_rate": 9.7e-06,
|
| 37 |
+
"loss": 0.2252,
|
| 38 |
+
"step": 125
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.56,
|
| 42 |
+
"learning_rate": 9.342857142857144e-06,
|
| 43 |
+
"loss": 0.1987,
|
| 44 |
+
"step": 150
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.66,
|
| 48 |
+
"learning_rate": 8.985714285714287e-06,
|
| 49 |
+
"loss": 0.1903,
|
| 50 |
+
"step": 175
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.75,
|
| 54 |
+
"learning_rate": 8.628571428571429e-06,
|
| 55 |
+
"loss": 0.1446,
|
| 56 |
+
"step": 200
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.75,
|
| 60 |
+
"eval_loss": 0.24177707731723785,
|
| 61 |
+
"eval_runtime": 349.354,
|
| 62 |
+
"eval_samples_per_second": 4.809,
|
| 63 |
+
"eval_steps_per_second": 0.601,
|
| 64 |
+
"eval_wer": 14.714424007744434,
|
| 65 |
+
"step": 200
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.84,
|
| 69 |
+
"learning_rate": 8.271428571428572e-06,
|
| 70 |
+
"loss": 0.135,
|
| 71 |
+
"step": 225
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"epoch": 0.94,
|
| 75 |
+
"learning_rate": 7.914285714285715e-06,
|
| 76 |
+
"loss": 0.1365,
|
| 77 |
+
"step": 250
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.03,
|
| 81 |
+
"learning_rate": 7.557142857142857e-06,
|
| 82 |
+
"loss": 0.1029,
|
| 83 |
+
"step": 275
|
| 84 |
+
},
|
| 85 |
+
{
|
| 86 |
+
"epoch": 1.12,
|
| 87 |
+
"learning_rate": 7.2000000000000005e-06,
|
| 88 |
+
"loss": 0.0364,
|
| 89 |
+
"step": 300
|
| 90 |
+
},
|
| 91 |
+
{
|
| 92 |
+
"epoch": 1.22,
|
| 93 |
+
"learning_rate": 6.842857142857143e-06,
|
| 94 |
+
"loss": 0.0495,
|
| 95 |
+
"step": 325
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 1.31,
|
| 99 |
+
"learning_rate": 6.485714285714286e-06,
|
| 100 |
+
"loss": 0.0403,
|
| 101 |
+
"step": 350
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.4,
|
| 105 |
+
"learning_rate": 6.128571428571429e-06,
|
| 106 |
+
"loss": 0.0384,
|
| 107 |
+
"step": 375
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 1.5,
|
| 111 |
+
"learning_rate": 5.771428571428572e-06,
|
| 112 |
+
"loss": 0.0402,
|
| 113 |
+
"step": 400
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 1.5,
|
| 117 |
+
"eval_loss": 0.21568405628204346,
|
| 118 |
+
"eval_runtime": 367.8332,
|
| 119 |
+
"eval_samples_per_second": 4.567,
|
| 120 |
+
"eval_steps_per_second": 0.571,
|
| 121 |
+
"eval_wer": 13.220854653574888,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.59,
|
| 126 |
+
"learning_rate": 5.414285714285715e-06,
|
| 127 |
+
"loss": 0.0418,
|
| 128 |
+
"step": 425
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.69,
|
| 132 |
+
"learning_rate": 5.057142857142857e-06,
|
| 133 |
+
"loss": 0.0365,
|
| 134 |
+
"step": 450
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.78,
|
| 138 |
+
"learning_rate": 4.7e-06,
|
| 139 |
+
"loss": 0.0357,
|
| 140 |
+
"step": 475
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.87,
|
| 144 |
+
"learning_rate": 4.342857142857143e-06,
|
| 145 |
+
"loss": 0.0419,
|
| 146 |
+
"step": 500
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.97,
|
| 150 |
+
"learning_rate": 3.985714285714286e-06,
|
| 151 |
+
"loss": 0.0349,
|
| 152 |
+
"step": 525
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 2.06,
|
| 156 |
+
"learning_rate": 3.628571428571429e-06,
|
| 157 |
+
"loss": 0.0221,
|
| 158 |
+
"step": 550
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.15,
|
| 162 |
+
"learning_rate": 3.2714285714285717e-06,
|
| 163 |
+
"loss": 0.0088,
|
| 164 |
+
"step": 575
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.25,
|
| 168 |
+
"learning_rate": 2.9142857142857146e-06,
|
| 169 |
+
"loss": 0.0086,
|
| 170 |
+
"step": 600
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 2.25,
|
| 174 |
+
"eval_loss": 0.2222386747598648,
|
| 175 |
+
"eval_runtime": 351.4131,
|
| 176 |
+
"eval_samples_per_second": 4.781,
|
| 177 |
+
"eval_steps_per_second": 0.598,
|
| 178 |
+
"eval_wer": 13.075646521919513,
|
| 179 |
+
"step": 600
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.34,
|
| 183 |
+
"learning_rate": 2.5571428571428576e-06,
|
| 184 |
+
"loss": 0.01,
|
| 185 |
+
"step": 625
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.43,
|
| 189 |
+
"learning_rate": 2.2e-06,
|
| 190 |
+
"loss": 0.0094,
|
| 191 |
+
"step": 650
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 2.53,
|
| 195 |
+
"learning_rate": 1.8428571428571428e-06,
|
| 196 |
+
"loss": 0.009,
|
| 197 |
+
"step": 675
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 2.62,
|
| 201 |
+
"learning_rate": 1.4857142857142858e-06,
|
| 202 |
+
"loss": 0.0073,
|
| 203 |
+
"step": 700
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 2.72,
|
| 207 |
+
"learning_rate": 1.1285714285714285e-06,
|
| 208 |
+
"loss": 0.0092,
|
| 209 |
+
"step": 725
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 2.81,
|
| 213 |
+
"learning_rate": 7.714285714285715e-07,
|
| 214 |
+
"loss": 0.0083,
|
| 215 |
+
"step": 750
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 2.9,
|
| 219 |
+
"learning_rate": 4.142857142857143e-07,
|
| 220 |
+
"loss": 0.007,
|
| 221 |
+
"step": 775
|
| 222 |
+
},
|
| 223 |
+
{
|
| 224 |
+
"epoch": 3.0,
|
| 225 |
+
"learning_rate": 5.714285714285715e-08,
|
| 226 |
+
"loss": 0.0083,
|
| 227 |
+
"step": 800
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 3.0,
|
| 231 |
+
"eval_loss": 0.2183152139186859,
|
| 232 |
+
"eval_runtime": 351.6231,
|
| 233 |
+
"eval_samples_per_second": 4.778,
|
| 234 |
+
"eval_steps_per_second": 0.597,
|
| 235 |
+
"eval_wer": 12.92352371732817,
|
| 236 |
+
"step": 800
|
| 237 |
+
}
|
| 238 |
+
],
|
| 239 |
+
"max_steps": 800,
|
| 240 |
+
"num_train_epochs": 3,
|
| 241 |
+
"total_flos": 3.69273878249472e+18,
|
| 242 |
+
"trial_name": null,
|
| 243 |
+
"trial_params": null
|
| 244 |
+
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 3643
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40aa8cde2009b6c026271957e82c37ba5779c3880ced0c387121acb310ffded0
|
| 3 |
size 3643
|