|
--- |
|
library_name: transformers |
|
license: mit |
|
base_model: microsoft/git-base |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: git-base-bdd100k |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# git-base-bdd100k |
|
|
|
This model is a fine-tuned version of [microsoft/git-base](https://huggingface.co/microsoft/git-base) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4505 |
|
- Wer Score: 2.0146 |
|
|
|
## Model description |
|
|
|
More information needed |
|
|
|
## Intended uses & limitations |
|
|
|
More information needed |
|
|
|
## Training and evaluation data |
|
|
|
More information needed |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 5e-05 |
|
- train_batch_size: 20 |
|
- eval_batch_size: 20 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 40 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 100 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer Score | |
|
|:-------------:|:-------:|:----:|:---------------:|:---------:| |
|
| 10.7599 | 0.9091 | 5 | 9.0648 | 7.4053 | |
|
| 7.8749 | 2.0 | 11 | 7.9846 | 5.4869 | |
|
| 8.5483 | 2.9091 | 16 | 7.3958 | 5.9978 | |
|
| 6.5899 | 4.0 | 22 | 6.8122 | 7.3117 | |
|
| 7.3362 | 4.9091 | 27 | 6.3566 | 5.4206 | |
|
| 5.6682 | 6.0 | 33 | 5.8240 | 2.6977 | |
|
| 6.2659 | 6.9091 | 38 | 5.3805 | 2.3248 | |
|
| 4.7822 | 8.0 | 44 | 4.8517 | 2.4497 | |
|
| 5.2042 | 8.9091 | 49 | 4.4194 | 2.3350 | |
|
| 3.9022 | 10.0 | 55 | 3.9023 | 2.0637 | |
|
| 4.1582 | 10.9091 | 60 | 3.4813 | 2.4832 | |
|
| 3.0413 | 12.0 | 66 | 2.9854 | 2.5313 | |
|
| 3.1438 | 12.9091 | 71 | 2.5871 | 2.4395 | |
|
| 2.2196 | 14.0 | 77 | 2.1313 | 2.5160 | |
|
| 2.199 | 14.9091 | 82 | 1.7799 | 2.4064 | |
|
| 1.4819 | 16.0 | 88 | 1.4052 | 2.3929 | |
|
| 1.3977 | 16.9091 | 93 | 1.1385 | 2.4009 | |
|
| 0.9006 | 18.0 | 99 | 0.8846 | 2.3711 | |
|
| 0.8222 | 18.9091 | 104 | 0.7261 | 2.5171 | |
|
| 0.5272 | 20.0 | 110 | 0.5892 | 2.5583 | |
|
| 0.4908 | 20.9091 | 115 | 0.5160 | 2.5098 | |
|
| 0.3346 | 22.0 | 121 | 0.4587 | 2.3434 | |
|
| 0.3306 | 22.9091 | 126 | 0.4197 | 2.3015 | |
|
| 0.2313 | 24.0 | 132 | 0.3966 | 2.0754 | |
|
| 0.237 | 24.9091 | 137 | 0.3828 | 2.2418 | |
|
| 0.1691 | 26.0 | 143 | 0.3792 | 1.7196 | |
|
| 0.1745 | 26.9091 | 148 | 0.3729 | 2.2782 | |
|
| 0.1261 | 28.0 | 154 | 0.3665 | 1.8682 | |
|
| 0.1294 | 28.9091 | 159 | 0.3745 | 1.8237 | |
|
| 0.0916 | 30.0 | 165 | 0.3762 | 2.3332 | |
|
| 0.0944 | 30.9091 | 170 | 0.3758 | 1.9060 | |
|
| 0.0682 | 32.0 | 176 | 0.3796 | 2.1471 | |
|
| 0.0703 | 32.9091 | 181 | 0.3846 | 1.8350 | |
|
| 0.0512 | 34.0 | 187 | 0.3891 | 2.0670 | |
|
| 0.0537 | 34.9091 | 192 | 0.3909 | 2.0998 | |
|
| 0.0392 | 36.0 | 198 | 0.3944 | 2.2658 | |
|
| 0.0418 | 36.9091 | 203 | 0.3999 | 2.1865 | |
|
| 0.0314 | 38.0 | 209 | 0.3970 | 2.2338 | |
|
| 0.0344 | 38.9091 | 214 | 0.4057 | 2.0838 | |
|
| 0.0252 | 40.0 | 220 | 0.4073 | 2.2542 | |
|
| 0.0285 | 40.9091 | 225 | 0.4079 | 2.2538 | |
|
| 0.022 | 42.0 | 231 | 0.4121 | 2.0579 | |
|
| 0.0237 | 42.9091 | 236 | 0.4097 | 2.1475 | |
|
| 0.0182 | 44.0 | 242 | 0.4185 | 2.1577 | |
|
| 0.0203 | 44.9091 | 247 | 0.4151 | 2.2378 | |
|
| 0.0157 | 46.0 | 253 | 0.4212 | 2.0703 | |
|
| 0.0177 | 46.9091 | 258 | 0.4212 | 2.0237 | |
|
| 0.0136 | 48.0 | 264 | 0.4208 | 1.9676 | |
|
| 0.0155 | 48.9091 | 269 | 0.4229 | 2.0262 | |
|
| 0.0123 | 50.0 | 275 | 0.4253 | 2.0612 | |
|
| 0.0144 | 50.9091 | 280 | 0.4284 | 2.0663 | |
|
| 0.0112 | 52.0 | 286 | 0.4315 | 2.0706 | |
|
| 0.0129 | 52.9091 | 291 | 0.4301 | 2.0568 | |
|
| 0.0107 | 54.0 | 297 | 0.4301 | 2.0087 | |
|
| 0.0121 | 54.9091 | 302 | 0.4311 | 2.0022 | |
|
| 0.0095 | 56.0 | 308 | 0.4313 | 1.9996 | |
|
| 0.0109 | 56.9091 | 313 | 0.4333 | 2.0546 | |
|
| 0.0086 | 58.0 | 319 | 0.4338 | 2.0787 | |
|
| 0.0102 | 58.9091 | 324 | 0.4359 | 2.0091 | |
|
| 0.0082 | 60.0 | 330 | 0.4369 | 2.0430 | |
|
| 0.0095 | 60.9091 | 335 | 0.4366 | 1.9592 | |
|
| 0.0076 | 62.0 | 341 | 0.4388 | 1.9905 | |
|
| 0.0089 | 62.9091 | 346 | 0.4395 | 2.0295 | |
|
| 0.0072 | 64.0 | 352 | 0.4404 | 2.0200 | |
|
| 0.0084 | 64.9091 | 357 | 0.4393 | 2.0641 | |
|
| 0.0067 | 66.0 | 363 | 0.4408 | 2.0798 | |
|
| 0.0078 | 66.9091 | 368 | 0.4422 | 2.0601 | |
|
| 0.0063 | 68.0 | 374 | 0.4420 | 2.0408 | |
|
| 0.0076 | 68.9091 | 379 | 0.4427 | 2.0273 | |
|
| 0.0063 | 70.0 | 385 | 0.4438 | 2.0306 | |
|
| 0.0072 | 70.9091 | 390 | 0.4436 | 2.0462 | |
|
| 0.006 | 72.0 | 396 | 0.4456 | 2.0160 | |
|
| 0.007 | 72.9091 | 401 | 0.4472 | 2.0382 | |
|
| 0.0057 | 74.0 | 407 | 0.4466 | 2.0532 | |
|
| 0.0066 | 74.9091 | 412 | 0.4459 | 2.0612 | |
|
| 0.0055 | 76.0 | 418 | 0.4469 | 2.0229 | |
|
| 0.0065 | 76.9091 | 423 | 0.4474 | 1.9632 | |
|
| 0.0054 | 78.0 | 429 | 0.4481 | 1.9519 | |
|
| 0.0064 | 78.9091 | 434 | 0.4475 | 1.9836 | |
|
| 0.0052 | 80.0 | 440 | 0.4475 | 2.0149 | |
|
| 0.0062 | 80.9091 | 445 | 0.4482 | 2.0197 | |
|
| 0.0052 | 82.0 | 451 | 0.4490 | 2.0208 | |
|
| 0.0061 | 82.9091 | 456 | 0.4496 | 2.0324 | |
|
| 0.0049 | 84.0 | 462 | 0.4498 | 2.0240 | |
|
| 0.006 | 84.9091 | 467 | 0.4496 | 2.0168 | |
|
| 0.0049 | 86.0 | 473 | 0.4499 | 2.0 | |
|
| 0.0059 | 86.9091 | 478 | 0.4505 | 1.9822 | |
|
| 0.005 | 88.0 | 484 | 0.4506 | 1.9978 | |
|
| 0.0058 | 88.9091 | 489 | 0.4505 | 2.0117 | |
|
| 0.0049 | 90.0 | 495 | 0.4505 | 2.0135 | |
|
| 0.0053 | 90.9091 | 500 | 0.4505 | 2.0146 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.45.2 |
|
- Pytorch 2.1.1+cu121 |
|
- Datasets 3.0.2 |
|
- Tokenizers 0.20.1 |
|
|
