|
--- |
|
license: mit |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: git-base-coco-dummy-temp100 |
|
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-coco-dummy-temp100 |
|
|
|
This model is a fine-tuned version of [microsoft/git-base-coco](https://huggingface.co/microsoft/git-base-coco) on an unknown dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.3244 |
|
- Wer Score: 2.5274 |
|
- Blue Score: 0.1539 |
|
|
|
## 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: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 4 |
|
- total_train_batch_size: 32 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- num_epochs: 50 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer Score | Blue Score | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------:|:----------:| |
|
| 8.003 | 1.01 | 35 | 6.0344 | 1.2184 | 0.0042 | |
|
| 4.5865 | 2.03 | 70 | 2.7898 | 2.0108 | 0.0150 | |
|
| 1.6831 | 3.04 | 105 | 0.7239 | 1.8867 | 0.0250 | |
|
| 0.5535 | 4.06 | 140 | 0.4567 | 1.9632 | 0.0377 | |
|
| 0.3898 | 5.07 | 175 | 0.4016 | 2.2010 | 0.0618 | |
|
| 0.3024 | 6.09 | 210 | 0.3712 | 1.7956 | 0.0892 | |
|
| 0.2411 | 7.1 | 245 | 0.3506 | 2.5006 | 0.0950 | |
|
| 0.2034 | 8.12 | 280 | 0.3388 | 2.1609 | 0.1094 | |
|
| 0.1721 | 9.13 | 315 | 0.3319 | 2.2919 | 0.1092 | |
|
| 0.1449 | 10.14 | 350 | 0.3236 | 2.1683 | 0.1184 | |
|
| 0.1245 | 11.16 | 385 | 0.3205 | 2.3594 | 0.1242 | |
|
| 0.1073 | 12.17 | 420 | 0.3160 | 2.4311 | 0.1343 | |
|
| 0.0956 | 13.19 | 455 | 0.3141 | 2.3027 | 0.1327 | |
|
| 0.0821 | 14.2 | 490 | 0.3121 | 2.3957 | 0.1369 | |
|
| 0.07 | 15.22 | 525 | 0.3131 | 2.2508 | 0.1407 | |
|
| 0.0678 | 16.23 | 560 | 0.3114 | 2.4291 | 0.1390 | |
|
| 0.0563 | 17.25 | 595 | 0.3113 | 2.5218 | 0.1428 | |
|
| 0.0518 | 18.26 | 630 | 0.3108 | 2.2964 | 0.1513 | |
|
| 0.0474 | 19.28 | 665 | 0.3138 | 2.2457 | 0.1492 | |
|
| 0.0407 | 20.29 | 700 | 0.3136 | 2.3072 | 0.1485 | |
|
| 0.0383 | 21.3 | 735 | 0.3138 | 2.4791 | 0.1426 | |
|
| 0.0373 | 22.32 | 770 | 0.3136 | 2.4541 | 0.1472 | |
|
| 0.03 | 23.33 | 805 | 0.3145 | 2.4218 | 0.1500 | |
|
| 0.0316 | 24.35 | 840 | 0.3141 | 2.4169 | 0.1466 | |
|
| 0.0255 | 25.36 | 875 | 0.3149 | 2.5450 | 0.1473 | |
|
| 0.0266 | 26.38 | 910 | 0.3159 | 2.4613 | 0.1475 | |
|
| 0.0245 | 27.39 | 945 | 0.3161 | 2.4809 | 0.1506 | |
|
| 0.0223 | 28.41 | 980 | 0.3172 | 2.4252 | 0.1516 | |
|
| 0.0189 | 29.42 | 1015 | 0.3173 | 2.6111 | 0.1501 | |
|
| 0.0204 | 30.43 | 1050 | 0.3184 | 2.5457 | 0.1518 | |
|
| 0.0194 | 31.45 | 1085 | 0.3191 | 2.6389 | 0.1493 | |
|
| 0.0154 | 32.46 | 1120 | 0.3188 | 2.5125 | 0.1518 | |
|
| 0.017 | 33.48 | 1155 | 0.3192 | 2.5197 | 0.1485 | |
|
| 0.0161 | 34.49 | 1190 | 0.3210 | 2.5103 | 0.1512 | |
|
| 0.0146 | 35.51 | 1225 | 0.3206 | 2.4992 | 0.1527 | |
|
| 0.0135 | 36.52 | 1260 | 0.3221 | 2.4620 | 0.1516 | |
|
| 0.0139 | 37.54 | 1295 | 0.3216 | 2.4769 | 0.1519 | |
|
| 0.0132 | 38.55 | 1330 | 0.3215 | 2.5613 | 0.1525 | |
|
| 0.0121 | 39.57 | 1365 | 0.3222 | 2.5648 | 0.1518 | |
|
| 0.0121 | 40.58 | 1400 | 0.3226 | 2.5601 | 0.1541 | |
|
| 0.0114 | 41.59 | 1435 | 0.3231 | 2.4888 | 0.1527 | |
|
| 0.0123 | 42.61 | 1470 | 0.3239 | 2.5037 | 0.1537 | |
|
| 0.0101 | 43.62 | 1505 | 0.3241 | 2.5378 | 0.1526 | |
|
| 0.0109 | 44.64 | 1540 | 0.3245 | 2.5312 | 0.1534 | |
|
| 0.0109 | 45.65 | 1575 | 0.3245 | 2.5692 | 0.1529 | |
|
| 0.0098 | 46.67 | 1610 | 0.3243 | 2.5583 | 0.1536 | |
|
| 0.0109 | 47.68 | 1645 | 0.3248 | 2.5498 | 0.1536 | |
|
| 0.0101 | 48.7 | 1680 | 0.3244 | 2.5274 | 0.1539 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.30.2 |
|
- Pytorch 2.0.1+cu118 |
|
- Datasets 2.14.3 |
|
- Tokenizers 0.13.3 |
|
|
