|
--- |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: led-large |
|
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. --> |
|
|
|
# led-large |
|
|
|
This model was trained from scratch on the None dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.1850 |
|
|
|
## 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: 3e-05 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 64 |
|
- total_train_batch_size: 64 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: polynomial |
|
- lr_scheduler_warmup_steps: 500 |
|
- training_steps: 20000 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:-----:|:---------------:| |
|
| 0.1479 | 0.11 | 500 | 0.1901 | |
|
| 0.1442 | 0.22 | 1000 | 0.1917 | |
|
| 0.1466 | 0.33 | 1500 | 0.1959 | |
|
| 0.1447 | 0.45 | 2000 | 0.1918 | |
|
| 0.1633 | 0.56 | 2500 | 0.1874 | |
|
| 0.171 | 0.67 | 3000 | 0.1849 | |
|
| 0.1662 | 0.78 | 3500 | 0.1843 | |
|
| 0.1743 | 0.89 | 4000 | 0.1837 | |
|
| 0.1492 | 1.0 | 4500 | 0.1842 | |
|
| 0.1515 | 1.11 | 5000 | 0.1849 | |
|
| 0.1497 | 1.23 | 5500 | 0.1840 | |
|
| 0.1515 | 1.34 | 6000 | 0.1839 | |
|
| 0.1482 | 1.45 | 6500 | 0.1841 | |
|
| 0.145 | 1.56 | 7000 | 0.1849 | |
|
| 0.1467 | 1.67 | 7500 | 0.1824 | |
|
| 0.1509 | 1.78 | 8000 | 0.1809 | |
|
| 0.15 | 1.89 | 8500 | 0.1832 | |
|
| 0.1383 | 2.01 | 9000 | 0.1831 | |
|
| 0.1331 | 2.12 | 9500 | 0.1820 | |
|
| 0.1406 | 2.23 | 10000 | 0.1830 | |
|
| 0.1362 | 2.34 | 10500 | 0.1844 | |
|
| 0.1373 | 2.45 | 11000 | 0.1836 | |
|
| 0.1269 | 2.56 | 11500 | 0.1842 | |
|
| 0.1362 | 2.67 | 12000 | 0.1819 | |
|
| 0.14 | 2.79 | 12500 | 0.1832 | |
|
| 0.1319 | 2.9 | 13000 | 0.1837 | |
|
| 0.1304 | 3.01 | 13500 | 0.1845 | |
|
| 0.1278 | 3.12 | 14000 | 0.1844 | |
|
| 0.1235 | 3.23 | 14500 | 0.1832 | |
|
| 0.1293 | 3.34 | 15000 | 0.1855 | |
|
| 0.1302 | 3.45 | 15500 | 0.1836 | |
|
| 0.1285 | 3.57 | 16000 | 0.1860 | |
|
| 0.1274 | 3.68 | 16500 | 0.1860 | |
|
| 0.1261 | 3.79 | 17000 | 0.1854 | |
|
| 0.1304 | 3.9 | 17500 | 0.1859 | |
|
| 0.1223 | 4.01 | 18000 | 0.1862 | |
|
| 0.1235 | 4.12 | 18500 | 0.1849 | |
|
| 0.1286 | 4.23 | 19000 | 0.1858 | |
|
| 0.1186 | 4.35 | 19500 | 0.1856 | |
|
| 0.1293 | 4.46 | 20000 | 0.1850 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.37.2 |
|
- Pytorch 2.2.2+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.1 |
|
|
