<!-- 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. -->
tiny-mlm-wikitext-custom-tokenizer
This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 6.4940
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: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 200
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 8.1543 | 0.4 | 500 | 7.6501 |
| 7.4342 | 0.8 | 1000 | 7.5531 |
| 7.3656 | 1.2 | 1500 | nan |
| 7.2844 | 1.6 | 2000 | 7.4543 |
| 7.2621 | 2.0 | 2500 | 7.4480 |
| 7.1668 | 2.4 | 3000 | 7.3456 |
| 7.1874 | 2.8 | 3500 | 7.3750 |
| 7.1284 | 3.2 | 4000 | nan |
| 7.1041 | 3.6 | 4500 | 7.2361 |
| 7.0693 | 4.0 | 5000 | 7.2836 |
| 7.0604 | 4.4 | 5500 | 7.2521 |
| 6.993 | 4.8 | 6000 | 7.2082 |
| 7.0014 | 5.2 | 6500 | 7.1960 |
| 6.9607 | 5.6 | 7000 | 7.1971 |
| 6.9514 | 6.0 | 7500 | nan |
| 6.9524 | 6.4 | 8000 | 7.0977 |
| 6.8999 | 6.8 | 8500 | 7.0787 |
| 6.8471 | 7.2 | 9000 | 7.1168 |
| 6.8511 | 7.6 | 9500 | 7.0589 |
| 6.8111 | 8.0 | 10000 | 7.0058 |
| 6.8131 | 8.4 | 10500 | 7.0089 |
| 6.717 | 8.8 | 11000 | 6.9681 |
| 6.7024 | 9.2 | 11500 | 6.9542 |
| 6.7567 | 9.6 | 12000 | 6.9008 |
| 6.7025 | 10.0 | 12500 | 6.8863 |
| 6.6509 | 10.4 | 13000 | 6.8794 |
| 6.6151 | 10.8 | 13500 | 6.8888 |
| 6.6348 | 11.2 | 14000 | 6.8106 |
| 6.6061 | 11.6 | 14500 | 6.8399 |
| 6.5637 | 12.0 | 15000 | 6.8289 |
| 6.5526 | 12.4 | 15500 | 6.7866 |
| 6.4899 | 12.8 | 16000 | 6.7108 |
| 6.5106 | 13.2 | 16500 | 6.7707 |
| 6.5022 | 13.6 | 17000 | 6.7289 |
| 6.429 | 14.0 | 17500 | 6.6883 |
| 6.4342 | 14.4 | 18000 | 6.6669 |
| 6.4385 | 14.8 | 18500 | 6.6722 |
| 6.4328 | 15.2 | 19000 | 6.6867 |
| 6.3802 | 15.6 | 19500 | 6.6403 |
| 6.375 | 16.0 | 20000 | 6.6141 |
| 6.332 | 16.4 | 20500 | 6.6759 |
| 6.3237 | 16.8 | 21000 | 6.5960 |
| 6.3551 | 17.2 | 21500 | 6.5551 |
| 6.2918 | 17.6 | 22000 | nan |
| 6.3 | 18.0 | 22500 | 6.5744 |
| 6.2555 | 18.4 | 23000 | 6.5212 |
| 6.2569 | 18.8 | 23500 | 6.5515 |
| 6.2658 | 19.2 | 24000 | 6.5763 |
| 6.2205 | 19.6 | 24500 | 6.4887 |
| 6.2022 | 20.0 | 25000 | 6.4955 |
| 6.1881 | 20.4 | 25500 | 6.4849 |
| 6.1479 | 20.8 | 26000 | 6.4727 |
| 6.1805 | 21.2 | 26500 | 6.4253 |
| 6.1439 | 21.6 | 27000 | 6.4397 |
| 6.1332 | 22.0 | 27500 | 6.4876 |
| 6.1379 | 22.4 | 28000 | 6.4940 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2
