| | --- |
| | tags: |
| | - generated_from_trainer |
| | model-index: |
| | - name: BERT_summer-6_tokenized |
| | 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. --> |
| |
|
| | # BERT_summer-6_tokenized |
| |
|
| | This model is a fine-tuned version of [armheb/DNA_bert_6](https://huggingface.co/armheb/DNA_bert_6) on an unknown dataset. |
| | It achieves the following results on the evaluation set: |
| | - Loss: 0.0351 |
| |
|
| | ## 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: 2e-05 |
| | - train_batch_size: 64 |
| | - eval_batch_size: 64 |
| | - seed: 42 |
| | - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
| | - lr_scheduler_type: linear |
| | - num_epochs: 10 |
| | - mixed_precision_training: Native AMP |
| |
|
| | ### Training results |
| |
|
| | | Training Loss | Epoch | Step | Validation Loss | |
| | |:-------------:|:-----:|:----:|:---------------:| |
| | | 0.0702 | 1.0 | 256 | 0.0383 | |
| | | 0.0392 | 2.0 | 512 | 0.0402 | |
| | | 0.0384 | 3.0 | 768 | 0.0370 | |
| | | 0.0377 | 4.0 | 1024 | 0.0361 | |
| | | 0.0372 | 5.0 | 1280 | 0.0392 | |
| | | 0.0372 | 6.0 | 1536 | 0.0373 | |
| | | 0.0375 | 7.0 | 1792 | 0.0351 | |
| | | 0.0371 | 8.0 | 2048 | 0.0364 | |
| | | 0.0369 | 9.0 | 2304 | 0.0354 | |
| | | 0.0359 | 10.0 | 2560 | 0.0371 | |
| |
|
| |
|
| | ### Framework versions |
| |
|
| | - Transformers 4.23.1 |
| | - Pytorch 1.12.1+cu113 |
| | - Datasets 2.6.1 |
| | - Tokenizers 0.13.1 |
| |
|
