metadata
license: bsd-3-clause
library_name: peft
tags:
- generated_from_trainer
base_model: hugohrban/progen2-small
metrics:
- precision
- recall
- accuracy
model-index:
- name: progen2-small-lora-64-remote-homology-filtered
results: []
progen2-small-lora-64-remote-homology-filtered
This model is a fine-tuned version of hugohrban/progen2-small on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1353
- Precision: 0.9499
- Recall: 0.9699
- F1-score: 0.9598
- Accuracy: 0.9595
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: 0.0002
- train_batch_size: 192
- eval_batch_size: 192
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 384
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1-score | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.3088 | 0.9985 | 332 | 0.2490 | 0.8692 | 0.9277 | 0.8975 | 0.8943 |
| 0.1806 | 2.0 | 665 | 0.1541 | 0.9259 | 0.9507 | 0.9382 | 0.9375 |
| 0.1036 | 2.9985 | 997 | 0.1190 | 0.9522 | 0.9541 | 0.9532 | 0.9532 |
| 0.0533 | 4.0 | 1330 | 0.1140 | 0.9538 | 0.9628 | 0.9582 | 0.9581 |
| 0.0272 | 4.9925 | 1660 | 0.1353 | 0.9499 | 0.9699 | 0.9598 | 0.9595 |
Framework versions
- PEFT 0.10.0
- Transformers 4.40.2
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1