|
--- |
|
library_name: peft |
|
license: llama3 |
|
base_model: aaditya/Llama3-OpenBioLLM-8B |
|
tags: |
|
- llama-factory |
|
- lora |
|
- generated_from_trainer |
|
model-index: |
|
- name: Llama3-OpenBioLLM-8B-PsyCourse-fold4 |
|
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. --> |
|
|
|
# Llama3-OpenBioLLM-8B-PsyCourse-fold4 |
|
|
|
This model is a fine-tuned version of [aaditya/Llama3-OpenBioLLM-8B](https://huggingface.co/aaditya/Llama3-OpenBioLLM-8B) on the course-train-fold4 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0349 |
|
|
|
## 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: 0.0001 |
|
- train_batch_size: 1 |
|
- eval_batch_size: 1 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 16 |
|
- total_train_batch_size: 16 |
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_ratio: 0.1 |
|
- num_epochs: 5.0 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 0.4799 | 0.0763 | 50 | 0.2933 | |
|
| 0.1053 | 0.1527 | 100 | 0.0766 | |
|
| 0.0594 | 0.2290 | 150 | 0.0722 | |
|
| 0.0586 | 0.3053 | 200 | 0.0537 | |
|
| 0.066 | 0.3816 | 250 | 0.0550 | |
|
| 0.0553 | 0.4580 | 300 | 0.0551 | |
|
| 0.0468 | 0.5343 | 350 | 0.0476 | |
|
| 0.0515 | 0.6106 | 400 | 0.0481 | |
|
| 0.0562 | 0.6870 | 450 | 0.0466 | |
|
| 0.0347 | 0.7633 | 500 | 0.0401 | |
|
| 0.0417 | 0.8396 | 550 | 0.0408 | |
|
| 0.0466 | 0.9159 | 600 | 0.0393 | |
|
| 0.0346 | 0.9923 | 650 | 0.0404 | |
|
| 0.035 | 1.0686 | 700 | 0.0389 | |
|
| 0.0408 | 1.1449 | 750 | 0.0390 | |
|
| 0.0324 | 1.2213 | 800 | 0.0427 | |
|
| 0.0331 | 1.2976 | 850 | 0.0382 | |
|
| 0.0335 | 1.3739 | 900 | 0.0433 | |
|
| 0.034 | 1.4502 | 950 | 0.0374 | |
|
| 0.0253 | 1.5266 | 1000 | 0.0399 | |
|
| 0.0299 | 1.6029 | 1050 | 0.0382 | |
|
| 0.0534 | 1.6792 | 1100 | 0.0424 | |
|
| 0.0318 | 1.7556 | 1150 | 0.0402 | |
|
| 0.0484 | 1.8319 | 1200 | 0.0385 | |
|
| 0.0263 | 1.9082 | 1250 | 0.0355 | |
|
| 0.0329 | 1.9845 | 1300 | 0.0349 | |
|
| 0.0171 | 2.0609 | 1350 | 0.0365 | |
|
| 0.0204 | 2.1372 | 1400 | 0.0374 | |
|
| 0.0248 | 2.2135 | 1450 | 0.0413 | |
|
| 0.0132 | 2.2899 | 1500 | 0.0397 | |
|
| 0.0152 | 2.3662 | 1550 | 0.0383 | |
|
| 0.0222 | 2.4425 | 1600 | 0.0387 | |
|
| 0.0187 | 2.5188 | 1650 | 0.0374 | |
|
| 0.0177 | 2.5952 | 1700 | 0.0415 | |
|
| 0.0164 | 2.6715 | 1750 | 0.0380 | |
|
| 0.0212 | 2.7478 | 1800 | 0.0395 | |
|
| 0.0248 | 2.8242 | 1850 | 0.0357 | |
|
| 0.0187 | 2.9005 | 1900 | 0.0384 | |
|
| 0.0315 | 2.9768 | 1950 | 0.0372 | |
|
| 0.006 | 3.0531 | 2000 | 0.0423 | |
|
| 0.0077 | 3.1295 | 2050 | 0.0459 | |
|
| 0.0073 | 3.2058 | 2100 | 0.0493 | |
|
| 0.0096 | 3.2821 | 2150 | 0.0523 | |
|
| 0.0086 | 3.3585 | 2200 | 0.0449 | |
|
| 0.0057 | 3.4348 | 2250 | 0.0469 | |
|
| 0.0098 | 3.5111 | 2300 | 0.0460 | |
|
| 0.0086 | 3.5874 | 2350 | 0.0493 | |
|
| 0.0073 | 3.6638 | 2400 | 0.0471 | |
|
| 0.0086 | 3.7401 | 2450 | 0.0468 | |
|
| 0.0079 | 3.8164 | 2500 | 0.0455 | |
|
| 0.0042 | 3.8928 | 2550 | 0.0473 | |
|
| 0.0088 | 3.9691 | 2600 | 0.0474 | |
|
| 0.004 | 4.0454 | 2650 | 0.0474 | |
|
| 0.0033 | 4.1217 | 2700 | 0.0500 | |
|
| 0.0009 | 4.1981 | 2750 | 0.0520 | |
|
| 0.0021 | 4.2744 | 2800 | 0.0532 | |
|
| 0.0042 | 4.3507 | 2850 | 0.0546 | |
|
| 0.0017 | 4.4271 | 2900 | 0.0566 | |
|
| 0.002 | 4.5034 | 2950 | 0.0577 | |
|
| 0.0021 | 4.5797 | 3000 | 0.0583 | |
|
| 0.0017 | 4.6560 | 3050 | 0.0585 | |
|
| 0.0036 | 4.7324 | 3100 | 0.0587 | |
|
| 0.0021 | 4.8087 | 3150 | 0.0588 | |
|
| 0.0028 | 4.8850 | 3200 | 0.0588 | |
|
| 0.0021 | 4.9614 | 3250 | 0.0588 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.12.0 |
|
- Transformers 4.46.1 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.1.0 |
|
- Tokenizers 0.20.3 |