metadata
license: llama2
library_name: peft
tags:
- generated_from_trainer
base_model: vilm/vinallama-7b-chat
model-index:
- name: experiments
results: []
experiments
This model is a fine-tuned version of vilm/vinallama-7b-chat on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.4668
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: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.3854 | 0.7 | 10 | 0.4145 |
| 0.3751 | 1.4 | 20 | 0.3999 |
| 0.3457 | 2.11 | 30 | 0.3857 |
| 0.324 | 2.81 | 40 | 0.3821 |
| 0.2329 | 3.51 | 50 | 0.3845 |
| 0.2332 | 4.21 | 60 | 0.3878 |
| 0.2211 | 4.91 | 70 | 0.3929 |
| 0.1682 | 5.61 | 80 | 0.4093 |
| 0.1503 | 6.32 | 90 | 0.4416 |
| 0.1439 | 7.02 | 100 | 0.4299 |
| 0.1299 | 7.72 | 110 | 0.4648 |
| 0.1074 | 8.42 | 120 | 0.4574 |
| 0.099 | 9.12 | 130 | 0.4655 |
| 0.1078 | 9.82 | 140 | 0.4668 |
Framework versions
- PEFT 0.11.1
- Transformers 4.36.0
- Pytorch 2.3.0+cu121
- Datasets 2.19.2
- Tokenizers 0.15.2