metadata
license: llama2
library_name: peft
tags:
- generated_from_trainer
base_model: TheBloke/Llama-2-13B-chat-GPTQ
model-index:
- name: qlora
results: []
qlora
This model is a fine-tuned version of TheBloke/Llama-2-13B-chat-GPTQ on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2281
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: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 2
- training_steps: 30
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 1.7256 | 1.0 | 1 | 2.3492 |
| 0.8422 | 2.0 | 3 | 2.0594 |
| 0.7278 | 3.0 | 5 | 1.6949 |
| 1.2299 | 4.0 | 6 | 1.5504 |
| 1.1219 | 5.0 | 7 | 1.4207 |
| 0.4884 | 6.0 | 9 | 1.2156 |
| 0.4229 | 7.0 | 11 | 0.9934 |
| 0.7082 | 8.0 | 12 | 0.8837 |
| 0.6263 | 9.0 | 13 | 0.7754 |
| 0.2502 | 10.0 | 15 | 0.5958 |
| 0.1915 | 11.0 | 17 | 0.4347 |
| 0.2841 | 12.0 | 18 | 0.3766 |
| 0.2395 | 13.0 | 19 | 0.3339 |
| 0.0959 | 14.0 | 21 | 0.2914 |
| 0.0808 | 15.0 | 23 | 0.2631 |
| 0.1391 | 16.0 | 24 | 0.2533 |
| 0.13 | 17.0 | 25 | 0.2434 |
| 0.0589 | 18.0 | 27 | 0.2357 |
| 0.0558 | 19.0 | 29 | 0.2298 |
| 0.1076 | 20.0 | 30 | 0.2281 |
Framework versions
- PEFT 0.11.1
- Transformers 4.41.2
- Pytorch 2.1.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1