|
--- |
|
library_name: peft |
|
license: llama3.2 |
|
base_model: NousResearch/Llama-3.2-1B |
|
tags: |
|
- axolotl |
|
- generated_from_trainer |
|
datasets: |
|
- teknium/GPT4-LLM-Cleaned |
|
model-index: |
|
- name: llama-fr-lora |
|
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. --> |
|
|
|
[<img src="https://raw.githubusercontent.com/axolotl-ai-cloud/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/axolotl-ai-cloud/axolotl) |
|
<details><summary>See axolotl config</summary> |
|
|
|
axolotl version: `0.6.0` |
|
```yaml |
|
adapter: lora |
|
base_model: NousResearch/Llama-3.2-1B |
|
bf16: auto |
|
dataset_prepared_path: last_run_prepared |
|
datasets: |
|
- path: teknium/GPT4-LLM-Cleaned |
|
type: alpaca |
|
eval_sample_packing: true |
|
evals_per_epoch: 4 |
|
flash_attention: true |
|
gradient_accumulation_steps: 2 |
|
gradient_checkpointing: true |
|
group_by_length: false |
|
hub_model_id: pandyamarut/llama-fr-lora |
|
learning_rate: 0.0002 |
|
load_in_4bit: false |
|
load_in_8bit: false |
|
logging_steps: 1 |
|
lora_alpha: 32 |
|
lora_dropout: 0.05 |
|
lora_r: 16 |
|
lora_target_modules: |
|
- gate_proj |
|
- down_proj |
|
- up_proj |
|
- q_proj |
|
- v_proj |
|
- k_proj |
|
- o_proj |
|
loss_watchdog_patience: 3 |
|
loss_watchdog_threshold: 5 |
|
lr_scheduler: cosine |
|
micro_batch_size: 2 |
|
num_epochs: 1 |
|
optimizer: adamw_8bit |
|
output_dir: /runpod-volume/fine-tuning/test-run |
|
pad_to_sequence_len: true |
|
run_name: test-run |
|
runpod_job_id: dd327f42-5f67-4830-b512-4561fa9a3d45-u1 |
|
sample_packing: true |
|
saves_per_epoch: 1 |
|
sequence_len: 2048 |
|
special_tokens: |
|
pad_token: <|end_of_text|> |
|
strict: false |
|
tf32: false |
|
train_on_inputs: false |
|
val_set_size: 0.1 |
|
wandb_entity: axo-test |
|
wandb_name: test-run-1 |
|
wandb_project: test-run-1 |
|
warmup_steps: 10 |
|
weight_decay: 0 |
|
|
|
``` |
|
|
|
</details><br> |
|
|
|
# llama-fr-lora |
|
|
|
This model is a fine-tuned version of [NousResearch/Llama-3.2-1B](https://huggingface.co/NousResearch/Llama-3.2-1B) on the teknium/GPT4-LLM-Cleaned dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 1.1018 |
|
|
|
## 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: 2 |
|
- eval_batch_size: 2 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 4 |
|
- optimizer: Use OptimizerNames.ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 10 |
|
- num_epochs: 1 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:------:|:----:|:---------------:| |
|
| 1.4537 | 0.0009 | 1 | 1.3971 | |
|
| 1.1978 | 0.2503 | 271 | 1.1561 | |
|
| 1.1637 | 0.5007 | 542 | 1.1131 | |
|
| 1.1894 | 0.7510 | 813 | 1.1018 | |
|
|
|
|
|
### Framework versions |
|
|
|
- PEFT 0.14.0 |
|
- Transformers 4.47.1 |
|
- Pytorch 2.5.1+cu124 |
|
- Datasets 3.2.0 |
|
- Tokenizers 0.21.0 |
