|
--- |
|
base_model: NousResearch/Llama-2-7b-hf |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: lora-out |
|
results: [] |
|
--- |
|
|
|
|
|
[<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>](https://github.com/OpenAccess-AI-Collective/axolotl) |
|
# lora-out |
|
|
|
This model is a fine-tuned version of [NousResearch/Llama-2-7b-hf](https://huggingface.co/NousResearch/Llama-2-7b-hf) on a synthetic recipe assistant dataset comprised of 2000 samples. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.8666 |
|
|
|
## 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: 6 |
|
- eval_batch_size: 6 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 8 |
|
- total_train_batch_size: 48 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 10 |
|
- num_epochs: 50 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | |
|
|:-------------:|:-----:|:----:|:---------------:| |
|
| 0.9548 | 8.0 | 20 | 0.9240 | |
|
| 0.8514 | 16.0 | 40 | 0.8523 | |
|
| 0.7774 | 24.0 | 60 | 0.8498 | |
|
| 0.7178 | 32.0 | 80 | 0.8597 | |
|
| 0.7103 | 40.0 | 100 | 0.8666 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.34.0.dev0 |
|
- Pytorch 2.0.1+cu118 |
|
- Datasets 2.14.5 |
|
- Tokenizers 0.14.0 |
|
|
