|
--- |
|
license: apache-2.0 |
|
tags: |
|
- dpo |
|
base_model: |
|
- CorticalStack/neurotic-crown-clown-7b-ties |
|
dataset: |
|
- CorticalStack/tak-stack-dpo |
|
--- |
|
|
|
<img src="neurotic-crown-clown-tak-stack.png" alt="Neurotic crown clown tak stack logo" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/> |
|
|
|
# neurotic-crown-clown-7b-tak-stack-dpo |
|
|
|
neurotic-crown-clown-7b-tak-stack-dpo is a DPO fine-tuned version of [CorticalStack/neurotic-crown-clown-7b-ties](https://huggingface.co/CorticalStack/neurotic-crown-clown-7b-ties) using the [CorticalStack/tak-stack-dpo](https://huggingface.co/datasets/CorticalStack/tak-stack-dpo) dataset. |
|
|
|
### LoRA |
|
- r: 32 |
|
- LoRA alpha: 32 |
|
- LoRA dropout: 0.05 |
|
|
|
### Training arguments |
|
- Batch size: 4 |
|
- Gradient accumulation steps: 4 |
|
- Optimizer: paged_adamw_32bit |
|
- Max steps: 100 |
|
- Learning rate: 5e-05 |
|
- Learning rate scheduler type: cosine |
|
- Beta: 0.1 |
|
- Max prompt length: 1024 |
|
- Max length: 1536 |
