|
--- |
|
base_model: migtissera/Tess-34B-v1.4 |
|
inference: false |
|
license: other |
|
license_name: yi-license |
|
license_link: https://huggingface.co/01-ai/Yi-34B/blob/main/LICENSE |
|
--- |
|
|
|
[<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) |
|
|
|
An instruct based fine tune of [migtissera/Tess-34B-v1.4](https://huggingface.co/migtissera/Tess-34B-v1.4). |
|
|
|
It works well with long system prompts. |
|
|
|
It works well for reasoning tasks. |
|
|
|
This model is trained on a private dataset. The high GSM8K score is **NOT** because of the MetaMath dataset. |
|
|
|
# Prompt Format: |
|
|
|
``` |
|
SYSTEM: <ANY SYSTEM CONTEXT> |
|
USER: |
|
ASSISTANT: |
|
``` |
