Instructions to use Maynx/Rex-Instruct-v0.1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Maynx/Rex-Instruct-v0.1 with Transformers:
# Load model directly from transformers import REX model = REX.from_pretrained("Maynx/Rex-Instruct-v0.1", dtype="auto") - Notebooks
- Google Colab
- Kaggle
| library_name: transformers | |
| tags: | |
| - causal-lm | |
| - chatml | |
| - custom-code | |
| # Rex-Instruct-v0.1 | |
| This repository contains the `genesis_smol` fine-tuned checkpoint exported from this project. | |
| ## Notes | |
| - Architecture: `REX` | |
| - Chat format: ChatML with `<|im_start|>` and `<|im_end|>` | |
| - Tokenizer base: `mistralai/Mistral-7B-v0.3` | |
| ## Local Loading | |
| This is a custom architecture, so load it with the local `REX` class from this repository: | |
| ```python | |
| from model.model import REX | |
| from transformers import AutoTokenizer | |
| tokenizer = AutoTokenizer.from_pretrained("Maynx/Rex-Instruct-v0.1", use_fast=False) | |
| model = REX.from_pretrained("Maynx/Rex-Instruct-v0.1") | |
| ``` | |