File size: 630 Bytes
60a1db1
 
 
 
 
 
 
 
 
d0302cd
 
62fb563
2b7daaa
62fb563
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
---
title: README
emoji: 📈
colorFrom: indigo
colorTo: green
sdk: static
pinned: false
---

LLaVa, a visual instruction tuned version of LLaMa and other large language models, can now be used natively with the Transformers library.

TRL now includes experimental support for fine-tuning! Find the script [here](https://github.com/huggingface/trl/blob/main/examples/scripts/vsft_llava.py). A blog post which goes more in detail can be found [here](https://huggingface.co/blog/vlms).

An inference demo notebook can be found [here](https://colab.research.google.com/drive/1_q7cOB-jCu3RExrkhrgewBR0qKjZr-Sx#scrollTo=PuWVAAOinC8q).