Xuyao Wang
commited on
Commit
·
624f7e7
1
Parent(s):
0a2912f
Add README
Browse files
README.md
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-4.0
|
| 3 |
+
language:
|
| 4 |
+
- en
|
| 5 |
+
pipeline_tag: any-to-any
|
| 6 |
+
tags:
|
| 7 |
+
- multimodal
|
| 8 |
+
library_name: transformers
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Align-Anything Chameleon 7B Plus
|
| 12 |
+
|
| 13 |
+
## Introduction
|
| 14 |
+
|
| 15 |
+
Repository for Align-Anything Chameleon 7B Plus, a powerful model for text-image interleaved input and output, with further alignment by [Align-Anything](https://github.com/PKU-Alignment/Align-Anything) algorithm. This model is based on the [Chameleon](https://huggingface.co/facebook/chameleon-7b) model, and is trained and aligned on the [Align-Anything](https://github.com/PKU-Alignment/Align-Anything) framework to further unlock its capability of image generation, and improve the alignment ability towards human preferences.
|
| 16 |
+
|
| 17 |
+
## Usage
|
| 18 |
+
|
| 19 |
+
To use this model, you can refer to the [Align-Anything](https://github.com/PKU-Alignment) repository for more details, including the training, inference and evaluation:
|
| 20 |
+
|
| 21 |
+
```bash
|
| 22 |
+
git clone https://github.com/PKU-Alignment/align-anything.git
|
| 23 |
+
cd align-anything/projects/text_image_to_text_image
|
| 24 |
+
```
|
| 25 |
+
|
| 26 |
+
Then follow the instructions in the README.md file to set up the environment and run the scripts.
|
| 27 |
+
|
| 28 |
+
Currently, the official Transformer repo does not support Chameleon model with image output (see [this PR](https://github.com/huggingface/transformers/pull/32013) for more details), so we rely on a certain fork of the repo.
|
| 29 |
+
|
| 30 |
+
After installing Align-Anything and correctly set up the envrionment, you can install the forked stable version of the repo by running:
|
| 31 |
+
|
| 32 |
+
```bash
|
| 33 |
+
pip install git+https://github.com/htlou/transformers.git@hantao_stable_cham
|
| 34 |
+
```
|
| 35 |
+
|
| 36 |
+
If you want to generate image (pure text generation can be directly done by `Transformers`), you can follow the instructions in the [mmsg_chameleon](https://github.com/htlou/mmsg_chameleon) repo to run the inference.
|
| 37 |
+
|
| 38 |
+
```bash
|
| 39 |
+
git clone https://github.com/htlou/mmsg_chameleon.git
|
| 40 |
+
cd mmsg_chameleon
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
+
Then set up the envrionment using
|
| 44 |
+
```bash
|
| 45 |
+
pip install -e .
|
| 46 |
+
```
|
| 47 |
+
|
| 48 |
+
After setting up the envrioment, set up the correct paths in `scripts/interleaved_gen.sh` and then run
|
| 49 |
+
```bash
|
| 50 |
+
bash scripts/interleaved_gen.sh
|
| 51 |
+
```
|