AA Chameleon 7b Base

🚀 Align-Anything Chameleon 7B Base

A powerful model for text-image interleaved input and output, enhancing image generation capabilities.

🚀 Quick Start

✨ Features

The Align-Anything Chameleon 7B Base is a potent model designed for text-image interleaved input and output. It builds upon the Chameleon model and is trained using the Align-Anything framework to further unleash its image generation potential.

📦 Installation

To use this model, refer to the Align-Anything repository for comprehensive details on training, inference, and evaluation.

First, clone the repository:

git clone https://github.com/PKU-Alignment/align-anything.git
cd align-anything/projects/text_image_to_text_image

Then, follow the instructions in the README.md file to set up the environment and execute the scripts.

Currently, the official Transformer repo does not support the Chameleon model with image output (see this PR for more details). Therefore, we rely on a specific fork of the repo.

After installing Align-Anything and properly setting up the environment, install the forked stable version of the repo by running:

pip install git+https://github.com/htlou/transformers.git@hantao_stable_cham

💻 Usage Examples

Basic Usage

If you want to generate images (pure text generation can be directly performed using Transformers), follow the instructions in the mmsg_chameleon repo to run the inference.

Clone the repository:

git clone https://github.com/htlou/mmsg_chameleon.git
cd mmsg_chameleon

Set up the environment:

pip install -e . 

After setting up the environment, configure the correct paths in scripts/interleaved_gen.sh and then run:

bash scripts/interleaved_gen.sh

📄 License

This project is licensed under the CC BY 4.0 license.