rainbowneko 0.5

Neural network training and inference framework

RainbowNeko Engine

📘English document 📘中文文档

Introduction

RainbowNeko Engine is a toolbox for pytorch based neural network training and inference. Multiple tasks and training strategies are built-in and highly expandable.

Install

1. Install pytorch

2. Install from source:

git clone https://github.com/IrisRainbowNeko/RainbowNekoEngine.git
cd RainbowNekoEngine
pip install -e .

3. To use xFormers to reduce VRAM usage and accelerate training:

# use conda
conda install xformers -c xformers

# use pip
pip install xformers>=0.0.17

User guidance

Start a new project

mkdir my_project
cd my_project
# Modified based on this project or start a new project and make initialization
nekoinit

Training

Training scripts based on 🤗 Accelerate or Colossal-AI are provided.

• For 🤗 Accelerate, you may need to configure the environment before launching the scripts.
• For Colossal-AI, you can use torchrun to launch the scripts.

# with Accelerate
neko_train --cfg cfgs/train/cfg_file.yaml
# with Accelerate and only one GPU
neko_train_1gpu --cfg cfgs/train/cfg_file.yaml

Inference

TODO

Tutorials

TODO

Contributing

You are welcome to contribute more models and features to this toolbox!