Getting Started¶

Python Installation¶

Download from the Official Website¶

Download Python and install it.

Install via Conda (Recommended)¶

Download Miniconda.

Open Miniconda and create a Python environment:

conda create --name neko python=3.11 -y
conda activate neko

PyTorch Installation¶

Install PyTorch:

conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia

RainbowNeko Engine Installation¶

Install from Python Package

pip install rainbowneko

Install from Source Code

Download and install:

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

Optional Installations¶

Note

Use xformers to reduce memory usage and accelerate training. Find the version corresponding to your PyTorch version on the xformers releases page, then install it:

pip3 install -U xformers==<version> --index-url https://download.pytorch.org/whl/<cuda_version> 
# Replace <cuda_version> with your CUDA version (e.g., cuda12.4 corresponds to cu124).

To monitor the training process with TensorBoard:

pip install tensorboard

To monitor the training process with WandB:

pip install wandb