# Palette-PL-sample

**Repository Path**: guyi2000/Palette-PL-sample

## Basic Information

- **Project Name**: Palette-PL-sample
- **Description**: A minimum exemplar of the Palette model implemented using the PyTorch Lightning framework.
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: main
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2024-06-02
- **Last Updated**: 2024-06-02

## Categories & Tags

**Categories**: Uncategorized

**Tags**: diffusion, AI, diffusion-models, pytorch-lightning

## README

# Palette: image-to-image diffusion model

A minimum exemplar of the [Palette](https://iterative-refinement.github.io/palette/) model implemented using the [PyTorch Lightning](https://lightning.ai/docs/pytorch/latest/) framework.

## Environment Setup

Setup the environment with the following commands:

```shell
conda create -n palette
conda activate palette
# Install pytorch & torchvision
conda install pytorch torchvision torchaudio pytorch-cuda=11.8 -c pytorch -c nvidia
# Install pytorch-lightning
conda install pytorch-lightning -c conda-forge
# Install pytorch-lightning CLO
pip install "pytorch-lightning[extra]"
# Install tensorboard
conda install tensorboard
# Install opencv for python
pip install opencv-python
```

Then, you need download the dataset [CelebaHQ here](https://www.kaggle.com/datasets/badasstechie/celebahq-resized-256x256). Put it in the `./data` directory and unzip.

```shell
unzip achieve.zip
```

## Usage

To **train** the model, run the following command:

```shell
python main.py fit -c ./conf/config.yaml
```

To **test** the model, run the following command:

```shell
python main.py test -c ./conf/config.yaml
```

## Acknowledgments

This code borrows heavily from [guided diffusion model](https://github.com/openai/guided-diffusion) and so on.