https://github.com/tcapelle/cloud_diffusion

Diffusion on the Clouds: Short-term solar energy forecasting with Diffusion Models
https://github.com/tcapelle/cloud_diffusion

Last synced: 3 days ago
JSON representation

Diffusion on the Clouds: Short-term solar energy forecasting with Diffusion Models

• Host: GitHub
• URL: https://github.com/tcapelle/cloud_diffusion
• Owner: tcapelle
• Created: 2023-03-10T11:37:45.000Z (over 1 year ago)
• Default Branch: master
• Last Pushed: 2024-03-21T20:24:13.000Z (8 months ago)
• Last Synced: 2024-10-01T12:37:41.817Z (about 1 month ago)
• Language: Python
• Size: 12.6 MB
• Stars: 37
• Watchers: 2
• Forks: 10
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
[![](https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-gradient.svg)](https://wandb.ai/capecape/ddpm_clouds/reports/Diffusion-on-the-Clouds-Short-term-solar-energy-forecasting-with-Diffusion-Models--VmlldzozNDMxNTg5)

[![PyPI version](https://badge.fury.io/py/cloud_diffusion.svg)](https://badge.fury.io/py/cloud_diffusion)

# Cloud Diffusion Experiment

You can check our GTC presentation on YouTube:

[![](assets/front.jpg)](https://www.youtube.com/watch?v=L5h9kbMMzZs)

Samples and training logs for the model generations can be found [here](https://wandb.me/gtc2023).

This codebase contains an implementation of a deep diffusion model applied to cloud images. It was developed as part of a research project exploring the potential of diffusion models for image generation and forecasting.

## Setup

1. Clone this repository and run `pip install -e .` or `pip install cloud_diffusion`

2. Set up your WandB account by signing up at [wandb.ai](https://wandb.ai/site).

3. Set up your WandB API key by running `wandb login` and following the prompts.

## Usage

To train the model, run `python train.py`. You can play with the parameters on top of the file to change the model architecture, training parameters, etc.

You can also override the configuration parameters by passing them as command-line arguments, e.g.

```bash

> python train.py --epochs=10 --batch_size=32

```

## Training a Simple Diffusion Model

This training is based on a Transformer based Unet (UViT), you can train the default model by running:

```bash

> python train_uvit.py

```

## Running Inference

If you are only interested on using the trained models, you can run inference by running:

```bash

> python inference.py  --future_frames 10 --num_random_experiments 2

```

This will generate 10 future frames for 2 random experiments.

## License

This code is released under the [MIT License](LICENSE).