Master deep generative models in PyTorch with ease!
Welcome to Diffusion-GAN-VAE-PyTorch! This repository is your ultimate resource for mastering deep generative models, implemented from scratch in PyTorch. It features Variational Autoencoders (VAE), Generative Adversarial Networks (GAN), Conditional GANs, Diffusion Models, and Conditional Diffusion Models, all crafted with clarity and precision.
| Source Code | Website |
|---|---|
| github.com/deepmancer/diffusion-gan-vae-pytorch | deepmancer.github.io/diffusion-gan-vae-pytorch |
- 🧩 Modular & Educational
- 🔍 Explore Cutting-Edge Models
- 💡 Beginner-Friendly Yet Research-Ready
- 📕 Fully documented
👉 Star this repo if you find it helpful, and join our community of AI enthusiasts!
Ensure the following dependencies are installed:
- Python 3: The programming language used.
- PyTorch: The deep learning framework for model building and training.
- NumPy: For numerical computations.
- Matplotlib: For result visualizations.
- tqdm: For progress tracking.
Install dependencies with pip:
pip install torch numpy matplotlib tqdmA VAE learns a probabilistic latent space, enabling smooth interpolation and robust generation of new data points.
GANs pit a generator against a discriminator in a game-like setup, creating highly realistic samples over time.
cGANs incorporate conditional inputs (like class labels) to control data generation, enabling targeted synthesis.
Diffusion Models simulate a stochastic process to progressively model complex distributions, resulting in high-quality generation.
Building on diffusion models, Conditional Diffusion Models allow for guided, condition-driven generation.
-
Clone the repository:
git clone https://github.com/deepmancer/diffusion-gan-vae-pytorch.git cd diffusion-gan-vae-pytorch -
Install dependencies.
-
Run the notebook!
This project is licensed under the MIT License. Feel free to use it in your projects while crediting the repository. See the LICENSE file for details.