README.md

---
license: mit
tags:
- gan
- pytorch
- vision
- cats
- dcgan
metrics:
- loss
datasets:
- huggan/cats
---

# CatGen v2 - 128px DCGAN

This model is a Deep Convolutional Generative Adversarial Network (DCGAN) trained to generate high-quality 128x128 images of cats. It was trained for 165 epochs on a curated dataset of feline images, pushing the boundaries of traditional GAN architectures at this resolution.

## Sample
Here's a sample after epoch 165:
![__results___8_0](https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/VV8AhZgJFA_dvsV1-ul7P.png)

## Best of - Cat Images

![best_of_1](https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/bOThglzoRcfy8nNnVjxGg.png)
![best_of_2](https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/OGAZijZhGyY4Ss1k2zRPo.png)
![best_of_3](https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/UATLmkIogTIZhyFJUTImA.png)

## Model Details
- **Architecture:** DCGAN (Deep Convolutional GAN)
- **Resolution:** 128x128 pixels (RGB)
- **Parameters:** ~186M (Generator)
- **Training Duration:** ~5 hours on NVIDIA T4 GPU
- **Framework:** PyTorch with Mixed Precision (AMP)

## Training Hyperparameters
- **Batch Size:** 128
- **Learning Rate:** 0.0002
- **Optimizer:** Adam (Beta1: 0.5, Beta2: 0.999)
- **Latent Vector (Z):** 128 dimensions

## Training details
The full training code can be found as `catgen-v2.ipynb` in this repo.
The training data we used was from HF: huggan/cats

## Intended Use
This model is intended for artistic and research purposes. It demonstrates how GANs can capture complex textures like fur and eye reflections at medium resolutions.

## How to use
To use this model, clone this repository and run the provided inference script. Ensure you have `matplotlib`, `torch` and `torchvision` installed.

```bash
python3 inference.py
```

--> Sample output:
![image](https://cdn-uploads.huggingface.co/production/uploads/697f2832c2c5e4daa93cece7/UA3btFlIlqwEhrTOEaqRe.png)

## Limitations & Bias
As a GAN, the model might occasionally produce "dream-like" artifacts or distorted anatomy (e.g., extra ears or eyes). It is not a diffusion model and generates images in a single forward pass.