https://github.com/vahid0001/tiny-pix2pix
Redesigning the Pix2Pix model for small datasets with fewer parameters and different PatchGAN architecture
https://github.com/vahid0001/tiny-pix2pix
gan generative-adversarial-network patchgan pix2pix pix2pix-model tinypix2pix unet
Last synced: 5 months ago
JSON representation
Redesigning the Pix2Pix model for small datasets with fewer parameters and different PatchGAN architecture
- Host: GitHub
- URL: https://github.com/vahid0001/tiny-pix2pix
- Owner: vahid0001
- Created: 2019-08-26T20:27:30.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2025-10-18T14:34:14.000Z (8 months ago)
- Last Synced: 2025-10-19T08:48:59.587Z (8 months ago)
- Topics: gan, generative-adversarial-network, patchgan, pix2pix, pix2pix-model, tinypix2pix, unet
- Language: Python
- Homepage:
- Size: 1.19 MB
- Stars: 23
- Watchers: 4
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# tiny-pix2pix
Tiny Pix2Pix is a compact PyTorch re-implementation of the pix2pix image-to-image translation model designed for small images such as CIFAR-10. The project provides a lightweight U-Net generator, a PatchGAN discriminator, and a training pipeline that follows modern PyTorch best practices.
## Getting started
### Installation
Create and activate a virtual environment, then install the project dependencies:
```
pip install -r requirements.txt
```
If you already have PyTorch and torchvision installed, you can skip reinstalling them by running:
```
pip install -r requirements.txt --no-deps
```
### Training
The training script downloads CIFAR-10 automatically and trains the Tiny Pix2Pix model by pairing each image with itself, replicating the behaviour of the original demonstration script:
```
python train.py --epochs 5 --batch-size 8 --dataset-root ./data --device cuda
```
Model checkpoints are stored in the `models/` directory by default. Use `python train.py --help` to inspect all available options.
## Important links
- [Original paper](https://arxiv.org/abs/1611.07004)
- [Project page](https://phillipi.github.io/pix2pix/)
## Model Architecture
### Generator
U-Net: The generator in pix2pix resembles an auto-encoder. The Skip Connections in the U-Net differentiate it from a standard Encoder-decoder architecture. The generator takes in the Image to be translated and compresses it into a low-dimensional, “Bottleneck”, vector representation. The generator then learns how to upsample this into the output image. The U-Net is similar to ResNets in the way that information from earlier layers are integrated into later layers. The U-Net skip connections are also interesting because they do not require any resizing, projections etc. since the spatial resolution of the layers being connected already match each other.
### Discriminator
PatchGAN: The discriminator used in pix2pix is another unique component to this design. The PatchGAN works by classifying individual (N x N) patches in the image as “real vs. fake”, opposed to classifying the entire image as “real vs. fake”. The authors reason that this enforces more constraints that encourage sharp high-frequency detail. Additionally, the PatchGAN has fewer parameters and runs faster than classifying the entire image.
### tiny-pix2pix
