TorchStyle is a Pytorch based framework for GAN based Neural Style Transfer.
Project description
TorchStyle is a Pytorch based framework for designing and developing GAN based Neural Style Transfer. This framework has been designed to provide building blocks for popular GANs and also to allow customization for cutting edge research.
System requirements
TorchStyle should work on
- all Linux distributions no earlier than Ubuntu 16.04
- macOS X
- Windows 10
Installation
Using pip (for stable release):
pip install torchstyle
Using pip (for latest master):
conda install -c tczhangzhi torchstyle
From source:
git clone https://github.com/tczhangzhi/torchstyle.git cd torchstyle python setup.py install
Getting Started
The network architectures of popular GANs and cutting edge GAN based Neural Style Transfer methods are out of the box:
from torchstyle.model.acgan import Generator, Discriminator g_model = Generator() d_model = Discriminator()
The diverse loss functions and complicated adversarial training methods are packed in different learners. It is convenient for users to reproduce the state-of-the-art training methods quickly without knowing implementation details:
from torchstyle.trainer.acgan import ACGANTrainer trainer = ACGANTrainer(g_model, d_model, dataloader, device=device) trainer.start(epoch=10)
Test
Setup environments:
export PYTHONPATH=./
Run the test case:
python tests/test_acgan.py
Documentation
Coming soon.
License
PyTorch is MIT-style licensed, as found in the LICENSE file.
Copyright (c) 2020-present, Zhang Zhi
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Filename, size | File type | Python version | Upload date | Hashes |
---|---|---|---|---|
Filename, size torchstyle-0.0.1.tar.gz (2.9 kB) | File type Source | Python version None | Upload date | Hashes View |