https://github.com/tomgorb/artistic-style-transfer
real-time style transfer in pyTorch and TensorFlow
https://github.com/tomgorb/artistic-style-transfer
Last synced: 10 months ago
JSON representation
real-time style transfer in pyTorch and TensorFlow
- Host: GitHub
- URL: https://github.com/tomgorb/artistic-style-transfer
- Owner: tomgorb
- Created: 2020-07-16T15:22:42.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2020-07-16T15:24:41.000Z (almost 6 years ago)
- Last Synced: 2025-01-09T05:25:33.209Z (over 1 year ago)
- Language: Python
- Homepage:
- Size: 30.8 MB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
Artistic Style Transfer
---
- [COCO dataset](https://cocodataset.org/#download)
- [TensorFlow 1.x models](https://drive.google.com/file/d/1G-J0KlSp9Z4QJR98T_aAbSruW5uHG42z/view?usp=sharing).
- convert content images to *jpg* before doing stylization (*png* have **4** channels).
- pyTorch
- requirements: torch==1.3.1+cpu
- usage: ```python neural_style.py eval --content-image 01.jpg --model legion.pth --output-image 01-legion.png --cuda 0```
- TensorFlow
- requirements: tensorflow==1.15.0, imageio==2.9.0
- usage: ```python stylize_image.py --content results/01.jpg --network-path models/dora-marr-network --output-path results/01-dora-marr.jpg```