Neural Artistic Style Transfer
This is a pytorch implementation of Neural Artistic Style Transfer, which contains both gradient-based methods (e.g., Gatys et al.) and feed-forward methods (e.g., Johnson et al.).
Setup
All code is implemented in Ubuntu 16.04 with following packages:
- Pytorch >= 1.0
- Python >= 3.6
- Cuda >= 9.0
Install pytorch-colors to preserve original colors of content image onto results. Please follow the README script to install it.
Examples
Examples from Gradient-based methods
1. Results produced by a normal gradient-based neural style transfer
2. Results produced by gradient-based method with multi-labels in corresponding masks.
3. Gradual transferred results
Examples from Feed-forward methods
1. Results produced with multi-labels in corresponding style masks
2. Results produced by normal feed-forward networks
