StyDiT: High-Quality Artistic Style Transfer Using Diffusion Transformers
This repository will host the official implementation of StyDiT, a diffusion-transformer-based framework for artistic style transfer.
The corresponding paper has been submitted to IEEE Transactions on Multimedia (TMM) and is currently under review.
Upon acceptance, we will release the full training and inference code, pretrained model weights, and the Aes4M dataset to facilitate reproducibility and future research.
Please stay tuned for updates.