dengyecode / T-former_image_inpainting

T-former: An Efficient Transformer for Image Inpainting (MM 2022)

This is the code for ACM multimedia 2022 “T-former: An Efficient Transformer for Image Inpainting”

visualization during training

python - m visdom.server

train:

python train.py --no_flip --no_rotation --no_augment --img_file your_data --lr 1e-4

fine_tune:

python train.py --no_flip --no_rotation --no_augment --img_file your_data --lr 1e-5 --continue_train

test:

python test.py --batchSize 1 --mask_type 3 --img_file your_data --mask_file your_mask your_data

Citation

If you are interested in this work, please consider citing:

@inproceedings{tformer_image_inpainting,
  author = {Deng, Ye and Hui, Siqi and Zhou, Sanping and Meng, Deyu and Wang, Jinjun},
  title = {T-former: An Efficient Transformer for Image Inpainting},
  year = {2022},
  isbn = {9781450392037},
  publisher = {Association for Computing Machinery},
  address = {New York, NY, USA},
  doi = {10.1145/3503161.3548446},
  pages = {6559–6568},
  numpages = {10},
  series = {MM '22}

}