Yuming Jiang

Hi @MarcuseXiao , thanks for your interest in our dataset. I am so sorry that we do not have links to the Baidu disk for now.

Hi @TianhaoFu , thanks for your interest in our work! Our dataset doesn't provide camera pose labels.

Thanks for your interest in our work. Our method is developed under Linux system. We don't test it under the windows system. We will consider adding colab demo later~

Sorry, the C2-Matching does not support distributed training.

Thanks for your interest in our work! For datasets without any references, we still need the ann_file. You may construct the ann_files like [this](https://github.com/yumingj/C2-Matching/blob/master/datasets/CUFED5_pairs.txt). For each line, the first one...

Hi, thanks for your interest in our work! Can you have a look at the size of img_in_up? It should have the same size as img_ref_up.

Yes. In the current implementation, if the reference images are smaller than 512x512, we will do the padding operation to make the reference images have that resolution.

你好，这个应该和GAN Loss weight有一些关系。如果想要有更好的视觉效果，可以对gan loss weight进行调整。

Hi, for now, the pre-trained model of the teacher contrastive network would not be released temporarily, but I will consider releasing the pre-trained teacher model if I am available. Since...

1. The alpha [here](https://github.com/yumingj/Talk-to-Edit/blob/d0248c6d6c4c44436f8dbb9a40a6a3ec91709a28/models/base_model.py#L496) controls the positive and negative direction of the attribute. 2. We allow to specify some fine-grained target attributes. You need to use the edit step to...