Shengwei An

A little fault in README and the python doc string: some unnecessary and extra `</li>` may occur.

1

Here are the codes in your instrcution: ``` {% for comment in comment_list|fill_tree|annotate_tree %} {% ifchanged comment.parent_id %}{% else %}{% endifchanged %} {% if not comment.open and not comment.close %}{%...

The following is the function [`fixed_image_standardization`](https://github.com/timesler/facenet-pytorch/blob/master/models/mtcnn.py#L508): ``` def fixed_image_standardization(image_tensor): processed_tensor = (image_tensor - 127.5) / 128.0 return processed_tensor ``` `fixed_image_standardization` seems to normalize images in the range `[0., 255.]` into...

For class-conditional cases, StyleGAN2-ADA and StyleGAN2 embed labels differently. StyleGAN2-ADA uses a Dense layer. The output's shape of the discriminator is different as well.

The pixel range of ULP in [tiny-imagenet/train_ULP.py](https://github.com/UMBCvision/Universal-Litmus-Patterns/blob/master/tiny-imagenet/train_ULP.py#L66) is [0,255]. However, when training the clean/poisoned models, the pixel range of inputs is [0, 1] because of the [transforms.ToTensor() in tiny-imagenet/train_clean_model.py](https://github.com/UMBCvision/Universal-Litmus-Patterns/blob/master/tiny-imagenet/train_clean_model.py#L59) See...

Solve the loading problems mentioned in this closed issue: https://github.com/my89/imSitu/issues/3#issuecomment-469077352 Because class `baseline_crf ` has a default parameter [`cnn_type='resenet_101'`](https://github.com/my89/imSitu/blob/master/baseline_crf.py#L138), the `eval.py` will always try to load the weights to the...

Is there any consistency model on face dataset?

1

It would be great to see how this consistency model performs on face datasets such as CelebA-HQ and FFHQ, since the paper didn't mention the face dataset.