search

ByChelsea / VAND-APRIL-GAN

[CVPR 2023 Workshop] VAND Challenge: 1st Place on Zero-shot AD and 4th Place on Few-shot AD
162 stars 20 forks source link

global image representations #31

Open TerryMelody opened 3 months ago

TerryMelody commented 3 months ago

Hello dear authors！In the code "image_features, patch_tokens = model.encode_image(image, features_list)", is "image_features" the global image representations? Just like the source code in openclip: "image_features = model.encode_image(image)", as we always do. But you change the transformer to add additional outputs of assigned "features_list". I don't know whether my understanding is right or not.

ByChelsea commented 3 months ago

Yes, image_features is the global image representation.

By the way, you can use "copy permalink" to help others locate your issue more quickly. Like this. image

TerryMelody commented 3 months ago

Copy link Thank you for your suggestions! Can I understand the “image features” in your code to be the same as those implemented in Openclip source codes. It seems you make a lot of modifications to openclip (https://github.com/mlfoundations/open_clip/tree/main/src/open_clip)?

ByChelsea commented 3 months ago

Yes, image features are the class tokens in the original implementation.

TerryMelody commented 3 months ago

Yes, image features are the class tokens in the original implementation.

Thank you so much!