Extract vision embeddings

[hep-raidium]

Hi there, First, thank you for sharing your work!

I would like to extract one embedding for each image from the vision encoder. I didn't succeed in doing so, have you ever done it ? If yes, do you have a few hints / code to share ?

Thx!!

[ljwztc]

After obtaining the feature map, you can scale up to the input size. Then you can find the feature vector for each category with ground truth.