Hi! You mentioned the numpy array in image embedding .npy files generated from clip_retrieval lost the index correspondence to the original image .tar files. I wonder are they completely irrelevant orders or the same sequential order but divided into different shards?
Hi! You mentioned the numpy array in image embedding
.npy
files generated from clip_retrieval lost the index correspondence to the original image.tar
files. I wonder are they completely irrelevant orders or the same sequential order but divided into different shards?