Hello. I'm currently using an image to verify the part of the code about high resolution processing.
When I run the code：image_tensor, hw_patch_nums = process_image(image, image_mean=image_mean,image_std=image_std )
hw_patch_nums According to this code is supposed to return a list containing three dimensions, slice_h_num, slice_w_num, and j(position).
But when I pass image_tensor and hw_patch_nums as arguments to the classes in adapt_clip.py, I run into a bug. that's where code I found that at this point, hw_patch_nums only has two dimensions and is missing information about the third dimension (j position).
May I ask if this has been processed in any other way in between?
Hello. I'm currently using an image to verify the part of the code about high resolution processing.
When I run the code：
image_tensor, hw_patch_nums = process_image(image, image_mean=image_mean,image_std=image_std )hw_patch_numsAccording to this code is supposed to return a list containing three dimensions, slice_h_num, slice_w_num, and j(position).But when I pass image_tensor and hw_patch_nums as arguments to the classes in adapt_clip.py, I run into a bug. that's where code I found that at this point, hw_patch_nums only has two dimensions and is missing information about the third dimension (j position).
May I ask if this has been processed in any other way in between?
Looking forward to your reply!