image input

Open Dpw506 opened this issue 4 years ago • 2 comments

Hi, I am very glad to discuss with you about Transformer. I would like to ask you how to input data into the network if it is a image.I want to see the effect of the specific picture here, and I hope to receive your reply as soon as possible. In addition, I tried to change the image into four dimensions by myself, and then presented the results, but the results were garbled. Have you ever tried to send the pictures into the SIT model? Thank you so much!

Apr 14 '21 09:04 Dpw506

Load images
Convert them to tensor or dataloader if you are training
Make sure the dimension of image is 3 and 4th dimension is batch dim
And then just pass that tensor to SiT model

Apr 23 '21 23:04 rishikksh20

Ok, Thank you for your reply!

Apr 28 '21 01:04 Dpw506