About reconstructed image and MS-SSIM for TF2 version model.
59 views
Skip to first unread message
shahid abbas
unread,
Aug 24, 2021, 12:09:07 AM8/24/21
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to tensorflow-compression
Hi Johannes,
I trained your model by myself, here I have two queries.
Reproducibility Problem: Once the training is complete and if I want to encode and decode the image, where sometimes the reconstructed image is not decoded properly. sample image is attached.
MS-SSIM: I use the ms-ssim function "train_loss=tf.reduce_mean(tf.image.ssim_multiscale(x,x_tilde,255))" in T2 dependent model but after completing training the reconstructed image doesn't seem like original (reconstructed image is either black or gray). but the same function works properly in the model which is dependent on TF1 version.