Closed Alleycat404 closed 2 years ago
Hi,
Thank you for your interest. I started with a simple implementation -- did not have the goal to fully reproduce the paper.
This is also mentioned in the wiki:
The models are inspired by [1], [2]. They are just experiments I have tried and do not fully implement the architectures described in the referenced papers. [1] Lossy Image Compression with Compressive Autoencoders, Theis et al. [2] Variable Rate Image Compression with Recurrent Neural Networks, Toderici et al.
I see, thanks for your timely reply
Thanks for sharing. But I believe the loss function used in the paper "Lossy Image Compression with Compressive Autoencoders" is supposed to be Rate-distortion, why is MSE loss used instead?