Open AlbertoPresta opened 2 years ago
Thank you for your interest in this work and yes, the scaling factors (a_i and b_i) are all trainable parameters. You may check the code and paper here(https://github.com/tongxyh/ImageCompression_VariableRate) for more details about variable rate learned image compression.
Hi,
first of all congratulations for your paper. it is a remarkable work in my opinion. I have just a question (maybe is a stupid one): how do you get the scaling factors (ai,bi)? are they trainable parameters or not?
Thanks in advance for your answer, Alberto