Mateen Ulhaq

> [!NOTE] > The above comment still applies. I took another look at the paper, and it says: > The architecture of $g_{\\text{ch}}$ network is frankly sketched from Minnen et...

Please see the following two sections: - https://yodaembedding.github.io/post/learned-image-compression/#probabilistic-modeling-for-data-compression - https://yodaembedding.github.io/post/learned-image-compression/#entropy-modeling Briefly, each element of y is encoded using a probability distribution p. The y_likelihoods represent the value p(y). > A...

1. The exact same scales need to be available at the decoder, so you will need to transmit them to the decoder as well. ``` ENCODER ==== Communication channel ====>...

For the $i$-th symbol, there is a probability distribution $f_i : \mathbb{Z} \to [0, 1]$. (I called it $p\_{{\hat{y}}\_i}$ earlier.) By definition, $\int_{-\infty}^{\infty} f_i(t) \\, dt = 1$. But how...

That looks like [subpixel convolution](https://distill.pub/2016/deconv-checkerboard/) upscaling gone wrong. I believe @chyomin06 has run into similar issues. I suspect it's because the encoder/decoder are not "in sync" due to non-deterministic CUDA...

My other theory is that the image lies far outside of the distribution of images (dataset) that it was trained on, so each of the 4 upscales within the synthesis...

One can probably just install [`bin2c`](https://manpages.ubuntu.com/manpages/focal/man1/bin2c.1.html): - https://ubuntu.pkgs.org/18.04/ubuntu-universe-amd64/hxtools_20170430-1_amd64.deb.html - https://aur.archlinux.org/packages/bin2c - https://sourceforge.net/projects/bin2c/

Please provide the input image file and model name / settings and commands/code used. Consider setting `torch.backends.cudnn.deterministic = True`, or other PyTorch determinism tricks.

Each mosaic block is probably happening because of some improperly decoded element of y. For instance, if an element of y is usually ~0, and we accidentally decode a value...

Please see: - https://github.com/InterDigitalInc/CompressAI/issues/235#issuecomment-1740940306 - https://github.com/InterDigitalInc/CompressAI/issues/279#issuecomment-2019388285