I was curious about VQ-VAE and found your model because not everybody has implemented the PixelCNN as well. However, just running a quick test on MNIST, it only generates grey images. Inspecting the generated tensor's values, they're all roughly 0.44. Any thoughts as to what might be going on?
I was curious about VQ-VAE and found your model because not everybody has implemented the PixelCNN as well. However, just running a quick test on MNIST, it only generates grey images. Inspecting the
generatedtensor's values, they're all roughly0.44. Any thoughts as to what might be going on?