Comparison of codebook vectors of autoencoders (DALLE's dVAE vs VQGAN) that map any image to a fixed vocabulary of vectors
Almost Any Image Is Only 8k Vectors. Post describing the comparison
Image above includes image of a bird from DIV2k dataset. Image used with permission by authors for the article above.
Visualization of DALL-E codebook
Visualization of VQGAN codebook
ajitrajasekharan / codebook_comparisons Goto Github PK
View Code? Open in Web Editor NEWComparison of codebook vectors of autoencoders (DALLE's dVAE vs VQGAN) that map any image to a fixed vocabulary of vectors
License: MIT License