Autoregressive Image Generation without Vector Quantization — mentioned by schopra909

Thu, 17 Sep 2026 00:43:42 GMT

Just skimmed the paper (haven't seen this before). Might need to read more carefully, but at first glance I don't understand their intuition why the "Markovian property limits the model’s ability to fully utilize the generation trajectory". When it comes to multi-resolution training (e.g. matryoshka training), there are precedents that don't require this AR formulation. They cite MAR (https://arxiv.org/pdf/2406.11838), which I think is a much clearer articulation of "autoregressive diffusion". MAR uses an autoregressive base (like an LLM) and staples on a small MLP on-top that's trained as a…

Paper and discussion history. A mention is not an endorsement.

Autoregressive Image Generation without Vector Quantization — mentioned by zaptrem

Wed, 26 Mar 2025 08:40:55 GMT

Token by token diffusion was done by MAR https://arxiv.org/abs/2406.11838 and Fluid (scaled up MAR) https://arxiv.org/abs/2410.13863 When you randomly pick the locations they found it worked okay, but doing it in raster order (left to right, top to bottom) they found it didn't work as well. We tried it for music and found it was vulnerable to compounding error and lots of oddness relating to the fragility of continuous space CFG.

Paper and discussion history. A mention is not an endorsement.

Autoregressive Image Generation without Vector Quantization · HackerPapers discussions

Autoregressive Image Generation without Vector Quantization — mentioned by schopra909

Autoregressive Image Generation without Vector Quantization — mentioned by zaptrem