Hi,
Thanks for your work.
I have a question about static decoding in SDAR. From the paper, my understanding is that when block size is B, static decoding uses B denoising steps for each block.
In my evaluation, when I set block size to 32 with 32 denoising steps, static decoding becomes extremely slow.
Have you tested static decoding with block size 32? If so, did you also use 32 denoising steps, or did you use a different setting in practice?
Thanks!