Linear transformers are faster after all(manifestai.com)94 points by JoelEinbinder 2 years ago | 16 comments
RenderFormer: Neural rendering of triangle meshes with global illumination(microsoft.github.io)259 points by klavinski 1 year ago | 51 comments
CODA: Rewriting Transformer Blocks as GEMM-Epilogue Programs(arxiv.org)96 points by matt_d 5 months ago | 12 comments
A Mathematical Perspective on Transformers(arxiv.org)63 points by georgehill 2 years ago | 4 comments