mysterious hhhh

[RFC]: Integrate Nunchaku to Support SVDQuant (W4A4) for Diffusion Models

2

### Motivation. We propose integrating SVDQuant (W4A4) to achieve significant inference speedup and reduced memory footprint for DiT models (e.g., FLUX, Wan2.2). Instead of re-implementing kernels, we choose to integrate...

Hi cuTile team, I have two specific questions regarding the support for Blackwell-specific hardware features: 1. Automatic Micro-block Scaling for FP8 When using fp8 with ct.matmul, how is the Micro-block...