Driss Guessous
where else to find them
what they talk about
where they keep going back
no series more than once — all 2 in the filters
on a stage with
- Boyuan Fengonce
- Vasily Kuznetsovonce
who talks about the same things
- Christian Keller5 talks
- Horace He4 talks
- Chien-Chin Huang3 talks
- Garrett Goon2 talks
- Philip Kiely6 talks
- Luca Antiga5 talks
| title | conference | date |
|---|---|---|
| FlexAttention + FlashAttention-4: Fast and Flexible | PyTorch Conference Europe 2026 | 2026-04-07 |
FlexAttention integrates with FlashAttention-4 to run Flex-style score and mask hooks as JIT-instantiated CuTeDSL kernels via PyTorch Inductor, converting… | ||
| mxfp8, mxfp4, nvfp4 formats and applications in PyTorch | PyTorch Conference 2025 | 2025-10-22 |
MicroScaling formats MXFP8, MXFP4 and NVFP4 appear as concrete PyTorch features: MXFP8 stores tensors ~2x smaller and achieves up to 2x speed on matrix… | ||
| Flex Attention for Inference | PyTorch Conference 2025 | 2025-10-22 |
FlexAttention ships in PyTorch 2.5.0 and implements most attention variants in a few lines of PyTorch; it adds a compiler-driven FlexDecoding backend that uses… | ||
| nothing matches | ||