FPTalks 2026 Focuses on Low-Precision LLM Pretraining
nmwsharp · x · 2026-08-06
FPTalks 2026 will feature several hardcore talks on floating-point and numerical analysis. Highlights include a presentation from NVIDIA on pretraining Large Language Models using NVFP4, alongside research on robust geometry for neural implicit surfaces and direction-preserving number representations.
More from Infra
- The Enterprise AI Question: Where Does Your AI Actually Run? — DavidLinthicum · 2026-08-06
- AI Inference Demand Growing 10x Yearly Will Make Compute Scarcity the Default — TansuYegen · 2026-08-06
- If Your Data Can't Move, Your AI Strategy Is Doomed — DavidLinthicum · 2026-08-06
- Nativ brings LiquidAI LFM2.5 to Mac locally: 82 tok/s decode, <8.5GB RAM for 128K context — JosephJacks_ · 2026-08-06
- Startup Panthalassa Develops Floating AI Data Centers Powered by Wave Energy — TinfoilTricorn · 2026-08-06
- Maple 20B Hits 9.8k Tokens/s on Single GH200, Opens Free Inference Endpoint — MaziyarPanahi · 2026-08-06