Federico Cassano
@ellev3n11
understanding the universe @spacexai - trained composer @cursor_ai
加入 September 2020
354 正在關注 7.8K 粉絲
most cracked intern
We’ve open-sourced the MoE megakernel we use to train models on NVL72s. A few of my favorite details: bf16 and mxfp8 support, pull-based dispatch, configurable overlap granularity, full determinism, and no CPU-GPU syncs. Read more in our blog and code!
顯示更多
