Crack the Code: MXFP in Triton with Hicham Badri скачать в хорошем качестве

Crack the Code: MXFP in Triton with Hicham Badri

MXFP is a new quantization format featured in Nvidia's Blackwell GPUs to accelerate AI workloads. Join us to: Deep dive into the MXFP format specification and its practical implementation in Triton. Learn how to write efficient kernels for various MXFP operations to harness the full power of the new architecture. Featuring: Hicham Badri (Co-Founder & Principal Research Scientist at Mobius Labs) Don’t miss this opportunity to power up your AI workloads with the next-gen GPU technology! ---- Enjoy the Quality and Efficiency! Want to take it further? Compress your models with Pruna: https://www.pruna.ai/ Join our live events: https://luma.com/pruna-ai?k=c Stay up to date with the latest AI efficiency research on our blog: https://www.pruna.ai/blog Join the conversation in our Discord:   / discord  . Why Pruna AI? Pruna provides a model optimization framework that enables the delivery of faster, more efficient models with minimal overhead, making the AI solutions smaller, cheaper, faster, and greener. #quantization #efficiency #ai #sustainability #triton