From DeepSpeed to FSDP and Back Again with Hugging Face Accelerate
•
45
Stas Bekman
stas
AI & ML interests
Toolmaker. Software creator, optimizer and harmonizer.
Makes things work and fly at Contextual.AI
Training LLM/RAG/Generative AI/Machine Learning/Scalability
Recent Activity
updated
a model
6 days ago
stas/ml-engineering-book
posted
an
update
22 days ago
If you remember my work on MAMF - to find the realistic TFLOPS achievable ceiling - the Intel AI team has shared their measurements and they scored ...
an incredible 99.4% TFLOPS efficiency for Gaudi 2!
That's quite amazing! Your ROI on these accelerators will be very high.
The full table is here: https://github.com/stas00/ml-engineering/tree/master/compute/accelerator#maximum-achievable-matmul-flops-comparison-table
As we have seen the competitors get their achievable efficiency worse with each new generation, I'm looking forward to see if Gaudi 3 will keep the high bar!
Thanks to Avi Rubin, Lakshman Chari, Imtiaz Sajwani, Ramy J and Zhiqi Tao for helping to get these numbers to the community.
Articles
Organizations
stas's activity
upvoted
an
article
4 months ago
Article
Mixture of Experts Explained
•
214
upvoted
an
article
7 months ago
Article
From DeepSpeed to FSDP and Back Again with Hugging Face Accelerate
•
45
upvoted
an
article
8 months ago
Article
Introducing Idefics2: A Powerful 8B Vision-Language Model for the community
•
170