Stas Bekman

stas

https://stasosphere.com/machine-learning/

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

posted an update about 1 month ago

Do you want ArcticTraining at @SnowflakeDB to add an ability to post-train DeepSeek V3/R1 models with DPO using just a few GPU nodes? Please vote here and tell others about it: https://github.com/snowflakedb/ArcticTraining/discussions/58 ArcticTraining is an open-source, easy to use post-training framework for NVIDIA GPUs built on top of DeepSpeed.

updated a model 2 months ago

stas/ml-engineering-book

posted an update 4 months 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.

View all activity

Organizations

stas's activity

published an article 10 months ago

Article

From DeepSpeed to FSDP and Back Again with Hugging Face Accelerate

and 3 others •

Jun 13, 2024

• 48

published an article over 1 year ago

Article

Introducing IDEFICS: An Open Reproduction of State-of-the-art Visual Language Model

and 10 others •

Aug 22, 2023

• 31

published an article over 2 years ago

Article

Incredibly Fast BLOOM Inference with DeepSpeed and Accelerate

and 1 other •

Sep 16, 2022

• 1

published an article over 2 years ago

Article

The Technology Behind BLOOM Training

•

Jul 14, 2022

• 29

published an article about 4 years ago

Article

Fit More and Train Faster With ZeRO via DeepSpeed and FairScale

•

Jan 19, 2021

• 4

published an article over 4 years ago

Article

Porting fairseq wmt19 translation system to transformers

•

Nov 3, 2020