“Stas is extremely discerning, wise, and is extraordinarily proficient at a vast number of things. He has given me excellent guidance over a number of years. He�...
Toolmaker. Author. Software creator, optimizer and harmonizer. Makes things work. Current domains: LLM/Retrieval/RAG/Scalability/Machine Learning - stas00.
Added Intel Gaudi3 and adding missing Gaudi2 specs via the paper www.intel.com/content/www… 1. TFLOPS: github.com/stas00/ml-engi… Note that the paper also disclosed a few Gaudi2 TFLOPS as well which were not disclosed until now! Thanks to @amitp_ai for letting me know about this new…
Twitter • 3 days ago
If your CI depends on orjson it's likely broken now as the new release is missing wheels for linux X86_64 github.com/ijl/orjson/iss… Workaround: pin to `orjson==3.10.6` thanks to github.com/smallsam for suggesting a workaround
Twitter • 3 days ago
Has github.com/Dao-AILab/flas… got faster than torch's SDPA? I'm seeing ~15-20% faster throughput with FA2@main Back in Feb-24 I clocked the 2 to give about the same training speed, so I switched to SDPA since it was a built in. I have just retested with torch-2.3.1 vs FA2 built…
Twitter • 3 days ago
If you need a working example of a cross-accelerator script, this script supports - NVIDIA: V100, A100, H100, ... - AMD: MI250, MI300X, ... - Intel Gaudi2+ github.com/stas00/ml-engi… It, of course, doesn't cover everything, but it's a good starting point. Thanks to Imtiaz…
Twitter • 4 days ago
Hear, hear, I'm excited to introduce a new performance metric: Maximum Achievable Matmul FLOPS (MAMF): github.com/stas00/ml-engi… Please read the notes at the url above to see what's what and I have the first measurements included (snapshot). As I get access to more accelerators…
Twitter • 4 days ago
Toolmaker. Software creator, optimizer and harmonizer. Makes things work and fly at Contextual.AI Training LLM/RAG/Generative AI/Machine�...
Lysandre Debut, Stas Bekman, Pierric Cistac, Thibault Goehringer, Victor Mustar, Fran�ois Lagunas, Alexander Rush, and Thomas Wolf. 2021. Datasets: A community�...
My name is Stas Bekman and I'm a software engineer who enjoys tinkering, building reliable systems and who excells at identifying and solving problems, and�...
$29.22
Follow Stas Bekman and explore their bibliography from Amazon.com's Stas Bekman Author Page.
This section includes downloadable material from the presentations and tutorials I've given over the years. - Photography: Enjoy Stas Bekman's photographic�...
Aug 31, 2023 � I'm super excited to start working at Contextual AI where I will be training LLMs w/ Retrieval to help businesses deploy AI that overcomes�...