AI - AI Dynamics

Thanks to compute providers and partners for model release support

By

–

27 September 2023 17h25

Huge thanks to our compute providers @CoreWeave and the EuroHPC, @tri_dao and @d_haziza for their help with FlashAttention and xFormers, and @huggingface
, vLLM, @skypilot_org
, FastChat for their help and support with this release.

GQA and Sliding Window Optimization for Efficient Decoding

By

–

27 September 2023 17h25

We trained it with GQA and a sliding window of 4096 tokens, resulting in constant cache size and a linear decoding speed. Our changes to FlashAttention v2 and xFormers to support sliding window are available to the community.

Mistral AI team builds advanced MLOps stack in three months

By

–

27 September 2023 17h25

Very proud of the Mistral AI team who rebuilt a top-performance MLops stack, and designed a very sophisticated data processing pipeline, from scratch, in less than 3 months.

Mistral 7B Released: Outperforms Llama 2 13B on All Benchmarks

By

–

27 September 2023 17h25

Mistral 7B is out. It outperforms Llama 2 13B on every benchmark we tried. It is also superior to LLaMA 1 34B in code, math, and reasoning, and is released under the Apache 2.0 licence. https://
mistral.ai/news/announcin
g-mistral-7b/
…

→ View original post on X — @soumithchintala

Serverless GPU Inference Now Available for AI

By

@soumithchintala

–

27 September 2023 17h22

serverless gpu inference is here!
things getting quite interesting now…

→ View original post on X — @groqinc

Groq Solutions Designed and Manufactured in North America

By

@groqinc

–

27 September 2023 17h13

Our solutions are designed, engineered, and manufactured in North America. Visit
http://
groq.com to explore our technical documents, videos, specs, and customer testimonials.