Lambda reposted this
The folks at Lambda fine-tuned Open-Sora for stop-motion Lego-style animations using 1,000 GPU hours on a 32-GPU Lambda Cluster. Explore the technical steps, hardware setup, training metrics, and future improvements with Chuan Li, Corey Lowman, David Hartmann, and Jeremy Hummel. Read more: https://lnkd.in/g4j5RzwC
Useful tips
A.G.I. is here! (A) lot of (G)PUs (I)nterconnected https://lnkd.in/eqdZGbmA
Our team will be walking the floor at #SIGGRAPH2024. DM Noah Spear or Joseph Boufford if you want to organize a meet-up! Coffee is on us.
We must say, Fierce Network did a fine job capturing the essence of our press release 👀: https://lnkd.in/dKt2tJf5
One Llama, two Llamas, three Llamas… 405B Llamas?! Meta’s Llama 3.1 brings top-tier performance to the open-source community—a game-changer for innovation. With Llama 3.1 405B, a multi-node cluster is a must for full-precision inferencing. Read vLLM's complete deployment guide using Lambda’s 1-Click Clusters: https://lnkd.in/g3qrQh4i Get your on-demand cluster today: https://lnkd.in/eu9xieKe
Cerebral Valley New York hosted 200 leaders in the AI community last month! Here’s Stephen Balaban’s full panel discussion on AI in the enterprise with Volley co-founder James Wilsterman, Vectara CEO Amr Awadallah, and Modal CEO Erik Bernhardsson: https://lnkd.in/e2TKSZGX
New NVIDIA GPU availability high score! Spin up an 8x NVIDIA H100 Tensor Core GPU instance in Lambda Cloud today: https://lnkd.in/ehVqfuBy
Join Lambda and VAST Data next week for a webinar on supercharging your AI training with the right infrastructure: https://hubs.ly/Q02DMZbw0
Extra lives for AI projects! Throughout the summer, Lambda is making on-demand 8x NVIDIA H100 Tensor Core GPU instances more available than ever. The first deployment is live: game on! https://hubs.ly/Q02DZ9cH0
Senior Product Manager • 2x 8-figure products from scratch • B2B | SaaS | Commerce | AI | Web 3 | Data Platforms
1wFascinating to see Lambda's innovation with Open-Sora for stop-motion animations—impressive use of GPU technology for creative applications.