Keep on to blur preview images; turn off to show them clearly
RT @tkhoury: Exciting to see further improvements to DiLoCo from the Google DeepMind team
Distributed Learning @ deepmind | DiLoCo, DiPaCo. Continual Learning PhD @ Sorbonne
RT @msharmavikram: 🚀 Want to become a CUDA ninja? Start with the new CUDA Programming Guide - Section 4 is your gold mine! It’s packed w…
gpu enjoyer at @modal. he/him. ex @full_stack_dl, @weights_biases (acq. @CoreWeave), phd Berkeley @Redwood_Neuro. try https://t.co/SYWVMCazZ3
under-rated economic benefit of self-hosting OSS models: the "lossless" performance improvements are rapidly discovered and disseminated, leaving only "lossy" tricks that have to be validated per workload not available for multi-tenant APIs -- "give me bf16 or give me death"
gpu enjoyer at @modal. he/him. ex @full_stack_dl, @weights_biases (acq. @CoreWeave), phd Berkeley @Redwood_Neuro. try https://t.co/SYWVMCazZ3
RT @NielsRogge: This is another really cool release by @Microsoft Try it here: https://t.co/WcExDggJjP
AI research paper tweets, ML @Gradio (acq. by @HuggingFace 🤗) dm for promo ,submit papers here: https://t.co/UzmYN5XOCi
Stoica at NeurIPS, on AI for Systems Research: largest advantage of LLMs is ability to transfer knowledge across domains I've always felt this was an under-rated use case: automatic remixes (not new tracks), paper discovery (not new papers), SDK translation (not generation), etc
gpu enjoyer at @modal. he/him. ex @full_stack_dl, @weights_biases (acq. @CoreWeave), phd Berkeley @Redwood_Neuro. try https://t.co/SYWVMCazZ3
RT @deviparikh: I had mentioned Humans of AI in the Business Insider article and some of you pinged me about it. It was a side pandemic pro…
Co-founder & Chief Scientist @yutori_ai. Prev: Senior Director leading FAIR Embodied AI @MetaAI and Professor @GeorgiaTech.