RT Adrien CarreiraHardest IR of my career: one narrow objective, endless parallel paths, machine speed. One takeaway, we fought back with open models, in the open.…
RT PoolsideToday we're releasing Laguna S 2.1, our most capable model to date.It's a 118B total parameter Mixture-of-Experts model with 8B activated per token, a context…
RT clem 🤗Plot twist: SF march permits take forever, and open-source AI can't wait. So we're going scrappy on Saturday: small march, sidewalk only, maximum fun.What…
RT PiHi kids, gramps here, taking over for the intern. Just released pi 0.81.0 which features first class integration with @ggerganov wonderful llama.cpp server. Professional video…
RT LeRobotLeRobot can now see in 3D 👀🤖Depth cameras feed straight into training data end-to-end in v0.6.0: 12-bit precision preserved through encoding, or fully lossless raw…
RT LysandreWe're having our first @huggingface event in Japan in Tokyo this Thursday, in Otemachi!We'll go over the latest tooling we put out and how you…
RT eliehuge open weight release, motif (korean company) just released a 13B active 314B total MoE performing on par with bigger models like minimax M3 and…
RT Bryan CatanzaroOpen models are more secure.Nicolas Pinto: The asymmetry problemWhen we started the log analysis, we first used "frontier" models behind commercial APIs. This did…
RT clem 🤗"Open weights are inherently secure" @sriramk
RT clem 🤗Banning open-source AI would hurt defenders 10x more than attackers, which would make the world 10x more dangerous and this is a good example…