“If this doesn’t make you bullish on Open Source – I don’t know what will! 🔥 That’s a 32B LLM that can easily fit on a ~0.8 USD/ hour GPU – spitting ungodly num of tokens Back of the napkin math: – fp16/ bf16 – 32GB VRAM (would fit on a L40S) – 8-bit – 16GB VRAM (L4) – 4-bit –
Hermes 3 – NOUS RESEARCH
“Can a tiny startup’s 70 billion parameter model beat OpenAI’s o1 model? Nous Research just launched the Forge Reasoning Engine, and it even managed to beat o1 on the American Invitational Math Exam. Forge uses a combination of: A) Monte Carlo Tree Search B) Chain of Code C)
Releasing the largest multilingual open pretraining dataset
“”Open-source developer platform to power your entire infra and turn scripts into webhooks, workflows and UIs. Fastest workflow engine (13x vs Airflow). Open-source alternative to Retool and Temporal.”
Releasing Common Corpus: the largest public domain dataset for training LLMs
“Check out how much CO₂ we spent doing model evaluations in the Open LLM Leaderboard 🔍 -> It’s an easy way to see which models have the best CO₂ inference cost to performance ratio! (not all big models are worth it, but @Alibaba_Qwen models are imo in a sweet spot 👏)” / X
“@AndrewYNg @guardrails_ai @ShreyaR That’s awesome! Congrats, @ShreyaR! Guardrails is one of my favorite open-source AI projects.” / X
“Good News! We convinced @mishig25 to decouple the hf(.co)/playground into a standalone open-source project! 🚀 Lets make this awesome together! 🤗
PleIAs/OCRonos-Vintage · Hugging Face
Hugging Face
“🚀 You might know @huggingface as a leader in the open-source ML/AI space, but did you know they’ve stepped into open-source robotics with @LeRobotHF? 🤖 Over the past couple of weeks, I’ve been experimenting with the 3D-printable robotic arm SO-ARM100 from TheRobotStudio and” / X
“Not many people know you can use `lms get` to download any local LLM from @huggingface. It’s secretly a full keyword search interface. Give it a try!
Meta/Llama
Exclusive: Chinese researchers develop AI model for military use on back of Meta’s Llama | Reuters
Qwen
“Early testing – this is a genuinely good model. Everything I’ve tested so far it’s indistinguishable from my o1-preview results. Code review, bug finding, writing new things, spatial reasoning – results coming soon
Qwen2.5-Coder Series: Powerful, Diverse, Practical. | Qwen
Qwen2.5 Coder Artifacts – a Hugging Face Space by Qwen
“Qwen2.5-Coder is a big deal! It’s been quiet in the open LLM space these days but this new release resurfaces an important question. Can open models close the gap with closed-source competitors? If the Qwen results are correct, then these are exciting times for open-source
“🚀 Qwen2.5 Coder skyrockets to #2 on the Hub within 24 hours of launch! 🔥 The AI coding race is heating up! 🧑💻 #AI #coding #MLops #Qwen
“🚀Now it is the time, Nov. 11 10:24! The perfect time for our best coder model ever! Qwen2.5-Coder-32B-Instruct! Wait wait… it’s more than a big coder! It is a family of coder models! Besides the 32B coder, we have coders of 0.5B / 1.5B / 3B / 7B / 14B! As usual, we not only
“”Qwen2.5-Coder is the code version of Qwen2.5, the large language model series developed by Qwen team, Alibaba Cloud.”
Snowflake
“We’re introducing Snowflake Intelligence to make it easier for everyone to drive value with data. Imagine asking a data agent: “give me a summary of this Google Doc” or “tell me how many deals we had in North America last quarter,” and instantly following up with the next steps” / X
Leave a Reply