“Thrilled to see Open Canvas hit 1k+ stars last night! If you want to help contribute to this project, we have a wishlist of features covering everything from frontend features to adding/improving the agent! Checkout the repo & feature requests here:
“📃Open Canvas Open Canvas is an open source web application for collaborating with agents to better write documents. It is inspired by OpenAI’s “Canvas”, but with a few key differences: 📂Open Source 🧠Built in memory ✒️Start from existing documents 📂Open Source: All the
“5 trillion tokens permissively licensed Zyda-2 dataset just dropped from @ZyphraAI 🗃️ Dataset Overview: – Composed of DCLM, FineWeb-Edu, Zyda-1, and Dolma v1.7’s Common Crawl – Processed using NVIDIA NeMo Curator – Powers Zamba2 series models – Available on HuggingFace 🔧
“LLMs have made exciting progress on hard tasks! But they still struggle to analyze complex, unstructured documents (including today’s Gemini 1.5 Pro 002). We (UC Berkeley) built 📜DocETL, an open-source, low-code system for LLM-powered data processing:
T2V Turbo V2 – a Hugging Face Space by TIGER-Lab
deepseek-ai/Janus-1.3B · Hugging Face
“My next Machine Learning Engineering cohort starts in 2 weeks with a big change: We’ll build a massive, end-to-end machine learning system using exclusively open-source tools. We’ll start from zero and build our way to a fully-fledged system with the following components: •
“Important report “Stopping Big Tech from becoming Big AI” from @maxvonthun @danielahanley & the @mozilla foundation! “Open source AI has an important role to play in countering a lack of interoperability and access, and fostering innovation, by lowering barriers to entry,
“Gap between open-source and closed source in LLMs is insignificant IMO now” / X
Pyramid Flow open source AI video generator launches | VentureBeat
Hugging Face
Llama-3.1-Nemotron-70B – a nvidia Collection
“You can now run *any* GGUF on the Hugging Face Hub directly with @ollama 🔥 This has been a constant ask from the community, starting today you can point to any of the 45,000 GGUF repos on the Hub* *Without any changes whatsoever! ⚡ All you need to do is: ollama run
“When @ClementDelangue sits down with @gilbert and @djrosent for an hour-long podcast episode, it results in this fascinating conversation not only about Hugging Face, but also about building a startup in the age of AI and the future of AI development. Sharing this video snippet
Ichigo Llama3.1 S Instruct – a Hugging Face Space by jan-hq
CoTracker – a Hugging Face Space by facebook
“The gradient accumulation fix is now in the main branch of transformers! Thank you to the entire @huggingface team, especially @TheZachMueller and @art_zucker for collabing with us to fix it! 🤗🦥” / X
Meta/Llama
Llama-3.1-Nemotron-70B – a nvidia Collection
“🤯I think i’ve cracked lead generation!🚀 Built an AI that generates leads for my business and it creates a google sheet with the lead data. thanks to @composiohq @llama_index @firecrawl_dev 100 likes and ill share the code 👉 For more examples:
Meta’s open AI hardware vision – Engineering at Meta
OCP Summit 2024: The open future of networking hardware for AI – Engineering at Meta
“📢 Meta Spirit: The first open source multimodal language model from @AIatMeta 👏 Meta Spirit LM offers: – Seamless integration of speech and text in a multimodal language model – Word-level interleaving of speech and text datasets – Cross-modality generation capabilities – Two” / X
Mistral
Mistral AI’s new language models bring AI power to your phone and laptop | VentureBeat
“🎉 We @MistralAI just dropped two new edge models – Ministral 3B and Ministral 8B!
Un Ministral, des Ministraux | Mistral AI | Frontier AI in your hands
NVIDIA
“We’re soo back!: Nvidia Nemotron 70B – beats Llama 3.1 405B, GPT4o & Claude 3.5 Sonnet! 🔥 Evals (Nemotron 70B vs Claude 3.5 vs GPT4o) > Arena Hard – 85.0 vs 79.2 vs 79.3 > AlpacaEval 2 LC – 57.6 vs 52.4 vs 57.5 > MT Bench – 8.98 vs 8.81 vs 8.74 Secret Sauce? RLHF (REINFORCE)
Llama-3.1-Nemotron-70B – a nvidia Collection
Nvidia just dropped a new AI model that crushes OpenAI’s GPT-4—no big launch, just big results | VentureBeat
nvidia/Llama-3.1-Nemotron-70B-Instruct · Hugging Face
[AINews] Nvidia Minitron: LLM Pruning and Distillation updated for Llama 3.1 • Buttondown
“The Nvidia Nemotron Fine-Tune Isn’t A Very Good 70b Model! While it improves on the base 70b Llama model on reasoning, it underperforms across several categories It’s worse than 405b and isn’t as good as the other SOTA models Detailed numbers coming soon on Livebench AI” / X
“🕵️♂️ Let’s investigate NVIDIA’s Llama-3.1-Nemotron-70B-Instruct-HF model performance (1/7) A long detailed thread! 🧶” / X
Leave a Reply