Arcee releases Arcee-SuperNova, better fine-tune of Llama-3.1-70B! versions: 70B and 8B Trained by distilling logits from Llama-3.1-405B Used a clever compression method to reduce dataset weight from 2.9 Petabytes down to 50GB (may share it in a paper) Not all
OPEN SOURCE
-
HuggingFace Model Updates: Fixed Version Released
By
–
Those are two HF models: the first uploaded version, and then a "fixed" version (confirmed by someone from the team).
-
Reflection 70B: New Open Source AI at Center of Controversies
By
–
[#Article] Open source AI: Just unveiled, Reflection 70B at the center of controversies https://actuia.com/actualite/ia-open-source-tout-juste-devoile-reflection-70b-au-coeur-des-controverses/
… #AI #ArtificialIntelligence -
Free ISP Network Bandwidth Issues Affecting Hugging Face Access
By
–
Colleagues making fun of me because from the Free ISP network I only get 30 MB/s downloads from @huggingface while they get 300+ MB/s on the same URLs Anything we can do to improve peering bandwidth, @xavier75
? You're holding back the AI community's progress speed! -
Upstage Releases SolarPro Preview LLM for Single GPU
By
–
Super happy to release @Upstageai
’s flagship model, #SolarProPreview – the most intelligent LLM on a single GPU: strong on key benchmarks such as MMLU-Pro and IFEval 1. Download weights directly from @huggingface https://
huggingface.co/upstage/solar-
pro-preview-instruct
… 2. FREE 'solar-pro' API for three months -
Code Data Improves LLM Generalization Beyond Coding Tasks
By
–
New from @CohereForAI and @cohere
! To Code, or Not To Code? This paper reveals that including code data in LLM pre-training is critical for generalization far beyond coding tasks. The authors @viraataryabumi @ahmetustun89 and @sarahookr are here to answer your questions! -
Own Your Models: Control, Privacy, and Cost in GenAI
By
–
Stop renting and start #owning your models As if #privacy and #cost aren't motivating enough, what about ownership? Who wants to get caught off guard when commercial #LLMs get deprecated Looking for a #GenAI stack that gives you the control you need without sacrificing
-
Open-Source Email Statistics Scripts Released
By
–
here, i open-sourced this: https://
github.com/yoheinakajima/
quick_email_scripts
… • email_stats_by_domain searches by domain
• email_stats_by_people searches by email both give you stats about emails sent/received/dates -
SambaNova Launches Fastest API for Llama 3.1 Models
By
–
We are stoked to launch the fastest API for developers — independently verified by @ArtificialAnlys Starting today, developers can access SambaNova Cloud to create their own generative AI applications using the open-source Llama 3.1 405B at 100 t/s and Llama 3.1 70B models at
-
vLLM Integration Brings Local Model Testing to Hugging Face
By
–
Excited to see @vllm_project as an option for local apps in the @huggingface hub! It comes with easy snippets to quickly test out the model.