Build a multi-agent LLM app with GPT-4o in just 15 lines of Python Code (step-by-step instructions):
LLMS
-
SolarLLM Hackathon in SF: New LLM Version and Layout Analysis
By
–
Join us in SF on 6/1 for our first hackathon with #solarllm's friends @MongoDB , @langchain , @togethercompute
, and @MindsDB . Meet our new #solaremb, Layout Analysis, and #solarllm ver 1.1.2. See you all there! -
Training GPT-3 equivalent models with FineWeb dataset for $500
By
–
GPT-3 model is GPT-2 but trained for longer (300B) tokens and yes on a better dataset. FineWeb is a good dataset, so you can train your own like this. It will cost ~$500. Use -b 32 -t 2048 instead to use the 2048 GPT-3 context length to be accurate.
-
GPT-3 4th Anniversary: Re-training Smallest Model Achievement
By
–
Apparently today is the 4th year anniversary of GPT-3! https://
arxiv.org/abs/2005.14165 Which I am accidentally celebrating by re-training the smallest model in the miniseries right now :). HellaSwag 33.7 (Appendix H) almost reached this a few steps ago (though this is only 45% of the -
Reverse Turing Test: Which AI Among Famous Figures?
By
–
They conducted a reverse Turing Test with 4 AIs. There were 5 famous figures on a train, and they had to figure out which among them was the human. The video has 140k views, was made in Unity and voiced through ElevenLabs, and pitted GPT-4-Turbo, Claude 3 Opus, Llama 3, and
-
FIM Performance with Instruct Models Unexplored
By
–
I've never tried FIM with an instruct model to be honest, I didn't expect it to work well. Maybe they all do and I didn't know?
-
Smaller Models Excel on Specialized Tasks with Quality Data
By
–
70B is general, this one is specifically trained for code tasks. With a smaller model, you can do better on a smaller task scope if you have a ton of great data.
-
Sharing perspectives on open-source AI and LLaMA 3
By
–
Enjoyed sharing my thoughts on open-source AI and LLaMA 3 for today’s @nytimes piece. Thanks @MikeIsaac for including my perspective!
-
Mistral AI dominates HF trending with 7B Instruct and Codestral
By
–
.
@MistralAI holds 3 spots in the top 10 trending models on HF including #1 with 7B Instruct. Congrats for the launch of Codestral, I'm sure it'll get #1 trending very soon! https://
huggingface.co/models -
Aligning LLMs for Enterprise Compliance and Custom Domain Solutions
By
–
Time to align LLMs to provide organizationally compliant and correct responses? Learn why even with RAG, LLMs need YOUR DATA and YOUR DOMAIN EXPERTISE to serve custom enterprise use cases. Read more: https://
snorkel.ai/introducing-en
terprise-alignment-path-to-superalignment/
… #EnterpriseA