Alex Reibman ๐Ÿ–‡๏ธ
Alex Reibman ๐Ÿ–‡๏ธ

@AlexReibman

8 Tweets 5 reads Apr 16, 2024
Never underestimate the open source AI community.
These cracked engineers are here to break the limits of whatโ€™s possible with local LLMs. We just witnessed some nutty inventions.
Hereโ€™s what we saw at the @ollama Open Source and Local AI meetup at @cerebral_valley (๐Ÿงต):
1/ Ollama: Concurrency mode
Run multiple local LLMs on one machine at the same time with shared resources
Even works with different models.
@ollama @dhiltgen
2/ Bootleg Amazon Kindle
Running local LLMs and StableDiffusion on Raspberry Pi to create generative ebooks
Swap out parts for faster performance
3/ Enchanted
MacOS native app for running local LLMs
Comes with built-in computer vision (LLaVA1.5 + BakLLaVA), chat completion, and the ability to interact with apps and windows
@amgauge
4/ Continue
Open source coding copilot powered by local LLMs. Available in VSCode
Also now includes continuous fine-tuning models on your codebase
@continuedev
@tylerjdunn
@NateSesti
5/ Uncensoring censored models
Methods and approaches for removing guardrails from censored LLMs
1. Remove prompts rewarding censorship from instruct fine-tuning
2. Fine-tuning on unaligned data
3. System prompts that instruct LLMs to behave improperly
@erhartford
6/ Whereโ€™s Waldo w/ Weaviate
Translating images into vectors and using local models to find items in the photo (i.e. waldo)
@weaviate_io @weaviate_jon
That's all for this time. Huge thanks to @ollama @cerebral_valley @SHACK15sf @mchiang0610 for hosting. More demos coming soon ๐Ÿ‘€
Want to learn more about agents, LLMs, and generative AI? Follow @AgentOpsAI @AlexReibman

Loading suggestions...