It may have some use in case they will make browsing available. Currently, it is purely targeting privacy concerned userbase I am also still not confident in the performance of OS models in terms of quality. Their responses are often way too off.
OPEN SOURCE
-
Codestral Excels as Web Agent Action Model
By
–
Codestral seems to to be really good as Action Model for Web Agents! Beating all other open models and even closed ones
-
MLCEngine unifies cloud and local LLM deployment across devices
By
–
MLCEngine The cool MLC project now supports both cloud and local deployment in a unified engine. It can deploy LLMs on servers with multi-GPU support or local devices like laptops, browsers (WebGPU), and phones Quick start: https://
llm.mlc.ai/docs/get_start
ed/quick_start.html
… -
Hugging Face Pollen Robotics Open Source Household Chores Robot
By
–
https://
venturebeat.com/ai/hugging-fac
e-and-pollen-robotics-show-off-first-project-an-open-source-robot-that-does-chores/
… "…train the robot to do a variety of household chores and safely interact with humans and dogs" understudied problem here imo (as a pet owner) -
P2P Technology as Decentralized Video Hosting Infrastructure
By
–
Peer To Peer technology has come a long way, could be a frontend for some of the existing P2P video websites — for which they could provide additional servers and help people host their own videos too.
-
Awesome LLM Apps with RAG: GitHub Repository Showcase
By
–
Find all the awesome LLM Apps demo with RAG in the following Github Repo. P.S: Don't forget to star the repo to show your support
-
Open Source Robotics and AI Tools: Reachy2, LeRobot, Dora-rs
By
–
Starting @matth_lapeyre
’s dog, @pollenrobotics upcoming Reachy2 robot, @huggingface lerobot library from @RemiCadene
, and @HaixuanT Dora-rs library -
Open-source autonomous robot policy without teleoperation
By
–
– full end-to-end policy
— Thomas Wolf (@Thom_Wolf) 7 juin 2024
– no teleoperation
– quasi-full open-source software and hardware
(even Reachy-1’s hardware is open-source CC-BY-SA CAO assembly)pic.twitter.com/u0yEE9y1Pb– full end-to-end policy
– no teleoperation
– quasi-full open-source software and hardware (even Reachy-1’s hardware is open-source CC-BY-SA CAO assembly) -
llm.c optimization: balancing historical accuracy and training speed
By
–
This is cool!! I'm not exactly sure how to upstream these changes to llm.c… Part of me wants to reproduce GPT-2/3 using their exact hyperparameters just for historical aesthetics, but part of me also wants to just train things as fast as possible. Probably both. – lr 3X is
