Distilling Llama3.1 8B into 1B in torchtune https://
bit.ly/3Z67K7l #AI #MachineLearning #DeepLearning #LLMs #DataScience
OPEN SOURCE
-
Distilling Llama 3.1 8B into 1B using torchtune
By
–
-
Release of Qwen-QwQ-32B Preview Model on Hugging Face
By
–
https://
huggingface.co/chat/models/Qw
en/QwQ-32B-Preview
… -
Stability AI Releases Three ControlNets for Stable Diffusion 3.5
By
–
ICYMI: We released three ControlNets for Stable Diffusion 3.5 Large this week: Blur, Canny, and Depth Model weights https://
huggingface.co/stabilityai/st
able-diffusion-3.5-controlnets
… Code https://
github.com/Stability-AI/s
d3.5/
… Stability AI Community License https://
stability.ai/community-lice
nse-agreement
… Comfy support -
Running AI NPCs Locally: Feasibility and Implementation
By
–
Yes, you can run local LLMs. But can you run AI NPCs locally?
-
Comprehensive llama.cpp Guide: Building and Performance Benchmarking
By
–
Cool guide to llama.cpp from building to perf benchmarking llama.cpp's documentation is notoriously difficult to work with. There's some more advanced stuff I didn't know about, it's great stuff from @steel_ph0enix https://
steelph0enix.github.io/posts/llama-cp
p-guide/
… -
ML Publishing Standards: Open Access and Protection Against Predatory Journals
By
–
Fortunately, in ML, this video is largely irrelevant. All reputable venues (NeurICMLR, T/JMLR, etc) don't charge publication costs, are open access, and allow arXiv posting.
— Gautam Kamath (@thegautamkamath) 28 novembre 2024
Unfortunately, some bad actors also try to enter the space and break this nice equilibrium, eg Nature MI https://t.co/wvYSAYmHBjFortunately, in ML, this video is largely irrelevant. All reputable venues (NeurICMLR, T/JMLR, etc) don't charge publication costs, are open access, and allow arXiv posting. Unfortunately, some bad actors also try to enter the space and break this nice equilibrium, eg Nature MI
-
Orange, OpenAI and Meta Integrate African Languages into AI
By
–
[#Article] Orange, OpenAI and Meta partner to integrate African regional languages into open source AI models https://actuia.com/actualite/orange-openai-et-meta-sassocient-pour-integrer-les-langues-regionales-africaines-dans-les-modeles-dia-open-source/
…
#AI #artificialintelligence -
Optimizing AWQ Deployment for Flexible Model Distribution
By
–
Think a more likely one would be to help people create the best, most optimised AWQs and then they can deploy wherever they want ofc you can have direct deployment options. But I’m 100% there’s value in this.
-
AGI Accessible: Install QwQ with Two Lines of Code
By
–
You too can have AGI in just a couple lines of code! `pip install transformers` & QwQ is all you need
-
Contribution guidance: Referencing Alpaca and LIMA code implementations
By
–
Thanks for your interest in contributing. I unfortunately currently don’t have access to a computer until likely next week (long story) but off the top of my head and as far as I can tell right now, I think you can copy most code from the Alpaca or LIMA (
https://
github.com/Lightning-AI/l
itgpt/blob/main/litgpt/data/lima.py
…)