MACHINE LEARNING

LLaMA Innovation: Quantization and LoRA Techniques for Developers

By

–

22 June 2023 19h28

But that sometimes breeds innovation—like in the case of LLaMA. And scrappy developers without access to massive training clusters like Google or OpenAI have to find ways around the problem. That's led to the widespread adoption of techniques like quantization and LoRA.

Startups Lower ML Model Training and Serving Barriers

By

–

22 June 2023 19h28

As the cultures change here—at Meta or otherwise—the net beneficiary is likely to be the startups. Companies like MosaicML/Modal/Anyscale are trying to drop the barrier to training and serving ML models, and a lot of very cool research is being done on compact models.

AI Hype Reshapes Research Talent Competition at Big Tech

By

–

22 June 2023 19h28

The hype around AI products is quickly changing the way research divisions work within larger orgs. Companies like Meta and Google have spent the last decade wooing some of the smartest minds in machine learning out of academia with the promise of working on cutting-edge tech.

ML Challenges Drive Development of Hyped Technologies

By

–

22 June 2023 19h28

They join those organizations with the hopes of getting direct exposure to some of the most difficult problems in ML, like reinforcement learning. But most of those problems are now directly applicable to one of the most-hyped technologies in decades.

→ View original post on X — @rasbt

Blocking ML and AI researchers is closed-minded and counterproductive

By

@rasbt

–

22 June 2023 19h20

No, not that I’m aware. But yeah, arbitrarily blocking ML & AI researchers is not a good look and quite close-minded. I honestly don’t want to waste time on people who behave like that.

→ View original post on X — @mustafasuleyman

Inflection-1 LLM Outperforms GPT-3.5 and Llama on Benchmarks

By

@mustafasuleyman

–

22 June 2023 18h34

We have amazing results to announce! Inflection-1 is our new best-in-class LLM powering Pi, outperforming GPT-3.5, Llama and PALM-540B on major benchmarks commonly used for comparing LLMs.

→ View original post on X — @jeande_d

Awesome Multimodal Large Language Models: Papers and Datasets

By

@jeande_d

–

22 June 2023 17h11

Awesome Multimodal Large Language Models A curated list of the latest papers and datasets in multimodal large models. Provides an excellent listing of multimodal papers, codes, and demos. Papers about topics like instruction tuning, in-context learning, chain-of-thought(CoT),

→ View original post on X — @baiduresearch

StyleSync: High-Fidelity Lip Synchronization Framework

By

@baiduresearch

–

22 June 2023 17h10

We propose StyleSync, an effective framework that enables high-fidelity lip synchronization. We identify that a style-based generator would sufficiently enable such a charming property on both one-shot and few-shot scenarios. Read more: https://t.co/IVP6DmaFFF #CVPR2023 pic.twitter.com/6r7on4uqUp
— Baidu Research (@BaiduResearch) 22 juin 2023

We propose StyleSync, an effective framework that enables high-fidelity lip synchronization. We identify that a style-based generator would sufficiently enable such a charming property on both one-shot and few-shot scenarios. Read more: https://
arxiv.org/abs/2305.05445 #CVPR2023