AI/ML

And then there were three...

In this article, we will look at the limitations of RAG and domain-specific Fine-tuning to adapt LLMs to existing knowledge and how a team of UC Berkeley..

Reverse Engineering: Decompiling Binary Code with Large Language Models - albertan017/LLM4Decompile

LLMs are trained to block harmful responses. Old-school images can override those rules.

‎Cubox is the ultimate tool for reading, learning, and research. With its powerful AI designed specifically for reading, Cubox can help you easily manage collections of web articles, make annotations and take notes while reading, review your knowledge, and share or organize your reading materials eff…

Dola: a powerful AI calendar assistant managing your personal calendar and group calendar by just chatting. Dola turns voice, texts, pictures and complicated context into schedules. You can also sync Dola with your Google calendar and Apple calendar.

Apple researchers achieve state-of-the-art results in multimodal AI with MM1 models, combining text and images for breakthroughs in image captioning, visual question answering, and few-shot learning, as the company invests heavily in AI to enhance Siri, Messages, and future products.

The fracking to AI pipeline is a crime. And the crime is homicide.

Tools for merging pretrained large language models. - arcee-ai/mergekit

Tools for merging pretrained large language models. - arcee-ai/mergekit

Four years ago, I did an analysis of the open source ML ecosystem. Since then, the landscape has changed, so I revisited the topic. This time, I focused exclusively on the stack around foundation models.

Automate browser-based workflows with LLMs and Computer Vision - Skyvern-AI/skyvern

This video is about mergekit, how to choose and blend models. It's non technical but links to technical papers are included. You need to know how to navigate...

Marking a major investment in Meta’s AI future, we are announcing two 24k GPU clusters. We are sharing details on the hardware, network, storage, design, performance, and software that help us extr…

How news organizations are using AI in good and bad ways.

We’re releasing an open source system, based on FSDP and QLoRA, that can train a 70b model on two 24GB GPUs.

Sometimes we all need a robot cheerleader.

Fine-tune LLM agents with online reinforcement learning - KhoomeiK/LlamaGym

AI note-taking app that runs models locally. Contribute to reorproject/reor development by creating an account on GitHub.

Has anybody looked into ways of using Zotero libraries in 'retrieval-augmented generation' pipelines?

The official implementation of RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval - parthsarthi03/raptor