This is pretty awesome - I’ve been using it for a few weeks. Never too late to appreciate a well designed tool :) #googlescholar

They should get rid of reposts too

TTS ecosystem has been booming lately: 1. Chat TTS - English + Chinese TTS model optimised for daily conversations/ dialogues + Voice Cloning 2. MARS5 TTS - English only but gives insane prosodic control paired with voice cloning 3. Parler TTS - Smol but powerful text prompt

Seattle Independence Day fireworks from above:

lol

Impressive work by XKDR Forum and coauthors! Survey.jl - An Efficient Framework for Analysing Complex Surveys xkdr.org/paper/surveyjl… GitHub: github.com/xKDR/Survey.jl CC: Sid Zãdéy

Releasing the full dataset, training code, and model weights for TF-ID models under mit License. It's time to finetune your own vision language models to work with documents! You only need *a few hundred more* domain specific bounding boxes to adapt TF-ID for your special use

Bookmarking: gwern.net/utext by 𝔊𝔴𝔢𝔯𝔫 Thanks for sharing suhit !

I sense a paradigm shift coming

We’re presenting the first AI to solve International Mathematical Olympiad problems at a silver medalist level.🥈 It combines AlphaProof, a new breakthrough model for formal reasoning, and AlphaGeometry 2, an improved version of our previous system. 🧵 dpmd.ai/imo-silver

You can tell just by looking at this progression that this was a labor of love for a lot of very smart people.

Another paper pointing out in details what we've known for a while: LLMs (used via prompting) cannot make sense of situations that substantially differ from the situations found in their training data. Which is to say, LLMs do not possess general intelligence to any meaningful

This

We trained the best Privacy AI model available on Hugging Face: Piiranha fully open source and thoroughly documented, so you can actually use it. recognizes 17 types of personal data trained on H100s from Akash Network

Just found an asycio replcament for celery (python library): taskiq-python.github.io/guide/getting-…

Classic linear programming problems are one of the most foundational problems in computer science, yet today’s approaches struggle to scale. Today on the blog read about PDLP, a new solver, based on first-order methods, for large-scale linear programming →goo.gle/3MSoHfD