AI Pulse by Inblix

AI Research

Cutting-edge AI research papers, breakthroughs, and academic developments.

601 articles

PyCharm now lets you drop an AI chatbot into your app in 10 minutes — no browser required — Inblix summary
Research

PyCharm now lets you drop an AI chatbot into your app in 10 minutes — no browser required

Hugging Face Blog · Nov 5, 2024 · 2 min read

JetBrains and Hugging Face just shipped an integration that feels almost mundane, and that's exactly why it matters. Ri...

Argilla 2.4 Lets You Build AI Datasets Directly From Hugging Face Hub, No Code Needed — Inblix summary
Research

Argilla 2.4 Lets You Build AI Datasets Directly From Hugging Face Hub, No Code Needed

Hugging Face Blog · Nov 4, 2024 · 2 min read

Argilla 2.4 just dropped, and it rips out the last big barrier between a domain expert and a high-quality fine-tuning d...

Intel and Hugging Face crack 1.5x-2x faster LLM decoding across any model family — Inblix summary
Research

Intel and Hugging Face crack 1.5x-2x faster LLM decoding across any model family

Hugging Face Blog · Oct 29, 2024 · 2 min read

The dirty secret of assisted generation—that speed-boosting technique also called speculative decoding—has always been...

Digital Green's chatbot serves 340K queries, but can LLM judges keep its farming advice grounded? — Inblix summary
Research

Digital Green's chatbot serves 340K queries, but can LLM judges keep its farming advice grounded?

Hugging Face Blog · Oct 28, 2024 · 2 min read

Digital Green's Farmer.chat has quietly scaled to over 20,000 smallholder farmers, fielding 340,000 queries with a RAG...

Cohere's Aya Expanse 8B beats models twice its size by pitting teacher AIs against each other — Inblix summary
Research

Cohere's Aya Expanse 8B beats models twice its size by pitting teacher AIs against each other

Hugging Face Blog · Oct 24, 2024 · 2 min read

Cohere For AI just dropped the Aya Expanse family, and the 32B model is punching so far above its weight class that it'...

CinePile 2.0 uses adversarial AI to fix broken video questions instead of trashing them — Inblix summary
Research

CinePile 2.0 uses adversarial AI to fix broken video questions instead of trashing them

Hugging Face Blog · Oct 23, 2024 · 2 min read

The original CinePile dataset launched in May 2024 with a genuinely impressive stat: humans outperformed the best comme...

Google DeepMind's SynthID Text Watermark Proves Imperceptible—But Not Impenetrable — Inblix summary
Research

Google DeepMind's SynthID Text Watermark Proves Imperceptible—But Not Impenetrable

Hugging Face Blog · Oct 23, 2024 · 3 min read

Google DeepMind has formally pulled back the curtain on SynthID Text, a production-ready watermarking technique for LLM...

Hugging Face kills HUGS microservice just 2 years after launch — Inblix summary
Research

Hugging Face kills HUGS microservice just 2 years after launch

Hugging Face Blog · Oct 23, 2024 · 2 min read

Hugging Face has quietly pulled the plug on HUGS, its Generative AI Services platform for deploying open models, markin...

Hugging Face taps Protect AI to scan its 1M+ model repos for hidden exploits — Inblix summary
Research

Hugging Face taps Protect AI to scan its 1M+ model repos for hidden exploits

Hugging Face Blog · Oct 22, 2024 · 2 min read

Hugging Face isn't waiting for a disaster to strike. The model-sharing giant has partnered with Protect AI, integrating...

Hugging Face’s new S2S pipeline turns speech into speech, but you’ll need rented GPUs to run it — Inblix summary
Research

Hugging Face’s new S2S pipeline turns speech into speech, but you’ll need rented GPUs to run it

Hugging Face Blog · Oct 22, 2024 · 2 min read

Hugging Face just dropped an open-source Speech-to-Speech pipeline that feels a bit like magic: you talk, the machine t...

Machine Learning in the Browser Just Got a 100x Speed Boost with WebGPU — Inblix summary
Research

Machine Learning in the Browser Just Got a 100x Speed Boost with WebGPU

Hugging Face Blog · Oct 22, 2024 · 2 min read

Running AI models directly in a web browser has always been a neat trick — but agonizingly slow. Transformers.js v3, th...

Outlines-core hits 0.1.0 with 2x speed boost as Rust replaces Numba — Inblix summary
Research

Outlines-core hits 0.1.0 with 2x speed boost as Rust replaces Numba

Hugging Face Blog · Oct 22, 2024 · 2 min read

The team behind Outlines just shipped a foundational rewrite that should make anyone building on top of LLMs sit up and...

Stable Diffusion 3.5 hits Diffusers with an 8B model and a 4-step turbo variant — Inblix summary
Research

Stable Diffusion 3.5 hits Diffusers with an 8B model and a 4-step turbo variant

Hugging Face Blog · Oct 22, 2024 · 2 min read

Stability AI's largest open image model just landed in Hugging Face Diffusers, and it's a chunky one. The release inclu...

Llama 3.2 Has Been Quietly Working in Keras Since Day 1 — Inblix summary
Research

Llama 3.2 Has Been Quietly Working in Keras Since Day 1

Hugging Face Blog · Oct 21, 2024 · 2 min read

If you've been waiting to use Meta's new Llama 3.2 models in Keras, you can stop. Martin Görner from Google confirmed t...

Hugging Face ships 24-hour fix for a silent training bug that broke gradient accumulation — Inblix summary
Research

Hugging Face ships 24-hour fix for a silent training bug that broke gradient accumulation

Hugging Face Blog · Oct 16, 2024 · 2 min read

For anyone training language models with Hugging Face’s Trainer, here’s a bug that may have been quietly corrupting you...

A security audit found a remote code execution bug in Gradio, the tool behind 470,000 AI apps — Inblix summary
Research

A security audit found a remote code execution bug in Gradio, the tool behind 470,000 AI apps

Hugging Face Blog · Oct 10, 2024 · 2 min read

The team behind Gradio, the Python library so ubiquitous it clocks over 6 million monthly PyPI installs, just did somet...

AMD's 192-core Turin chip doubles Llama 3 speed, no GPU required — Inblix summary
Research

AMD's 192-core Turin chip doubles Llama 3 speed, no GPU required

Hugging Face Blog · Oct 10, 2024 · 2 min read

The era of running serious LLM inference without a GPU just got a lot more interesting. AMD's 5th Gen EPYC "Turin" proc...

Gradio 5 ships with streaming, AI playground, and a design overhaul — Inblix summary
Research

Gradio 5 ships with streaming, AI playground, and a design overhaul

Hugging Face Blog · Oct 9, 2024 · 2 min read

The launch of Gradio 5 is a direct shot at the friction that's been quietly frustrating machine learning developers for...

Hugging Face and Dask team up to score 211 million web pages for educational value — Inblix summary
Research

Hugging Face and Dask team up to score 211 million web pages for educational value

Hugging Face Blog · Oct 9, 2024 · 2 min read

Processing a few hundred rows of data on your laptop is a fine way to start a project. But what happens when you need t...

Hugging Face squeezes 52% faster LLM inference by guessing when a draft model is wrong — Inblix summary
Research

Hugging Face squeezes 52% faster LLM inference by guessing when a draft model is wrong

Hugging Face Blog · Oct 8, 2024 · 2 min read

The whole premise of speculative decoding is clever but wasteful. You let a tiny, fast draft model spit out a bunch of...

A single row change in a 2GB Parquet file just cost 230MB—here’s the fix Hugging Face found — Inblix summary
Research

A single row change in a 2GB Parquet file just cost 230MB—here’s the fix Hugging Face found

Hugging Face Blog · Oct 5, 2024 · 2 min read

Most of us assume that tweaking a single row in a massive dataset should be a lightweight operation. Hugging Face engin...

Open FinLLM Leaderboard drops: 7-task gauntlet tests if AI can actually forecast stocks — Inblix summary
Research

Open FinLLM Leaderboard drops: 7-task gauntlet tests if AI can actually forecast stocks

Hugging Face Blog · Oct 4, 2024 · 2 min read

A new independent benchmark is throwing down the gauntlet for financial AI, and it’s not asking for polite summaries. T...

China's AI firms flood overseas markets as 78,612 rivals collapse at home — Inblix summary
Research

China's AI firms flood overseas markets as 78,612 rivals collapse at home

Hugging Face Blog · Oct 3, 2024 · 2 min read

A wave of Chinese AI companies is crashing into international markets, and it's not just about ambition — for many, it'...

BenCzechMark exposes 25 LLMs: most still mangle Czech grammar and culture — Inblix summary
Research

BenCzechMark exposes 25 LLMs: most still mangle Czech grammar and culture

Hugging Face Blog · Oct 1, 2024 · 2 min read

The best open-source large language models still stumble badly over Czech, and now we have the receipts. A new evaluati...