195 stories
Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers. If you’d like to support this, please subscribe.
A 48-minute video walkthrough of token sampling, watermark detection, and removal
We read the AGENTS.md files of the 100 most-starred repos on GitHub that have one. Here's what they instruct their coding agents, what they ban, and a copy-paste template distilled from all of it
🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch. - Release Diffusers 0.40.0: New pipelines, tensor-parallel support, improved CLI, and more · huggingface/diffusers
OpenRecall failing with a torch and torchvision conflict or a PyObjC error? Here are the four workarounds from the open issues that actually install it.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
An End-to-End Project With Dataset Construction, Model Training, Local Deployment, and RLVR
You have an agent that can already record a demonstration and push it to the Hugging Face Hub. Now you want to run that loop continuously: collect episodes through the day, train a policy on the growing dataset, deploy it, and pull the next batch back to improve it.
NVIDIA Nemotron 3.5 Lightning is now available in Cline, bringing a 30B MoE model with 3B active parameters, up to 1M context, and high throughput for agentic coding workflows.
CodeRabbit worked with NVIDIA and Baseten to post-train NVIDIA Nemotron 3.5 Lightning for one of CodeRabbit's highest-volume routing tasks, improving route agreement while reducing estimated inference cost.
NVIDIA Nemotron 3.5 Lightning is now available on Ollama. It's a 30 billion parameter (3B active) open model built for agents that stay running, gathering context, calling tools, and working through multi-step tasks on your own hardware.
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training. - Release Release: v5.15.0 · huggingface/transformers
Kimi K3 is the first open 3T-class model. See how it benchmarks, what it costs, and how to call it on the Together AI API, with copy-paste code examples.
Figure 1: CUDA-to-MLX optimization translation map. CUDA optimization knowledge can be translated into architecture-native MLX strategies rather than copied instruction-for-instruction. We face a new epoch in computing.
Short correction note for the random seed in Listing 6.5 on page 198 of Build a Reasoning Model From Scratch.
We propose a measure of an AI agent’s optimization ability with an "expenditure horizon." We give an empirical illustration from the NanoGPT speedrun.
Go ahead and axolotl questions. Contribute to axolotl-ai-cloud/axolotl development by creating an account on GitHub.
🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training. - Release Release v5.14.0 · huggingface/transformers
Showing 126 of 195 stories
Load more storiesGet the top AI news and tools straight to your inbox every week.
Weekly · AI news
Join 15,000+ AI enthusiasts
