A natural language interface for computers

Lightweight coding agent that runs in your terminal

Train the smallest LM you can that fits in 16MB. Best model wins!

Official Codebase for "Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights"

Agent framework and applications built upon Qwen>=3.0, featuring Function Calling, MCP, Code Interpreter, RAG, Chrome extension, etc.

Minimal, clean code for the Byte Pair Encoding (BPE) algorithm commonly used in LLM tokenization.

🚀🤖 Crawl4AI: Open-source LLM Friendly Web Crawler & Scraper. Don't be shy, join here: https://discord.gg/jP8KfhDhyN

Test your prompts, agents, and RAGs. Red teaming/pentesting/vulnerability scanning for AI. Compare performance of GPT, Claude, Gemini, Llama, and more. Simple declarative configs with command line …

🚀 Efficient implementations for emerging model architectures

Accessible large language models via k-bit quantization for PyTorch.

A large collection of system log datasets for AI-driven log analytics [ISSRE'23]

A Flexible Framework for Experiencing Heterogeneous LLM Inference/Fine-tune Optimizations

The Infibench variant of bigcode-evaluation-harness --- a framework for the evaluation of autoregressive code generation language models.

New repo collection for NVIDIA Cosmos: https://github.com/nvidia-cosmos

Embodied Agent Interface (EAI): Benchmarking LLMs for Embodied Decision Making (NeurIPS D&B 2024 Oral)

[ICLR 2024] Codes for COLEP: Certifiably Robust Learning-Reasoning Conformal Prediction via Probabilistic Circuits

a lighter, faster browser kernel of blink to integrate HTML UI in your app. 一个小巧、轻量的浏览器内核，用来取代wke和libcef

Leetcode solutions, code skeletons, and unit tests in Java (in progress)

DeepSeek Coder: Let the Code Write Itself

The evaluation framework for the InfiCoder-Eval benchmark.

MetaRoom Benchmark in VNN-COMP 2023

[NeurIPS 2022] Code for Certifying Some Distributional Fairness with Subpopulation Decomposition

[ICSE 2023] Differentiable interpretation and failure-inducing input generation for neural network numerical bugs.

[NeurIPS 2022] ClimbQ: Class Imbalanced Quantization Enabling Robustness on Efficient Inferences

Travel marker created following「中国制霸生成器」by 卜卜口 (@itorr)

Code for Double Sampling Randomized Smoothing [ICML 2022]