Force DeepSeek r1 models to think for as long as you wish

Forces DeepSeek R1 models to engage in extended reasoning by intercepting early termination tokens.

Automatic architecture search and hyperparameter optimization for PyTorch

HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusal

[ACL 2024] An Easy-to-use Knowledge Editing Framework for LLMs.

Implementation of InstructEdit

A straightforward method for training your LLM, from downloading data to generating text.

Movement trajectory classes and functions built on top of GeoPandas

Minimal reproduction of DeepSeek R1-Zero

LLM inference in C/C++

Editing Models with Task Arithmetic

Fully open reproduction of DeepSeek-R1

[CVPR2025 Highlight] Video Generation Foundation Models: https://saiyan-world.github.io/goku/

aider is AI pair programming in your terminal

SGLang is a fast serving framework for large language models and vision language models.

Official repository of the xLSTM.

A visual playground for agentic workflows: Iterate over your agents 10x faster

Fully open data curation for reasoning models

A collection of MCP servers.

The official Python SDK for Model Context Protocol servers and clients

OpenAPI Tool Servers

A simple, secure MCP-to-OpenAPI proxy server

Model Context Protocol Servers

A flexible HTTP fetching Model Context Protocol server.

Ein Model Context Protocol (MCP) Server für die Deutsche Bahn Timetable API

This project provides a modular Python wrapper for the SNCF API, with an MCP server interface that integrates seamlessly with Claude Desktop for intelligent journey planning and train information r…

Stop re-explaining your project to AI every session. Automatic context memory for Claude, VS Code, Cursor, and 13+ AI tools.

A Model Context Protocol (MCP) server that provides tools for fetching and analyzing Reddit content.

A fast, forgiving GTFS reader built on pandas DataFrames

MCP server to give client the ability to access Wikipedia pages