Stars
🔥 Turn entire websites into LLM-ready markdown or structured data. Scrape, crawl and extract with a single API.
SGLang is a fast serving framework for large language models and vision language models.
800,000 step-level correctness labels on LLM solutions to MATH problems
A very simple GRPO implement for reproducing r1-like LLM thinking.
Train your grpo with zero dataset and low resources, 8bit/4bit/lora/qlora supported, multi-gpu supported ...
From Llama to Deepseek, grpo/mtp implemented. With pt/sft/lora/qlora included
🚀🚀 「大模型」2小时完全从0训练26M的小参数GPT！🌏 Train a 26M-parameter GPT from scratch in just 2h!
Python tool for converting files and office documents to Markdown.
整理目前开源的最优表格识别模型，完善前后处理，模型转换为ONNX Organize the currently open-source optimal table recognition models, improve pre-processing and post-processing, and convert the models to ONNX.
Efficient Triton Kernels for LLM Training
Supercharge Your LLM Application Evaluations 🚀
Official implementation of "Hydra: Bidirectional State Space Models Through Generalized Matrix Mixers"
收录NLP竞赛策略实现、各任务baseline、相关竞赛经验贴（当前赛事、往期赛事、训练赛）、NLP会议时间、常用自媒体、GPU推荐等，持续更新中
A Comprehensive Toolkit for High-Quality PDF Content Extraction
Neuro-Symbolic Integration Brings Causal and Reliable Reasoning Proofs
Official implementation for "iTransformer: Inverted Transformers Are Effective for Time Series Forecasting" (ICLR 2024 Spotlight), https://openreview.net/forum?id=JePfAI8fah
MiniCPM-o 2.6: A GPT-4o Level MLLM for Vision, Speech and Multimodal Live Streaming on Your Phone
TrustRAG：The RAG Framework within Reliable input,Trusted output
Function to calculate mAP for set of detected boxes and annotated boxes.
Code release for "COTR: Correspondence Transformer for Matching Across Images"(ICCV 2021)
All deep learning-based infrared and visible image fusion algorithms in a whole framework
MedicalGPT: Training Your Own Medical GPT Model with ChatGPT Training Pipeline. 训练医疗大模型，实现了包括增量预训练(PT)、有监督微调(SFT)、RLHF、DPO、ORPO、GRPO。