ZikaiXiao

new-api. A unified AI model hub for aggregation & distribution. It supports cross-converting various LLMs into OpenAI-compatible, Claude-compatible, or Gemini-compatible formats. A centralized gateway for personal and enterprise model management. 🍥

sub2api. Sub2API 一站式开源中转服务，让 Claude、Openai 、Gemini、Grok订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。

Qwen-AgentWorld. Qwen-AgentWorld: Language World Models for General Agents

baidupan. Fast Baidu Pan CLI tool using official xpan API. Concurrent upload/download, rapid upload, resume support.

science-taste-skills. Two-layer Codex-native science taste skills: a persona layer plus a reusable gate.

Awesome-Agent-Environments. Awesome Agent Environments

claude-code-py. Python

everyone-can-use-english. 人人都能用英语

claw-code. An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.

learn-claude-code. Learn Claude Code — 基于源码的完整技术分析文档集，15章深度解析 Agent Loop、工具系统、权限系统等核心机制

TorchCode. 🔥 LeetCode for PyTorch — practice implementing softmax, attention, GPT-2 and more from scratch with instant auto-grading. Jupyter-based, self-hosted or try online.

Scrapling. 🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl!

neo. Turn any web app into an API. Chrome extension captures browser traffic, auto-generates schemas, lets AI replay APIs directly. No official API needed.

Agent-Reach. Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

MedEthicsQA. Python

SurgBench. Python

UniVBench. [CVPR 2026]The official code and datasets for "UniVBench: Towards Unified Evaluation for Video Foundation Models"

nanochat. The best ChatGPT that $100 can buy.

minimind. 🧠「大模型」2小时完全从0训练64M的小参数LLM！Train a 64M-parameter LLM from scratch in just 2h!

TradingAgents. TradingAgents: Multi-Agents LLM Financial Trading Framework

learning_research. 本人的科研经验

AgentsMeetRL. Awesome List for Agentic RL

claude-code-proxy. Claude Code to OpenAI API Proxy

claude-relay-service. CRS-自建Claude Code镜像，一站式开源中转服务，让 Claude、OpenAI、Gemini、Droid 订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。

claude-code-router. One local control plane for every AI agent: route across models, fuse new capabilities, orchestrate tools, and stay fully in control.

LongWeave. Python

crypto_info. （原创）全网最全-币圈区块链各类常用工具与相关信息资料大全-虚拟加密货币-欧易OKX币安Binace芝麻开门Gate-交易所App注册-NFT-Defi-加密钱包-比特币-新手入门教程 -持续更新

akshare. AKShare is an elegant and simple financial data interface library for Python, built for human beings! 开源财经数据接口库

freqtrade. Free, open source crypto trading bot

QuantsPlaybook. 量化研究-券商金工研报复现

vnpy. 基于Python的开源量化交易平台开发框架

aioquant. Asynchronous event I/O driven quantitative trading framework.

bitcoinjs-lib. A javascript Bitcoin library for node.js and browsers.

ccf-deadlines. ⏰ Agenticly track worldwide conference deadlines (Website, Python Cli, Wechat Applet)

BrowserGym. 🌎💪 BrowserGym, a Gym environment for web task automation

Awesome-World-Models. A comprehensive list of papers for the definition of World Models and using World Models for General Video Generation, Embodied AI, and Autonomous Driving, including papers, codes, and related websites.

json_repair. Repair malformed JSON from LLMs, APIs, logs, and user input in Python.

smolagents. 🤗 smolagents: a barebones library for agents that think in code.

Awesome-World-Model. Collect some World Models for Autonomous Driving (and Robotic, etc.) papers.

Awesome-MCP-ZH. MCP 资源精选， MCP指南，Claude MCP，MCP Servers, MCP Clients

alfworld. ALFWorld: Aligning Text and Embodied Environments for Interactive Learning

SkyRL. SkyRL: A Modular Full-stack RL Library for LLMs

verl-agent. verl-agent is an extension of veRL, designed for training LLM/VLM agents via RL. verl-agent is also the official code for paper "Group-in-Group Policy Optimization for LLM Agent Training"

Awesome-Agent-RL.

verl. verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

Awesome-Surgical-Video-Understanding. There are compilations of surgery-related tasks, datasets, and papers.

ReCall. ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning & ReCall: Learning to Reason with Tool Call for LLMs via Reinforcement Learning

loong. 🐉 Loong: Synthesize Long CoTs at Scale through Verifiers.

owl. 🦉 OWL: Optimized Workforce Learning for General Multi-Agent Assistance in Real-World Task Automation

A-Comprehensive-Survey-For-Long-Context-Language-Modeling. A Comprehensive Survey on Long Context Language Modeling

OneRuler. Python

COMET. A Neural Framework for MT Evaluation

open-deep-research. Open source alternative to Gemini Deep Research. Generate reports with AI based on search results.

Awesome-Text2SQL. Curated tutorials and resources for Large Language Models, Text2SQL, Text2DSL、Text2API、Text2Vis and more.

awesome-ai-agents. A list of AI autonomous agents

aimo-progress-prize. Jupyter Notebook

LongProc. LongProc: Benchmarking Long-Context Language Models on Long Procedural Generation

FActScore. A package to evaluate factuality of long-form generation. Original implementation of our EMNLP 2023 paper "FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation"

long-form-factuality. Benchmarking long-form factuality in large language models. Original code for our paper "Long-form factuality in large language models".

miniforge. A conda-forge distribution.

agentinstruct. Code repo for "Agent Instructs Large Language Models to be General Zero-Shot Reasoners"

Qwen3. Qwen3 is the large language model series developed by Qwen team, Alibaba Cloud.

qwen-vllm. 通义千问VLLM推理部署DEMO

Endo-FM. [MICCAI'23] Foundation Model for Endoscopy Video Analysis via Large-scale Self-supervised Pre-train

InternVideo. [ECCV2024] Video Foundation Models & Data for Multimodal Understanding

LLM-Synthetic-Data. A live reading list for LLM data synthesis (Updated to July, 2025).

Awesome-LLM-Synthetic-Data. A reading list on LLM based Synthetic Data Generation 🔥

RWKV-LM. RWKV (pronounced RwaKuv) is an RNN with great LLM performance, which can also be directly trained like a GPT transformer (parallelizable). We are at RWKV-7 "Goose". So it's combining the best of RNN and transformer - great performance, linear time, constant space (no kv-cache), fast training, infinite ctx_len, and free sentence embedding.

RL_Learning. Python

Book-Mathematical-Foundation-of-Reinforcement-Learning. This is the homepage of a new book entitled "Mathematical Foundations of Reinforcement Learning."

Toolkit-for-Prompt-Compression. Toolkit for Prompt Compression

EgoSurgery. [MICCAI 2024] Official dataset release for "EgoSurgery: A Dataset for Surgical Video Understanding from Egocentric Open Surgery Videos"

LLMLingua. [EMNLP'23, ACL'24] To speed up LLMs' inference and enhance LLM's perceive of key information, compress the prompt and KV-Cache, which achieves up to 20x compression with minimal performance loss.

fluent-reader. Modern desktop RSS reader built with Electron, React, and Fluent UI

awesome-newsCN-feeds. 优质的「中文新闻媒体」订阅列表

VideoMAEv2. [CVPR 2023] VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking

VideoMAE. [NeurIPS 2022 Spotlight] VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training

Awesome-LLMs-for-Video-Understanding. 🔥🔥🔥 [IEEE TCSVT] Latest Papers, Codes and Datasets on Vid-LLMs.

OphNet-benchmark. [ECCV 2024] Official Implementation of "OphNet: A Large-Scale Video Benchmark for Ophthalmic Surgical Workflow Understanding"

SAM-2_Surgical_Video. Segment Anything 2 for Surgical Video Segmentation

list-of-surgical-tool-datasets. List of surgical tool datasets organised by task.

Awesome-Surgical-Video-Analysis. Papers of ComputerVision x Surgery

needle-threading. ICLR 2025: Needle Threading

GPT-Book-Summarizer. Automatically distilling comprehensive book summaries by leveraging recursive meta summarization across book sections.

Awesome-Video-Language-Understanding. A Survey on video and language understanding.

Awesome_Long_Form_Video_Understanding. Awesome papers & datasets specifically focused on long-term videos.

ai-for-grant-writing. A curated list of resources for using LLMs to develop more competitive grant applications.

Awesome-Dataset-Distillation. A curated list of awesome papers on dataset distillation and related applications.

LLM-ToolMaker. Jupyter Notebook

UltraChat. Large-scale, Informative, and Diverse Multi-round Chat Data (and Models)

lingua. Meta Lingua: a lean, efficient, and easy-to-hack codebase to research LLMs.

OpenRLHF. An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO & REINFORCE++ & VLM & TIS & vLLM & Ray & Async RL)

Megatron-LM. Ongoing research training transformer models at scale

ring-flash-attention. Ring attention implementation with flash attention

FlagScale. FlagScale is a large model toolkit based on open-sourced projects.

Odysseus-Transformer. Odysseus: Playground of LLM Sequence Parallelism

ProLong. Homepage for ProLong (Princeton long-context language models) and paper "How to Train Long-Context Language Models (Effectively)"

MedicalGPT. 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。

long-context-attention. USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference

LongRecipe. LongRecipe: Recipe for Efficient Long Context Generalization in Large Language Models

ms-swift. Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, Phi4, ...) (AAAI 2025).

FABLES. Python

opencompass. OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

meditron. Meditron is a suite of open-source medical Large Language Models (LLMs).

LMOps. General technology for enabling AI capabilities w/ LLMs and MLLMs

EasyContext. Memory optimization and training recipes to extrapolate language models' context length to 1 million tokens, with minimal hardware.

Firefly. Firefly: 大模型训练工具，支持训练Qwen2.5、Qwen2、Yi1.5、Phi-3、Llama3、Gemma、MiniCPM、Yi、Deepseek、Orion、Xverse、Mixtral-8x7B、Zephyr、Mistral、Baichuan2、Llma2、Llama、Qwen、Baichuan、ChatGLM2、InternLM、Ziya2、Vicuna、Bloom等大模型

MetaGPT. 🌟 The Multi-Agent Framework: First AI Software Company, Towards Natural Language Programming

LongWriter. [ICLR 2025] LongWriter: Unleashing 10,000+ Word Generation from Long Context LLMs

awesome-LLM-resources. 🧑‍🚀 全世界最好的LLM资料总结（多模态生成、Agent、辅助编程、AI审稿、数据处理、模型训练、模型推理、o1 模型、MCP、小语言模型、视觉语言模型） | Summary of the world's best LLM resources.

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

OLAPH. OLAPH: Improving Factuality in Biomedical Long-form Question Answering

Awesome-Multimodal-Large-Language-Models. :sparkles::sparkles:Latest Advances on Multimodal Large Language Models

Awesome-Medical-Pretraining. Literature reviews of (Unsupervised/self-supervised) pretraining on medical datasets

the-pile. Python

mimic-code. MIMIC Code Repository: Code shared by the research community for the MIMIC family of databases

LongHealth. LongHealth: A Question Answering Benchmark with Long Clinical Documents

Clinical-Longformer. Python

MedOdyssey. Python

ContrastiveDecoding. contrastive decoding

distil-cd. Distillation Contrastive Decoding: Improving LLMs Reasoning with Contrastive Decoding and Distillation

Awesome-LLM-Strawberry. A collection of LLM papers, blogs, and projects, with a focus on OpenAI o1 🍓 and reasoning techniques.

Awesome-LLM-hallucination. LLM hallucination paper list

LLMs_interview_notes. 该仓库主要记录 大模型（LLMs） 算法工程师相关的面试题

Awesome-Mamba-Papers. Awesome Papers related to Mamba.

self-correction-llm-papers. This is a collection of research papers for Self-Correcting Large Language Models with Automated Feedback.

awesome-hallucination-detection. List of papers on hallucination detection in LLMs.

llm-action. 本项目旨在分享大模型相关技术原理以及实战经验（大模型工程化、大模型应用落地）

quantgpt-agents. A Stock Price prediction system using LLM and Multi-agent-system

VCD. [CVPR 2024 Highlight] Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decoding

CAD. Unofficial re-implementation of "Trusting Your Evidence: Hallucinate Less with Context-aware Decoding"

Awesome-LLM-Uncertainty-Reliability-Robustness. Awesome-LLM-Robustness: a curated list of Uncertainty, Reliability and Robustness in Large Language Models

Awesome-LLM-Inference. 📚A curated list of Awesome LLM/VLM Inference Papers with Codes: Flash-Attention, Paged-Attention, WINT8/4, Parallelism, etc.🎉

GEAR. GEAR: An Efficient KV Cache Compression Recipefor Near-Lossless Generative Inference of LLM

keyformer-llm. Keyformer proposes KV Cache reduction through key tokens identification and without the need for fine-tuning

pyramidinfer. Python

Awesome-KV-Cache-Compression. 📰 Must-read papers on KV Cache Compression (constantly updating 🤗).

SnapKV. Python

LLM-Attributor. LLM Attributor: Attribute LLM's Generated Text to Training Data

semantic_uncertainty. Codebase for reproducing the experiments of the semantic uncertainty paper (short-phrase and sentence-length experiments).

llama3-sft. Supervised FineTuning with Llama-3

PositionalHidden. To mitigate position bias in LLMs, especially in long-context scenarios, we scale only one dimension of LLMs, reducing position bias and improving average performance by up to 15.4%

Entropy-ABF. Official implementation for 'Extending LLMs’ Context Window with 100 Samples'

LongLM. [ICML'24 Spotlight] LLM Maybe LongLM: Self-Extend LLM Context Window Without Tuning

perm-sc. Official codebase for permutation self-consistency.

llama3-Chinese-chat. Llama3-中文后训练版

MoICE. Official implementation for "Mixture of In-Context Experts Enhance LLMs’ Awareness of Long Contexts" (Accepted by Neurips2024)

Mooncake. Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

know-dont-tell. Python

moe_attention. Official repository for the paper "SwitchHead: Accelerating Transformers with Mixture-of-Experts Attention"

LoRA. Code for loralib, an implementation of "LoRA: Low-Rank Adaptation of Large Language Models"

LoRAMoE. LoRAMoE: Revolutionizing Mixture of Experts for Maintaining World Knowledge in Language Model Alignment

xlora. X-LoRA: Mixture of LoRA Experts

grokfast. Official repository for the paper "Grokfast: Accelerated Grokking by Amplifying Slow Gradients"

eda_nlp. Data augmentation for NLP, presented at EMNLP 2019

data-selection-survey. A Survey on Data Selection for Language Models

Awesome-FL. Comprehensive and timely academic information on federated learning (papers, frameworks, datasets, tutorials, workshops)

needle-in-a-haystack. Doing simple retrieval from LLM models at various context lengths to measure accuracy

proxy_based_uncertainty. Python

ProLong. [ACL 2024 (Oral)] A Prospector of Long-Dependency Data for Large Language Models

FedLoGe. Python

SkipAlign.

CLEX. [ICLR 2024] CLEX: Continuous Length Extrapolation for Large Language Models

DAPE. The this is the official implementation of "DAPE: Data-Adaptive Positional Encoding for Length Extrapolation"

CaR. Clustering and Ranking: Diversity-preserved Instruction Selection through Expert-aligned Quality Estimation

MoDS. Python

Active-IT. Code for our EMNLP-2023 paper: "Active Instruction Tuning: Improving Cross-Task Generalization by Training on Prompt Sensitive Tasks"

WizardLM. LLMs build upon Evol Insturct: WizardLM, WizardCoder, WizardMath

Superfiltering. [ACL'24] Superfiltering: Weak-to-Strong Data Filtering for Fast Instruction-Tuning

Cherry_LLM. [NAACL'24] Self-data filtering of LLM instruction-tuning data using a novel perplexity-based difficulty score, without using any other models

LESS. [ICML 2024] LESS: Selecting Influential Data for Targeted Instruction Tuning

Long-Context-Data-Engineering. Implementation of paper Data Engineering for Scaling Language Models to 128K Context

llama3. The official Meta Llama 3 GitHub site

LEval. [ACL'24 Outstanding] Data and code for L-Eval, a comprehensive long context language models evaluation benchmark

AlpacaDataCleaned. Alpaca dataset from Stanford, cleaned and curated

Self-instruct. A repository to perform self-instruct with a model on HF Hub

self-instruct. Aligning pretrained language models with instruction data generated by themselves.

LongBench. LongBench v2 and LongBench (ACL 25'&24')

long-range-arena. Long Range Arena for Benchmarking Efficient Transformers

RedPajama-Data. The RedPajama-Data repository contains code for preparing large datasets for training large language models.

llm-training-calculator. Python

LongAlign. [EMNLP 2024] LongAlign: A Recipe for Long Context Alignment of LLMs

performer-pytorch. An implementation of Performer, a linear attention-based transformer, in Pytorch

reformer-pytorch. Reformer, the efficient Transformer, in Pytorch

memory-transformer-xl. A variant of Transformer-XL where the memory is updated not with a queue, but with attention

Retrieval_Head. open-source code for paper: Retrieval Head Mechanistically Explains Long-Context Factuality

gputasker. An awesome gpu tasks scheduler. 轻量好用的GPU机群任务调度工具。觉得有用可以点个star

transformer-xl. Python

Infini-Attention. Efficient Infinite Context Transformers with Infini-attention Pytorch Implementation + QwenMoE Implementation + Training Script + 1M context keypass retrieval

LongQLoRA. LongQLoRA: Extent Context Length of LLMs Efficiently

Awesome-LLM-Long-Context-Modeling. 📰 Must-read papers and blogs on LLM based Long Context Modeling 🔥

GaLore. GaLore: Memory-Efficient LLM Training by Gradient Low-Rank Projection

Chinese-LLaMA-Alpaca-2. 中文LLaMA-2 & Alpaca-2大模型二期项目 + 64K超长上下文模型 (Chinese LLaMA-2 & Alpaca-2 LLMs with 64K long context models)

long_llama. LongLLaMA is a large language model capable of handling long contexts. It is based on OpenLLaMA and fine-tuned with the Focused Transformer (FoT) method.

infini-transformer. PyTorch implementation of Infini-Transformer from "Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention" (https://arxiv.org/abs/2404.07143)

LlamaFactory. Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)

Chinese-LLaMA-Alpaca. 中文LLaMA&Alpaca大语言模型+本地CPU/GPU训练部署 (Chinese LLaMA & Alpaca LLMs)

roformer. Rotary Transformer

lloco. The official repo for "LLoCo: Learning Long Contexts Offline"

rho. Repo for Rho-1: Token-level Data Selection & Selective Pretraining of LLMs.