Reproduce R1 Zero on Logic Puzzle

Distilling the Essence: Efficient Reasoning Distillation via Sequence Truncation

Step-DeepResearch

A Real-Time Fault-tolerant In-Memory Distributed Message Queue

LlamBERT implements a hybrid approach approach for text classification that leverages LLMs to annotate a small subset of large, unlabeled databases and uses the results for fine-tuning transformer …

EFFICIENT AND OPTIMIZED TOKENIZER ENGINE FOR LLM INFERENCE SERVING

Official Repository for "See, Rank and Filter: Important Word-Aware Clip Filtering via Scene Understanding for Moment Retrieval and Highlight Detection" (AAAI 2026 Oral)

Counter-factual reward ranking

Ambrosia is a Python library for A/B tests design, split and result measurement

Effective LLM Alignment Toolkit

A lightweight, high-performance microservice for forwarding browser-side logs to server-side log aggregation systems (ELK, Loki, Splunk, etc.).

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

Source code for the paper "Fast Offline Policy Optimization for Large Scale Recommendation" published at the Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI-23).

Materials for the "Reward Optimising Recommendation using Deep Learning and Fast Maximum Inner Product Search" tutorial delivered at the 28th SIGKDD Conference on Knowledge Discovery and Data Minin…

Source code for the paper "Logarithmic Smoothing for Pessimistic Off-Policy Evaluation, Selection and Learning" published at NeuRIPS '24.

A GPipe implementation in PyTorch

Context-Adaptive and Consistency-Aware Multi-Modal Outfit Compatibility Modeling

FB (Facebook) + GEMM (General Matrix-Matrix Multiplication) - https://code.fb.com/ml-applications/fbgemm/

PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning

1688 taobao jd image search products

Official implementation for "SocialNav: Training Human-Inspired Foundation Model for Socially-Aware Embodied Navigation"

Official Implementation of paper: [Nav-R2:Dual‑Relation Reasoning for Generalizable Open‑Vocabulary Object‑Goal Navigation]

Guided Proximal Policy Optimization with Structured Action Graph

Lifelong Generative Recommendation Unlearning via Dual-Process Memory and Hierarchical Preference Alignment

Open-source platform to build and deploy AI agent workflows.

DreamPRM tackles the dataset quality imbalance and distribution shift that plague multimodal PRM training by domain-reweighting.

本项目旨在提供一个微调酒店推荐垂直领域大模型并应用的完整闭环案例作为大家的参考案例。本项目使用的基础大模型为Qwen2.5-7B-Instruct。项目特色:完整的垂直应用案例闭环、项目源码剖析开源共享、详实的图文指导手册、手把手全流程实操演示视频