My learning notes/codes for ML SYS.

[NeurIPS 2025] Official Implementation of ViSpec: Accelerating Vision-Language Models with Vision-Aware Speculative Decoding.

A multilingual and multimodal LLM E-Commerce benchmark.

Code from "Exploring optimal transport-based multi-grained alignments for text-molecule retrieval" (IEEE BIBM 2024)

SWE-bench: Can Language Models Resolve Real-world Github Issues?

Demystifying Reinforcement Learning in Agentic Reasoning

The development and future prospects of large multimodal reasoning models.

🚀 EvoAgentX: Building a Self-Evolving Ecosystem of AI Agents

[ACL 2025] Optimal Transport-Based Token Weighting scheme for Enhanced Preference Optimization

SPEC-RL: Accelerating On-Policy Reinforcement Learning via Speculative Rollouts

DCPO: Dynamic Adaptive Clipping for RL

EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL

Interactive Pytorch forward pass visualization in notebooks

Awesome-GraphRAG: A curated list of resources (surveys, papers, benchmarks, and opensource projects) on graph-based retrieval-augmented generation.

About Awesome things towards foundation agents. Papers / Repos / Blogs / ...

诺亚盘古大模型研发背后的真正的心酸与黑暗的故事。

Fully open reproduction of DeepSeek-R1

A series of math-specific large language models of our Qwen2 series.

Recipes to train reward model for RLHF.

verl: Volcano Engine Reinforcement Learning for LLMs

Official repository for ACL 2025 paper "ProcessBench: Identifying Process Errors in Mathematical Reasoning"

An Open Large Reasoning Model for Real-World Solutions

InstantIR: Blind Image Restoration with Instant Generative Reference 🔥

GenRM-CoT: Data release for verification rationales

A compact LLM pretrained in 9 days by using high quality data

ACL 2024 | LooGLE: Long Context Evaluation for Long-Context Language Models