AP2O
AAAI'26, AP2O-Coder: Adaptively Progressive Preference Optimization for Reducing Compilation and Runtime Errors in LLM-Generated Code
Install / Use
/learn @TsingZ0/AP2OREADME
Introduction
This is the official implementation of our paper AP2O-Coder: Adaptively Progressive Preference Optimization for Reducing Compilation and Runtime Errors in LLM-Generated Code. Accepted by AAAI'26.
Adaptive Progressive Preference Optimization
Coding Error Reduction
Data Efficiency
Reward Curves
Requirements
- deepspeed 0.17.2
- python 3.11.11
- torch 2.7.0
- trl 0.14.0
- transformers 4.51.3
- vllm 0.9.2
Usage
To initiate the preference data self-generation and preference optimization processes, use the following command:
sh pipe-qwen2.5-coder.sh
Related Skills
node-connect
349.0kDiagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps
frontend-design
109.4kCreate distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.
openai-whisper-api
349.0kTranscribe audio via OpenAI Audio Transcriptions API (Whisper).
qqbot-media
349.0kQQBot 富媒体收发能力。使用 <qqmedia> 标签，系统根据文件扩展名自动识别类型（图片/语音/视频/文件）。
