microsoft/monitors4codegen
Code and Data artifact for NeurIPS 2023 paper - "Monitor-Guided Decoding of Code LMs with Static Analysis of Repository Context". `multispy` is a lsp client library in Python intended to be used to build applications around language servers.
This project helps software developers who use large language models (LLMs) to generate code. It takes partially written code and guides the LLM to produce completions that are syntactically and semantically correct, leveraging static analysis of the entire codebase. The output is more reliable and compilable code, reducing common errors like using undefined variables or incorrect method arguments.
280 stars. No commits in the last 6 months.
Use this if you are a software developer using AI code generation tools and need to improve the correctness and compilation rates of the generated code, especially for tasks requiring global context from your repository.
Not ideal if you are looking for a general-purpose code generation tool that does not rely on static analysis or language servers, or if your primary concern is not code correctness.
Stars
280
Forks
34
Language
Python
License
MIT
Category
Last pushed
Aug 08, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/microsoft/monitors4codegen"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
aalok-sathe/surprisal
A unified interface for computing surprisal (log probabilities) from language models! Supports...
EvolvingLMMs-Lab/lmms-engine
A simple, unified multimodal models training engine. Lean, flexible, and built for hacking at scale.
FunnySaltyFish/Better-Ruozhiba
【逐条处理完成】人为审核+修改每一条的弱智吧精选问题QA数据集
reasoning-machines/pal
PaL: Program-Aided Language Models (ICML 2023)
YutongWang1216/DocMTAgent
Code and data releases for the paper -- DelTA: An Online Document-Level Translation Agent Based...