Code-LMs

Visit Tool

Code-LMs is a Coding & Development tool that provides a guide and resources for using pre-trained large language models of source code. It includes models like PolyCoder for code generation and evaluation.

Claim this tool

1View

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is Code-LMs?

Code-LMs offers a comprehensive guide for leveraging pre-trained large language models (LLMs) specifically designed for source code. It details how to use models such as PolyCoder, which are available on the Huggingface Hub, for tasks like code generation. The guide covers setup instructions, model retrieval from public Zenodo repositories, and methods for running models either via Docker or from source using a modified GPT NeoX toolkit. It also provides information on datasets used for training, evaluation methodologies including perplexity and HumanEval benchmarks, and important caveats regarding model limitations and whitespace sensitivity. This resource is ideal for developers and researchers looking to integrate or experiment with code-specific LLMs.

Best used for

Ideal for developers and researchers who need to generate code using large language models, evaluate the performance of code LLMs on benchmarks like HumanEval, and understand the training and deployment processes. Especially valuable for those working with PolyCoder or similar code-specific models.

Common actions

generate code

evaluate models

research language models

deploy code models

low-code/no-codeautomated workflowdeepfakeopen-sourceworkflowscollaboration"AI Agents"github copilotface swapping

Capabilities

Key features

Pre-trained code LLMs
Code generation
Model evaluation benchmarks
Huggingface Hub integration
Docker deployment
Multilingual code corpus

Target Audience

developer

Integrations

hugging-face

Pricing & Plans

Open Source

Free

FAQs

What are the primary models available through Code-LMs?

Code-LMs primarily features PolyCoder models, including 160M, 0.4B, and 2.7B parameter versions. These models are trained on a large multilingual corpus of source code and are accessible via the Huggingface Hub for easy integration and use in code generation tasks.

How can I deploy and run the Code-LMs models?

You can deploy and run the models either by using a provided Docker image, which contains a modified GPT NeoX repository, or by setting up the environment from source using a public fork of the NeoX repository. Checkpoint files for the models are hosted on Zenodo.

What are the known limitations of the Code-LMs models?

The models have limitations such as not being trained to solve programming problems, potential issues with generating new files after an end-of-document token, and high sensitivity to whitespace. They primarily learn from comments in code, not natural language prompts.

Trending

Subcategories trending in Coding & Development

Open Source & Models DevOps & Infrastructure No-Code / Low-Code Testing & QA Backend & APIs Prompt Engineering

Trending

Also listed in

This tool also appears in

AI Agents & Automation › AI Frameworks & Infra

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce