OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Latent Program Network (LPN)

Latent Program Network (LPN) is an Apache-2.0 research implementation of a neural architecture for inductive program synthesis. It learns a latent space that can be searched at test time, and provides model code, re-ARC dataset utilities, Hydra configurations and a training script.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Latent Program Network (LPN) — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use LPN when you want a research-codebase starting point for training or adapting a neural approach to inductive program synthesis, rather than a general-purpose AI service.

Who it is for

  • Researchers studying inductive program synthesis and latent-space search.
  • Developers adapting neural training and dataset workflows for related program-synthesis experiments.

Documented uses

  • Train a model implementing the LPN architecture using a repository configuration.
  • Explore research workflows for searching latent program spaces.
  • Adapt the included model, dataset-processing and configuration code for related experiments.
Before you choose

Requirements & limits

The code is publicly available on GitHub under the Apache-2.0 license. The README also links to an interactive Hugging Face demo.

  • The documented training workflow requires JAX and the packages in requirements.txt.
  • The README asks users to configure Hugging Face and Weights & Biases credentials.
  • The repository includes model, dataset, configuration and training components rather than only a model description.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start a documented LPN training run

Use the repository’s example command to launch training with the pattern_2d configuration.

Before you start

  • The LPN repository and dependencies set up.
  • Hugging Face and Weights & Biases credentials configured as described in the README.

Steps

  1. From the repository directory, run the command below.

    Documented command
    python src/train.py --config-name pattern_2d

  2. To use a different configuration, replace pattern_2d with the desired configuration name.

Expected outcome

The documented training script is invoked with the selected configuration.

Guide limitations

  • The README documents the command but does not specify training duration, resource requirements or expected output.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
clement-bonnet/lpn · Official homepage title · retained in its source language
How we review sourcesSuggest a correction