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

Lamp

LaMP is a Python and PyTorch research implementation of Label Message Passing networks for multi-label classification. It represents labels as graph nodes and uses message passing to model label dependencies conditioned on an input, with repository scripts for preprocessing, training and evaluation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Lamp — 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 LaMP when you want to experiment with graph-based label dependencies in a multi-label classification workflow and can work with its listed Python and PyTorch requirements.

Who it is for

  • Researchers studying graph neural networks and multi-label classification.
  • Developers who want to adapt a research implementation for multi-label text or image classification.
  • Practitioners comparing label-graph configurations on multi-label datasets.

Documented uses

  • Train and evaluate a multi-label classifier using label relationships.
  • Prepare custom input and label data using the repository’s preprocessing example.
  • Compare prior, fully connected and edgeless label-graph configurations.
  • Choose an encoder appropriate to the input representation.
Before you choose

Requirements & limits

The repository is public and MIT-licensed.

  • The README lists Python 3.4 or later, PyTorch 0.2.0, tqdm and NumPy as requirements.
  • Preprocessed example data is available in a 745 MB archive; custom data requires following the example in utils/preprocess.py.
  • The README specifies an MLP encoder for nuswide_vector, while the graph encoder is used for the other listed datasets.
  • Per-epoch evaluation uses a default threshold; final results require selecting an optimal threshold for each metric on the validation set.
  • The README points to the separate C-Tran project as an updated model for multi-label image classification.

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

Prepare the included LaMP dataset

Download and unpack the preprocessed dataset archive used by the repository workflow.

Before you start

  • Python 3.4 or later, PyTorch 0.2.0, tqdm and NumPy are listed as project requirements.
  • A shell with wget and tar available.
  • The dataset archive is listed as 745 MB.

Steps

  1. Download the archive with wget using the URL shown in the README.

  2. Unpack it into the current directory with tar, as shown in the README.

Expected outcome

The provided preprocessed data is unpacked for the repository’s training workflow.

Guide limitations

  • This guide covers data preparation only, not running training or evaluating a model.
  • The README directs users with custom data to the example in utils/preprocess.py.
  • The archive size is listed as 745 MB.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
16 October 2026
Original website title
QData/LaMP · Official homepage title · retained in its source language
How we review sourcesSuggest a correction

Related options

C-Tran — A separate project linked in the LaMP README as an updated model for multi-label image classification.