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

Texting

TextING is a Python and TensorFlow research implementation of inductive text classification using graph neural networks. It builds an individual graph for each document, learns word representations from local structure, and aggregates word nodes into a document representation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Texting — 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 TextING when you want a research implementation and training workflow for inductive graph-based text classification, including experiments with a custom labeled corpus.

Who it is for

  • Researchers exploring graph neural networks for text classification
  • Practitioners training classifiers on labeled document collections

Documented uses

  • Train a classifier on a labeled text corpus.
  • Adapt the repository workflow to custom text and label files.
  • Create document representations by aggregating graph-encoded word nodes.
  • Explore classification of new documents containing unseen words.
Before you choose

Requirements & limits

The implementation is in a public GitHub repository. License status is not specified in the project information.

  • The method constructs a separate graph for each document rather than relying on one shared corpus graph.
  • The documented stack is Python 3.6+, TensorFlow/TensorFlow-GPU 1.12.0 and SciPy 1.5.1.
  • The workflow requires pretrained GloVe 6B, 300-dimensional embeddings.
  • The README identifies four included datasets: MR, OHSUMED, R8 and R52.
  • Larger hidden and batch sizes may increase memory demands.

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 a custom corpus and start training

Use the documented repository workflow to preprocess custom text, build document graphs, and launch training and inference.

Before you start

  • Python 3.6 or later
  • TensorFlow or TensorFlow-GPU 1.12.0
  • SciPy 1.5.1
  • Pretrained GloVe 6B, 300-dimensional embeddings placed in the repository
  • Custom text file and corresponding label file arranged in the repository’s data structure

Steps

  1. Place the custom text file under data/corpus/ and its label file under data/.

  2. Preprocess the text by running remove_words.py.

  3. Build graphs with python build_graph.py [DATASET] [WINSIZE].

  4. Start training and inference with python train.py, using the documented optional parameters as needed.

Expected outcome

The repository workflow builds graphs for the custom corpus and starts model training and inference.

Guide limitations

  • The README does not provide a complete command for remove_words.py or prescribe a custom dataset identifier format.
  • Training memory needs depend on parameter choices; the README advises adjusting hidden size and batch size to available memory.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
CRIPAC-DIG/TextING · Official homepage title · retained in its source language
How we review sourcesSuggest a correction