OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Developer Tools & SDKs

TensorSpace.js

TensorSpace.js is a JavaScript framework for building interactive 3D visualizations of neural-network structure and behavior in a web browser. It provides Keras-like layer APIs and can display pre-trained TensorFlow, Keras and TensorFlow.js models after preprocessing.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

TensorSpace.js — OpenViST project profile in AI Developer Tools & SDKs.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when a web project needs an interactive 3D view of a neural network or a pre-trained model’s intermediate behavior and predictions.

Who it is for

  • JavaScript and front-end developers building browser-based neural-network visualizations.
  • Educators and demo creators who want to present model structure, intermediate information and predictions interactively.

Documented uses

  • Build an interactive 3D representation of neural-network layers.
  • Load and visualize a pre-trained model after converting it to TensorSpace-compatible form.
  • Show intermediate model information and final inference in a browser-based demo.
Before you choose

Requirements & limits

  • The framework uses JavaScript with TensorFlow.js, Three.js and Tween.js; the basic browser setup also lists TrackballControls.
  • Pre-trained models need TensorSpace preprocessing before visualization; the project identifies TensorSpace Converter as a conversion option.
  • Installation is documented through browser build files or npm/Yarn for progressive-framework projects.
  • The repository specifies the Apache-2.0 license.

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

Install TensorSpace in a progressive-framework project

Add TensorSpace as a project dependency and import its JavaScript module.

Before you start

  • A progressive-framework project using npm or Yarn.

Steps

  1. Install the package with the command below or yarn add tensorspace.

    Documented command
    npm install tensorspace

  2. Import it in your JavaScript code with import * as TSP from 'tensorspace';.

Expected outcome

TensorSpace is installed and available through the `TSP` module import.

Guide limitations

  • This covers installation and importing only; it does not build or render a visualization.
Documentation 1 ↗
Source-based practical guide

Visualize a preprocessed LeNet model and sample input

Create a browser-based TensorSpace visualization of the documented LeNet example and run its sample prediction.

Before you start

  • TensorSpace installed and available in the page.
  • A preprocessed TensorSpace-compatible LeNet model and sample input image; the documented HelloWorld resources include these.
  • A TensorSpace-Converter runtime environment and the tensorspacejs Python package are required if preprocessing the model yourself.

Steps

  1. If needed, preprocess the model with TensorSpace Converter; for a tf.keras model, the documented conversion uses tensorspacejs_converter with the TensorFlow source, tf_keras format, selected output layer names, input model path, and save directory.

  2. Create a model with new TSP.models.Sequential(document.getElementById("container")).

  3. Add the LeNet layers in order: GreyscaleInput, Padding2d, Conv2d, Pooling2d, Conv2d, Pooling2d, two Dense layers, and Output1d with outputs 0 through 9.

  4. Load the compatible model with model.load({ type: "tensorflow", url: './PATH/TO/MODEL/model.json' }) and call model.init(...).

  5. After initialization, call model.predict(image_5) in the initialization callback to display the sample prediction.

Expected outcome

The browser displays the LeNet model, with the sample image prediction after initialization.

Guide limitations

  • The model must be preprocessed into TensorSpace-compatible format before loading.
  • The paths, container, and image_5 must be supplied by the application; the guide does not provide their actual values.
  • The documented example uses a preprocessed LeNet model and sample input, not an arbitrary model or input.
Documentation 1 ↗Documentation 2 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
TensorSpace.jsBuild an interactive 3D representation of neural-network layers.Web
HyperposeCustomize pose-estimation architectures and datasets.Docker, Linux, macOS, Windows
Multi Task NlpDefine multiple tasks and their data and model settings in a YAML file.Linux, macOS, WindowsThe project is an open-source library under the Apache-2.0 license, distributed through its public GitHub repository.
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
15 October 2026
Original website title
tensorspace-team/tensorspace · Official homepage title · retained in its source language
How we review sourcesSuggest a correction