tensorgrid-platform 1.19.7

TensorGrid: multimodal data platform for building high-quality datasets and computer vision models

Building intelligence. Powering tomorrow.

TensorGrid is the open-source multimodal data platform for teams building computer vision and physical AI. Explore datasets visually, curate the samples that matter, annotate in the browser, and evaluate models — all from one interactive workspace and a powerful Python SDK.

Install · Docs · GitHub · Releases

---

Highlights

Visual-first	Interactive app at localhost:5151 — no more blind data wrangling
Multimodal	Images, video, 3D, LiDAR, audio, geolocation, and time series
Python-native	Full SDK for loading, filtering, embedding, and evaluating datasets
Extensible	Plugins, custom operators, and integrations with your ML stack
Open source	Apache 2.0 — run locally, on-prem, or in your cloud

---

Install

pip install tensorgrid-platform

Requirement	Details
Python	3.10, 3.11, or 3.12
Database	MongoDB (auto-started on first launch)
Optional	pip install "tensorgrid-platform[multimodal]" for MCAP codecs

---

Quick start

Launch from the CLI

pip install tensorgrid-platform
tg quickstart                       # downloads the demo dataset and launches the app

Or use the Python SDK

import tensorgrid as tg
import tensorgrid.zoo as tgz

dataset = tgz.load_zoo_dataset("quickstart")
session = tg.launch_app(dataset)
session.wait(-1)  # open http://localhost:5151

Filter a subset and explore

view = dataset.match_tags("train").limit(100)
session = tg.launch_app(view)

---

Platform capabilities

Module	Description
Explore	Slice, search, and filter datasets with visual queries
Curate	Surface hard examples, duplicates, and edge cases
Annotate	Label and review samples in an interactive workflow
Evaluate	Compare models and drill into failure modes
Brain	Embeddings, similarity, and automated dataset insights
Zoo	Built-in sample datasets and models to get started fast

---

Command line

tg zoo datasets list                # browse available zoo datasets
tg zoo datasets load quickstart     # download a dataset by name
tg app launch quickstart            # launch the app with a loaded dataset
tg --help                           # full CLI reference

Tip: tg app launch <name> opens a dataset that is already downloaded. Run tg zoo datasets load <name> first (or use tg quickstart, which downloads and launches the demo dataset in one step).

---

Resources

Documentation	User guide & tutorials
Source code	github.com/rksaklani/TensorGrid
Bug reports	GitHub Issues
Plugins	Plugin directory

---

Development

git clone https://github.com/rksaklani/TensorGrid.git
cd TensorGrid

python3 -m venv .venv && source .venv/bin/activate
pip install -e core/

cd core/app && corepack yarn install && corepack yarn build && cd ../..
./scripts/start-tensorgrid.sh

---

License

Apache 2.0 — free for commercial and research use.

See NOTICE for third-party attributions.