Viewer for neural network, deep learning and machine learning models, supporting ONNX, TensorFlow, PyTorch, Core ML and many other formats.
33.7K+26 this week3.2KJavaScriptMITWell maintainedUpdated yesterday
meta-pytorch/tnt
TNT (torchtnt) is a lightweight library of training tools and utilities for PyTorch models.
- Stars
- 1,723
- Forks
- 313
- Open issues
- 96
- Last commit
- 9 days ago
What is it?
TNT, published on PyPI as torchtnt, is a library of training tools and utilities built on top of PyTorch. It is meant to help developers assemble the common parts of a training pipeline without writing them from scratch.
The package can be installed with pip or conda-forge, and the latest development version can be installed directly from the master branch on GitHub. PyTorch should be installed first if installation issues occur. A nightly build is also published to PyPI under the name torchtnt-nightly.
The project is licensed under a BSD-style license per its PyPI metadata and is tagged for deep learning, reinforcement learning and neural network work.
Key features
- 01Provides reusable utilities for PyTorch training workflows.
- 02Installable via pip or conda-forge.
- 03Offers a nightly build for trying the latest changes.
- 04Can be installed directly from the GitHub master branch.
Use cases
- Speeding up the setup of a PyTorch training loop with shared utilities.
- Experimenting with deep learning or reinforcement learning models in PyTorch.
- Tracking the latest library changes through the nightly or master builds.
Quick start
- 1
Install PyTorch first if you run into issues
- 2
Install torchtnt with pip
pip install torchtnt
Things to know
- The README does not state a license; metadata shows only an unidentified (NOASSERTION) license, so check the LICENSE file before use.
In collections
Topics
FAQ
- What is tnt?
- TNT (torchtnt) is a lightweight library of training tools and utilities for PyTorch models. TNT, published on PyPI as torchtnt, is a library of training tools and utilities built on top of PyTorch. It is meant to help developers assemble the common parts of a training pipeline without writing them from scratch.
- Is tnt free to use? What license does it have?
- tnt is open source under the Other license. Check the license terms before using it in commercial products.
- Is tnt still maintained?
- Actively maintained. The last commit to tnt was on Oct 29, 2026.
- What can I use tnt for?
- Speeding up the setup of a PyTorch training loop with shared utilities. Experimenting with deep learning or reinforcement learning models in PyTorch. Tracking the latest library changes through the nightly or master builds.
- How do I install tnt?
- Install PyTorch first if you run into issues Install torchtnt with pip: pip install torchtnt
- Where can I get tnt?
- The source code and installation instructions are on GitHub: https://github.com/meta-pytorch/tnt
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
Rust bindings for the C++ API of PyTorch (libtorch), letting Rust code build, train and run deep learning models.
5.5K+4 this week463RustApache-2.0Updated 3 months ago
An engine-agnostic deep learning framework for Java that lets developers build, train, and run models with a native Java API.
4.9K+2 this week761JavaApache-2.0Updated 6 days ago
A curated awesome-list cataloging PyTorch-related libraries, models, implementations, tutorials and papers on GitHub.
16.7K+14 this week2.8KNo licenseUpdated 2 months ago
A short introductory book on reinforcement learning, from basics to applied algorithms, with PyTorch implementations in the repo.
1.7K+7 this week96PythonNo licenseUpdated 4 months ago
An end-to-end open source platform for building and deploying machine learning models, with Python and C++ APIs.
201.4K+133 this week78.4KC++Apache-2.0Well maintainedUpdated 5 hours ago