metaopt/torchopt

TorchOpt is an efficient library for differentiable optimization built upon PyTorch.

59
/ 100
Established

TorchOpt helps machine learning engineers and researchers build and train complex neural networks more efficiently. It takes network parameters and loss functions, and outputs optimized parameters, speeding up training for advanced models like those in meta-learning or reinforcement learning. This is designed for practitioners who develop and fine-tune machine learning models.

625 stars. Available on PyPI.

Use this if you are a machine learning practitioner who needs to implement advanced optimization techniques, especially for bi-level optimization problems or when experimenting with functional programming styles in PyTorch.

Not ideal if you are looking for a simple, off-the-shelf deep learning library and are not comfortable with advanced optimization concepts or functional programming paradigms.

deep-learning-optimization meta-learning reinforcement-learning neural-network-training scientific-machine-learning
Maintenance 10 / 25
Adoption 10 / 25
Maturity 25 / 25
Community 14 / 25

How are scores calculated?

Stars

625

Forks

42

Language

Python

License

Apache-2.0

Last pushed

Mar 02, 2026

Commits (30d)

0

Dependencies

5

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/metaopt/torchopt"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.