median-research-group/LibMTL
A PyTorch library implementing multi-task learning algorithms including MMoE, PLE, and gradient-based multi-objective optimization.
Not currently ranked — collecting fresh signals.
star history
LibMTL is an open-source PyTorch library for multi-task learning that provides standardized implementations of various MTL architectures and optimization algorithms. It includes popular models like Multi-Gate Mixture of Experts (MMoE) and Progressive Layered Extraction (PLE), as well as gradient manipulation techniques for multi-objective optimization. The library serves as a research framework and practical tool for developing and comparing multi-task deep learning models.
Frequently asked
- What is median-research-group/LibMTL?
- A PyTorch library implementing multi-task learning algorithms including MMoE, PLE, and gradient-based multi-objective optimization.
- Is LibMTL open source?
- Yes — median-research-group/LibMTL is open source, released under the MIT license.
- What language is LibMTL written in?
- median-research-group/LibMTL is primarily written in Python.
- How popular is LibMTL?
- median-research-group/LibMTL has 2.6k stars on GitHub.
- Where can I find LibMTL?
- median-research-group/LibMTL is on GitHub at https://github.com/median-research-group/LibMTL.