This repository contains the implementation of the paper "Task weighting in meta-learning with trajectory optimisation" published on Transactions on Machine Learning Research (August 2023).
The implementation relies on PyTorch. The code is tested on the following Python packages:
- torch 2.0.1
- torchvision
- torcheval
- higher
- aim (for experiment management and result visualisation)
One can run the code following the arguments specified in run.sh. The result can be visualised by running the following command in the terminal:
aim up --repo logs
and follows the prompt in the terminal to open the browser at the corresponding port.