kelvin34501
commented
2 weeks ago How can I use reinforcement learning algorithms to train my model?
This repository only provides a framework for implementing reinforcement learning/imitation learning algorithms and does not include specific algorithm implementations (e.g., PPO); you can select any algorithm to suit the model learning process.
How can I evaluate my model?
In the challenge's evaluation, some particles are initialized in the source container, and the model's success is determined by calculating the proportion of particles transferred to the target container.
Should I only focus on the "pour" task?
The challenge considers only the "pour" task for ranking.
how can I use reinforcement learning algorithms to train my model. After that, how can I evaluate my model, should I only focus on the "pour"task?