Implement skeleton based on stable-baselines
dniku opened this issue · 1 comments
dniku commented
Approximate requirements:
- Script that trains an agent with one of the existing algorithms in
stable-baselines
- Runs locally, no need to adapt for cloud
- Fixes random seed
- Specifies network architecture (might be a copy of a predefined one), not imports it from the library
- Saves model after training
- Saves metrics after training (similar to ones reported in LTIAA paper)
Let's make attention visualization a part of another issue (inb4: gym has video recording).
dniku commented
Remains:
- Import code to repo
- Write instructions for launching