This pull request includes several improvements and fixes related to Project Tau. The main changes are:
Implemented a state machine to replace the coroutine for managing the training process, resulting in a significant speed increase and more stable training.
Optimized PCA values, leading to faster convergence on generalization with only 3 columns.
Fixed issues related to inconsistent tokens during the token table build phase.
[x] Bug fix (non-breaking change which fixes an issue)
[x] New feature (non-breaking change which adds functionality)
[ ] Breaking change (fix or feature that would cause existing functionality to not work as expected)
[ ] Documentation update
How Has This Been Tested?
The changes have been tested through extensive debugging and logging. The following tests were conducted:
Verified that the state machine transitions correctly through all states.
Ensured that the training process runs smoothly without crashes.
Monitored the performance and reward metrics to confirm improvements.
Test Configuration:
Firmware version: N/A
Hardware: Standard development machine
Toolchain: Unity, C#
SDK: Unity ML-Agents
Checklist:
[x] My code follows the style guidelines of this project
[x] I have performed a self-review of my own code
[x] I have commented my code, particularly in hard-to-understand areas
[x] I have made corresponding changes to the documentation
[x] My changes generate no new warnings
[x] I have added tests that prove my fix is effective or that my feature works
[x] New and existing unit tests pass locally with my changes
[x] Any dependent changes have been merged and published in downstream modules
Additional Notes
The implementation of the state machine has not only improved the speed but also the stability of the training process. The optimized PCA values have shown significant improvements in the convergence rate.
Description
This pull request includes several improvements and fixes related to Project Tau. The main changes are:
Fixes:
Type of Change
Please delete options that are not relevant.
How Has This Been Tested?
The changes have been tested through extensive debugging and logging. The following tests were conducted:
Test Configuration:
Checklist:
Additional Notes
The implementation of the state machine has not only improved the speed but also the stability of the training process. The optimized PCA values have shown significant improvements in the convergence rate.