NuCon/scripts
Dominik Roth 646399dcc7 feat: improve NN dynamics model and SAC training
- ReactorDynamicsNet: add dropout (0.3) for regularisation
- ReactorDynamicsModel: z-score normalisation of inputs/outputs, predict
  per-second rates of change, forward_with_uncertainty() stub
- rl.py: misc SAC training improvements
- sim.py: minor fixes
- train_sac.py: updated training loop

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-15 11:18:15 +01:00
..
collect_dataset.py feat: SAC+HER training on kNN-GP sim with direct bypass and scripts/ 2026-03-12 20:43:37 +01:00
reactor_control.py feat: hand-written PID reactor controller with curses TUI 2026-03-15 00:18:00 +01:00
train_sac.py feat: improve NN dynamics model and SAC training 2026-03-15 11:18:15 +01:00