jgrusewski
ba8fb8eedd
feat(ppo): wire symlog, adaptive entropy, percentile scaling into training loop
...
Integrate all PPO improvement modules into the core training paths:
- Symlog value predictions in compute_value_loss (MLP + LSTM)
- Adaptive entropy auto-tuning replaces fixed entropy_coeff
- Percentile P5/P95 advantage scaling for heavy-tailed returns
- DAPO clip_epsilon_high wired in all 7 PPOConfig construction sites
- Shape mismatch fix in adaptive_entropy (unsqueeze scalar)
2726 tests pass, 0 clippy errors.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com >
2026-03-03 20:12:00 +01:00
..
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-26 00:50:25 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-26 00:50:25 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-03 00:57:14 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-26 00:50:25 +01:00
2026-02-26 00:05:38 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-03 20:12:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-03 20:12:00 +01:00
2026-03-02 11:02:56 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-26 00:50:25 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-26 00:50:25 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-03-02 15:41:23 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00
2026-02-25 11:56:00 +01:00