Khalil Meftah 64613d8f03 fix(rewards): harden distributional value target projection ...

- Clamp HL-Gauss targets
- Handle terminal mask shapes safely
- Align the default smoothing ratio with Stop Regressing

2026-07-25 17:17:35 +02:00

..

__init__.py

Reward models refactor (#3142)

2026-04-28 17:56:24 +02:00

test_classifier_processor.py

Reward models refactor (#3142)

2026-04-28 17:56:24 +02:00

test_distributional_value_function.py

fix(rewards): harden distributional value target projection

2026-07-25 17:17:35 +02:00

test_modeling_classifier.py

RL stack refactoring (#3075)

2026-05-12 15:49:54 +02:00

test_modeling_robometer.py

feat(rewards): add ROBOMETER reward model (#3627)

2026-05-29 21:45:39 +02:00

test_modeling_topreward.py

feat(rewards): add TOPReward reward model (#3629)

2026-05-27 14:24:31 +02:00

test_nanovlm_value_function.py

fix: nanoVLM processing

2026-07-23 22:50:29 +02:00

test_reward_factory_features.py

fix: populate config from dataset metadata

2026-07-23 22:17:39 +02:00

test_reward_model_base.py

fix(train): migrate legacy RA-BC fields in train config loading (#3480)

2026-04-29 16:17:00 +02:00

test_robometer_processor.py

feat(rewards): add ROBOMETER reward model (#3627)

2026-05-29 21:45:39 +02:00

test_sarm_processor.py

Reward models refactor (#3142)

2026-04-28 17:56:24 +02:00

test_sarm_utils.py

Reward models refactor (#3142)

2026-04-28 17:56:24 +02:00

test_temporal_siglip_value_function.py

fix(rewards): harden distributional value target projection

2026-07-25 17:17:35 +02:00

test_topreward_processor.py

feat(rewards): add TOPReward reward model (#3629)

2026-05-27 14:24:31 +02:00

test_topreward.py

feat(rewards): add TOPReward reward model (#3629)

2026-05-27 14:24:31 +02:00