Commit Graph

5 Commits

Author SHA1 Message Date
Khalil Meftah 0944b84279 feat(rl): consolidate HIL-SERL checkpoint into HF-style components
Make  and  s, add abstract
 /  for algorithm-owned tensors (critics,
target nets, ), and persist them as a sibling
component next to . Replace the pickled
 side-file with an enriched
carrying both  and , so resume restores actor +
critics + target nets + temperature + optimizers + RNG + counters from
plain HF-standard files.
2026-05-08 21:24:23 +02:00
Khalil Meftah e298474bf3 fix(tests): gate RL tests on the datasets extra 2026-04-27 16:53:34 +02:00
Khalil Meftah 577f14337a refactor(tests): remove grpc import checks from test files for cleaner code 2026-04-27 16:20:13 +02:00
Khalil Meftah 7a1c9e74c3 fix: skip tests that require grpc if not available 2026-04-15 15:18:04 +02:00
Khalil Meftah e022207c75 refactor: RL stack refactoring — RLAlgorithm, RLTrainer, DataMixer, and SAC restructuring 2026-04-13 11:39:48 +02:00