Commit Graph

  • 71f39f6912 chore(video backend): renaming codec into video_backend in get_safe_default_video_backend() CarolinePascal 2026-04-22 20:14:39 +02:00
  • e6e54391bd fix(missing method): adding missing add_microphones_recordings method in LeRobotDataset feat/audio_dataset CarolinePascal 2026-04-29 16:17:40 +02:00
  • b5f65e5332 Expose sarm package API and ship reward model card template (#3477) Khalil Meftah 2026-04-29 16:17:16 +02:00
  • cd6b43ea7a fix(train): migrate legacy RA-BC fields in train config loading (#3480) Khalil Meftah 2026-04-29 16:17:00 +02:00
  • 2236bbe7a3 fix(rollout): propagate policy-specific CLI config paramaters (#3483) Steven Palma 2026-04-29 16:13:10 +02:00
  • a201b33d20 chore(typos): fixing uncaught typos CarolinePascal 2026-04-29 16:13:03 +02:00
  • 5c444302c1 feat(so_follower): synchronize goal position with present position to prevent positional error during torque re-enablement feat/leader_activate_teleop Khalil Meftah 2026-04-28 18:40:48 +02:00
  • cb0a944941 refactor(datasets): replace untyped dict with typed DatasetInfo dataclass (#3472) Maxime Ellerbach 2026-04-28 18:40:30 +02:00
  • 8a3d64033f Reward models refactor (#3142) Khalil Meftah 2026-04-28 17:56:24 +02:00
  • c868f874f1 feat(teleop): enhance leader-follower behavior and torque management in SO101 teleoperation Khalil Meftah 2026-04-28 17:46:06 +02:00
  • e228f0880f feat(teleop): add SO100/SO101 leader-follower teleoperation example fix: update import for SO101Leader in so101_leader_follower.py chore: include SO101LeaderFollower in exports Khalil Meftah 2026-04-28 17:28:15 +02:00
  • fe2c32d9e7 add so leader arm Khalil Meftah 2026-04-28 16:53:36 +02:00
  • 03ee50e08f chore(ci): bump docs workflows (#3476) Steven Palma 2026-04-28 15:06:44 +02:00
  • 05764da0f1 disable rotation feat/leader-arm-intervention-pr2596 Khalil Meftah 2026-04-28 15:04:05 +02:00
  • f40a36fe82 diable rotation Khalil Meftah 2026-04-28 13:29:31 +02:00
  • 9b538d6cbf fix follower shakiness and space-bar trigger Khalil Meftah 2026-04-28 13:11:59 +02:00
  • a59900a339 fix missing kwargs Khalil Meftah 2026-04-28 12:17:24 +02:00
  • 6ed80f5a59 Merge remote-tracking branch 'origin/main' into user/khalil-meftah/2026-02-16-rl-stack-refactor Khalil Meftah 2026-04-28 12:04:13 +02:00
  • 5cea61708d add so100 leader as hil teleoperation Khalil Meftah 2026-04-28 11:46:21 +02:00
  • ef6b3b5b0f refactor: simplify docstrings for clarity and conciseness across multiple files Khalil Meftah 2026-04-28 11:11:02 +02:00
  • ca87ccd941 feat(rollout): decouple policy deployment from data recording with new lerobot-rollout CLI (#3413) Steven Palma 2026-04-28 00:57:35 +02:00
  • 77352c495c chore(dependencies): update uv.lock (#3437) Steven Palma 2026-04-27 23:15:46 +02:00
  • 7b82a5c381 feat(rl): add bus-control primitives and smooth move functionality for leader intervention feat/rl-leader-arm-intervention Khalil Meftah 2026-04-27 18:31:13 +02:00
  • 13418dcd7b feat(rl): port haptic follow + torque toggle from #2596 to leader intervention Khalil Meftah 2026-04-27 17:50:29 +02:00
  • a3cb9f5317 feat(rl): leader arm as HIL-SERL intervention device (position-only) Khalil Meftah 2026-04-27 17:26:29 +02:00
  • e298474bf3 fix(tests): gate RL tests on the datasets extra Khalil Meftah 2026-04-27 16:53:34 +02:00
  • 577f14337a refactor(tests): remove grpc import checks from test files for cleaner code Khalil Meftah 2026-04-27 16:20:13 +02:00
  • 47be90f040 refactor(rl): make RLAlgorithmConfig an abstract base class for better extensibility Khalil Meftah 2026-04-27 15:59:59 +02:00
  • 47dd65347e refactor(rl): add type property to RLAlgorithmConfig for better clarity Khalil Meftah 2026-04-27 15:57:24 +02:00
  • fd5a788120 refactor(rl): add make_algorithm_config function for RLAlgorithmConfig instantiation Khalil Meftah 2026-04-27 15:55:16 +02:00
  • 0b06790da0 feat(language): add motion (persistent) and trace (event-only) styles Pepijn 2026-04-27 14:21:49 +02:00
  • b43dc39ba4 Add docstrings to all new helpers; revert uv.lock Pepijn 2026-04-27 14:15:03 +02:00
  • 9ce9e01469 refactor(rl): make algorithm a nested config so all SAC hyperparameters are JSON-addressable Khalil Meftah 2026-04-27 13:39:03 +02:00
  • 2b71221194 Address review: split persistent/event schemas, drop event timestamps Pepijn 2026-04-27 13:38:23 +02:00
  • 21c16a27f0 Revert "perf(observation_processor): add CUDA support for image processing" Khalil Meftah 2026-04-27 11:52:19 +02:00
  • b3164543f4 fix(rl): enhance intervention handling in actor and learner Khalil Meftah 2026-04-26 23:09:33 +02:00
  • f3993cbbb1 fix(rl): improve action processing for discrete and continuous actions Khalil Meftah 2026-04-26 22:47:52 +02:00
  • c278cfa026 fix(rl): postprocess action in actor Khalil Meftah 2026-04-26 18:15:04 +02:00
  • 77d18659b1 fix(rl): mirror gym_manipulator in actor Khalil Meftah 2026-04-26 18:11:26 +02:00
  • 6347edefb1 fix(rl): merge environment and action-processor info in transition processing Khalil Meftah 2026-04-26 17:12:37 +02:00
  • eda47eca18 fix(rl): update neutral gripper action Khalil Meftah 2026-04-26 16:42:53 +02:00
  • a64e6f5070 fix(rl): clarify discrete gripper action mapping in GripperVelocityToJoint for SO100 Khalil Meftah 2026-04-26 16:41:55 +02:00
  • 3def86c2c3 fix(rl): add time limit processor to environment pipeline Khalil Meftah 2026-04-26 16:38:20 +02:00
  • 356a64d8c4 fix(rl): correctly wire HIL-SERL gripper penalty through processor pipeline Khalil Meftah 2026-04-26 16:36:21 +02:00
  • 8833d735a1 Add extensive language support Pepijn 2026-04-27 10:56:32 +02:00
  • ef8bfffbd7 fix(rl): enhance intervention handling in actor and learner test/rl-processor Khalil Meftah 2026-04-26 23:09:33 +02:00
  • f887ab3f6a fix(rl): improve action processing for discrete and continuous actions Khalil Meftah 2026-04-26 22:47:52 +02:00
  • c2556439e5 fix(rl): postprocess action in actor Khalil Meftah 2026-04-26 18:15:04 +02:00
  • d2a046dfc5 fix(rl): mirror gym_manipulator in actor Khalil Meftah 2026-04-26 18:11:26 +02:00
  • 613d581f6c remove debug Khalil Meftah 2026-04-26 18:08:13 +02:00
  • 58b6d844c4 debug Khalil Meftah 2026-04-26 17:33:15 +02:00
  • 30e1886b64 fix(rl): merge environment and action-processor info in transition processing Khalil Meftah 2026-04-26 17:12:37 +02:00
  • 9c9064e5be fix(rl): update neutral gripper action Khalil Meftah 2026-04-26 16:42:53 +02:00
  • 494f469a2b fix(rl): clarify discrete gripper action mapping in GripperVelocityToJoint for SO100 Khalil Meftah 2026-04-26 16:41:55 +02:00
  • cd105f65cb fix(rl): add time limit processor to environment pipeline Khalil Meftah 2026-04-26 16:38:20 +02:00
  • 9c2af818ff fix(rl): correctly wire HIL-SERL gripper penalty through processor pipeline Khalil Meftah 2026-04-26 16:36:21 +02:00
  • 05a5223885 fix(pi): avoid peak RAM in PiGemma construction by freeing replaced submodules (#3454) Steven Palma 2026-04-24 17:50:12 +02:00
  • 6495bb9706 add processor to main Khalil Meftah 2026-04-24 17:06:57 +02:00
  • 727ca1a92c chore(rollout): remove speculative action order fix codex/fix-rollout-relative-actions Pepijn 2026-04-24 16:24:44 +02:00
  • ee737b72d0 fix(rollout): avoid oversampling observations during interpolation Pepijn 2026-04-24 15:42:36 +02:00
  • cf2e42f557 fix(rollout): preserve relative action chunks Pepijn 2026-04-23 13:52:58 +02:00
  • e9f3f88377 fix(rollout): preserve relative action chunks codex/rollout-relative-action-fixes Pepijn 2026-04-23 13:52:58 +02:00
  • 580d818aa9 fix(dataset): no default overwrite in lerobot tool recompute stats (#3452) Steven Palma 2026-04-24 15:07:19 +02:00
  • 38b88c414c perf(observation_processor): add CUDA support for image processing Khalil Meftah 2026-04-24 13:36:26 +02:00
  • 1ed32210c7 refactor(rl/sac): consolidate hyperparameter ownership and clean up discrete critic Khalil Meftah 2026-04-24 13:18:33 +02:00
  • 587aa82021 fix(imports): realsense import name is platform dependent (#3451) Steven Palma 2026-04-24 12:55:38 +02:00
  • 12b88fce02 not use dataclass (#3414) Chuyao Shen 2026-04-24 05:26:59 -04:00
  • 06255996ea refactor(policies): rename policies/sac → policies/gaussian_actor Khalil Meftah 2026-04-23 19:13:18 +02:00
  • fc6c94c82a fix(sarm): handle BaseModelOutputWithPooling from transformers 5.x in… (#3419) masato-ka 2026-04-23 23:26:58 +09:00
  • 1add460678 fix(policy): loss normalization for padded actions in ACT, Diffusion, and MultiTaskDiT (#3442) Steven Palma 2026-04-23 15:23:54 +02:00
  • 4587c2b648 fix xvla docs (#3291) Qi Jia 2026-04-24 00:50:32 +12:00
  • a23ebf9d35 fix(profiling): address review feedback codex/model-profiling Pepijn 2026-04-23 13:23:09 +02:00
  • 2236cdb302 fix(smolvla): correct loss normalization for padded actions (#3434) whats2000 2026-04-23 16:34:11 +08:00
  • eb9519eb91 fix(tests): default value RTCConfig Steven Palma 2026-04-22 23:06:01 +02:00
  • f29ce39f39 fix(rollout): enable default Steven Palma 2026-04-22 18:47:15 +02:00
  • 682d5f2f95 Merge branch 'main' into feat/lerobot-rollout Steven Palma 2026-04-22 16:39:18 +02:00
  • 7c2466979e chore(dependencies): update uv.lock (#3408) Steven Palma 2026-04-22 16:38:51 +02:00
  • 6ee093db9a Merge branch 'main' into feat/lerobot-rollout Steven Palma 2026-04-22 16:38:13 +02:00
  • 4499519dbf chore(rollout): address coments + minor improvements Steven Palma 2026-04-22 16:37:19 +02:00
  • bfff81fd4b perf(smolvla): remove redundant img_emb identity assignment in embed_prefix Pepijn 2026-04-22 16:34:19 +02:00
  • d70c3baf7c chore(rollout): multiple minor improvements Steven Palma 2026-04-22 15:18:50 +02:00
  • bff2d50dc1 docs(rollout): edit rename_map instructions Steven Palma 2026-04-22 13:59:55 +02:00
  • 8d1401abe3 fix(rollout): device and task propagation + feature pos + warn fps + move rename_map config Steven Palma 2026-04-22 12:55:22 +02:00
  • 39b966e20a docs(agents): add AGENT_GUIDE.md for user facing agent (#3430) Pepijn 2026-04-22 11:54:19 +02:00
  • ba27aab79c fix(robotwin): pin compatible curobo in benchmark image (#3427) Pepijn 2026-04-21 19:51:44 +02:00
  • 475b8da7d6 device + task + warn fix feat/lerobot-rollout-fix Steven Palma 2026-04-21 19:06:53 +02:00
  • 51025c6cad fix(robotwin): pin compatible curobo in benchmark image codex/robotwin-curobo-fix Pepijn 2026-04-21 18:41:16 +02:00
  • 929400cd44 style(profiling): satisfy pre-commit checks Pepijn 2026-04-21 18:16:00 +02:00
  • fe78f8fee9 fix(profiling): handle datasets without metadata in forward artifacts Pepijn 2026-04-21 18:04:35 +02:00
  • ce9bfa754d Merge branch 'main' into codex/model-profiling Pepijn 2026-04-21 17:59:39 +02:00
  • 5adad11128 feat(sim): VLABench benchmark integration (#3396) Pepijn 2026-04-21 17:54:11 +02:00
  • 8065bf15c7 fix test for flat dict structure Khalil Meftah 2026-04-21 12:06:25 +02:00
  • 8191d2d87f remove unused type alias Khalil Meftah 2026-04-21 11:56:27 +02:00
  • 6b93f31238 fix docstring Khalil Meftah 2026-04-21 11:55:17 +02:00
  • a4c0c9e358 update losses names in tests Khalil Meftah 2026-04-21 11:53:32 +02:00
  • b86935c64b Merge branch 'main' into codex/model-profiling Pepijn 2026-04-21 11:23:26 +02:00
  • a2f72e42f6 fix(profiling): convert uint8 images to float32 in deterministic forward Pepijn 2026-04-20 23:33:24 +02:00
  • a515eadc96 refactor(profiling): consolidate into single module Pepijn 2026-04-20 21:31:17 +02:00
  • a07f22e22c feat(envs): add LIBERO-plus robustness benchmark (#3313) Pepijn 2026-04-20 21:07:21 +02:00
  • 282c31cfef feat(envs): add RoboMME benchmark (#3311) Pepijn 2026-04-20 20:21:27 +02:00