admin

admin synced commits to refs/heads/main at admin/lerobot from mirror 2026-05-26 14:09:48 +00:00

5c98e80430 fix(gr00t): fix Eagle25VL model and processor crash in transformers>=5.4.0, <5.6.0 (#3652)

f65f3f7a4a Fix policy.path in YAML configs (PR #3145 followup) (#3597)

Compare 2 commits »

admin synced commits to refs/heads/feat/vla-jepa at admin/lerobot from mirror 2026-05-26 14:09:47 +00:00

5e2c4c22c8 Merge branch 'main' into feat/vla-jepa

1a536d1f71 smol fix to avoid having default CPU device when training

5c98e80430 fix(gr00t): fix Eagle25VL model and processor crash in transformers>=5.4.0, <5.6.0 (#3652)

f65f3f7a4a Fix policy.path in YAML configs (PR #3145 followup) (#3597)

d09410882d adding instructions for different embodiement + fixing some tests

Compare 10 commits »

admin synced commits to refs/heads/auto/update-uv-lock at admin/lerobot from mirror 2026-05-26 05:59:53 +00:00

31ae19ccbf chore(dependencies): update uv.lock

admin synced commits to refs/heads/feat/language-annotation-pipeline at admin/lerobot from mirror 2026-05-26 05:59:53 +00:00

1e7c0d6aa1 annotate(plan): force composite-action subtasks; ban ultra-fine splits

920c6ef5a2 docs(annotate): disable phase-0 vocabulary discovery by default in run_hf_job

Compare 2 commits »

admin synced commits to refs/heads/feat/smolvla-on-steerable at admin/lerobot from mirror 2026-05-26 05:59:53 +00:00

2686450d68 annotate(plan): force composite-action subtasks; tune run_hf_job for robocasa_smoke

4913356564 pi052: SDPA attention port + selective AC + bench harness

Compare 2 commits »

admin synced commits to refs/heads/feat/language-annotation-pipeline at admin/lerobot from mirror 2026-05-25 21:50:04 +00:00

c37b1fc7d0 Merge origin/feat/language-annotation-pipeline (8 fix(annotate) commits + vocabulary phase)

9020635b14 Merge branch 'main' into feat/language-annotation-pipeline

8194897994 fix(deps): cap placo below 0.9.16 and harden kinematics import (#3647)

9f437d86b6 fix(groot): align GR00TN15Config with transformers config dataclasses (#3606)

b74a551d38 fix(pi0, pi05): stabilize torch.compile and expand test coverage (#3610)

Compare 34 commits »

admin synced commits to refs/heads/feat/smolvla-on-steerable at admin/lerobot from mirror 2026-05-25 21:50:03 +00:00

673cc6b0fe pi052: opt-in Liger fused kernels (rope + geglu + layer_norm)

2ed6519a93 ema: enable by default (matches openpi JAX behavior)

72ea531017 train: switch EMA from custom ModelEMA to ema-pytorch

56a934ec55 train: EMA of policy parameters (opt-in via --ema.enable=true)

738e317caa pi052: PaLM-style z-loss on text CE (default weight 1e-4)

Compare 45 commits »

admin synced commits to refs/heads/feat/add-topreward-rm at admin/lerobot from mirror 2026-05-25 21:50:03 +00:00

3b5b94dbd6 optmize topreward input processing (#3660)

admin synced commits to refs/heads/feat/smolvla-on-steerable at admin/lerobot from mirror 2026-05-25 13:40:03 +00:00

83d0c390da pi052: drop debug scaffolding left over from training/inference bug hunts

1ff10b935c Merge branch 'feat/language-annotation-pipeline' into feat/smolvla-on-steerable

67bdf4690e examples(port_datasets): rewrite RoboCasa composite_seen builder

8085feab6e pi052(runtime): factor out shared observation-prep boilerplate

a088c10c80 examples(port_datasets): SLURM+datatrove RoboCasa composite_seen build

Compare 76 commits »

admin synced commits to refs/heads/auto/update-uv-lock at admin/lerobot from mirror 2026-05-25 05:29:55 +00:00

b14587bada chore(dependencies): update uv.lock

admin synced commits to refs/heads/feat/language-annotation-pipeline at admin/lerobot from mirror 2026-05-23 20:50:03 +00:00

471b2b1b1d fix(annotate): bump same-frame subtasks onto distinct frames

admin synced commits to refs/heads/auto/update-uv-lock at admin/lerobot from mirror 2026-05-23 12:40:09 +00:00

bf9d518527 chore(dependencies): update uv.lock

admin synced commits to refs/heads/feat/language-annotation-pipeline at admin/lerobot from mirror 2026-05-23 12:40:08 +00:00

a15e16c072 fix(annotate): replace fuzzy subtask snapping with strict match + one-shot retry

admin synced commits to refs/heads/feat/add-topreward-rm at admin/lerobot from mirror 2026-05-23 04:30:11 +00:00

616663cd9f fix(rewards/topreward): fix pyproject extra typo and simplify processor (#3653)

admin synced commits to refs/heads/feat/depth-integration at admin/lerobot from mirror 2026-05-22 20:19:43 +00:00

3ab08a5318 fix(imports): fixing av import in test_depth.py

admin synced commits to refs/heads/feat/vla-jepa at admin/lerobot from mirror 2026-05-22 20:19:43 +00:00

b75b3ce02d fix qwen norm layer output libero eval is now as expected

admin synced commits to refs/heads/feat/language-annotation-pipeline at admin/lerobot from mirror 2026-05-22 20:19:43 +00:00

336af85c09 fix(annotate): never leave an episode with zero canonical subtasks

admin synced commits to refs/heads/auto/update-uv-lock at admin/lerobot from mirror 2026-05-22 12:09:42 +00:00

20158fa385 chore(dependencies): update uv.lock

8194897994 fix(deps): cap placo below 0.9.16 and harden kinematics import (#3647)

9f437d86b6 fix(groot): align GR00TN15Config with transformers config dataclasses (#3606)

b74a551d38 fix(pi0, pi05): stabilize torch.compile and expand test coverage (#3610)

c0a2e9814d fix examples (#3623)

Compare 7 commits »

admin synced commits to refs/heads/main at admin/lerobot from mirror 2026-05-22 12:09:42 +00:00

8194897994 fix(deps): cap placo below 0.9.16 and harden kinematics import (#3647)

9f437d86b6 fix(groot): align GR00TN15Config with transformers config dataclasses (#3606)

b74a551d38 fix(pi0, pi05): stabilize torch.compile and expand test coverage (#3610)

Compare 3 commits »

admin synced commits to refs/heads/feat/smolvla-on-steerable at admin/lerobot from mirror 2026-05-22 12:09:42 +00:00

77a16db529 fix(smolvla2): make HighLevelSubtaskFwd actually fire at low hz + quiet startup log

ca1b951e7b feat(pi05): expose lm_head_lr_scale for stronger text-CE gradient

9d30d91021 fix(pi052,smolvla2): unblock text generation when LM head drifted to <loc>

Compare 3 commits »