[RyanPCo]

Some Mecka episodes have failed head tracking. This PR adds a way to exclude episodes by hash and a script that finds the Mecka ones.

Problem. Of 41,607 local Mecka episodes, 2,866 (6.9%) carry a single obs_head_pose for the whole episode even though the video shows the head moving; 142 more have single-frame head jumps of 0.3–3.7 m. Each action chunk is inv(head_pose_t) · world_keypoints_{t+k}, so with a frozen pose, head motion leaks into the hand trajectory labels. mecka_to_zarr passes the vendor's egomotion through unchanged, so the fault is upstream; reported to the vendor.

Changes

• DatasetFilter(exclude_hashes=...): a list of hashes or a path to a text file of hashes; a missing file raises. Excluded episodes never match. A pinned hash that is excluded is dropped with a warning instead of raising PinError; excluding every pin raises, since an empty pin set means "no filter". Part of cache_key.
• scripts/mecka_process/scan_egomotion.py classifies each local episode's head pose (ok / constant / jump / ...) and writes a blocklist in the format exclude_hashes reads.
• data/mecka_all_6d gains exclude_hashes: null with a comment on how to pass a blocklist. Nothing is excluded by default, so splits are unchanged.

The generated blocklist and the repinned 20% split are kept out of this PR.

tests/unit/test_dataset_filter.py: exclusion by list and file, a missing file raising, dropping excluded pins, and the cache key.

🤖 Generated with Claude Code

https://claude.ai/code/session_01HjCQBqQCr2igFEMR7yUovw

[RyanPCo]

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

• norm stats: cache sampled rows per episode so splits share them #687 : 2 dependent PRs (#688 , #689 )
• data: exclude episodes by hash; scan Mecka egomotion for failed head tracking #686 👈 (View in Graphite)
• eval: log validation videos to wandb #684 : 1 other dependent PR (#685 )
• slurm: stop cancelled jobs from leaking whole GPUs #682 : 1 other dependent PR (#683 )
• algo: RDT (Robotics Diffusion Transformer) on the HPT chassis #680 : 1 other dependent PR (#681 )
• pi: initialize from PaliGemma with a scratch action expert #677 : 1 other dependent PR (#678 )
• perf: vectorized image augs and a torch.compile knob for HPT and Pi; Pi clips gradients like openpi #676
• loader: one zarr read per key, not one per keymap entry #674
• eval: build the pinned-episode video batches only where they are used #673
• checkpoints: last.ckpt as a symlink, not a second full copy #672
• trainHydra: metric_frames_per_episode: auto, and warn on empty prompts #671
• transforms: per-step wrist framing for proprio history #670
• viz: one keypoint renderer, the horizon trace #669
• configs: history and backbone-LR knobs on the keypoint wrist HPT model #668
• trainer: ddp_bf16_compress, DDP with a bf16 gradient all-reduce #667
• eval: overlay videos play at the source rate under DDP #666
• hpt: build the policy on CPU and let Lightning place it #664 : 1 other dependent PR (#665 )
• hpt: never re-initialise pretrained stems; hash-check them; mask text padding #645 : 1 other dependent PR (#646 )
• obs: proprio history (K proprio steps, one token each, learned time embedding) #650
• loss: mask the repeat-last action padding out of the HPT loss #649
• resnet: keep a frozen backbone in eval mode; give backbones their own LR group #652
• augs: per-sample image augmentation for HPT train-mode augs #651
• eval: deterministic validation (seeded prompts + flow sampling); log-only grad-norm MAD #644
• data: mecka_fold_flagship_top3_hpt_6d, the 3-operator flagship train split #643
• val: subsampled metric loaders + pinned contiguous video loaders #642
• infra: mecka 6D / keypoint recipes, 300M models, defaults, and the rollout revert #640
• eval: layout-aware frame-invariant metrics, and a train-split viz head #659
• data: stop letting one bad episode or a rotation channel take a run down #658
• norm: guard degenerate channels, horizon pooling, and no rot6d normalization #657
• keypoints: the 144-D wrist-first hand-keypoint action becomes the human default #656
• mecka: fix the double-mirrored LEFT wrist frame, and correct existing zarrs #655
• 6d: PI goes 6D-only; Eva and Human gain the continuous-rotation modes #654
• 6d: continuous rotation representation, bimanual layouts, shared norm helpers #653
• train: give each Slurm slurm job gets its own cache dir to avoid conflict #638 : 1 other dependent PR (#639 )
• write git version into the run dir for reproducibility #637
• train: don't load PI base weights on resume and verify that checkpoint path exists #636
• configs: declare proprio and actions dims once per HPT model config #634 : 1 other dependent PR (#635 )
• cache norm stats for data configs to minimize recomputation #633
• import cloud packages lazily to support local training #631 : 1 other dependent PR (#632 )
• cache dataset resolver for faster training initialization #630
• validate episode hashes in data config before training #628
• modify pi camera map to match data #629
• add tests to ensure training works between every model and data vendor #626
• run CI tests on every PR #625
• add env variables for user-specific paths #624
• build: set pyproject and uv as the dependency managers #623
• fix(configs): fix paths for data hydra yaml configs #622
• refactor(algo): delete ACT #627 : 1 other dependent PR (#621 )
• tests: create dedicated test folder with unit and integration tests #620
• add pyright type checker #619
• main

This stack of pull requests is managed by Graphite. Learn more about stacking.