chore(claude): learn from #229#231
Closed
claude[bot] wants to merge 9 commits into
Closed
Conversation
- addresses @claude[bot] (low-level is_pad defaults, low_level_planner modeling_pi07_low_level.py:984-986): switched the *_is_pad fallbacks to torch.ones to match the high-level planner — missing speed/quality/mistake no longer fabricate "Speed: 0.0" entries in the prompt. - addresses @claude[bot] (prepare_metadata docstrings): updated both planners' docstrings to reference Gemma 3 (not PaliGemma) and to enumerate robot_type / control_mode (string-valued, empty-string-as-pad). - addresses @claude[bot] (high-level all-empty guard, high_level_planner/modeling_pi07_high_level.py:715): mirrored the low-level "if segments else ''" guard so an all-padded sample emits "" instead of the literal "Metadata: ". Both planners now agree. - addresses @claude[bot] (CPU coverage for prepare_metadata): added TestPrepareMetadataSegments to tests/policies/test_pi07_cpu.py covering (a) robot+control populated, (b) both absent, (c) one populated and one empty, (d) per-sample all-empty emits "" (regression for both planners), and (e) low-level missing speed/quality/mistake never produces a fabricated "Speed: 0.0" segment. tests: passed — pytest -m "not gpu" tests/policies/test_pi07_cpu.py Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
- addresses @claude[bot] (subgoal opt-in docstring): updated _load_subgoal_frames docstring to match the always-on behaviour; replaced the stale test_missing_subgoals_key_in_info_returns_empty test (which encoded the old opt-in gate) with two pinned cases — no-cameras → {}, and "no info.subgoals key still loads subgoals". Added camera_keys / image_keys / episode_data_index attrs to the SimpleNamespace meta in the two existing video tests so they match the new attribute reads. - addresses @claude[bot] (image-dtype fallback row index): new test_image_dtype_fallback_uses_absolute_row_index in tests/datasets/ test_optional_keys.py stubs hf_dataset.__getitem__ and pins that the parquet-row lookup uses ep_start + subgoal_frame, never the within-episode index. - addresses @claude[bot] ("fully 0 padded" comment): rewrote the comment at modeling_pi07_low_level.py:946 to call out -1 ([-1, 1] SigLIP range) and the False-mask role of the placeholder. - addresses @claude[bot] (_action_indicator_len recompute): cached the Action-indicator length at PI07LowLevelPlannerFlowMatching.__init__; both forward sites now read self._action_indicator_len. - addresses @claude[bot] (embed_prefix CPU coverage): added TestEmbedPrefixConditionalGuards in tests/policies/test_pi07_cpu.py with a fake Gemma3WithExpert + tokenizer + state_proj + embed_video so the three guards (response_masks.any(), subgoal availability, metadata_masks.any()) are exercised without GPU. Cases: all-False optional masks → no spurious causal boundary; mixed-availability subgoal batch → header/footer pad mask zeroes the pad-only sample; response.any() → exactly one boundary. - addresses @claude[bot] (.base_layer. stale comment in video_encoder.py): rewrote the wrap comment at video_encoder.py:422 to explain that the wrapper adopts submodules by reference, so wrapped-layer state-dict keys are byte-for-byte identical to a vanilla SiglipEncoderLayer (no .base_layer. prefix). tests: passed — pytest tests/policies tests/datasets -m "not gpu" -n auto (452 passed, 12 skipped; the 1 collection error in tests/policies/test_pi07_paligemma_low_level_planner.py is pre-existing and unrelated — pi07_paligemma still imports VJEPA2VideoEncoder from pi05_mem, which was removed in #171). Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Add hard rule 4: pin training-path layout fixes with a CPU unit test. PR #229 review surfaced that the `[1]+[0]*(N-1)` -> `[1]*N` att_masks fix — the most material correctness change in the PR, since it shifts the cumsum at the indicator -> first-discrete boundary and changes the discrete-action CE loss — initially shipped without a CPU test pinning the pattern. It would have only been caught by deferred GPU and nightly regression tests. Author added a CPU assertion in dda4446 after review. Rule 3 (determinism) doesn't cover this case: two seeded runs of the new code agree, but determinism cannot tell you the layout is correct. A pinned CPU test does. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Member
|
Doesn't worth an entry in CLAUDE.md |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this does
Adds hard rule 4 to
CLAUDE.md: pin training-path layout fixes with a CPU unit test that asserts the exact pattern.PR #229 review (claude-bot, discussion_r3176226835) flagged that the
[1]+[0]*(N-1)→[1]*Natt_maskschange inpi07/low_level_planner/modeling_pi07_low_level.py— the most material correctness fix in the PR, since it shifts the cumsum at the indicator → first-discrete boundary and changes the discrete-action CE loss — shipped without a CPU test pinning the pattern. Without one, the divergence would only have been caught by deferred GPU/nightly regression tests. The author addedtest_discrete_actions_indicator_uses_per_token_causal_blocksintests/policies/test_pi07_cpu.py(commitdda4446) after review.The lesson generalizes beyond this PR: rule 3 (determinism) only proves two seeded runs agree, not that the layout is correct. A pinned CPU assertion catches a regression on the same PR; GPU + nightly do not.
No stale CLAUDE.md content surfaced by #229's diff —
cpu_test.yml,gpu_test.yml,regression_test.yml, and the file paths in rule 3 all still exist.How it was tested
Documentation-only change. No code paths affected.
How to checkout & try? (for the reviewer)
Checklist
🤖 Generated with Claude Code