Repository navigation
fix(swebench): preserve per-test grading across pytest formats - #4475
Closed
giuliolovisotto wants to merge 1 commit into
Closed
giuliolovisotto wants to merge 1 commit into
giuliolovisotto wants to merge 1 commit into
Conversation
Signed-off-by: Giulio Lovisotto <glovisotto@nvidia.com>
|
Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually. Contributors can view more details about this message here. |
khazic
pushed a commit
to khazic/Gym-lao
that referenced
this pull request
Oct 15, 2026
## Summary - Add the SA-V tracking verifier and its server configuration. - Declare the SA-V environment servers used by video RL. - Add portable five-row example data, schema, metrics, and checked rollout artifacts. ## Why This is the Gym-side verifier foundation for mixed SA-V tracking and CapRL video training in NeMo-RL. No separate issue is needed; this is part of the Super 3.5 video RL upstreaming stack. ## Validation - Runtime evidence: the [healthy 16-node async `r18` W&B run](https://wandb.ai/adlr/Nemotron-omni-RL-debug/runs/oci-hsg-cs-001_n3post_main_1d76d243b_vllm029_super35_video_step120_async16n_trainmb45056_expandable_uncapped_r18_20261003) shows repeated GRPO cycles, stable TMPE, and increasing reward on the equivalent video implementation. - `gym dataset collate ... +mode=example_validation`: 5 examples; generated `data/example_metrics.json`. - Focused validation: 25 verifier tests and all 5 example rows against `task_data.py` passed. - Ruff format and lint passed for the changed Python files. - Both Gym stack ranges pass `git diff --check`. ## Rollouts `data/example_rollouts.jsonl` contains one checked rollout for each of the five example rows. Representative task checks: - `pt`: `<answer>{"tracks":[{"point":1,"frame":1,"xy":[500,500],"visible":true}]}</answer>` — `format_ok=true`, reward `1.0`, visible-frame mask membership `true`. - `box`: `<answer>{"tracks":[{"box":1,"frame":1,"bbox":[250,250,750,750],"visible":true}]}</answer>` — `format_ok=true`, reward `1.0`, frame IoU `1.0`. - `ref`: `<answer>{"tracks":[{"id":7,"frame":1,"xy":[500,500]}]}</answer>` — `format_ok=true`, reward `1.0`, predicted id `7` assigned to the GT object and mask membership `true`. The artifact also contains visible/invisible `pt` coverage and an all-absent `ref` negative case, preserving the verifier's current unmatched-ID behavior. ## Compatibility Adds a new verifier/configuration and does not change unrelated environments. Example media is embedded as a small inline JPEG, so no employee or cluster-local paths are required. ## Stack Gym PR 1 of 2. Consolidates and supersedes the implementation scope of NVIDIA-NeMo#3504 after review history is preserved. ## Consolidated stack links ### Gym 1. [Gym NVIDIA-NeMo#3999 — SA-V tracking verifier](NVIDIA-NeMo#3999) 2. [Gym NVIDIA-NeMo#4000 — R3 routing replay and rollout identity](NVIDIA-NeMo#4000) ### NeMo-RL 1. [NeMo-RL NVIDIA-NeMo#4472 — Super video data and model loading](NVIDIA-NeMo/RL#4472) 2. [NeMo-RL NVIDIA-NeMo#4473 — RL/MOPD video frame contract](NVIDIA-NeMo/RL#4473) 3. [NeMo-RL NVIDIA-NeMo#4474 — sync/async Gym routing runtime](NVIDIA-NeMo/RL#4474) 4. [NeMo-RL NVIDIA-NeMo#4475 — SA-V pin and CI fixtures](NVIDIA-NeMo/RL#4475) 5. [NeMo-RL NVIDIA-NeMo#4476 — final dependency pins](NVIDIA-NeMo/RL#4476) ### Child dependency refs - [MCore integration ref](https://github.com/ehosseiniasl/Megatron-LM/tree/ehsan/video-r3-mtp-gradient-integration-20261004), adapting [MCore #7278](NVIDIA/Megatron-LM#7278) and [MCore #7279](NVIDIA/Megatron-LM#7279) to the pinned base. - [MBridge integration ref](https://github.com/NVIDIA-NeMo/Megatron-Bridge/tree/ehsan/video-r3-mcore-pin-integration-20261004). --------- Signed-off-by: ehosseiniasl <ehsan.hosseiniasl@gmail.com> Co-authored-by: Pulkit Kumar <pulkitk@nvidia.com>
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
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.
SWE-bench Verified can mark passing patches unresolved when the test runner's output differs from the expected format. This fixes two reproduced cases:
-rAthrough tox so the grader receives named pass/fail results even whentox.inilacks the image-time reporting edit. Tox's own-vdoes not enable pytest reporting. Only the generated test command changes; the embedded test patch remains intact.astropy__astropy-7606, match the expectedtest_compose_roundtrip[]to the observedtest_compose_roundtrip[unit0]only when the original ID is absent. Preserve the observed status, raw log, and caller's test specification. Apply the same correction to live verification andrewrite_reports.Missing tests, failed tests, invalid logs and unrelated instances retain their existing grading behavior. No migration is required. Scores can increase for affected passing patches: archived Astropy output can be regraded; Sphinx output without named results requires a verifier replay. No pass is inferred from dot output or an aggregate pass count.
No separate issue is needed for these focused, reproduced verifier defects.
Validation:
gym env test --resources-server swebench +should_validate_data=true ++skip_venv_if_present=true: 45 passed; example-data validation passed.pre-commit run --all-filesandgit diff --cached --check: passed.