Clarify LEAPP export backend selection - #7494
Conversation
There was a problem hiding this comment.
Isaac Lab Review Bot
Clarifies LEAPP export-backend selection in the user documentation and argparse help, while adding the existing pt2 option to the documented backend list. The reviewed changes do not alter export behavior.
- Design and architecture: Backend selection remains centralized in
add_common_export_args; no module boundaries, ownership, or export pipeline responsibilities change. The documentation now aligns its backend list with the parser choices. - API: The
--export_methodname, parser default, accepted choices, and forwarding behavior are unchanged. Only its help text changes, and the documentation now includes the already-supportedpt2value. - Implementation: The documented choices were checked against the unchanged argparse
choiceslist and now match. The patch only changes documentation and CLI-facing explanatory text, with no package source change requiring a changelog fragment. Backend API descriptions may need future maintenance as PyTorch export APIs evolve.
No blocking issues. No inline issue met the actionable-evidence threshold; the assessment above records the review feedback.
Automated review; human maintainers own approval decisions.
Greptile SummaryClarifies how users should select a LEAPP policy-export backend and documents the relationship between backend choice and artifact format.
Confidence Score: 5/5The PR appears safe to merge because it only clarifies existing export-backend behavior and no actionable defect was identified. The documentation and CLI help consistently recommend the existing default, explain backend-dependent compatibility, and describe the available artifact formats without changing export execution. Important Files Changed
Reviews (1): Last reviewed commit: "Clarify LEAPP export backend selection" | Re-trigger Greptile |
|
run-ci |
Signed-off-by: Frank Lai NV <frlai@nvidia.com>
|
run-ci |
Backports #7494 to `release/3.0.0`. The original cherry-pick conflicted. An NVIDIA inference model proposed this resolution, and deterministic validation confirmed that it changes no paths outside the original PR. Because conflict resolution cannot be certified as an exact patch replay, this PR is intentionally a draft and requires release-maintainer review. | Field | Commit | |---|---| | Original merged change | `4440f7a67d7483cc6d3ffd87b7e6143e3d64794a` | | Release base used | `36794478ba81f4442a044983ad6e93821f5c01bf` | | Proposed backport | `93fd5eed7c43f54c87b61ab864f46e66d9b8406f` | Co-authored-by: Frank Lai NV <frlai@nvidia.com>
# Description Backports the following merged `develop` PRs to `release/3.0.0`: - #7220 — prune unused Docker volumes on package-test runners - #7169 — cache the resolved OVRTX Warp device - #6889 — move the OVPhysX collider cache out of the Python interpreter directory - #7345 — document the registered XR camera PiP tasks - #7050 — update and reorganize the LEAPP documentation - #7473 — register external RSL-RL tasks before agent discovery - #7467 — update the China storage profile to Isaac Sim 6.1 - #7385 — accept legal float scale values when seeding Fabric frame views from USD - #7517 — use lazy exports in `isaaclab_rl` The duplicate #7517 in the original request is included once. Seven source commits replayed cleanly. Two required release-specific reconciliation: - #7169 retains the release branch's mapped OVRTX write path, which was removed separately on `develop`, while resolving and caching one Warp device for render-product IDs, mappings, and CUDA streams. - #7050 preserves the #7494 export-backend clarification that is already present on the release branch while applying the broader LEAPP documentation overhaul. No new dependencies are added. ## Type of change - Bug fix (non-breaking change which fixes an issue) - Performance improvement - Documentation update - CI/infrastructure update ## Release backport - [ ] <!-- backport-active-release --> This PR already targets the active release branch; do not backport it again. ## Validation - `uv run --no-project --with pytest python -m pytest -q .github/actions/run-package-tests/test_cleanup_docker_storage.py` — 10 passed - Nine isolated `isaaclab_rl` namespace/lazy-import regression cases passed - Resolved-device OVRTX mapping behavior check passed with Warp 1.16.0 - `uv run --no-project python -m compileall -q` on all changed Python modules and tests - `uv run --no-project python tools/changelog/cli.py check` against the exact `upstream/release/3.0.0` base - `pre-commit==4.6.2 run --all-files` passed all available hooks - `git diff --check upstream/release/3.0.0..HEAD` The full project test environment and documentation build cannot resolve on this macOS/arm64 host because the release lockfile supports Linux and Windows only. The Git LFS hook was also unavailable because `git-lfs` is not installed; none of the changed paths are LFS-tracked. Linux CI remains authoritative for simulator, OVRTX, Fabric, and full documentation validation. ## Checklist - [x] I have read and understood the contribution guidelines - [x] I have run the available formatting and static checks - [x] I have included the corresponding documentation changes - [x] Each touched source package includes its original changelog fragment - [x] Every backported commit records its source commit with `cherry picked from commit` - [x] The original contributors are preserved as commit authors --------- Signed-off-by: Zeng Qingcheng <60593302+NeoZng@users.noreply.github.com> Co-authored-by: myurasov-nv <168484206+myurasov-nv@users.noreply.github.com> Co-authored-by: Peter Verswyvelen <pverswyvelen@nvidia.com> Co-authored-by: pv-nvidia <197907000+pv-nvidia@users.noreply.github.com> Co-authored-by: hujc <jichuanh@nvidia.com> Co-authored-by: hougantc-nvda <127865892+hougantc-nvda@users.noreply.github.com> Co-authored-by: Frank Lai NV <frlai@nvidia.com> Co-authored-by: peterd-NV <peterd@nvidia.com> Co-authored-by: Sheikh Dawood <7774242+sheikh-nv@users.noreply.github.com> Co-authored-by: Zeng Qingcheng <60593302+NeoZng@users.noreply.github.com> Co-authored-by: Antoine RICHARD <antoiner@nvidia.com> Co-authored-by: Mustafa H <34825877+StafaH@users.noreply.github.com>
Description
Previously it was unclear that there is no guarantee export is supported with all possible backends. This doc change aims to make that more clear.
Type of change
Release backport
develop