Repository navigation
Add twelve published CLIP-ReID model configurations - #33
Conversation
Preserve existing alias and update semantics across 334 configs and 595 variants. Add ordered resolution and CPU preprocessing regression tests without changing shipped model definitions. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Verify exact runtime board paths using metadata only, preserve local-only restrictions and distinguish live commits from approval provenance. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
…tocol scoring Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Keep existing marker-free configs unchanged and preserve Hub MXQ loading. Skip only marked ONNX in metadata verification, with positive and negative CPU guard tests. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Exclude configuration-only marker from MXQ runtime kwargs, guard compilation downloads and cover local/unmarked positive paths. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Use official timm defaults and no-network pending status for new sibling configurations only. Preserve all pre-existing YAML bytes and allow trusted local artifacts without inventing HF publication metadata. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
…figs Merge PR31 aaf0c14 into the approved PR32 lineage; guide and README additions auto-merge preserving both utility and artifact policies. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
Security findingsAdvisory findings (1)ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 6bc2db8f41
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
🛡️ Codex Security Review · Automatically triggered
Here are some automated security review suggestions for this pull request.
Reviewed commit: 6bc2db8f41
ℹ️ About Codex security reviews in GitHub
This is an experimental Codex feature. Security reviews are triggered when:
- You comment "@codex security review"
- A regular code review gets triggered (for example, "@codex review" or when a PR is opened), and you’re opted in so security review runs alongside code review
Once complete, Codex will leave suggestions, or a comment if no findings are found.
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Retain pending and ONNX-only safety behavior plus new configured-board precedence from main. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Preserve current main retrieval validation fixes and apply only original face verification changes. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Retain upstream Hub, retrieval rank validation, negative identity and symlink protections; preserve classifier markers, face utility dispatch and wrapper-only published exports. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
… precedence Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
|
@codex review Review exact integrated head Fixed/resolved code threads 4465211325 and 4465211336: no unsupported compile promise, explicit unsupported regression, and only model wrappers become top-level lazy exports. Security thread4465226206 addressed under project policy: this utility intentionally tracks main; strict revision/digest pinning is not required; no live tip was invented as historical approval provenance. Please inspect integration regressions, model-only exports, maintained artifact guards and source preprocessing. Full CPU suite: 1601 passed/19 deselected including 595 baseline identity, d1 marker/classifier/runtime/compile tests, c1/c2 symlink and protocol tests, ReID12 byte parity(atol0). Depends on #31/#32; no PR merge performed. |
|
Codex Review: Didn't find any major issues. Already looking forward to the next diff. Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Retain main version 0.0.7 and tiled OBB behavior; apply exact face-specific patch without guards duplication. Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
…ir behavior Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
🛡️ Codex Security Review · Automatically triggeredSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
|
@codex review Review exact head Previous integrated head Please check latest dependency integration and unchanged ReID12 model/YAML behavior. Full CPU suite from the intended worktree: 1688 passed, 15 skipped, 19 deselected, including 595 baseline variants and classifier marker, face/ReID protocol, symlink and concrete preprocessing tests. Merge is coordinated after #32; no PR merge performed here. |
|
@codex review Re-request after transient GitHub push error/retry: branch refs now point to exact head |
|
Codex Review: Didn't find any major issues. 👍 Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
🛡️ Codex Security Review · Automatically triggeredSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
Resolve face-verification conflicts to the squash-merged main versions and retain the twelve published CLIP-ReID configurations and documentation. Co-Authored-By: Claude Code <noreply@anthropic.com>
|
@codex review Please review exact head This is a main-sync merge after #32 was squash-merged as Scope: twelve published CLIP-ReID configs, person_reid processing base and model-only exports/discovery, metadata/parity tests, and related guides (20 files). Check RN50/ViT preprocessing, dataset and embedding dimensions, MXQ-only ONNX guard, registry behavior, and that no #32 or unrelated behavior regresses. Keep Validation: full offline CPU pytest 1695 passed, 15 skipped, 19 deselected; targeted regression subset 676 passed; pinned full-tree Ruff lint and both changed-file pre-commit hooks pass. Main has nine unrelated baseline formatting deviations, intentionally left outside this PR. |
|
Codex Review: Didn't find any major issues. Another round soon, please! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
🛡️ Codex Security Review · Automatically triggeredSecurity review completed. No security issues were found in this pull request. Reviewed commit: Only the user who started this review can view the report in Codex. ℹ️ About Codex security reviews in GitHubThis is an experimental Codex feature. Security reviews are triggered when:
Once complete, Codex will leave suggestions, or a comment if no findings are found. |
Summary
Stage d2, now based directly on
main. Dependencies #31 and #32 have landed. Main-sync merge45fac2076191f3abb94228bdd83e64bff6d49dffintegrates #32's squash7c8ce909974f35fcd23e4c278976a62c74190351without force-pushing. Face-verification conflicts use main's versions; the ReID-only binary diff is byte-identical to the previously reviewed #33 scope (87ff416..c765f8a). The final diff contains only twelve model YAMLs, the person_reid base/exports, registry metadata tests, and related documentation (20 files).Source and parity
Ops
ec537764e36ca10da963fcb7afab157ea60ca022: every model distribution.yaml and _source/predict_float.input.json were directly compared. All use single-input image-embedding.v1. RN50: 3072D, 256x128, torch stats. ViT: 1280D, 256x128, 0.5 stats. Existing HWC/no-op SetOrder semantics are preserved through PIL resize then explicit float32 normalization.Each of 12 concrete pipelines has synthetic RGB byte parity (atol=0). Postprocessing validates dimensions and unit norms; no accuracy claims or inference run.
Validation
CPU-only full offline suite on exact head
45fac2076191f3abb94228bdd83e64bff6d49dff: 1695 passed, 15 skipped, 19 deselected. Targeted ReID, face-verification, manifest and inheritance tests: 676 passed. Repository-pinned Ruff v0.6.9 lint passes across all files; both pre-commit hooks pass on all 20 PR-changed files;git diff --checkpasses. Previous wheel validation confirmed all twelve YAMLs and no development assets/tests.The main baseline has formatting drift in nine unrelated files; those files are intentionally unchanged by #33. A standalone newer system Ruff also picks up external parent configuration, so the passing lint gate is the repository's pinned pre-commit Ruff hook.
No weights/dataset downloads, HF writes, GPU/NPU runs, --execute, OBB/DOTA, CLI routing, version changes, releases or tags.
revision: mainremains unchanged. Awaiting fresh exact-head Codex review before squash merge.🤖 Generated with Claude Code