mule-image

dtoro/mule-image

Fork 0

Commit Graph

Author	SHA1	Message	Date
dtoro	fbeefb24a0	fix: vision tasks inherit user_id, admin owns mount root - detect_objects, classify_content, recluster_faces now look up the photo's user_id and set it on created Tag rows — fixes tags being invisible to the owning user due to NULL user_id - Initial admin setup creates source root at the mount root (/photos) instead of a subdirectory, since the admin owns the entire library - Revert to OpenCLIP ViT-B/32 (512-d) as default embedder — SigLIP requires transformers version alignment not yet available in the Docker image. SigLIP2 code remains for future enablement. - Add transformers to requirements for future SigLIP support Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-13 00:01:28 +02:00
dtoro	94c07b1d0d	feat: upgrade to SigLIP2 ViT-B/16 for semantic search Replace OpenCLIP ViT-B/32 (512-d, ~78% recall) with SigLIP2 ViT-B/16 (768-d, ~84% recall) as the default embedding model for significantly better image-text retrieval quality. - New SigLIP2Embedder class with 384px input and SigLIP normalization - ONNX export pipeline for SigLIP2 visual + textual encoders - Migration 0010: resize embeddings.vector from 512 to 768 dimensions - Config-driven model selection: "siglip2_vitb16" (default) or "openclip_vitb32" (legacy) — both models can coexist - Content classifier follows the configured embedder family - Existing embeddings cleared on migration; vision backfill regenerates Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-12 22:09:16 +02:00
dtoro	fa9b21856f	feat: replace face pipeline with InsightFace, add content classifier Face detection/recognition: - Replace YuNet + SFace with InsightFace buffalo_l (RetinaFace + ArcFace) - 512-d ArcFace embeddings (was 128-d SFace), migration 0006 resizes column - Remove YOLO person-bbox workaround — RetinaFace is accurate enough - Detection threshold 0.65 cleanly separates real faces (0.72+) from false positives on dogs/paintings (0.56-0.61) Content-type classification: - CLIP zero-shot classifier using native PyTorch text encoder + ONNX image encoder for high-quality text-image similarity - Categories: photograph, screenshot, document, receipt, meme, artwork - Writes Tag(kind=content_type) per photo via photo_tags - Margin-based confidence: top-1 vs top-2 score difference - New ClassifierSettings in config (enabled, min_confidence) - Wired into vision_fanout pipeline Tested: 6 real faces from 4 photos (zero false positives), 11/13 photos classified (8 photograph, 2 artwork, 1 meme). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>	2026-04-10 13:49:02 +02:00

Author

SHA1

Message

Date

dtoro

fbeefb24a0

fix: vision tasks inherit user_id, admin owns mount root

- detect_objects, classify_content, recluster_faces now look up the
  photo's user_id and set it on created Tag rows — fixes tags being
  invisible to the owning user due to NULL user_id
- Initial admin setup creates source root at the mount root (/photos)
  instead of a subdirectory, since the admin owns the entire library
- Revert to OpenCLIP ViT-B/32 (512-d) as default embedder — SigLIP
  requires transformers version alignment not yet available in the
  Docker image. SigLIP2 code remains for future enablement.
- Add transformers to requirements for future SigLIP support

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-13 00:01:28 +02:00

dtoro

94c07b1d0d

feat: upgrade to SigLIP2 ViT-B/16 for semantic search

Replace OpenCLIP ViT-B/32 (512-d, ~78% recall) with SigLIP2 ViT-B/16
(768-d, ~84% recall) as the default embedding model for significantly
better image-text retrieval quality.

- New SigLIP2Embedder class with 384px input and SigLIP normalization
- ONNX export pipeline for SigLIP2 visual + textual encoders
- Migration 0010: resize embeddings.vector from 512 to 768 dimensions
- Config-driven model selection: "siglip2_vitb16" (default) or
  "openclip_vitb32" (legacy) — both models can coexist
- Content classifier follows the configured embedder family
- Existing embeddings cleared on migration; vision backfill regenerates

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-12 22:09:16 +02:00

dtoro

fa9b21856f

feat: replace face pipeline with InsightFace, add content classifier

Face detection/recognition:
- Replace YuNet + SFace with InsightFace buffalo_l (RetinaFace + ArcFace)
- 512-d ArcFace embeddings (was 128-d SFace), migration 0006 resizes column
- Remove YOLO person-bbox workaround — RetinaFace is accurate enough
- Detection threshold 0.65 cleanly separates real faces (0.72+) from
  false positives on dogs/paintings (0.56-0.61)

Content-type classification:
- CLIP zero-shot classifier using native PyTorch text encoder + ONNX
  image encoder for high-quality text-image similarity
- Categories: photograph, screenshot, document, receipt, meme, artwork
- Writes Tag(kind=content_type) per photo via photo_tags
- Margin-based confidence: top-1 vs top-2 score difference
- New ClassifierSettings in config (enabled, min_confidence)
- Wired into vision_fanout pipeline

Tested: 6 real faces from 4 photos (zero false positives), 11/13 photos
classified (8 photograph, 2 artwork, 1 meme).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-10 13:49:02 +02:00

3 Commits