Manifest v13: Generative removal / Generative Remove (LaMa)

Manifest v12: add LaMa inpainting (issue #67)

Add LaMa inpainting (Apache-2.0 code, CC-BY weights via Places) — issue #67

Manifest v11: add SigLIP text tower (issue #66)

Add SigLIP text tower + tokenizer (Apache-2.0) — issue #66 typed search

Manifest v10: add SigLIP image tower (issue #66)

Add SigLIP image tower (Apache-2.0) — issue #66 semantic search

Manifest v9: add NAFNet denoise (issue #65)

Add NAFNet-SIDD width32 denoise (MIT) — issue #65 Phase A

Manifest v8: add SPAN x2 (issue #61 Phase 1e)

Add SPAN x2 (Apache-2.0) — the faithful upscaler for large sources (issue #61 Phase 1e)

manifest v7: add upscale-realesrgan-x2plus (issue #61)

Add Real-ESRGAN x2plus Core ML (fp16, 256px tile) for issue #61

manifest v6: auraface-v1 live (rev b8414800)

AuraFace-v1 face embedder (ArcFace R100) Core ML fp16 — fal.ai, Apache-2.0

manifest v5: depth-anything-v2-small live (rev 28d40084)

Depth Anything V2 Small (Apple Core ML F16, Apache-2.0)

manifest v4: face-parsing live (rev 3f008e26)

Face parsing (BiSeNet CelebAMask-HQ 19 classes) Core ML, MIT — zllrunning upstream

manifest v3: sam2-tiny live (rev 61e99b0a)

SAM 2 Tiny Core ML (Apple conversion, Apache-2.0): image+prompt encoders, mask decoder

manifest v2: sky-u2net live (pinned rev 81481be3, sha256-verified)

README: models, licenses, attribution

Sky U2-Net 512 Core ML fp16 (xiongzhu666 weights, MIT)

initial commit