Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- coreml
|
| 5 |
+
- image-segmentation
|
| 6 |
+
- cloth-segmentation
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
# Cloth Segmentation — Core ML (U²-Net)
|
| 10 |
+
|
| 11 |
+
On-device clothing segmentation for iOS, converted to Core ML. Segments a person's
|
| 12 |
+
clothing into **upper-body**, **lower-body**, and **full-body (dress)** + background.
|
| 13 |
+
|
| 14 |
+
- **Input:** RGB image, 768×768 (Core ML `ImageType`; normalization baked in).
|
| 15 |
+
- **Output:** `probs` — softmax probabilities, shape `(1, 4, 768, 768)`, channels = `[background, upper, lower, full]`.
|
| 16 |
+
- **Format:** FP16 mlprogram, ~84 MB. `ClothSegmentation.mlpackage.zip` (download + unzip + compile on device).
|
| 17 |
+
|
| 18 |
+
## Provenance & License
|
| 19 |
+
- Architecture: **U²-Net** (Xuebin Qin et al.) — Apache-2.0.
|
| 20 |
+
- Model/code: **levindabhi/cloth-segmentation** — MIT.
|
| 21 |
+
- Weights trained on **iMaterialist (Fashion) 2019 @ FGVC6** (competition dataset).
|
| 22 |
+
- Checkpoint mirror: `maiti/cloth-segmentation` (`cloth_segm_u2net_latest.pth`).
|
| 23 |
+
- Converted with the included `convert.py` (PyTorch 2.2 → coremltools 8.3).
|
| 24 |
+
|
| 25 |
+
Used by Chromaform (chromaform.art) for its Clothing Recolor tool.
|