| --- |
| title: README |
| emoji: 🧵 |
| colorFrom: gray |
| colorTo: blue |
| sdk: static |
| pinned: false |
| --- |
| |
| <p> |
| <strong>Resoa</strong> is a small Santa Barbara studio. This org holds models for |
| <strong>apparel manufacturing QC</strong> — construction attributes and garment |
| parts for a tech pack — not ecommerce tagging. |
| </p> |
|
|
| <p> |
| Weights are Apache 2.0. Training data is |
| <a href="https://huggingface.co/datasets/huggan/fashionpedia">Fashionpedia</a> |
| (CC BY 4.0). They were trained on street photographs of worn garments. |
| Factory flat-lays and on-hanger photos are a different domain and are not |
| measured here. |
| </p> |
|
|
| <p>Models in this org:</p> |
| <ul> |
| <li> |
| <a href="https://huggingface.co/resoa/garment-detector-seg">garment-detector-seg</a> |
| — instance segmentation of garments, accessories, and parts (Fashionpedia, 46 classes). |
| </li> |
| <li> |
| <a href="https://huggingface.co/resoa/garment-attributes">garment-attributes</a> |
| — multi-label construction attributes from a <em>single-garment crop</em> (SigLIP2, 93.1M). |
| Full-scene inputs degrade the scores sharply; that is a crop problem, not a model-card footnote. |
| </li> |
| <li> |
| <a href="https://huggingface.co/resoa/garment-attributes-v2">garment-attributes-v2</a> |
| — same architecture and label space as v1, retrained with crop-scale augmentation. |
| </li> |
| <li> |
| <a href="https://huggingface.co/resoa/garment-crop-gate-nano">garment-crop-gate-nano</a> |
| — 47K pre-flight: is this crop tight enough to trust the attributes model? |
| A full scene keeps about 14% of trained-crop mAP; the gate refuses those inputs. |
| </li> |
| </ul> |
|
|
| <p> |
| Collection: |
| <a href="https://huggingface.co/collections/resoa/apparel-qc-6a8b69d88b6595bfb1851782">Apparel QC</a>. |
| There is also a dataset, |
| <a href="https://huggingface.co/datasets/resoa/triclock">triclock</a>, |
| which is a LogLens memory-schedule benchmark — not apparel. |
| </p> |
|
|
| <p> |
| Same studio, two namespaces: this org is the product models. |
| Tiny edge models (camera, audio, forensics) live on |
| <a href="https://huggingface.co/resoajoe">resoajoe</a> — same person, |
| kept there so research nanos stay easy to browse separately. |
| </p> |
|
|
