| --- |
| model_summary: null |
| license: mit |
| datasets: |
| - detection-datasets/fashionpedia |
| language: |
| - en |
| base_model: |
| - Ultralytics/YOLOv8 |
| tags: |
| - clothes |
| - second |
| - hand |
| - fashion |
| - repair |
| - tshirt |
| - fashionpedia |
| - pants |
| - socks |
| - shoes |
| - dress |
| --- |
| # Yolo 8 trained on Fashionpedia dataset |
| This was originally developed for a second hand startup vetement.ai that generated product sheets from pictures. |
| LLVM such as Mistral or ChatGPT are now better and more cost effective. |
|  |
|
|
|
|
| ## Model Details |
| YOLO8 with size S and N finetunned. It was a head traning then a full training for each size. |
|
|
| ### Model Description |
|
|
| ## Uses |
| Ultralytics YOLO8 |
|
|
| ## Training Details |
|
|
| ### Training Data |
| It was trained on Fashionpedia dataset |
|
|
| ### Training Procedure |
| It was trained ~10 hours on a 4070 Ti SUPER |
|
|
| #### Preprocessing |
| No modification, just format conversion. |
|
|
| #### Training Hyperparameters |
| Default |
|
|
| ## Contact |
| ai@vetement.ai |
|
|