Returns the clothes category with about 78% accuracy based on an image.
See https://www.kaggle.com/code/dima806/clothes-image-detection-vit for details.
Classification report:
precision recall f1-score support
Blazer 0.7419 0.6900 0.7150 200
Coat 0.7512 0.7550 0.7531 200
Denim Jacket 0.8592 0.9150 0.8862 200
Dresses 0.8603 0.7700 0.8127 200
Hoodie 0.6985 0.9500 0.8051 200
Jacket 0.7686 0.4650 0.5794 200
Jeans 0.8657 0.8700 0.8678 200
Long Pants 0.8112 0.7950 0.8030 200
Polo 0.7929 0.5550 0.6529 200
Shirt 0.7430 0.7950 0.7681 200
Shorts 0.9149 0.8600 0.8866 200
Skirt 0.8102 0.8750 0.8413 200
Sports Jacket 0.6562 0.7350 0.6934 200
Sweater 0.7758 0.8650 0.8180 200
T-shirt 0.7743 0.8750 0.8216 200
accuracy 0.7847 3000
macro avg 0.7883 0.7847 0.7803 3000
weighted avg 0.7883 0.7847 0.7803 3000
- Downloads last month
- 261
Model tree for dima806/clothes_image_detection
Base model
google/vit-base-patch16-224-in21k