README.md

---
license: apache-2.0
metrics:
- accuracy
- f1
base_model:
- google/vit-base-patch16-224-in21k
pipeline_tag: image-classification
---
Returns the hair type (with 93% accuracy) based on facial image.

See https://www.kaggle.com/code/dima806/hair-type-image-detection-vit for details.

![image/png](https://cdn-uploads.huggingface.co/production/uploads/6449300e3adf50d864095b90/F0mBCM82WYy4ApWA5ESFm.png)

```
Classification report:

              precision    recall  f1-score   support

       curly     0.9109    0.8932    0.9020       206
  dreadlocks     0.9667    0.9902    0.9783       205
       kinky     0.9423    0.9561    0.9492       205
    straight     0.9585    0.8981    0.9273       206
        wavy     0.8651    0.9029    0.8836       206

    accuracy                         0.9280      1028
   macro avg     0.9287    0.9281    0.9281      1028
weighted avg     0.9287    0.9280    0.9280      1028
```