Predicts face emotion based on facial image.

See https://www.kaggle.com/code/dima806/face-emotions-image-detection-vit for more details.

Classification report:

              precision    recall  f1-score   support

      Ahegao     0.9738    0.9919    0.9828      1611
       Angry     0.8439    0.6580    0.7394      1611
       Happy     0.8939    0.9261    0.9098      1611
     Neutral     0.6056    0.7635    0.6755      1611
         Sad     0.6661    0.5140    0.5802      1611
    Surprise     0.7704    0.8733    0.8186      1610

    accuracy                         0.7878      9665
   macro avg     0.7923    0.7878    0.7844      9665
weighted avg     0.7923    0.7878    0.7844      9665

Downloads last month: 83

Safetensors

Model size

85.8M params

Tensor type

F32

Model tree for dima806/face_emotions_image_detection

Base model

google/vit-base-patch16-224-in21k

Finetuned

(2526)

this model