dima806's picture
Update README.md
50452c3 verified
|
raw
history blame
1.68 kB
metadata
license: apache-2.0
metrics:
  - accuracy
  - f1

Returns age bin based on a facial image.

See https://www.kaggle.com/code/dima806/facial-age-years-detection-vit for more details.

Classification report:

              precision    recall  f1-score   support

          01     0.6771    0.9102    0.7765       334
          02     0.4271    0.6156    0.5043       333
          03     0.5965    0.2036    0.3036       334
          04     0.7500    0.0988    0.1746       334
          05     0.5301    0.9491    0.6803       334
       06-07     0.7292    0.6048    0.6612       334
       08-09     0.5958    0.6796    0.6350       334
       10-12     0.6401    0.6497    0.6449       334
       13-15     0.7386    0.7297    0.7341       333
       16-20     0.6047    0.7027    0.6500       333
       21-25     0.6161    0.5976    0.6067       333
       26-30     0.4328    0.5210    0.4728       334
       31-35     0.4575    0.2102    0.2881       333
       36-40     0.4128    0.2904    0.3409       334
       41-45     0.4636    0.6677    0.5472       334
       46-50     0.4464    0.5375    0.4877       333
       51-55     0.4620    0.4731    0.4675       334
       56-60     0.4573    0.3204    0.3768       334
       61-65     0.6260    0.2305    0.3370       334
       66-70     0.4526    0.7006    0.5499       334
       71-80     0.7108    0.7087    0.7098       333
       81-90     0.8354    0.9910    0.9066       333
         90+     0.9881    1.0000    0.9940       333

    accuracy                         0.5822      7673
   macro avg     0.5935    0.5823    0.5587      7673
weighted avg     0.5934    0.5822    0.5586      7673