README.md
1.2 KB · 37 lines · markdown Raw
1 ---
2 license: apache-2.0
3 metrics:
4 - accuracy
5 - f1
6 base_model:
7 - google/vit-base-patch16-224-in21k
8 pipeline_tag: image-classification
9 library_name: transformers
10 datasets:
11 - nateraw/fairface
12 ---
13 Detects age group with about 59% accuracy based on an image.
14
15 See https://www.kaggle.com/code/dima806/age-group-image-classification-vit for details.
16
17 ![image/png](https://cdn-uploads.huggingface.co/production/uploads/6449300e3adf50d864095b90/gvzsgTtWDOE4vxwugZF4P.png)
18
19 ```
20 Classification report:
21
22 precision recall f1-score support
23
24 0-2 0.7803 0.7500 0.7649 180
25 3-9 0.7998 0.7998 0.7998 1249
26 10-19 0.5361 0.4236 0.4733 1086
27 20-29 0.6402 0.7221 0.6787 3026
28 30-39 0.4935 0.5083 0.5008 2099
29 40-49 0.4848 0.4386 0.4606 1238
30 50-59 0.5000 0.4814 0.4905 725
31 60-69 0.4497 0.4685 0.4589 286
32 more than 70 0.6897 0.1802 0.2857 111
33
34 accuracy 0.5892 10000
35 macro avg 0.5971 0.5303 0.5459 10000
36 weighted avg 0.5863 0.5892 0.5844 10000
37 ```