ResNet18 - HAM10000 (bkl, mel, nv)

  • Dataset: Kaggle kmader/skin-cancer-mnist-ham10000, 1000 ảnh/lớp, split 80/10/10 theo lesion_id
  • Pretrained ImageNet: True | img_size: 256 | input: RGB trong [0,1] (chuẩn hoá ImageNet nằm trong model)
  • Classes (thứ tự logit): ['bkl', 'mel', 'nv']
  • Test: acc=0.7766, bal_acc=0.7757, macro-F1=0.7761, macro-AUC=0.9280
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support