vit-base-15classes
This model is a vit-base-15classes trained on the NIH Chest X-ray dataset for multi-label classification of 14 pathologies.
Metrics
- auc_macro: 0.7464
- auc_micro: 0.8365
- mAP: 0.2165
Training Configuration
- learning_rate: 0.0001
- batch_size: 16
- num_epochs: 50
Usage
import torch
# Load model logic here
- Downloads last month
- 19