DongDem / variant_loss_function_vipl

Variant_loss_function_vipl

variant_loss_function_vipl

Variant_loss_function_vipl

Tensorflow version (old, maybe conflict)

• Comparison softmax loss, center loss and proposed loss in alex_net model
• The difference in the function get_center_loss

Training

• Prepare the image path and label path (train set)
• Run file MMI_with_proposed_loss_alex_net.py

Testing

• Get the accuracy and confusion matrix in the test set
• Run file test_MMI_with_proposed_loss_alex_net.py

Pytorch version (Recommend)

• The difference in the function VariantCenterLoss in variantcenterloss.py
• Only in mnist dataset (example)

Training

• Comment test phase and run file pytorch_test.py

Testing

• Comment train phase and run file pytorch_test.py