Pre-trained Models for AudioFuse

This release contains the pre-trained model weights for the experiments in our paper.

• best_model_AudioFuse_v1.0.0.keras: The main AudioFuse fusion model trained on PhysioNet.
• best_model_spectrogram_physionet.keras: The spectrogram-only ViT baseline trained on PhysioNet.
• best_model_waveform_baseline.keras: The waveform-only 1D-CNN baseline trained on PhysioNet.