update

examples/clean_unet_aishell/run.sh

@@ -12,7 +12,7 @@ sh run.sh --stage 2 --stop_stage 2 --system_version centos --file_folder_name fi
 --noise_dir "/data/tianxing/HuggingDatasets/nx_noise/data/noise" \
 --speech_dir "/data/tianxing/HuggingDatasets/aishell/data_aishell/wav/train"
 
-sh run.sh --stage 1 --stop_stage 2 --system_version centos --file_folder_name file_dir \
+sh run.sh --stage 2 --stop_stage 2 --system_version centos --file_folder_name file_dir \
 --noise_dir "/data/tianxing/HuggingDatasets/nx_noise/data/noise" \
 --speech_dir "/data/tianxing/HuggingDatasets/aishell/data_aishell/wav/train" \
 --max_count 10000

examples/clean_unet_aishell/step_2_train_model.py

@@ -225,7 +225,11 @@ def main():
             clean_audios = clean_audios.to(device)
             noisy_audios = noisy_audios.to(device)
 
+            print(f"clean_audios shape: {clean_audios.shape}, dtype: {clean_audios.dtype}")
+            print(f"noisy_audios shape: {noisy_audios.shape}, dtype: {noisy_audios.dtype}")
+
             enhanced_audios = model.forward(noisy_audios)
+            print(f"enhanced_audios shape: {enhanced_audios.shape}, dtype: {enhanced_audios.dtype}")
 
             ae_loss = ae_loss_fn(enhanced_audios, clean_audios)
             sc_loss, mag_loss = mr_stft_loss_fn(enhanced_audios.squeeze(1), clean_audios.squeeze(1))