Update app.py

app.py

@@ -142,7 +142,7 @@ def tse(gt_file_input, text_input, num_infer_steps, eta, seed, guidance_scale, g
             padding = target_length - current_length
             mixture = np.pad(mixture, (0, padding), mode='constant')
         mixture = torch.tensor(mixture).unsqueeze(0).to(device)
-        mixture = autoencoder(audio=mixture.unsqueeze(1))
+        # mixture = autoencoder(audio=mixture.unsqueeze(1))
 
         text_inputs = processor(
             text=[text_input],
@@ -158,7 +158,7 @@ def tse(gt_file_input, text_input, num_infer_steps, eta, seed, guidance_scale, g
         inputs = {key: value.to(device) for key, value in inputs.items()}
         timbre = clapmodel.get_text_features(**inputs)
 
-    mixture = autoencoder(embedding=mixture).squeeze(1)
+    # mixture = autoencoder(embedding=mixture).squeeze(1)
     # pred = sample_diffusion(mixture, timbre, num_infer_steps, eta, seed, guidance_scale, guidance_rescale)
     # return sample_rate, pred.squeeze().cpu().numpy()
     return sample_rate, mixture.squeeze().cpu().numpy()