Update app.py

app.py

@@ -154,9 +154,10 @@ 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)
 
-    
-    pred = sample_diffusion(mixture, timbre, num_infer_steps, eta, seed, guidance_scale, guidance_rescale)
-    return sample_rate, pred.squeeze().cpu().numpy()
+    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()
 
 
 # CSS styling (optional)