Update custom_interface_app.py

custom_interface_app.py

@@ -132,10 +132,14 @@ class ASR(Pretrained):
 
             # Add the last segment
             segments.append([current_start, current_end])
-            
-            outputs = []
+
             # Process each segment
+            outputs = []
             for i, segment in enumerate(segments):
+                start, end = segment
+                start = int(start * sr)
+                end = int(end * sr)
+                segment = waveform[start:end]
                 print(f"Processing segment {i + 1}/{len(segments)}, length: {len(segment) / sr:.2f} seconds")
 
                 # import soundfile as sf
@@ -159,4 +163,4 @@ class ASR(Pretrained):
             rel_length = torch.tensor([1.0]).to(device)
             # outputs.append(self.encode_batch_whisper(device, batch, rel_length))
             outputs = self.encode_batch_whisper(device, batch, rel_length)
-            yield outputs
+            yield outputs