Update app.py

app.py

@@ -66,7 +66,8 @@ def transcribe(audio: tuple[int, np.ndarray], transformers_convo: list[dict], gr
                             sr=processor.feature_extractor.sampling_rate)[0]
                         )
         inputs = processor(text=text, audios=audios, return_tensors="pt", padding=True)
-        inputs.input_ids = inputs.input_ids.to("cuda:0")
+        inputs = dict(**inputs)
+        inputs["input_ids"] = inputs["input_ids"].to("cuda:0")
 
         generate_ids = model.generate(**inputs, max_length=256)
         generate_ids = generate_ids[:, inputs.input_ids.size(1):]