Update app.py

app.py

@@ -26,6 +26,14 @@ def translate(audio):
     return outputs["text"]
 
 
+# Added to Base to translate from Language X to any Language Y using "task": "transcribe"
+# At Inference. it should use translate(sample["audio"].copy())
+
+def translate(audio):
+    outputs = pipe(audio, max_new_tokens=256, generate_kwargs={"task": "transcribe", "language": "es"})
+    return outputs["text"]
+
+
 def synthesise(text):
     inputs = processor(text=text, return_tensors="pt")
     speech = model.generate_speech(inputs["input_ids"].to(device), speaker_embeddings.to(device), vocoder=vocoder)