preproccessing the text

app.py

@@ -20,6 +20,22 @@ vocoder = SpeechT5HifiGan.from_pretrained("microsoft/speecht5_hifigan").to(devic
 embeddings_dataset = load_dataset("Matthijs/cmu-arctic-xvectors", split="validation")
 speaker_embeddings = torch.tensor(embeddings_dataset[7306]["xvector"]).unsqueeze(0)
 
+replacements = [
+    ("à", "a"),
+    ("ç", "c"),
+    ("è", "e"),
+    ("ë", "e"),
+    ("í", "i"),
+    ("ï", "i"),
+    ("ö", "o"),
+    ("ü", "u"),
+]
+
+def cleanup_text(text):
+    for src, dst in replacements:
+        text = text.replace(src, dst)
+    return text
+
 
 def translate(audio):
     outputs = asr_pipe(audio, max_new_tokens=256, generate_kwargs={"task": "transcribe", "language": "nl"})
@@ -27,6 +43,7 @@ def translate(audio):
 
 
 def synthesise(text):
+    text = cleanup_text(text)
     inputs = processor(text=text, return_tensors="pt")
     speech = model.generate_speech(inputs["input_ids"].to(device), speaker_embeddings.to(device), vocoder=vocoder)
     return speech.cpu()