yes

app.py

@@ -9,28 +9,29 @@ tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
 model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME)
 model.eval()
 
-def text_to_ipa(text: str) -> str:
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, T5ForConditionalGeneration
+
+MODEL_NAME = "google/byt5-small"
+
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME)
+model.eval()
+
+def text_to_ipa(text):
     prompt = f"Text: {text}\nIPA:"
     inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512)
     with torch.no_grad():
         outputs = model.generate(**inputs, max_new_tokens=64)
     return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
-
-app = FastAPI()
-
-@app.post("/phonetic")
-async def phonetic(payload: dict):
-    text = payload.get("text", "")
-    ipa = text_to_ipa(text)
-    return {"ipa": ipa}
-
-# Create Gradio UI
 iface = gr.Interface(
     fn=text_to_ipa,
     inputs=gr.Textbox(),
-    outputs=gr.Textbox()
+    outputs=gr.Textbox(),
+    api_name="predict"
 )
 
-# Mount Gradio inside FastAPI
-app = gr.mount_gradio_app(app, iface, path="/")
+iface.launch()
+