Update app.py

app.py

@@ -1,33 +1,17 @@
 import gradio as gr
-from fastapi import FastAPI
 import onnxruntime as ort
 from transformers import AutoTokenizer
-from pydantic import BaseModel
-import numpy as np
-
-# Initialize FastAPI and Gradio
-app = FastAPI()
 
 # Load ONNX model and tokenizer
 MODEL_FILE = "./model.onnx"
 session = ort.InferenceSession(MODEL_FILE)
 tokenizer = AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-en-fr")
 
-# Define input model for FastAPI
-class TranslationInput(BaseModel):
-    input_text: str
-
-# FastAPI endpoint for model prediction
-@app.post("/predict")
-async def predict(translation_input: TranslationInput):
-    """ 
-    Endpoint for inference.
-    :param translation_input: Text input in English.
-    :return: Translated text in French.
-    """
+# Gradio prediction function
+def gradio_predict(input_text):
     # Tokenize input text
     tokenized_input = tokenizer(
-        translation_input.input_text,
+        input_text,
         return_tensors="np",
         padding=True
     )
@@ -41,20 +25,12 @@ async def predict(translation_input: TranslationInput):
 
     # Decode output and return translated text
     translated_text = tokenizer.decode(outputs[0][0], skip_special_tokens=True)
-    return {"translated_text": translated_text}
-
-# Gradio Interface
-def gradio_predict(input_text):
-    response = predict(TranslationInput(input_text=input_text))
-    return response["translated_text"]
+    return translated_text
 
 # Gradio interface for the web app
-gradio_interface = gr.Interface(
+gr.Interface(
     fn=gradio_predict,
     inputs="text",
     outputs="text",
     live=True
-)
-
-# Launch Gradio app
-gradio_interface.launch(inline=True, server_name="0.0.0.0", server_port=7860)
+).launch(share=True)