Update app.py

app.py

@@ -1,32 +1,15 @@
 import gradio as gr
 from transformers import T5Tokenizer, T5ForConditionalGeneration
 import torch
-from fastapi import FastAPI
 from pydantic import BaseModel
-from threading import Thread
-import uvicorn
 
 # Initialize FastAPI and Gradio
-app = FastAPI()
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
 # Load the tokenizer and model once for use in both FastAPI and Gradio
 tokenizer = T5Tokenizer.from_pretrained("alpeshsonar/lot-t5-small-filter", legacy=False)
 model = T5ForConditionalGeneration.from_pretrained("alpeshsonar/lot-t5-small-filter").to(device)
 
-# Health check endpoint
-@app.get("/health")
-async def health_check():
-    return {"status": "API is running"}
-
-# FastAPI endpoint
-@app.post("/generate")
-async def generate_text_api(input_text: str):
-    inputs = tokenizer.encode("Extract lots from given text.\n" + input_text, return_tensors="pt").to(device)
-    outputs = model.generate(inputs, max_new_tokens=1024)
-    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
-    return {"output": result}
-
 # Gradio interface
 def generate_text(input_text):
     inputs = tokenizer.encode("Extract lots from given text.\n" + input_text, return_tensors="pt").to(device)
@@ -38,15 +21,8 @@ iface = gr.Interface(fn=generate_text, inputs="text", outputs="text", title="Lin
 
 # Function to run both FastAPI and Gradio
 def run():
-    # Start FastAPI in a separate thread
-    def start_fastapi():
-        uvicorn.run(app, host="0.0.0.0", port=7860)
-
-    t = Thread(target=start_fastapi)
-    t.start()
-
     # Launch Gradio interface
-    iface.launch()
+    iface.launch(server_name="0.0.0.0", server_port=7860)
 
 if __name__ == "__main__":
     run()