app/main.py

from fastapi import FastAPI
from transformers import BartForConditionalGeneration, BartTokenizer
import pickle
from pydantic import BaseModel

app = FastAPI()
model_name = "facebook/bart-large-cnn"
tokenizer = BartTokenizer.from_pretrained(model_name)

# Load the Bart model
model = BartForConditionalGeneration.from_pretrained(model_name)

async def mymodel(text):
    inputs = tokenizer(text, return_tensors="pt")
    summary_ids = model.generate(inputs["input_ids"], num_beams=4, max_length=100, min_length=30, length_penalty=2.0)
    summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
    return summary_text

class Request(BaseModel):
    input_text: str

class Response(BaseModel):
    summ_text: str

@app.post("/summarise", response_model=Response)
async def predict(request: Request) -> Response:
    x = request.input_text
    summary_text = await mymodel(x)
    return Response(summ_text=summary_text)