app.py

from fastapi import FastAPI
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "mynuddin/chatbot"

# Load model and tokenizer without setting a custom cache
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(model_name).to("cpu")

app = FastAPI()

@app.post("/generate")
def generate_text(prompt: str):
    inputs = tokenizer(prompt, return_tensors="pt")
    output = model.generate(**inputs, max_length=128)
    generated_text = tokenizer.decode(output[0], skip_special_tokens=True)
    return {"generated_query": generated_text}