main.py

# Simple implementation for translation using the BART model
from fastapi import FastAPI
from pydantic import BaseModel
from transformers import BartTokenizer, BartForConditionalGeneration

app = FastAPI()

# Define request model
class TranslationRequest(BaseModel):
    text: str
    max_length: int = 150
    min_length: int = 40

# Download and cache the model during initialization
# This happens only once when the app starts
try:
    # Explicitly download to a specific directory with proper error handling
    cache_dir = "./model_cache"
    model_name = "facebook/bart-large-cnn"
    
    print(f"Loading tokenizer from {model_name}...")
    tokenizer = BartTokenizer.from_pretrained(model_name, cache_dir=cache_dir, local_files_only=False)
    
    print(f"Loading model from {model_name}...")
    model = BartForConditionalGeneration.from_pretrained(model_name, cache_dir=cache_dir, local_files_only=False)
    
    print("Model and tokenizer loaded successfully!")
except Exception as e:
    print(f"Error loading model: {str(e)}")
    raise

@app.post("/summarize/")
async def translate_text(request: TranslationRequest):
    # Process the input text
    inputs = tokenizer(request.text, return_tensors="pt", max_length=1024, truncation=True)
    
    # Generate summary
    summary_ids = model.generate(
        inputs["input_ids"],
        max_length=request.max_length, 
        min_length=request.min_length,
        num_beams=4,
        length_penalty=2.0,
        early_stopping=True
    )
    
    # Decode the generated summary
    translation = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
    
    return {"summary": translation}

# Basic health check endpoint
@app.get("/health")
async def health_check():
    return {"status": "healthy", "model": "facebook/bart-large-cnn"}