app/main.py

from fastapi import FastAPI, UploadFile, File
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import torch
import io

app = FastAPI(title="QuickCare Captioning Endpoint")

blip_model = BlipForConditionalGeneration.from_pretrained(
    "Salesforce/blip-image-captioning-large"
)
blip_processor = BlipProcessor.from_pretrained(
    "Salesforce/blip-image-captioning-large"
)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
blip_model.to(device)


@app.post("/caption")
async def generate_caption(image: UploadFile = File(...)):
    image_bytes = await image.read()
    img = Image.open(io.BytesIO(image_bytes)).convert("RGB")

    inputs = blip_processor(images=img, return_tensors="pt").to(device)

    with torch.no_grad():
        caption_ids = blip_model.generate(**inputs, max_new_tokens=60)
    caption = blip_processor.decode(caption_ids[0], skip_special_tokens=True)

    return {"caption": caption}