pipeline.py

from typing import  Dict, List, Any
from PIL import Image
import requests
import torch
import base64
import os
from io import BytesIO
from models.blip_decoder import blip_decoder
from torchvision import transforms
from torchvision.transforms.functional import InterpolationMode

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

class PreTrainedPipeline():
    def __init__(self, path=""):
        # load the optimized model
        self.model_path = 'model_large_caption.pth'
        self.model = blip_decoder(
            pretrained=self.model_path, 
            image_size=384, 
            vit='large',
            med_config=os.path.join(path, 'configs/med_config.json')
        )
        self.model.eval()
        self.model = self.model.to(device)
        
        image_size = 384
        self.transform = transforms.Compose([
            transforms.Resize((image_size,image_size),interpolation=InterpolationMode.BICUBIC),
            transforms.ToTensor(),
            transforms.Normalize((0.48145466, 0.4578275, 0.40821073), (0.26862954, 0.26130258, 0.27577711))
            ]) 
     


    def __call__(self, data: Any) -> Dict[str, Any]:
        """
        Args:
            data (:obj:):
                includes the input data and the parameters for the inference.
        Return:
            A :obj:`dict`:. The object returned should be a dict of one list like [[{"label": 0.9939950108528137}]] containing :
                - "caption": A string corresponding to the generated caption.
        """
        inputs = data.pop("inputs", data)
        parameters = data.pop("parameters", {})

        # decode base64 image to PIL
        image = Image.open(BytesIO(base64.b64decode(inputs['image'])))
        image = self.transform(image).unsqueeze(0).to(device)   
        with torch.no_grad():
            caption = self.model.generate(
                image, 
                sample=parameters.get('sample',True),
                top_p=parameters.get('top_p',0.9), 
                max_length=parameters.get('max_length',20), 
                min_length=parameters.get('min_length',5)
            )
        # postprocess the prediction
        return {"caption": caption}