| from typing import Dict, List, Any |
| import numpy as np |
| from transformers import CLIPProcessor, CLIPModel |
| from PIL import Image |
| from io import BytesIO |
| import base64 |
|
|
| class EndpointHandler(): |
| def __init__(self, path=""): |
| |
| self.model = CLIPModel.from_pretrained("openai/clip-vit-base-patch32") |
| self.processor = CLIPProcessor.from_pretrained("openai/clip-vit-base-patch32") |
|
|
|
|
| def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]: |
| print("** data: ", data) |
| inputs = data.get("inputs") |
| print("** inputs: ", inputs) |
|
|
| text = inputs.get("text") |
| print("** text: ", text) |
|
|
| imageData = inputs.get("image") |
| print("** imageData: ", imageData) |
|
|
| image = None |
| if imageData: |
| try: |
| image = Image.open(BytesIO(base64.b64decode(imageData))) |
| print("** image: ", image) |
| except Exception as e: |
| raise ValueError(f"Error decoding image: {e}") |
|
|
| if not text and not image: |
| raise ValueError("Both text and image cannot be None. Provide at least one.") |
|
|
|
|
| inputs = self.processor(text=text, images=image, return_tensors="pt", padding=True) |
| outputs = self.model(**inputs) |
| embeddings = outputs.image_embeds.detach().numpy().flatten().tolist() |
| return { "embeddings": embeddings } |
