|
from typing import Dict, List, Any |
|
from PIL import Image |
|
from io import BytesIO |
|
from transformers import AutoProcessor, OmDetTurboForObjectDetection |
|
import base64 |
|
import logging |
|
|
|
class EndpointHandler(): |
|
def __init__(self, path=""): |
|
self.processor = AutoProcessor.from_pretrained("Blueway/inference-endpoint-for-omdet-turbo-swin-tiny-hf") |
|
self.model = OmDetTurboForObjectDetection.from_pretrained("Blueway/inference-endpoint-for-omdet-turbo-swin-tiny-hf") |
|
|
|
def __call__(self, data: Dict[str, Any]) -> List[Dict[str, Any]]: |
|
""" |
|
data args: |
|
image (:obj:`string`) |
|
candidates (:obj:`list`) |
|
Return: |
|
A :obj:`list`:. The list contains items that are dicts should be liked {"label": "XXX", "score": 0.82} |
|
""" |
|
inputs_request = data.pop("inputs", data) |
|
|
|
|
|
image = Image.open(BytesIO(base64.b64decode(inputs_request['image']))) |
|
|
|
|
|
inputs = self.processor(image, text=inputs_request["candidates"], return_tensors="pt") |
|
outputs = self.model(**inputs) |
|
results = self.processor.post_process_grounded_object_detection( |
|
outputs, |
|
classes=inputs_request["candidates"], |
|
target_sizes=[image.size[::-1]], |
|
score_threshold=0.3, |
|
nms_threshold=0.3, |
|
)[0] |
|
|
|
serializable_results = { |
|
'boxes': results['boxes'].tolist(), |
|
'scores': results['scores'].tolist(), |
|
'candidates': results['classes'] |
|
} |
|
return serializable_results |
|
|
|
|
|
|
|
|