File size: 2,730 Bytes
a7ce59e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
import json
from typing import Any, Dict, List

import tensorflow as tf
from tensorflow import keras
from app.pipelines import Pipeline
from huggingface_hub import from_pretrained_keras, hf_hub_download
from PIL import Image
import base64


MODEL_FILENAME = "saved_model.pb"
CONFIG_FILENAME = "config.json"


class PreTrainedPipeline(Pipeline):
    def __init__(self, model_id: str):


        # Reload Keras SavedModel
        self.model = from_pretrained_keras(model_id)

        # Number of labels
        self.num_labels = self.model.output_shape[1]

        # Config is required to know the mapping to label.
        config_file = hf_hub_download(model_id, filename=CONFIG_FILENAME)
        with open(config_file) as config:
            config = json.load(config)

        self.id2label = config.get(
            "id2label", {str(i): f"LABEL_{i}" for i in range(self.num_labels)}
        )

    def __call__(self, inputs: "Image.Image") -> List[Dict[str, Any]]:
        """
        Args:
            inputs (:obj:`PIL.Image`):
                The raw image representation as PIL.
                No transformation made whatsoever from the input. Make all necessary transformations here.
        Return:
            A :obj:`list`:. The list contains items that are dicts should be liked {"label": "XXX" (str), mask: "base64 encoding of the mask" (str), "score": float}
                It is preferred if the returned list is in decreasing `score` order
        """
        # Resize image to expected size

        expected_input_size = self.model.input_shape
        if expected_input_size[-1] == 1:  
            inputs = inputs.convert("L")

        target_size = (expected_input_size[1], expected_input_size[2])
        img = tf.image.resize(inputs, target_size)

        img_array = tf.keras.preprocessing.image.img_to_array(img)
        img_array = img_array[tf.newaxis, ...]

        predictions = self.model.predict(img_array, axis=-1)

        self.single_output_unit = (
            self.model.output_shape[1] == 1
        )  # if there are two classes

        if self.single_output_unit:
            score = predictions[0][0]
            labels = [
                {"label": str(self.id2label["1"]), "score": float(score)},
                {"label": str(self.id2label["0"]), "score": float(1 - score)},
            ]
        else:
            labels = [
                {
                    "label": str(self.id2label[str(i)]),
                    "mask": base64.b64encode(predictions[0][i]),
                    "score": float(score),
                }
                for i, score in enumerate(predictions[0])
            ]
        return sorted(labels, key=lambda tup: tup["score"], reverse=True)[: self.top_k]