SetVisPrototypeData/huggingfacePython.py at master · VisDunneRight/SetVisPrototypeData · GitHub

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
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
import os
from transformers import pipeline

from transformers import DetrImageProcessor, DetrForObjectDetection
from transformers import OwlViTProcessor, OwlViTForObjectDetection
from transformers import YolosFeatureExtractor, YolosForObjectDetection
from transformers import AutoImageProcessor, ConditionalDetrForObjectDetection

import torch
from PIL import Image
import requests


# models = ["facebook/detr-resnet-50", "facebookresearch/detr", "facebookresearch/maskrcnn-benchmark", "fvcore/mask-rcnn-fpn", "fvcore/detr", "fvcore/retinanet-fpn"]

def run_Facebook_OneHundredOne(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)

        image = Image.open(image_path)

        processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-101")
        model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-101")

        inputs = processor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        # convert outputs (bounding boxes and class logits) to COCO API
        # let's only keep detections with score > 0.9
        target_sizes = torch.tensor([image.size[::-1]])
        results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:

                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")


def run_Facebook_Fifty(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)
        image = Image.open(image_path)

        processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
        model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")

        inputs = processor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        # convert outputs (bounding boxes and class logits) to COCO API
        # let's only keep detections with score > 0.9
        target_sizes = torch.tensor([image.size[::-1]])
        results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:
                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")

def run_Google_Owl(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)
        image = Image.open(image_path)

        processor = OwlViTProcessor.from_pretrained("google/owlvit-base-patch32")
        model = OwlViTForObjectDetection.from_pretrained("google/owlvit-base-patch32")

        inputs = processor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        # convert outputs (bounding boxes and class logits) to COCO API
        # let's only keep detections with score > 0.9
        target_sizes = torch.tensor([image.size[::-1]])
        results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:
                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")

def run_YOLO_tiny(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)
        image = Image.open(image_path)

        feature_extractor = YolosFeatureExtractor.from_pretrained('hustvl/yolos-tiny')
        model = YolosForObjectDetection.from_pretrained('hustvl/yolos-tiny')

        inputs = feature_extractor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        logits = outputs.logits
        bboxes = outputs.pred_boxes

        # print(logits)
        # print(bboxes)
        target_sizes = torch.tensor([image.size[::-1]])
        results = feature_extractor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:
                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")

def run_YOLO_base(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)
        image = Image.open(image_path)

        feature_extractor = YolosFeatureExtractor.from_pretrained('hustvl/yolos-base')
        model = YolosForObjectDetection.from_pretrained('hustvl/yolos-base')

        inputs = feature_extractor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        logits = outputs.logits
        bboxes = outputs.pred_boxes

        # print(logits)
        # print(bboxes)
        target_sizes = torch.tensor([image.size[::-1]])
        results = feature_extractor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:
                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")

def run_Microsoft_detr(folder_path, folderName):
    if not os.path.exists(folderName):
        os.makedirs(folderName)
    for image_file in os.listdir(folder_path):
        image_path = os.path.join(folder_path, image_file)
        image = Image.open(image_path)

        processor = AutoImageProcessor.from_pretrained("microsoft/conditional-detr-resnet-50")
        model = ConditionalDetrForObjectDetection.from_pretrained("microsoft/conditional-detr-resnet-50")

        inputs = processor(images=image, return_tensors="pt")
        outputs = model(**inputs)

        # convert outputs (bounding boxes and class logits) to COCO API
        # let's only keep detections with score > 0.9
        target_sizes = torch.tensor([image.size[::-1]])
        results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.5)[0]

        with open(os.path.join(folderName, image_file.split(".")[0] + ".txt"), 'w') as f:
                for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
                    box = box.tolist()
                    box = [round(box[0] / image.size[0], 4), round(box[1] / image.size[1], 4),
                        round(box[2] / image.size[0], 4), round(box[3] / image.size[1], 4)]
                    f.write(f"{model.config.id2label[label.item()]} {box[0]} {box[1]} {box[2]} {box[3]} {round(score.item(), 3)}\n")

# these work
# run_Facebook_OneHundredOne("./oneHundredImages/images", "./oneHundredImages/detr-resnet-101")
# run_Facebook_Fifty("./oneHundredImages/images", "./oneHundredImages/detr-resnet-50")
# run_YOLO_tiny("./oneHundredImages/images", "./oneHundredImages/yolos-tiny")
# run_YOLO_base("./oneHundredImages/images", "./oneHundredImages/yolos-base")
# run_Microsoft_detr("./oneHundredImages/images", "./oneHundredImages/conditional-detr-resnet-50")


run_Facebook_OneHundredOne("./twentyFive/images", "./twentyFive/detr-resnet-101")
run_Facebook_Fifty("./twentyFive/images", "./twentyFive/detr-resnet-50")
run_YOLO_tiny("./twentyFive/images", "./twentyFive/yolos-tiny")
run_YOLO_base("./twentyFive/images", "./twentyFive/yolos-base")
run_Microsoft_detr("./twentyFive/images", "./twentyFive/conditional-detr-resnet-50")

# run_Google_Owl("./twentyFiveImages/images", "./twentyFiveImages/owlvit-base-patch32")