Hugging Face's logo

Datasets:
TrainingDataPro
/
license_plates

Tasks:
Image-to-Text
Image Feature Extraction
Feature Extraction
Languages:
English
Tags:
finance
code
Not-For-All-Audiences
License:
Dataset card Files Community
vkashko commited on
Commit
167bbdf
·
1 Parent(s): c5c8005

fix: script

Browse files
Files changed (2) hide show
  1. README.md +281 -0
  2. license_plates.py +25 -27
README.md CHANGED
@@ -6,6 +6,287 @@ language:
6
  - en
7
  tags:
8
  - finance
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9
  ---
10
 
11
  # License Plates
 
6
  - en
7
  tags:
8
  - finance
9
+ dataset_info:
10
+ - config_name: Brazil_youtube
11
+ features:
12
+ - name: image
13
+ dtype: image
14
+ - name: labeled_image
15
+ dtype: image
16
+ - name: bbox
17
+ dtype: string
18
+ - name: license_plate.id
19
+ dtype: string
20
+ - name: license_plate.visibility
21
+ dtype: string
22
+ - name: license_plate.rows_count
23
+ dtype: uint8
24
+ - name: license_plate.number
25
+ dtype: string
26
+ - name: license_plate.serial
27
+ dtype: string
28
+ - name: license_plate.country
29
+ dtype: string
30
+ - name: license_plate.mask
31
+ dtype: string
32
+ splits:
33
+ - name: train
34
+ num_bytes: 173536648
35
+ num_examples: 72
36
+ download_size: 22606962
37
+ dataset_size: 173536648
38
+ - config_name: Estonia_platesmania
39
+ features:
40
+ - name: image
41
+ dtype: image
42
+ - name: labeled_image
43
+ dtype: image
44
+ - name: bbox
45
+ dtype: string
46
+ - name: license_plate.id
47
+ dtype: string
48
+ - name: license_plate.visibility
49
+ dtype: string
50
+ - name: license_plate.rows_count
51
+ dtype: uint8
52
+ - name: license_plate.number
53
+ dtype: string
54
+ - name: license_plate.serial
55
+ dtype: string
56
+ - name: license_plate.country
57
+ dtype: string
58
+ - name: license_plate.mask
59
+ dtype: string
60
+ splits:
61
+ - name: train
62
+ num_bytes: 7990452
63
+ num_examples: 10
64
+ download_size: 7863164
65
+ dataset_size: 7990452
66
+ - config_name: Finland_platesmania
67
+ features:
68
+ - name: image
69
+ dtype: image
70
+ - name: labeled_image
71
+ dtype: image
72
+ - name: bbox
73
+ dtype: string
74
+ - name: license_plate.id
75
+ dtype: string
76
+ - name: license_plate.visibility
77
+ dtype: string
78
+ - name: license_plate.rows_count
79
+ dtype: uint8
80
+ - name: license_plate.number
81
+ dtype: string
82
+ - name: license_plate.serial
83
+ dtype: string
84
+ - name: license_plate.country
85
+ dtype: string
86
+ - name: license_plate.mask
87
+ dtype: string
88
+ splits:
89
+ - name: train
90
+ num_bytes: 9650579
91
+ num_examples: 10
92
+ download_size: 9485725
93
+ dataset_size: 9650579
94
+ - config_name: Kazakhstan_platesmania
95
+ features:
96
+ - name: image
97
+ dtype: image
98
+ - name: labeled_image
99
+ dtype: image
100
+ - name: bbox
101
+ dtype: string
102
+ - name: license_plate.id
103
+ dtype: string
104
+ - name: license_plate.visibility
105
+ dtype: string
106
+ - name: license_plate.rows_count
107
+ dtype: uint8
108
+ - name: license_plate.number
109
+ dtype: string
110
+ - name: license_plate.serial
111
+ dtype: string
112
+ - name: license_plate.country
113
+ dtype: string
114
+ - name: license_plate.mask
115
+ dtype: string
116
+ splits:
117
+ - name: train
118
+ num_bytes: 14064541
119
+ num_examples: 19
120
+ download_size: 7265915
121
+ dataset_size: 14064541
122
+ - config_name: Kazakhstan_youtube
123
+ features:
124
+ - name: image
125
+ dtype: image
126
+ - name: labeled_image
127
+ dtype: image
128
+ - name: bbox
129
+ dtype: string
130
+ - name: license_plate.id
131
+ dtype: string
132
+ - name: license_plate.visibility
133
+ dtype: string
134
+ - name: license_plate.rows_count
135
+ dtype: uint8
136
+ - name: license_plate.number
137
+ dtype: string
138
+ - name: license_plate.serial
139
+ dtype: string
140
+ - name: license_plate.country
141
+ dtype: string
142
+ - name: license_plate.mask
143
+ dtype: string
144
+ splits:
145
+ - name: train
146
+ num_bytes: 6324396
147
+ num_examples: 22
148
+ download_size: 2852873
149
+ dataset_size: 6324396
150
+ - config_name: Lithuania_platesmania
151
+ features:
152
+ - name: image
153
+ dtype: image
154
+ - name: labeled_image
155
+ dtype: image
156
+ - name: bbox
157
+ dtype: string
158
+ - name: license_plate.id
159
+ dtype: string
160
+ - name: license_plate.visibility
161
+ dtype: string
162
+ - name: license_plate.rows_count
163
+ dtype: uint8
164
+ - name: license_plate.number
165
+ dtype: string
166
+ - name: license_plate.serial
167
+ dtype: string
168
+ - name: license_plate.country
169
+ dtype: string
170
+ - name: license_plate.mask
171
+ dtype: string
172
+ splits:
173
+ - name: train
174
+ num_bytes: 8127614
175
+ num_examples: 10
176
+ download_size: 7940839
177
+ dataset_size: 8127614
178
+ - config_name: Serbia_platesmania
179
+ features:
180
+ - name: image
181
+ dtype: image
182
+ - name: labeled_image
183
+ dtype: image
184
+ - name: bbox
185
+ dtype: string
186
+ - name: license_plate.id
187
+ dtype: string
188
+ - name: license_plate.visibility
189
+ dtype: string
190
+ - name: license_plate.rows_count
191
+ dtype: uint8
192
+ - name: license_plate.number
193
+ dtype: string
194
+ - name: license_plate.serial
195
+ dtype: string
196
+ - name: license_plate.country
197
+ dtype: string
198
+ - name: license_plate.mask
199
+ dtype: string
200
+ splits:
201
+ - name: train
202
+ num_bytes: 10000777
203
+ num_examples: 10
204
+ download_size: 9808356
205
+ dataset_size: 10000777
206
+ - config_name: Serbia_youtube
207
+ features:
208
+ - name: image
209
+ dtype: image
210
+ - name: labeled_image
211
+ dtype: image
212
+ - name: bbox
213
+ dtype: string
214
+ - name: license_plate.id
215
+ dtype: string
216
+ - name: license_plate.visibility
217
+ dtype: string
218
+ - name: license_plate.rows_count
219
+ dtype: uint8
220
+ - name: license_plate.number
221
+ dtype: string
222
+ - name: license_plate.serial
223
+ dtype: string
224
+ - name: license_plate.country
225
+ dtype: string
226
+ - name: license_plate.mask
227
+ dtype: string
228
+ splits:
229
+ - name: train
230
+ num_bytes: 26535839
231
+ num_examples: 67
232
+ download_size: 4044272
233
+ dataset_size: 26535839
234
+ - config_name: UAE_platesmania
235
+ features:
236
+ - name: image
237
+ dtype: image
238
+ - name: labeled_image
239
+ dtype: image
240
+ - name: bbox
241
+ dtype: string
242
+ - name: license_plate.id
243
+ dtype: string
244
+ - name: license_plate.visibility
245
+ dtype: string
246
+ - name: license_plate.rows_count
247
+ dtype: uint8
248
+ - name: license_plate.number
249
+ dtype: string
250
+ - name: license_plate.serial
251
+ dtype: string
252
+ - name: license_plate.country
253
+ dtype: string
254
+ - name: license_plate.mask
255
+ dtype: string
256
+ splits:
257
+ - name: train
258
+ num_bytes: 8236358
259
+ num_examples: 10
260
+ download_size: 8028800
261
+ dataset_size: 8236358
262
+ - config_name: UAE_youtube
263
+ features:
264
+ - name: image
265
+ dtype: image
266
+ - name: labeled_image
267
+ dtype: image
268
+ - name: bbox
269
+ dtype: string
270
+ - name: license_plate.id
271
+ dtype: string
272
+ - name: license_plate.visibility
273
+ dtype: string
274
+ - name: license_plate.rows_count
275
+ dtype: uint8
276
+ - name: license_plate.number
277
+ dtype: string
278
+ - name: license_plate.serial
279
+ dtype: string
280
+ - name: license_plate.country
281
+ dtype: string
282
+ - name: license_plate.mask
283
+ dtype: string
284
+ splits:
285
+ - name: train
286
+ num_bytes: 41202317
287
+ num_examples: 162
288
+ download_size: 2666314
289
+ dataset_size: 41202317
290
  ---
291
 
292
  # License Plates
license_plates.py CHANGED
@@ -27,13 +27,16 @@ _DATA = f"https://huggingface.co/datasets/TrainingDataPro/{_NAME}/resolve/main/d
27
  class LicensePlates(datasets.GeneratorBasedBuilder):
28
 
29
  BUILDER_CONFIGS = [
30
- datasets.BuilderConfig(name="Brazil"),
31
- datasets.BuilderConfig(name="Estonia"),
32
- datasets.BuilderConfig(name="Finland"),
33
- datasets.BuilderConfig(name="Kazakhstan"),
34
- datasets.BuilderConfig(name="Lithuania"),
35
- datasets.BuilderConfig(name="Serbia"),
36
- datasets.BuilderConfig(name="UAE"),
 
 
 
37
  ]
38
 
39
  DEFAULT_CONFIG_NAME = "Brazil"
@@ -64,34 +67,29 @@ class LicensePlates(datasets.GeneratorBasedBuilder):
64
  def _split_generators(self, dl_manager):
65
  data = dl_manager.download(f"{_DATA}{self.config.name}.tar.gz")
66
  data = dl_manager.iter_archive(data)
 
67
  return [
68
  datasets.SplitGenerator(name=datasets.Split.TRAIN,
69
- gen_kwargs={"data": data}),
 
 
 
70
  ]
71
 
72
- def _generate_examples(self, data):
73
 
74
- annotations = []
75
  images = {}
76
 
77
  for idx, (file_path, file) in enumerate(data):
78
- if file_path.endswith('.csv'):
79
- print(file_path)
80
- df = pd.read_csv(file.read(), sep=',')
81
- annotations.append(df)
82
-
83
- else:
84
- file_name = file_path.split('/')[-1]
85
- images[file_name] = (file_path, file.read())
86
-
87
- if len(annotations) > 1:
88
- annotations_df = pd.concat(annotations)
89
- else:
90
- annotations_df = annotations[0]
91
-
92
- annotations_df.drop(
93
- columns=['license_plate.region', 'license_plate.color'],
94
- inplace=True)
95
 
96
  for row in annotations_df.itertuples(index=False):
97
  image = images[row[0]]
 
27
  class LicensePlates(datasets.GeneratorBasedBuilder):
28
 
29
  BUILDER_CONFIGS = [
30
+ datasets.BuilderConfig(name="Brazil_youtube"),
31
+ datasets.BuilderConfig(name="Estonia_platesmania"),
32
+ datasets.BuilderConfig(name="Finland_platesmania"),
33
+ datasets.BuilderConfig(name="Kazakhstan_platesmania"),
34
+ datasets.BuilderConfig(name="Kazakhstan_youtube"),
35
+ datasets.BuilderConfig(name="Lithuania_platesmania"),
36
+ datasets.BuilderConfig(name="Serbia_platesmania"),
37
+ datasets.BuilderConfig(name="Serbia_youtube"),
38
+ datasets.BuilderConfig(name="UAE_platesmania"),
39
+ datasets.BuilderConfig(name="UAE_youtube")
40
  ]
41
 
42
  DEFAULT_CONFIG_NAME = "Brazil"
 
67
  def _split_generators(self, dl_manager):
68
  data = dl_manager.download(f"{_DATA}{self.config.name}.tar.gz")
69
  data = dl_manager.iter_archive(data)
70
+ annotations = dl_manager.download(f'{_DATA}{self.config.name}.csv')
71
  return [
72
  datasets.SplitGenerator(name=datasets.Split.TRAIN,
73
+ gen_kwargs={
74
+ "data": data,
75
+ 'annotations': annotations
76
+ }),
77
  ]
78
 
79
+ def _generate_examples(self, data, annotations):
80
 
81
+ annotations_df = pd.read_csv(annotations, sep=',', index_col=0)
82
  images = {}
83
 
84
  for idx, (file_path, file) in enumerate(data):
85
+ file_name = file_path.split('/')[-1]
86
+ images[file_name] = (file_path, file.read())
87
+
88
+ annotations_df.drop(columns=['license_plate.region', 'license_plate.color'],
89
+ inplace=True,
90
+ errors='ignore')
91
+
92
+ annotations_df.fillna(0, inplace=True)
 
 
 
 
 
 
 
 
 
93
 
94
  for row in annotations_df.itertuples(index=False):
95
  image = images[row[0]]