Hugging Face's logo

Datasets:
Wakals
/
CoVT-Dataset

Modalities:
Image
Text
Formats:
parquet
Size:
1M - 10M
ArXiv:
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files
xet
Community
2

Improve dataset card: Add description, links, tasks, and sample usage

#2
by nielsr HF Staff - opened
base: refs/heads/main
←
from: refs/pr/2
Discussion Files changed
+539
-485
Files changed (1) hide show
  1. README.md +539 -485
README.md CHANGED
@@ -1,485 +1,539 @@
1
- ---
2
- license: apache-2.0
3
- dataset_info:
4
- - config_name: part1
5
- features:
6
- - name: id
7
- dtype: string
8
- - name: conversations
9
- list:
10
- - name: from
11
- dtype: string
12
- - name: value
13
- dtype: string
14
- - name: image_relpath
15
- dtype: string
16
- - name: image
17
- dtype: image
18
- splits:
19
- - name: train
20
- num_bytes: 446733439.675
21
- num_examples: 44475
22
- download_size: 17934592717
23
- dataset_size: 446733439.675
24
- - config_name: part10
25
- features:
26
- - name: id
27
- dtype: string
28
- - name: conversations
29
- list:
30
- - name: from
31
- dtype: string
32
- - name: value
33
- dtype: string
34
- - name: image_relpath
35
- dtype: string
36
- - name: image
37
- dtype: image
38
- splits:
39
- - name: train
40
- num_bytes: 17744642822.628
41
- num_examples: 44474
42
- download_size: 17769519559
43
- dataset_size: 17744642822.628
44
- - config_name: part11
45
- features:
46
- - name: id
47
- dtype: string
48
- - name: conversations
49
- list:
50
- - name: from
51
- dtype: string
52
- - name: value
53
- dtype: string
54
- - name: image_relpath
55
- dtype: string
56
- - name: image
57
- dtype: image
58
- splits:
59
- - name: train
60
- num_bytes: 17749640543.646
61
- num_examples: 44474
62
- download_size: 40708430498
63
- dataset_size: 17749640543.646
64
- - config_name: part12
65
- features:
66
- - name: id
67
- dtype: string
68
- - name: conversations
69
- list:
70
- - name: from
71
- dtype: string
72
- - name: value
73
- dtype: string
74
- - name: image_relpath
75
- dtype: string
76
- - name: image
77
- dtype: image
78
- splits:
79
- - name: train
80
- num_bytes: 16126888186.176
81
- num_examples: 44474
82
- download_size: 13512012899
83
- dataset_size: 16126888186.176
84
- - config_name: part13
85
- features:
86
- - name: id
87
- dtype: string
88
- - name: conversations
89
- list:
90
- - name: from
91
- dtype: string
92
- - name: value
93
- dtype: string
94
- - name: image_relpath
95
- dtype: string
96
- - name: image
97
- dtype: image
98
- splits:
99
- - name: train
100
- num_bytes: 12426535004.444
101
- num_examples: 44474
102
- download_size: 17473903640
103
- dataset_size: 12426535004.444
104
- - config_name: part14
105
- features:
106
- - name: id
107
- dtype: string
108
- - name: conversations
109
- list:
110
- - name: from
111
- dtype: string
112
- - name: value
113
- dtype: string
114
- - name: image_relpath
115
- dtype: string
116
- - name: image
117
- dtype: image
118
- splits:
119
- - name: train
120
- num_bytes: 18713928974.662
121
- num_examples: 44474
122
- download_size: 15650533270
123
- dataset_size: 18713928974.662
124
- - config_name: part15
125
- features:
126
- - name: id
127
- dtype: string
128
- - name: conversations
129
- list:
130
- - name: from
131
- dtype: string
132
- - name: value
133
- dtype: string
134
- - name: image_relpath
135
- dtype: string
136
- - name: image
137
- dtype: image
138
- splits:
139
- - name: train
140
- num_bytes: 5895768695.898
141
- num_examples: 44474
142
- download_size: 4400917224
143
- dataset_size: 5895768695.898
144
- - config_name: part16
145
- features:
146
- - name: id
147
- dtype: string
148
- - name: conversations
149
- list:
150
- - name: from
151
- dtype: string
152
- - name: value
153
- dtype: string
154
- - name: image_relpath
155
- dtype: string
156
- - name: image
157
- dtype: image
158
- splits:
159
- - name: train
160
- num_bytes: 6486479074.432
161
- num_examples: 44474
162
- download_size: 6271251512
163
- dataset_size: 6486479074.432
164
- - config_name: part17
165
- features:
166
- - name: id
167
- dtype: string
168
- - name: conversations
169
- list:
170
- - name: from
171
- dtype: string
172
- - name: value
173
- dtype: string
174
- - name: image_relpath
175
- dtype: string
176
- - name: image
177
- dtype: image
178
- splits:
179
- - name: train
180
- num_bytes: 6971738527.692
181
- num_examples: 44474
182
- download_size: 6469686725
183
- dataset_size: 6971738527.692
184
- - config_name: part18
185
- features:
186
- - name: id
187
- dtype: string
188
- - name: conversations
189
- list:
190
- - name: from
191
- dtype: string
192
- - name: value
193
- dtype: string
194
- - name: image_relpath
195
- dtype: string
196
- - name: image
197
- dtype: image
198
- splits:
199
- - name: train
200
- num_bytes: 7515615259.548
201
- num_examples: 44474
202
- download_size: 7563582116
203
- dataset_size: 7515615259.548
204
- - config_name: part19
205
- features:
206
- - name: id
207
- dtype: string
208
- - name: conversations
209
- list:
210
- - name: from
211
- dtype: string
212
- - name: value
213
- dtype: string
214
- - name: image_relpath
215
- dtype: string
216
- - name: image
217
- dtype: image
218
- splits:
219
- - name: train
220
- num_bytes: 7607522167.776
221
- num_examples: 44474
222
- download_size: 7377834021
223
- dataset_size: 7607522167.776
224
- - config_name: part2
225
- features:
226
- - name: id
227
- dtype: string
228
- - name: conversations
229
- list:
230
- - name: from
231
- dtype: string
232
- - name: value
233
- dtype: string
234
- - name: image_relpath
235
- dtype: string
236
- - name: image
237
- dtype: image
238
- splits:
239
- - name: train
240
- num_bytes: 33365826763.55
241
- num_examples: 44475
242
- download_size: 33245002109
243
- dataset_size: 33365826763.55
244
- - config_name: part20
245
- features:
246
- - name: id
247
- dtype: string
248
- - name: conversations
249
- list:
250
- - name: from
251
- dtype: string
252
- - name: value
253
- dtype: string
254
- - name: image_relpath
255
- dtype: string
256
- - name: image
257
- dtype: image
258
- splits:
259
- - name: train
260
- num_bytes: 7179023719.416
261
- num_examples: 44474
262
- download_size: 6145041547
263
- dataset_size: 7179023719.416
264
- - config_name: part3
265
- features:
266
- - name: id
267
- dtype: string
268
- - name: conversations
269
- list:
270
- - name: from
271
- dtype: string
272
- - name: value
273
- dtype: string
274
- - name: image_relpath
275
- dtype: string
276
- - name: image
277
- dtype: image
278
- splits:
279
- - name: train
280
- num_bytes: 24289174922.225
281
- num_examples: 44475
282
- download_size: 15536031938
283
- dataset_size: 24289174922.225
284
- - config_name: part4
285
- features:
286
- - name: id
287
- dtype: string
288
- - name: conversations
289
- list:
290
- - name: from
291
- dtype: string
292
- - name: value
293
- dtype: string
294
- - name: image_relpath
295
- dtype: string
296
- - name: image
297
- dtype: image
298
- splits:
299
- - name: train
300
- num_bytes: 6699346145.65
301
- num_examples: 44475
302
- download_size: 6646535744
303
- dataset_size: 6699346145.65
304
- - config_name: part5
305
- features:
306
- - name: id
307
- dtype: string
308
- - name: conversations
309
- list:
310
- - name: from
311
- dtype: string
312
- - name: value
313
- dtype: string
314
- - name: image_relpath
315
- dtype: string
316
- - name: image
317
- dtype: image
318
- splits:
319
- - name: train
320
- num_bytes: 6651458829.325
321
- num_examples: 44475
322
- download_size: 4896128944
323
- dataset_size: 6651458829.325
324
- - config_name: part6
325
- features:
326
- - name: id
327
- dtype: string
328
- - name: conversations
329
- list:
330
- - name: from
331
- dtype: string
332
- - name: value
333
- dtype: string
334
- - name: image_relpath
335
- dtype: string
336
- - name: image
337
- dtype: image
338
- splits:
339
- - name: train
340
- num_bytes: 18039837306.175
341
- num_examples: 44475
342
- download_size: 17786918533
343
- dataset_size: 18039837306.175
344
- - config_name: part7
345
- features:
346
- - name: id
347
- dtype: string
348
- - name: conversations
349
- list:
350
- - name: from
351
- dtype: string
352
- - name: value
353
- dtype: string
354
- - name: image_relpath
355
- dtype: string
356
- - name: image
357
- dtype: image
358
- splits:
359
- - name: train
360
- num_bytes: 17852309246.925
361
- num_examples: 44475
362
- download_size: 17774641127
363
- dataset_size: 17852309246.925
364
- - config_name: part8
365
- features:
366
- - name: id
367
- dtype: string
368
- - name: conversations
369
- list:
370
- - name: from
371
- dtype: string
372
- - name: value
373
- dtype: string
374
- - name: image_relpath
375
- dtype: string
376
- - name: image
377
- dtype: image
378
- splits:
379
- - name: train
380
- num_bytes: 17873005196.425
381
- num_examples: 44475
382
- download_size: 16131787376
383
- dataset_size: 17873005196.425
384
- - config_name: part9
385
- features:
386
- - name: id
387
- dtype: string
388
- - name: conversations
389
- list:
390
- - name: from
391
- dtype: string
392
- - name: value
393
- dtype: string
394
- - name: image_relpath
395
- dtype: string
396
- - name: image
397
- dtype: image
398
- splits:
399
- - name: train
400
- num_bytes: 34314021153.775
401
- num_examples: 44475
402
- download_size: 41828939793
403
- dataset_size: 34314021153.775
404
- configs:
405
- - config_name: part1
406
- data_files:
407
- - split: train
408
- path: part1/train-*
409
- - config_name: part10
410
- data_files:
411
- - split: train
412
- path: part10/train-*
413
- - config_name: part11
414
- data_files:
415
- - split: train
416
- path: part11/train-*
417
- - config_name: part12
418
- data_files:
419
- - split: train
420
- path: part12/train-*
421
- - config_name: part13
422
- data_files:
423
- - split: train
424
- path: part13/train-*
425
- - config_name: part14
426
- data_files:
427
- - split: train
428
- path: part14/train-*
429
- - config_name: part15
430
- data_files:
431
- - split: train
432
- path: part15/train-*
433
- - config_name: part16
434
- data_files:
435
- - split: train
436
- path: part16/train-*
437
- - config_name: part17
438
- data_files:
439
- - split: train
440
- path: part17/train-*
441
- - config_name: part18
442
- data_files:
443
- - split: train
444
- path: part18/train-*
445
- - config_name: part19
446
- data_files:
447
- - split: train
448
- path: part19/train-*
449
- - config_name: part2
450
- data_files:
451
- - split: train
452
- path: part2/train-*
453
- - config_name: part20
454
- data_files:
455
- - split: train
456
- path: part20/train-*
457
- - config_name: part3
458
- data_files:
459
- - split: train
460
- path: part3/train-*
461
- - config_name: part4
462
- data_files:
463
- - split: train
464
- path: part4/train-*
465
- - config_name: part5
466
- data_files:
467
- - split: train
468
- path: part5/train-*
469
- - config_name: part6
470
- data_files:
471
- - split: train
472
- path: part6/train-*
473
- - config_name: part7
474
- data_files:
475
- - split: train
476
- path: part7/train-*
477
- - config_name: part8
478
- data_files:
479
- - split: train
480
- path: part8/train-*
481
- - config_name: part9
482
- data_files:
483
- - split: train
484
- path: part9/train-*
485
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ task_categories:
4
+ - image-segmentation
5
+ - depth-estimation
6
+ - image-to-image
7
+ - image-feature-extraction
8
+ - image-text-to-text
9
+ language:
10
+ - en
11
+ tags:
12
+ - vlm
13
+ - multimodal
14
+ - visual-reasoning
15
+ - chain-of-thought
16
+ dataset_info:
17
+ - config_name: part1
18
+ features:
19
+ - name: id
20
+ dtype: string
21
+ - name: conversations
22
+ list:
23
+ - name: from
24
+ dtype: string
25
+ - name: value
26
+ dtype: string
27
+ - name: image_relpath
28
+ dtype: string
29
+ - name: image
30
+ dtype: image
31
+ splits:
32
+ - name: train
33
+ num_bytes: 446733439.675
34
+ num_examples: 44475
35
+ download_size: 17934592717
36
+ dataset_size: 446733439.675
37
+ - config_name: part10
38
+ features:
39
+ - name: id
40
+ dtype: string
41
+ - name: conversations
42
+ list:
43
+ - name: from
44
+ dtype: string
45
+ - name: value
46
+ dtype: string
47
+ - name: image_relpath
48
+ dtype: string
49
+ - name: image
50
+ dtype: image
51
+ splits:
52
+ - name: train
53
+ num_bytes: 17744642822.628
54
+ num_examples: 44474
55
+ download_size: 17769519559
56
+ dataset_size: 17744642822.628
57
+ - config_name: part11
58
+ features:
59
+ - name: id
60
+ dtype: string
61
+ - name: conversations
62
+ list:
63
+ - name: from
64
+ dtype: string
65
+ - name: value
66
+ dtype: string
67
+ - name: image_relpath
68
+ dtype: string
69
+ - name: image
70
+ dtype: image
71
+ splits:
72
+ - name: train
73
+ num_bytes: 17749640543.646
74
+ num_examples: 44474
75
+ download_size: 40708430498
76
+ dataset_size: 17749640543.646
77
+ - config_name: part12
78
+ features:
79
+ - name: id
80
+ dtype: string
81
+ - name: conversations
82
+ list:
83
+ - name: from
84
+ dtype: string
85
+ - name: value
86
+ dtype: string
87
+ - name: image_relpath
88
+ dtype: string
89
+ - name: image
90
+ dtype: image
91
+ splits:
92
+ - name: train
93
+ num_bytes: 16126888186.176
94
+ num_examples: 44474
95
+ download_size: 13512012899
96
+ dataset_size: 16126888186.176
97
+ - config_name: part13
98
+ features:
99
+ - name: id
100
+ dtype: string
101
+ - name: conversations
102
+ list:
103
+ - name: from
104
+ dtype: string
105
+ - name: value
106
+ dtype: string
107
+ - name: image_relpath
108
+ dtype: string
109
+ - name: image
110
+ dtype: image
111
+ splits:
112
+ - name: train
113
+ num_bytes: 12426535004.444
114
+ num_examples: 44474
115
+ download_size: 17473903640
116
+ dataset_size: 12426535004.444
117
+ - config_name: part14
118
+ features:
119
+ - name: id
120
+ dtype: string
121
+ - name: conversations
122
+ list:
123
+ - name: from
124
+ dtype: string
125
+ - name: value
126
+ dtype: string
127
+ - name: image_relpath
128
+ dtype: string
129
+ - name: image
130
+ dtype: image
131
+ splits:
132
+ - name: train
133
+ num_bytes: 18713928974.662
134
+ num_examples: 44474
135
+ download_size: 15650533270
136
+ dataset_size: 18713928974.662
137
+ - config_name: part15
138
+ features:
139
+ - name: id
140
+ dtype: string
141
+ - name: conversations
142
+ list:
143
+ - name: from
144
+ dtype: string
145
+ - name: value
146
+ dtype: string
147
+ - name: image_relpath
148
+ dtype: string
149
+ - name: image
150
+ dtype: image
151
+ splits:
152
+ - name: train
153
+ num_bytes: 5895768695.898
154
+ num_examples: 44474
155
+ download_size: 4400917224
156
+ dataset_size: 5895768695.898
157
+ - config_name: part16
158
+ features:
159
+ - name: id
160
+ dtype: string
161
+ - name: conversations
162
+ list:
163
+ - name: from
164
+ dtype: string
165
+ - name: value
166
+ dtype: string
167
+ - name: image_relpath
168
+ dtype: string
169
+ - name: image
170
+ dtype: image
171
+ splits:
172
+ - name: train
173
+ num_bytes: 6486479074.432
174
+ num_examples: 44474
175
+ download_size: 6271251512
176
+ dataset_size: 6486479074.432
177
+ - config_name: part17
178
+ features:
179
+ - name: id
180
+ dtype: string
181
+ - name: conversations
182
+ list:
183
+ - name: from
184
+ dtype: string
185
+ - name: value
186
+ dtype: string
187
+ - name: image_relpath
188
+ dtype: string
189
+ - name: image
190
+ dtype: image
191
+ splits:
192
+ - name: train
193
+ num_bytes: 6971738527.692
194
+ num_examples: 44474
195
+ download_size: 6469686725
196
+ dataset_size: 6971738527.692
197
+ - config_name: part18
198
+ features:
199
+ - name: id
200
+ dtype: string
201
+ - name: conversations
202
+ list:
203
+ - name: from
204
+ dtype: string
205
+ - name: value
206
+ dtype: string
207
+ - name: image_relpath
208
+ dtype: string
209
+ - name: image
210
+ dtype: image
211
+ splits:
212
+ - name: train
213
+ num_bytes: 7515615259.548
214
+ num_examples: 44474
215
+ download_size: 7563582116
216
+ dataset_size: 7515615259.548
217
+ - config_name: part19
218
+ features:
219
+ - name: id
220
+ dtype: string
221
+ - name: conversations
222
+ list:
223
+ - name: from
224
+ dtype: string
225
+ - name: value
226
+ dtype: string
227
+ - name: image_relpath
228
+ dtype: string
229
+ - name: image
230
+ dtype: image
231
+ splits:
232
+ - name: train
233
+ num_bytes: 7607522167.776
234
+ num_examples: 44474
235
+ download_size: 7377834021
236
+ dataset_size: 7607522167.776
237
+ - config_name: part2
238
+ features:
239
+ - name: id
240
+ dtype: string
241
+ - name: conversations
242
+ list:
243
+ - name: from
244
+ dtype: string
245
+ - name: value
246
+ dtype: string
247
+ - name: image_relpath
248
+ dtype: string
249
+ - name: image
250
+ dtype: image
251
+ splits:
252
+ - name: train
253
+ num_bytes: 33365826763.55
254
+ num_examples: 44475
255
+ download_size: 33245002109
256
+ dataset_size: 33365826763.55
257
+ - config_name: part20
258
+ features:
259
+ - name: id
260
+ dtype: string
261
+ - name: conversations
262
+ list:
263
+ - name: from
264
+ dtype: string
265
+ - name: value
266
+ dtype: string
267
+ - name: image_relpath
268
+ dtype: string
269
+ - name: image
270
+ dtype: image
271
+ splits:
272
+ - name: train
273
+ num_bytes: 7179023719.416
274
+ num_examples: 44474
275
+ download_size: 6145041547
276
+ dataset_size: 7179023719.416
277
+ - config_name: part3
278
+ features:
279
+ - name: id
280
+ dtype: string
281
+ - name: conversations
282
+ list:
283
+ - name: from
284
+ dtype: string
285
+ - name: value
286
+ dtype: string
287
+ - name: image_relpath
288
+ dtype: string
289
+ - name: image
290
+ dtype: image
291
+ splits:
292
+ - name: train
293
+ num_bytes: 24289174922.225
294
+ num_examples: 44475
295
+ download_size: 15536031938
296
+ dataset_size: 24289174922.225
297
+ - config_name: part4
298
+ features:
299
+ - name: id
300
+ dtype: string
301
+ - name: conversations
302
+ list:
303
+ - name: from
304
+ dtype: string
305
+ - name: value
306
+ dtype: string
307
+ - name: image_relpath
308
+ dtype: string
309
+ - name: image
310
+ dtype: image
311
+ splits:
312
+ - name: train
313
+ num_bytes: 6699346145.65
314
+ num_examples: 44475
315
+ download_size: 6646535744
316
+ dataset_size: 6699346145.65
317
+ - config_name: part5
318
+ features:
319
+ - name: id
320
+ dtype: string
321
+ - name: conversations
322
+ list:
323
+ - name: from
324
+ dtype: string
325
+ - name: value
326
+ dtype: string
327
+ - name: image_relpath
328
+ dtype: string
329
+ - name: image
330
+ dtype: image
331
+ splits:
332
+ - name: train
333
+ num_bytes: 6651458829.325
334
+ num_examples: 44475
335
+ download_size: 4896128944
336
+ dataset_size: 6651458829.325
337
+ - config_name: part6
338
+ features:
339
+ - name: id
340
+ dtype: string
341
+ - name: conversations
342
+ list:
343
+ - name: from
344
+ dtype: string
345
+ - name: value
346
+ dtype: string
347
+ - name: image_relpath
348
+ dtype: string
349
+ - name: image
350
+ dtype: image
351
+ splits:
352
+ - name: train
353
+ num_bytes: 18039837306.175
354
+ num_examples: 44475
355
+ download_size: 17786918533
356
+ dataset_size: 18039837306.175
357
+ - config_name: part7
358
+ features:
359
+ - name: id
360
+ dtype: string
361
+ - name: conversations
362
+ list:
363
+ - name: from
364
+ dtype: string
365
+ - name: value
366
+ dtype: string
367
+ - name: image_relpath
368
+ dtype: string
369
+ - name: image
370
+ dtype: image
371
+ splits:
372
+ - name: train
373
+ num_bytes: 17852309246.925
374
+ num_examples: 44475
375
+ download_size: 17774641127
376
+ dataset_size: 17852309246.925
377
+ - config_name: part8
378
+ features:
379
+ - name: id
380
+ dtype: string
381
+ - name: conversations
382
+ list:
383
+ - name: from
384
+ dtype: string
385
+ - name: value
386
+ dtype: string
387
+ - name: image_relpath
388
+ dtype: string
389
+ - name: image
390
+ dtype: image
391
+ splits:
392
+ - name: train
393
+ num_bytes: 17873005196.425
394
+ num_examples: 44475
395
+ download_size: 16131787376
396
+ dataset_size: 17873005196.425
397
+ - config_name: part9
398
+ features:
399
+ - name: id
400
+ dtype: string
401
+ - name: conversations
402
+ list:
403
+ - name: from
404
+ dtype: string
405
+ - name: value
406
+ dtype: string
407
+ - name: image_relpath
408
+ dtype: string
409
+ - name: image
410
+ dtype: image
411
+ splits:
412
+ - name: train
413
+ num_bytes: 34314021153.775
414
+ num_examples: 44475
415
+ download_size: 41828939793
416
+ dataset_size: 34314021153.775
417
+ configs:
418
+ - config_name: part1
419
+ data_files:
420
+ - split: train
421
+ path: part1/train-*
422
+ - config_name: part10
423
+ data_files:
424
+ - split: train
425
+ path: part10/train-*
426
+ - config_name: part11
427
+ data_files:
428
+ - split: train
429
+ path: part11/train-*
430
+ - config_name: part12
431
+ data_files:
432
+ - split: train
433
+ path: part12/train-*
434
+ - config_name: part13
435
+ data_files:
436
+ - split: train
437
+ path: part13/train-*
438
+ - config_name: part14
439
+ data_files:
440
+ - split: train
441
+ path: part14/train-*
442
+ - config_name: part15
443
+ data_files:
444
+ - split: train
445
+ path: part15/train-*
446
+ - config_name: part16
447
+ data_files:
448
+ - split: train
449
+ path: part16/train-*
450
+ - config_name: part17
451
+ data_files:
452
+ - split: train
453
+ path: part17/train-*
454
+ - config_name: part18
455
+ data_files:
456
+ - split: train
457
+ path: part18/train-*
458
+ - config_name: part19
459
+ data_files:
460
+ - split: train
461
+ path: part19/train-*
462
+ - config_name: part2
463
+ data_files:
464
+ - split: train
465
+ path: part2/train-*
466
+ - config_name: part20
467
+ data_files:
468
+ - split: train
469
+ path: part20/train-*
470
+ - config_name: part3
471
+ data_files:
472
+ - split: train
473
+ path: part3/train-*
474
+ - config_name: part4
475
+ data_files:
476
+ - split: train
477
+ path: part4/train-*
478
+ - config_name: part5
479
+ data_files:
480
+ - split: train
481
+ path: part5/train-*
482
+ - config_name: part6
483
+ data_files:
484
+ - split: train
485
+ path: part6/train-*
486
+ - config_name: part7
487
+ data_files:
488
+ - split: train
489
+ path: part7/train-*
490
+ - config_name: part8
491
+ data_files:
492
+ - split: train
493
+ path: part8/train-*
494
+ - config_name: part9
495
+ data_files:
496
+ - split: train
497
+ path: part9/train-*
498
+ ---
499
+
500
+ # CoVT Dataset
501
+
502
+ This repository contains the training data for **Chain-of-Visual-Thought (CoVT)**, a framework proposed in the paper "[Chain-of-Visual-Thought: Teaching VLMs to See and Think Better with Continuous Visual Tokens](https://huggingface.co/papers/2511.19418)".
503
+
504
+ CoVT enables Vision-Language Models (VLMs) to reason not only in words but also through **continuous visual tokens**—compact latent representations that encode rich perceptual cues. This framework addresses the limitations of current VLMs in capturing dense visual information across spatial dimensions, which is essential for perceptual understanding, spatial reasoning, and geometric awareness.
505
+
506
+ - **Paper:** [https://huggingface.co/papers/2511.19418](https://huggingface.co/papers/2511.19418)
507
+ - **Project Page:** https://wakalsprojectpage.github.io/comt-website
508
+ - **Code:** https://github.com/Wakals/CoMT
509
+
510
+ ## Dataset Composition
511
+
512
+ The CoVT dataset is designed to provide dense supervision signals for training VLMs. It utilizes some subsets of LLaVA-OneVision and merges the filtered TallyQA dataset and ADE20K-Depth from Aurora.
513
+
514
+ During training, a VLM equipped with CoVT autoregressively predicts these visual tokens to reconstruct dense supervision signals such as depth, segmentation, edges, and DINO features.
515
+
516
+ ![Data Image](https://github.com/Wakals/CoMT/raw/main/assets/data.png)
517
+
518
+ ## Sample Usage
519
+
520
+ To load a specific part of the CoVT dataset, you can use the Hugging Face `datasets` library:
521
+
522
+ ```python
523
+ from datasets import load_dataset
524
+
525
+ # Load a specific configuration (e.g., 'part1')
526
+ dataset = load_dataset("Wakals/CoVT-Dataset", "part1")
527
+
528
+ # Access the training split
529
+ train_data = dataset["train"]
530
+
531
+ # Print the first example
532
+ print(train_data[0])
533
+
534
+ # Example of accessing image and conversations from a sample
535
+ # first_example = train_data[0]
536
+ # image_obj = first_example["image"] # This will be a PIL Image object
537
+ # print(f"Image format: {image_obj.format}, size: {image_obj.size}")
538
+ # print(f"Conversations: {first_example['conversations']}")
539
+ ```