Hugging Face's logo

Datasets:
silentlooop
/
revt2v-data

Modalities:
Tabular
Text
Video
Formats:
json
Size:
< 1K
Tags:
video
text-to-video
video-diffusion
latents
Libraries:
Datasets
pandas
Polars
Dataset card Data Studio Files
xet
 Community
1
Dataset Viewer
Auto-converted to Parquet Duplicate
shard
stringlengths
20
20
prompt
stringlengths
19
68
prompt_index
int64
0
148
seed
int64
42
488
motion_score
float64
0.05
16.1
kept
bool
2 classes
negative_prompt
stringclasses
1 value
00000_s0_299e0555.pt
A person is archery
0
42
1.18216
true
watermark, text
00000_s1_299e0555.pt
A person is archery
0
43
0.844077
true
watermark, text
00000_s2_299e0555.pt
A person is archery
0
44
2.227724
true
watermark, text
00001_s0_91558930.pt
A person is arranging flowers
1
45
1.136687
true
watermark, text
00001_s1_91558930.pt
A person is arranging flowers
1
46
0.616208
true
watermark, text
00001_s2_91558930.pt
A person is arranging flowers
1
47
2.29818
true
watermark, text
00002_s0_c7ef4e69.pt
A person is bungee jumping
2
48
2.156478
true
watermark, text
00002_s1_c7ef4e69.pt
A person is bungee jumping
2
49
2.001419
true
watermark, text
00002_s2_c7ef4e69.pt
A person is bungee jumping
2
50
8.467844
true
watermark, text
00003_s0_7c9925c5.pt
A person is catching or throwing baseball
3
51
0.582844
true
watermark, text
00003_s1_7c9925c5.pt
A person is catching or throwing baseball
3
52
0.539787
true
watermark, text
00003_s2_7c9925c5.pt
A person is catching or throwing baseball
3
53
0.851243
true
watermark, text
00004_s0_311eeece.pt
A person is catching or throwing frisbee
4
54
0.881821
true
watermark, text
00004_s1_311eeece.pt
A person is catching or throwing frisbee
4
55
4.200144
true
watermark, text
00004_s2_311eeece.pt
A person is catching or throwing frisbee
4
56
0.633325
true
watermark, text
00005_s0_d7f4ef83.pt
A person is clay pottery making
5
57
1.112597
true
watermark, text
00005_s1_d7f4ef83.pt
A person is clay pottery making
5
58
1.868756
true
watermark, text
00005_s2_d7f4ef83.pt
A person is clay pottery making
5
59
3.958992
true
watermark, text
00006_s0_3b3c9c5c.pt
A person is climbing a rope
6
60
2.954818
true
watermark, text
00006_s1_3b3c9c5c.pt
A person is climbing a rope
6
61
5.534921
true
watermark, text
00006_s2_3b3c9c5c.pt
A person is climbing a rope
6
62
1.683119
true
watermark, text
00007_s0_f85956af.pt
A person is cutting watermelon
7
63
1.3639
true
watermark, text
00007_s1_f85956af.pt
A person is cutting watermelon
7
64
1.626607
true
watermark, text
00007_s2_f85956af.pt
A person is cutting watermelon
7
65
0.953176
true
watermark, text
00008_s0_b93536c9.pt
A person is digging
8
66
2.561192
true
watermark, text
00008_s1_b93536c9.pt
A person is digging
8
67
2.529052
true
watermark, text
00008_s2_b93536c9.pt
A person is digging
8
68
2.937348
true
watermark, text
00009_s0_bd3ef054.pt
A person is drawing
9
69
0.417732
true
watermark, text
00009_s1_bd3ef054.pt
A person is drawing
9
70
0.934517
true
watermark, text
00009_s2_bd3ef054.pt
A person is drawing
9
71
0.744535
true
watermark, text
00010_s0_00af929e.pt
A person is dunking basketball
10
72
0.154806
true
watermark, text
00010_s1_00af929e.pt
A person is dunking basketball
10
73
0.241118
true
watermark, text
00010_s2_00af929e.pt
A person is dunking basketball
10
74
0.047387
false
watermark, text
00011_s0_0978ff36.pt
A person is eating watermelon
11
75
1.439049
true
watermark, text
00011_s1_0978ff36.pt
A person is eating watermelon
11
76
0.228446
true
watermark, text
00011_s2_0978ff36.pt
A person is eating watermelon
11
77
2.564956
true
watermark, text
00012_s0_ce49e205.pt
A person is getting a haircut
12
78
1.286358
true
watermark, text
00012_s1_ce49e205.pt
A person is getting a haircut
12
79
1.496405
true
watermark, text
00012_s2_ce49e205.pt
A person is getting a haircut
12
80
3.254119
true
watermark, text
00013_s0_c514cce4.pt
A person is making bed
13
81
1.794641
true
watermark, text
00013_s1_c514cce4.pt
A person is making bed
13
82
1.208933
true
watermark, text
00013_s2_c514cce4.pt
A person is making bed
13
83
1.039699
true
watermark, text
00014_s0_032c2216.pt
A person is peeling apples
14
84
1.364456
true
watermark, text
00014_s1_032c2216.pt
A person is peeling apples
14
85
3.101135
true
watermark, text
00014_s2_032c2216.pt
A person is peeling apples
14
86
1.900577
true
watermark, text
00015_s0_2c01aba5.pt
A person is planting trees
15
87
0.915653
true
watermark, text
00015_s1_2c01aba5.pt
A person is planting trees
15
88
0.497404
true
watermark, text
00015_s2_2c01aba5.pt
A person is planting trees
15
89
2.100971
true
watermark, text
00016_s0_f5585c59.pt
A person is pushing cart
16
90
2.889263
true
watermark, text
00016_s1_f5585c59.pt
A person is pushing cart
16
91
0.489518
true
watermark, text
00016_s2_f5585c59.pt
A person is pushing cart
16
92
2.888967
true
watermark, text
00017_s0_f8743813.pt
A person is riding a bike
17
93
0.767804
true
watermark, text
00017_s1_f8743813.pt
A person is riding a bike
17
94
1.811525
true
watermark, text
00017_s2_f8743813.pt
A person is riding a bike
17
95
3.500055
true
watermark, text
00018_s0_958fcc38.pt
A person is rock climbing
18
96
0.268922
true
watermark, text
00018_s1_958fcc38.pt
A person is rock climbing
18
97
2.138784
true
watermark, text
00018_s2_958fcc38.pt
A person is rock climbing
18
98
0.924958
true
watermark, text
00019_s0_02649c7d.pt
A person is shooting basketball
19
99
1.1005
true
watermark, text
00019_s1_02649c7d.pt
A person is shooting basketball
19
100
0.072547
false
watermark, text
00019_s2_02649c7d.pt
A person is shooting basketball
19
101
0.918074
true
watermark, text
00020_s0_3cc3f0bf.pt
A person is shooting goal (soccer)
20
102
0.35517
true
watermark, text
00020_s1_3cc3f0bf.pt
A person is shooting goal (soccer)
20
103
1.898778
true
watermark, text
00020_s2_3cc3f0bf.pt
A person is shooting goal (soccer)
20
104
0.095024
false
watermark, text
00021_s0_7ec4d6f7.pt
A person is skydiving
21
105
8.18767
true
watermark, text
00021_s1_7ec4d6f7.pt
A person is skydiving
21
106
8.804457
true
watermark, text
00021_s2_7ec4d6f7.pt
A person is skydiving
21
107
6.806562
true
watermark, text
00022_s0_a87b972c.pt
A person is throwing axe
22
108
2.364505
true
watermark, text
00022_s1_a87b972c.pt
A person is throwing axe
22
109
1.533345
true
watermark, text
00022_s2_a87b972c.pt
A person is throwing axe
22
110
3.100005
true
watermark, text
00023_s0_e36ef2a0.pt
A person is writing
23
111
0.501566
true
watermark, text
00023_s1_e36ef2a0.pt
A person is writing
23
112
0.247149
true
watermark, text
00023_s2_e36ef2a0.pt
A person is writing
23
113
2.600271
true
watermark, text
00024_s0_d2390135.pt
a bear climbing a tree
24
114
0.586744
true
watermark, text
00024_s1_d2390135.pt
a bear climbing a tree
24
115
0.864273
true
watermark, text
00024_s2_d2390135.pt
a bear climbing a tree
24
116
0.145573
false
watermark, text
00025_s0_50957f66.pt
a bicycle accelerating to gain speed
25
117
3.225526
true
watermark, text
00025_s1_50957f66.pt
a bicycle accelerating to gain speed
25
118
2.203533
true
watermark, text
00025_s2_50957f66.pt
a bicycle accelerating to gain speed
25
119
5.405633
true
watermark, text
00026_s0_f54a6b63.pt
a bicycle slowing down to stop
26
120
1.513406
true
watermark, text
00026_s1_f54a6b63.pt
a bicycle slowing down to stop
26
121
1.171897
true
watermark, text
00026_s2_f54a6b63.pt
a bicycle slowing down to stop
26
122
0.360997
true
watermark, text
00027_s0_b1b76cc9.pt
a bird building a nest from twigs and leaves
27
123
0.181227
true
watermark, text
00027_s1_b1b76cc9.pt
a bird building a nest from twigs and leaves
27
124
0.126944
false
watermark, text
00027_s2_b1b76cc9.pt
a bird building a nest from twigs and leaves
27
125
0.163744
true
watermark, text
00028_s0_500c0406.pt
a boat accelerating to gain speed
28
126
2.91222
true
watermark, text
00028_s1_500c0406.pt
a boat accelerating to gain speed
28
127
5.778992
true
watermark, text
00028_s2_500c0406.pt
a boat accelerating to gain speed
28
128
0.953442
true
watermark, text
00029_s0_a8a1220f.pt
a boat slowing down to stop
29
129
8.283382
true
watermark, text
00029_s1_a8a1220f.pt
a boat slowing down to stop
29
130
0.55352
true
watermark, text
00029_s2_a8a1220f.pt
a boat slowing down to stop
29
131
4.72313
true
watermark, text
00030_s0_f83b19ef.pt
a bus accelerating to gain speed
30
132
5.86517
true
watermark, text
00030_s1_f83b19ef.pt
a bus accelerating to gain speed
30
133
7.165332
true
watermark, text
00030_s2_f83b19ef.pt
a bus accelerating to gain speed
30
134
1.022537
true
watermark, text
00031_s0_79354338.pt
a bus turning a corner
31
135
0.794208
true
watermark, text
00031_s1_79354338.pt
a bus turning a corner
31
136
1.278783
true
watermark, text
00031_s2_79354338.pt
a bus turning a corner
31
137
0.802404
true
watermark, text
00032_s0_d921378f.pt
a car slowing down to stop
32
138
0.939909
true
watermark, text
00032_s1_d921378f.pt
a car slowing down to stop
32
139
1.266399
true
watermark, text
00032_s2_d921378f.pt
a car slowing down to stop
32
140
2.123797
true
watermark, text
00033_s0_b34f37a2.pt
a car turning a corner
33
141
3.082647
true
watermark, text
End of preview. Expand in Data Studio

revt2v-data

Precomputed latent dataset for revT2V (reverse-time video diffusion distillation). Each .pt file is one (prompt, seed) clip generated from the frozen ali-vilab/text-to-video-ms-1.7b teacher pipeline. Every record is a dict with:

  • prompt -- the text prompt
  • latents -- forward-time latents, shape (C, F, H, W)
  • reversed_latents -- the same latents flipped along the frame axis (the training target)
  • prompt_embeds / negative_prompt_embeds -- cached text-encoder outputs

Structure

  • dataset1/ -- 154 shards (the original dataset)
  • dataset2/ -- 416 shards + manifest.jsonl (prompt, seed, motion_score, kept)

Loading 

from revt2v.data import LatentDataset

ds = LatentDataset("path/to/dataset1")  # or dataset2
record = ds[0]
Downloads last month
274
Number of rows:
447
Total file size:
331 MB

Models trained or fine-tuned on silentlooop/revt2v-data