Hugging Face's logo

Datasets:
Tera-AI
/
STRIDE

Tasks:
Image-to-Text
Modalities:
Image
Text
Formats:
parquet
Size:
1M - 10M
ArXiv:
Tags:
autonomous-driving
road-images
spatio-temporal
Libraries:
Datasets
Dask
License:
Dataset card Data Studio Files
xet
Community
18
Dataset Viewer
Auto-converted to Parquet
image
imagewidth (px)
512
512
file_name
stringclasses
16 values
path_index
int64
10k
108k
node_index
int64
0
15
split
stringclasses
1 value
latitude
float64
40.7
40.8
longitude
float64
-73.89
-73.79
month
int64
1
12
year
int64
2.01k
2.02k
move
float64
0
48.6
heading
float64
0
360
month_action
int64
0
11
year_action
int64
-12
12
node_0.jpg
10,000
0
train
40.73795
-73.79939
7
2,019
0.2
93.3
1
3
node_1.jpg
10,000
1
train
40.73795
-73.79939
8
2,022
0.5
92.7
1
-9
node_2.jpg
10,000
2
train
40.73795
-73.79938
9
2,013
0.2
92.2
0
-2
node_3.jpg
10,000
3
train
40.73795
-73.79938
9
2,011
0.2
94
11
7
node_4.jpg
10,000
4
train
40.73795
-73.79938
8
2,018
9
92.7
0
4
node_5.jpg
10,000
5
train
40.73795
-73.79927
8
2,022
9.8
92.7
0
0
node_6.jpg
10,000
6
train
40.73794
-73.79915
8
2,022
10.1
93.2
2
0
node_7.jpg
10,000
7
train
40.73794
-73.79903
10
2,022
11.3
233
0
0
node_8.jpg
10,000
8
train
40.73788
-73.79914
10
2,022
7.4
351.8
10
0
node_9.jpg
10,000
9
train
40.73794
-73.79915
8
2,022
10.1
93.2
2
0
node_10.jpg
10,000
10
train
40.73794
-73.79903
10
2,022
11.3
233
0
0
node_11.jpg
10,000
11
train
40.73788
-73.79914
10
2,022
7.4
351.8
10
0
node_12.jpg
10,000
12
train
40.73794
-73.79915
8
2,022
9.8
272.6
0
0
node_13.jpg
10,000
13
train
40.73795
-73.79927
8
2,022
9
272.8
0
-4
node_14.jpg
10,000
14
train
40.73795
-73.79938
8
2,018
0.2
274.1
1
-7
node_15.jpg
10,000
15
train
40.73795
-73.79938
9
2,011
0
0
0
0
node_0.jpg
100,000
0
train
40.74612
-73.87728
8
2,024
10.4
302.3
0
0
node_1.jpg
100,000
1
train
40.74617
-73.87738
8
2,024
1.1
225.7
8
-12
node_2.jpg
100,000
2
train
40.74616
-73.87739
4
2,012
0.2
299.6
6
9
node_3.jpg
100,000
3
train
40.74617
-73.87739
10
2,021
1.1
300.5
0
-7
node_4.jpg
100,000
4
train
40.74617
-73.87741
10
2,014
0.6
300.7
1
3
node_5.jpg
100,000
5
train
40.74617
-73.87741
11
2,017
0.6
300.3
9
-4
node_6.jpg
100,000
6
train
40.74618
-73.87742
8
2,013
0.1
299.1
1
9
node_7.jpg
100,000
7
train
40.74618
-73.87742
9
2,022
8.8
309.6
11
2
node_8.jpg
100,000
8
train
40.74623
-73.8775
8
2,024
6
299.8
0
0
node_9.jpg
100,000
9
train
40.74626
-73.87756
8
2,024
9.9
49.2
0
0
node_10.jpg
100,000
10
train
40.74631
-73.87747
8
2,024
10.6
18.9
0
0
node_11.jpg
100,000
11
train
40.7464
-73.87743
8
2,024
13.5
23.1
0
0
node_12.jpg
100,000
12
train
40.74652
-73.87737
8
2,024
10.1
24.1
0
0
node_13.jpg
100,000
13
train
40.7466
-73.87732
8
2,024
12.9
24.1
0
0
node_14.jpg
100,000
14
train
40.74671
-73.87726
8
2,024
10.6
24.1
0
0
node_15.jpg
100,000
15
train
40.74679
-73.8772
8
2,024
0
0
0
0
node_0.jpg
100,001
0
train
40.74612
-73.87728
8
2,024
10.4
302.3
0
0
node_1.jpg
100,001
1
train
40.74617
-73.87738
8
2,024
1.1
225.7
8
-12
node_2.jpg
100,001
2
train
40.74616
-73.87739
4
2,012
0.2
299.6
6
9
node_3.jpg
100,001
3
train
40.74617
-73.87739
10
2,021
1.1
300.5
0
-7
node_4.jpg
100,001
4
train
40.74617
-73.87741
10
2,014
0.6
300.7
1
3
node_5.jpg
100,001
5
train
40.74617
-73.87741
11
2,017
0.6
300.3
9
-4
node_6.jpg
100,001
6
train
40.74618
-73.87742
8
2,013
0.1
299.1
1
9
node_7.jpg
100,001
7
train
40.74618
-73.87742
9
2,022
8.8
309.6
11
2
node_8.jpg
100,001
8
train
40.74623
-73.8775
8
2,024
6
299.8
0
0
node_9.jpg
100,001
9
train
40.74626
-73.87756
8
2,024
11.3
299.3
0
0
node_10.jpg
100,001
10
train
40.7463
-73.87768
8
2,024
10.5
299.3
0
0
node_11.jpg
100,001
11
train
40.74635
-73.87779
8
2,024
9.9
299.3
0
0
node_12.jpg
100,001
12
train
40.74639
-73.87789
8
2,024
9.8
299.3
0
0
node_13.jpg
100,001
13
train
40.74644
-73.87799
8
2,024
9.8
299.3
0
0
node_14.jpg
100,001
14
train
40.74648
-73.87809
8
2,024
10
299.3
0
0
node_15.jpg
100,001
15
train
40.74653
-73.8782
8
2,024
0
0
0
0
node_0.jpg
100,002
0
train
40.74612
-73.87728
8
2,024
9
122.3
0
0
node_1.jpg
100,002
1
train
40.74608
-73.87719
8
2,024
9.5
122.3
0
0
node_2.jpg
100,002
2
train
40.74603
-73.87709
8
2,024
9.5
120.9
0
0
node_3.jpg
100,002
3
train
40.74599
-73.87699
8
2,024
9.5
120.5
0
0
node_4.jpg
100,002
4
train
40.74594
-73.8769
8
2,024
9.6
120.4
0
0
node_5.jpg
100,002
5
train
40.7459
-73.8768
8
2,024
6.7
120.4
2
-10
node_6.jpg
100,002
6
train
40.74587
-73.87673
10
2,014
1.3
120.4
6
-2
node_7.jpg
100,002
7
train
40.74586
-73.87671
4
2,012
0.1
120.3
6
10
node_8.jpg
100,002
8
train
40.74586
-73.87671
10
2,022
0
100.6
1
-5
node_9.jpg
100,002
9
train
40.74586
-73.87671
11
2,017
0.4
120.7
10
-4
node_10.jpg
100,002
10
train
40.74586
-73.87671
9
2,013
1
120.5
1
8
node_11.jpg
100,002
11
train
40.74586
-73.8767
10
2,021
0.5
122.2
10
3
node_12.jpg
100,002
12
train
40.74585
-73.87669
8
2,024
9.9
120.4
0
0
node_13.jpg
100,002
13
train
40.74581
-73.87659
8
2,024
10.1
120.4
0
0
node_14.jpg
100,002
14
train
40.74576
-73.87649
8
2,024
6.7
120
2
-10
node_15.jpg
100,002
15
train
40.74573
-73.87642
10
2,014
0
0
0
0
node_0.jpg
100,003
0
train
40.74617
-73.87738
8
2,024
1.1
225.7
8
-12
node_1.jpg
100,003
1
train
40.74616
-73.87739
4
2,012
0.2
299.6
6
9
node_2.jpg
100,003
2
train
40.74617
-73.87739
10
2,021
1.1
300.5
0
-7
node_3.jpg
100,003
3
train
40.74617
-73.87741
10
2,014
0.6
300.7
1
3
node_4.jpg
100,003
4
train
40.74617
-73.87741
11
2,017
0.6
300.3
9
-4
node_5.jpg
100,003
5
train
40.74618
-73.87742
8
2,013
0.1
299.1
1
9
node_6.jpg
100,003
6
train
40.74618
-73.87742
9
2,022
8.8
309.6
11
2
node_7.jpg
100,003
7
train
40.74623
-73.8775
8
2,024
15.4
207.5
0
0
node_8.jpg
100,003
8
train
40.7461
-73.87758
8
2,024
10.3
203.8
0
0
node_9.jpg
100,003
9
train
40.74602
-73.87763
8
2,024
9.4
203.7
0
0
node_10.jpg
100,003
10
train
40.74594
-73.87768
8
2,024
9.5
203.8
0
0
node_11.jpg
100,003
11
train
40.74586
-73.87773
8
2,024
9.9
203.8
0
0
node_12.jpg
100,003
12
train
40.74578
-73.87777
8
2,024
10.5
203.7
0
0
node_13.jpg
100,003
13
train
40.7457
-73.87782
8
2,024
8.9
203.8
0
0
node_14.jpg
100,003
14
train
40.74562
-73.87787
8
2,024
13.6
283.8
0
0
node_15.jpg
100,003
15
train
40.74565
-73.87802
8
2,024
0
0
0
0
node_0.jpg
100,004
0
train
40.74617
-73.87738
8
2,024
1.1
225.7
8
-12
node_1.jpg
100,004
1
train
40.74616
-73.87739
4
2,012
0.2
299.6
6
9
node_2.jpg
100,004
2
train
40.74617
-73.87739
10
2,021
1.1
300.5
0
-7
node_3.jpg
100,004
3
train
40.74617
-73.87741
10
2,014
0.6
300.7
1
3
node_4.jpg
100,004
4
train
40.74617
-73.87741
11
2,017
0.6
300.3
9
-4
node_5.jpg
100,004
5
train
40.74618
-73.87742
8
2,013
0.1
299.1
1
9
node_6.jpg
100,004
6
train
40.74618
-73.87742
9
2,022
8.8
309.6
11
2
node_7.jpg
100,004
7
train
40.74623
-73.8775
8
2,024
15.4
207.5
0
0
node_8.jpg
100,004
8
train
40.7461
-73.87758
8
2,024
10.3
203.8
0
0
node_9.jpg
100,004
9
train
40.74602
-73.87763
8
2,024
9.4
203.7
0
0
node_10.jpg
100,004
10
train
40.74594
-73.87768
8
2,024
9.5
203.8
0
0
node_11.jpg
100,004
11
train
40.74586
-73.87773
8
2,024
9.9
203.8
0
0
node_12.jpg
100,004
12
train
40.74578
-73.87777
8
2,024
10.5
203.7
0
0
node_13.jpg
100,004
13
train
40.7457
-73.87782
8
2,024
8.9
203.8
0
0
node_14.jpg
100,004
14
train
40.74562
-73.87787
8
2,024
8.5
203.6
2
-2
node_15.jpg
100,004
15
train
40.74555
-73.87791
10
2,022
0
0
0
0
node_0.jpg
100,005
0
train
40.74617
-73.87738
8
2,024
1.1
225.7
8
-12
node_1.jpg
100,005
1
train
40.74616
-73.87739
4
2,012
0.2
299.6
6
9
node_2.jpg
100,005
2
train
40.74617
-73.87739
10
2,021
1.1
300.5
0
-7
node_3.jpg
100,005
3
train
40.74617
-73.87741
10
2,014
0.6
300.7
1
3
End of preview. Expand in Data Studio

STRIDE

We develop STRIDE (Spatio-Temporal Road Image Dataset for Exploration), that consists of approximately 82B tokens which were arranged into a total of 6M visual "sentences" or token sequences. The sequences are generated from a relatively small set of 131k panoramic images, along with their metadata and openly available highway system data. Our approach enables a 27x increase in information, allowing for generative world model training. In essence, STRIDE turns real street-view panoramic observations into a navigable, interactive environment suitable for free-play across space and time.

For ease of use and data exploration, we prepared this sample of 10k detokenized paths, which amounts to about 100k projected panoramic images along with its corresponding metadata.

Breakdown

Dataset (STRIDE)

The full tokenized dataset is made available through two downloadable files in a public GCS bucket:

gsutil -m cp gs://tera-tardis/STRIDE-1/training.jsonl . # ~327GB
gsutil -m cp gs://tera-tardis/STRIDE-1/testing.jsonl . # ~9GB

In the future, the fully detokenized dataset will be made available. Should you need it, feel free to contact the authors.

San Mateo Coverage Map

image/png

Above is the 70km^2 area we selected for putting together the Google StreetView data using openly available road data. Each directly connected component of the graph is represented with a distinct color, for ease of visualization.

Queens Coverage Map

image/png

Above is an additional Queens, NY 27km^2 area. Each directly connected component of the graph is represented with a distinct color, for ease of visualization.

Code (TARDIS)

The code used for training of the model is available on GitHub.

Checkpoints (TARDIS)

The checkpoint/state used for evaluation of the model was saved in MessagePack format and is made available through this downloadable file:

gsutil -m cp gs://tera-tardis/STRIDE-1/checkpoint.msgpack . # ~10GB

Should you need other checkpoints, feel free to contact the authors.

Project Website

The project website is available here.

Contacts

Paper

TARDIS STRIDE: A Spatio-Temporal Road Image Dataset for Exploration and Autonomy

Downloads last month
2,227
Size of downloaded dataset files:
50.7 GB
Size of the auto-converted Parquet files:
50.7 GB
Number of rows:
1,704,003