Dataset Viewer
Auto-converted to Parquet Duplicate
task_id
stringlengths
5
8
kind
stringclasses
4 values
type
stringclasses
5 values
frames
int64
144
1.44k
prompt
stringlengths
918
6.8k
media
listlengths
0
5
output_spec
stringlengths
97
270
entities
stringlengths
61
2.97k
B250_135
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Student bedroom / study"}]}
B250_136
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Classroom"}]}
B250_137
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Informal learning lounge"}]}
B250_139
image
spatial_arkit
240
Rebuild the place shown in the 4 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg", "media/photo4.jpg" ]
{"fps": 24, "frame_count": 240, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [49, 240], "pan_360": [49, 144], "top_down_orbit": [145, 240], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Home study / guest room"}]}
B250_140
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one place by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view m...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Children's playground in a residential courtyard"}]}
B250_141
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Bright student bedroom and study"}]}
B250_142
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Attic loft apartment"}]}
B250_143
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Residential galley kitchen"}]}
B250_144
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Shared-flat kitchen"}]}
B250_145
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a phone moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view mat...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 480], "final_resolution": [640, 480], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Period sitting room"}]}
B250_146
image
spatial_arkit
240
Rebuild the place shown in the 4 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg", "media/photo4.jpg" ]
{"fps": 24, "frame_count": 240, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [49, 240], "pan_360": [49, 144], "top_down_orbit": [145, 240], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Small meeting room"}]}
B250_147
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Hotel double room"}]}
B250_148
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Electronics lab"}]}
B250_149
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Seminar classroom"}]}
B250_150
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Wood-panelled lodge / hostel lounge and hallway"}]}
B250_151
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a camera moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view ma...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Office-building break room"}]}
B250_152
image
spatial_arkit
240
Rebuild the place shown in the 4 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken on one street by the cameras of a car driving through it, each from a different position and a moment apart. Build the place and everything in it once, at real-world size, and place one ca...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg", "media/photo4.jpg" ]
{"fps": 24, "frame_count": 240, "preview_resolution": [640, 360], "final_resolution": [640, 360], "orbit": {"frames": [49, 240], "pan_360": [49, 144], "top_down_orbit": [145, 240], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Downtown San Francisco intersection by a supermarket"}]}
B250_153
image
spatial_arkit
228
Rebuild the place shown in the 3 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken on one street by the cameras of a car driving through it, each from a different position and a moment apart. Build the place and everything in it once, at real-world size, and place one ca...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg" ]
{"fps": 24, "frame_count": 228, "preview_resolution": [640, 360], "final_resolution": [640, 360], "orbit": {"frames": [37, 228], "pan_360": [37, 132], "top_down_orbit": [133, 228], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Roadworks lane at a historic stone building corner"}]}
B250_154
image
spatial_arkit
252
Rebuild the place shown in the 5 attached photos as one static 3D proxy blockout, with one camera per photo. The photos were taken in one room by a phone moving through it, each from a different position. Build the place and everything in it once, at real-world size, and place one camera per photo so that its view mat...
[ "media/photo1.jpg", "media/photo2.jpg", "media/photo3.jpg", "media/photo4.jpg", "media/photo5.jpg" ]
{"fps": 24, "frame_count": 252, "preview_resolution": [640, 480], "final_resolution": [640, 480], "orbit": {"frames": [61, 252], "pan_360": [61, 156], "top_down_orbit": [157, 252], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Cottage living room"}]}
B250_155
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Renovated open-plan living room"}]}
B250_156
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x480. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 480], "final_resolution": [640, 480], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Sitting-room window wall"}]}
B250_157
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x480. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 480], "final_resolution": [640, 480], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Child bedroom"}]}
B250_158
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x480. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 480], "final_resolution": [640, 480], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "U-shaped home kitchen"}]}
B250_159
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Office"}]}
B250_160
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Lodge / hostel lounge"}]}
B250_161
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Office"}]}
B250_162
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Hotel double room"}]}
B250_163
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Small office"}]}
B250_164
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Informal learning lounge"}]}
B250_165
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. The photo was taken on a street from a car. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 360], "final_resolution": [640, 360], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "San Francisco arterial"}]}
B250_167
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Home study / guest room"}]}
B250_168
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x426. The photo is seen through the camera fo...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 426], "final_resolution": [640, 426], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Children's playground in a residential courtyard"}]}
B250_169
image
spatial_arkit
204
Rebuild the place shown in the attached photo as one static 3D proxy blockout, with one camera. The photo was taken on a street from a car. Build the place and everything in it at real-world size, and place the camera so that its view matches the photo: position, direction and lens. Output: 204 frames at 24 fps, 640x...
[ "media/photo.jpg" ]
{"fps": 24, "frame_count": 204, "preview_resolution": [640, 360], "final_resolution": [640, 360], "orbit": {"frames": [13, 204], "pan_360": [13, 108], "top_down_orbit": [109, 204], "eye_height_m": 1.6, "pan_hfov_deg": 70, "orbit_elevation_deg": 55, "hide_above_m": 2.2}}
{"entities": [], "spaces": [{"id": "S01", "name": "Street corner at a downtown intersection"}]}
B250_170
continuation
continuation
336
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 336 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 336, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "The seated woman beside the bed, with short reddish hair and a loose pale pink long-sleeved garment with a broad light-colored collar."}, {"id": "P02", "class": "person", "description": "The standing man to her right, with short dark hair and a receding hai...
B250_172
continuation
continuation
480
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 480 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 480, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman on the left in the photo, with short curled blonde hair, a shimmering white dress with puffed sleeves and a wide flared skirt, and light-colored heeled shoes."}, {"id": "P02", "class": "person", "description": "Man on the right wearing a jeweled crown...
B250_174
continuation
continuation
367
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 367 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 367, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Adult between the two firefighters at the opening of the right-hand cage, with shoulder-length dark brown hair, a tan one-piece suit, black shoulder harness, red rectangular waist tag and a number 2 patch on the leg."}, {"id": "P02", "class": "person", "des...
B250_175
continuation
continuation
336
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 336 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 336, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person nearest the camera in a burgundy shirt, loose light-blue jeans, and gray-and-white lace-up sneakers with thick white soles."}, {"id": "P02", "class": "person", "description": "Person running behind in a loose black short-sleeved shirt, light-blue jea...
B250_176
continuation
continuation
353
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 353 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 353, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Light-skinned, apparently clean-shaven person near the centre-left of the photo, wearing a backward brown-tan cap, pale short-sleeved shirt, tan tactical vest, camouflage trousers and dark gloves."}, {"id": "P02", "class": "person", "description": "Light-sk...
B250_178
continuation
continuation
240
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 240 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 240, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "The man on the right in the photo, with thick, straight dark hair, a black uniform-style jacket with gold-trimmed cuffs and shoulder details, and black trousers."}, {"id": "P02", "class": "person", "description": "The man on the left in the photo, with shor...
B250_179
continuation
continuation
360
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 360 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 360, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Pale-skinned person with long auburn curls, a cream open cardigan, pale pink top, light blue flared jeans and dark shoes with white soles."}, {"id": "O01", "class": "other:object", "description": "Shallow rectangular turquoise object with a dark lower secti...
B250_180
continuation
continuation
384
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 384 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 384, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person with medium-length black hair and a red headband, wearing a blue basketball jersey with red-and-white edging over a white short-sleeved shirt."}, {"id": "P02", "class": "person", "description": "Short-haired basketball player in a black sleeveless je...
B250_181
continuation
continuation
216
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 216 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 216, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Skateboarder slightly right of centre on the upper deck, behind the railing, wearing an orange short-sleeved T-shirt, dark trousers and a light gray cap with a dark brim."}, {"id": "V01", "class": "skateboard", "description": "Gray skateboard held in front ...
B250_182
continuation
continuation
144
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 144 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 144, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman at center-right beside the hood, with dark hair pulled back and an open gray jacket over a dark high-neck top."}, {"id": "P02", "class": "person", "description": "Man behind the crossover, left of P01, with short dark hair, a tan long-sleeved uniform ...
B250_183
continuation
continuation
288
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 288 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 288, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Adult woman in the right foreground beside P02, with shoulder-length brown hair, a cream cardigan over a magenta top with a decorated neckline, dark trousers and brown open-toe shoes."}, {"id": "P02", "class": "person", "description": "Young girl beside P01...
B250_184
continuation
continuation
276
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 276 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 276, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Barefoot person with short dark hair, a bright blue short-sleeved shirt and red, white and blue plaid shorts; center-left beside the net, with an arm and knee raised in frame 1."}, {"id": "P02", "class": "person", "description": "Barefoot person with short ...
B250_185
continuation
continuation
168
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 168 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 168, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Boy in a blue-gray and red plaid shirt over a pale T-shirt, standing right of center by the open fridge in frame 1."}, {"id": "P02", "class": "person", "description": "Person seen as a blurred forearm and hand in the left foreground of frame 1, and walking ...
B250_186
continuation
continuation
288
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 288 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 288, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "The foreground racer airborne at centre-right, with dark hair tied back, a charcoal sleeveless top, bright pink shorts and bright pink athletic shoes."}, {"id": "P02", "class": "person", "description": "The short-dark-haired racer behind and to the left of ...
B250_187
continuation
continuation
240
Continue the attached photo as a proxy previs. Frame 1 must reproduce the photo: the camera's position, direction and lens, and every visible subject and object where it is in the photo. Then play out what happens next as described below, with the camera behaviour and the cuts at the stated times. Output: 240 frames ...
[ "media/frame1.jpg" ]
{"fps": 24, "frame_count": 240, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman filling the center-left of frame 1 in left-facing profile, with dark hair in a high ponytail and a loose black pullover hoodie with black drawstrings and a small chest logo; black trousers."}, {"id": "O01", "class": "ball", "description": "Orange bask...
B250_189
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Blonde woman with her hair tied in a ponytail, wearing a black short-sleeved polo shirt and dark trousers."}, {"id": "O01", "class": "dog", "description": "Medium-sized dog with grey, black, and white fur and semi-erect ears."}], "spaces": [{"id": "S01", "n...
B250_190
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman with long blonde hair, wearing a white short-sleeved t-shirt and dark patterned leggings, barefoot."}, {"id": "O01", "class": "dog", "description": "Black dog with a white patch on its chest, wearing a green collar."}], "spaces": [{"id": "S01", "name"...
B250_191
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Male rider wearing a grey jacket, white breeches, black riding boots, and a black helmet."}, {"id": "P02", "class": "person", "description": "Woman with brown hair in a ponytail wearing a dark quilted vest over a grey long-sleeved shirt."}, {"id": "P03", "c...
B250_192
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Female with curly dark brown hair, appearing from childhood through adulthood."}, {"id": "P02", "class": "person", "description": "Woman with dark shoulder-length hair, wearing an orange-brown long cardigan."}, {"id": "P03", "class": "person", "description"...
B250_193
video
video
1,008
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1008, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Young man wearing a white graphic t-shirt, black pants, black-rimmed glasses, and a black beanie."}, {"id": "P02", "class": "person", "description": "Young man wearing a black baseball cap, a grey raglan t-shirt with black sleeves, and grey pants."}, {"id":...
B250_194
video
video
1,200
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1200, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Young man wearing a light blue short-sleeved t-shirt with a chest pocket, khaki pants, red sneakers, and a grey cap."}, {"id": "P02", "class": "person", "description": "Man with short dark hair and stubble, wearing a white t-shirt with black text, black pan...
B250_195
video
video
601
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 601, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "V01", "class": "race_car", "description": "Stock race car painted bright yellow and red with the number 22 on the roof and sides."}, {"id": "V02", "class": "race_car", "description": "Stock race car painted bright yellow and light blue with the number 21 on the roof."}, {"id": "V03", "class": "rac...
B250_196
video
video
793
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 793, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing an orange and yellow tie-dye short-sleeved t-shirt, bright red knee-length shorts, and a large black metal frame stacked with pink boxes strapped to their back."}, {"id": "P02", "class": "person", "description": "Woman wearing a blue cardigan...
B250_197
video
video
600
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 600, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman in a yellow-to-orange gradient Hanfu robe with wide sleeves and hair styled up with traditional ornaments."}, {"id": "P02", "class": "person", "description": "Woman in a maroon Hanfu robe with gold belt and gold embroidery along the skirt hem."}, {"id...
B250_198
video
video
600
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 600, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing a bright blue ski jacket, grey snow pants, white ski boots, a red helmet, and holding ski poles."}, {"id": "P02", "class": "person", "description": "Person wearing a red ski suit with white accents, a black and white helmet, dark goggles, and...
B250_199
video
video
1,009
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1009, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing a black and white patterned hooded winter jacket, black snow pants, black helmet, and goggles."}, {"id": "P02", "class": "person", "description": "Snowboarder wearing a bright yellow hooded sweatshirt with a blue chest emblem, black snow pant...
B250_202
video
video
673
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 673, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man wearing a backwards grey baseball cap, a dark grey hoodie, grey sweatpants, and dark sneakers, with a dark beard."}, {"id": "P02", "class": "person", "description": "Man with a mustache wearing a black cap, sunglasses, a white short-sleeved polo shirt, ...
B250_203
video
video
829
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 829, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Young man with short dark hair wearing a black t-shirt with white text on the front and light grey sweatpants."}, {"id": "P02", "class": "person", "description": "Young man with dark hair with bangs wearing a black tracksuit jacket with white stripes along ...
B250_204
video
video
888
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 888, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Young man with short black hair, wearing a black long-sleeve sweatshirt with a circular green camouflage graphic on the chest, a white undershirt showing at the collar, and dark pants."}, {"id": "P02", "class": "person", "description": "Person wearing an ol...
B250_205
video
video
1,104
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1104, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "A Black man with short hair and a goatee, wearing a light tan jacket with red stripes along the sleeves, tan trousers, white sneakers, and a tan backpack."}, {"id": "P02", "class": "person", "description": "A short-haired fair-skinned man wearing a short-sl...
B250_206
video
video
1,056
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1056, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short dark hair, wearing a black padded jacket, dark cargo pants, and dark boots."}, {"id": "P02", "class": "person", "description": "Police officer wearing a navy uniform with shoulder patches and a navy baseball cap with police insignia."}, {"id"...
B250_207
video
video
1,068
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1068, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing a purple short-sleeved t-shirt, black trousers, and a black baseball cap."}, {"id": "P02", "class": "person", "description": "Person wearing a grey short-sleeved t-shirt, tan trousers, and a black baseball cap."}, {"id": "P03", "class": "pers...
B250_208
video
video
625
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 625, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short greying hair, wearing a black suit jacket, white dress shirt, black necktie, black trousers, and holding a handgun."}, {"id": "P02", "class": "person", "description": "Man with short dark hair, wearing an olive-green jacket over a dark t-shir...
B250_209
video
video
1,104
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1104, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "A young woman with shoulder-length dark hair and bangs, wearing a beige trench coat over a black turtleneck top, black leggings, black heeled boots, and carrying a shoulder bag with a red, white, and blue striped strap."}, {"id": "P02", "class": "person", "...
B250_210
video
video
579
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 579, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short dark hair and beard stubble, wearing a brown suit jacket over a dark shirt, holding a black handgun."}, {"id": "P02", "class": "person", "description": "Man with short blonde hair, wearing a light grey suit, white collared shirt, grey pattern...
B250_211
video
video
768
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 768, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short dark hair wearing a white short-sleeved t-shirt with black text on the front and back, grey sweatpants, white wristbands, and white sneakers."}, {"id": "P02", "class": "person", "description": "Man wearing a white short-sleeved t-shirt with b...
B250_212
video
video
648
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 648, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short dark hair wearing a black zip jacket over a black shirt and dark trousers."}, {"id": "P02", "class": "person", "description": "Police officer wearing a black tactical uniform, vest, and a peaked cap with a red-and-white checkered band, positi...
B250_213
video
video
310
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 310, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Woman with long dark hair wearing a black face mask, a dark jacket, and white wired earphones."}, {"id": "P02", "class": "person", "description": "Person with reddish hair holding a small silver/white toy gun with an orange tip."}], "spaces": [{"id": "S01",...
B250_214
video
video
1,104
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1104, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing a black hooded jacket over a black t-shirt with red lettering, black pants, and dark sneakers with light-colored soles."}, {"id": "P02", "class": "person", "description": "Man with wavy light-brown hair wearing an open dark jacket over a dark...
B250_215
video
video
1,152
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1152, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with dark swept-back hair, wearing a black leather jacket, dark trousers, and black gloves, riding without a helmet."}, {"id": "P02", "class": "person", "description": "Police officer wearing a white helmet and dark blue jacket with reflective stripes, ...
B250_216
video
video
346
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 346, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man wearing a white short-sleeved collared shirt, dark brown shorts, a brown baseball cap, and brown shoes."}, {"id": "P02", "class": "person", "description": "Man with short black hair wearing an oversized black short-sleeved t-shirt, a silver necklace, li...
B250_217
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man with short brown hair wearing a light blue sleeveless shirt, black athletic shorts, and blue running shoes."}, {"id": "P02", "class": "person", "description": "Man wearing a dark helmet, light blue sleeveless shirt, camouflage shorts, and a bright blue ...
B250_218
video
video
1,080
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 1080, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Person wearing a dark green hooded coat with red trim and lining, beige trousers, and brown boots."}, {"id": "P02", "class": "person", "description": "Person wearing a dark blue hooded coat, grey trousers, and brown boots."}, {"id": "P03", "class": "person"...
B250_219
video
video
336
Rebuild the attached reference video as a proxy previs. Match what the video shows: the places and what stands where, which way things face, how every listed subject moves and what it does when, and the camera of every shot (where it stands, how it frames and moves), with the cuts and transitions at the same times as ...
[ "media/reference.mp4" ]
{"fps": 24, "frame_count": 336, "preview_resolution": [640, 360], "final_resolution": [640, 360]}
{"entities": [{"id": "P01", "class": "person", "description": "Man wearing a solid black helmet, black long-sleeved shirt, and black trousers."}, {"id": "P02", "class": "person", "description": "Man wearing a patterned helmet, black long-sleeved shirt, and dark trousers."}, {"id": "V01", "class": "bicycle", "descriptio...
End of preview. Expand in Data Studio

VibePrevisBench

VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender

🌐 Website · 📄 Paper · 📑 arXiv · 💻 GitHub · 📊 Benchmark overview · 🏆 Leaderboard · 🎬 Case studies

Evaluating Agentic Coding for 4D Scene Generation in Blender

VibePrevisBench evaluates whether coding agents can turn a brief into an executable 4D Blender scene: who is where, when they move, how each shot is framed, and where the cuts happen. Given a script, photos, an opening frame to continue, or a reference video, an agent writes and runs Blender Python, renders previews, and delivers a greybox scene and its proxy video. Evaluation checks both the underlying scene and what is visible on screen.

The benchmark contains 242 English tasks, spanning four input kinds, seven scene domains and 33 places, with 84,406 scorable hidden ground-truth facts covering spatial layout, actions, cameras and physics. This Hugging Face repository hosts task briefs, annotations and source manifests. The GitHub repository provides the generation harness, media retrieval tools, evaluation pipeline and results.

Dataset at a glance

Task distribution across four input kinds and seven scene domains, with example inputs and generated scenes

Inner ring: input kinds. Outer ring: scene domains. Example cards pair task inputs with frames from GPT-6 Astra's generated proxy videos.

Input kind kind Agent input Tasks
Script text Script / text brief 94
Photos image Reference photos and text 39
Continuation continuation Opening frame and continuation brief 23
Reference video video Reference video and text 86
Total 242

The seven scene domains are homes, offices & schools, streets & cities, shops & hotels, industry & transit, sports & arts, and nature & water.

Task statistics

Distributions of brief length, output length, task requirements and checks per task

The panels summarize brief lengths, output lengths, common task requirements and check counts. 212 tasks require two or more shots; 194 require a moving camera; 192 involve two or more people.

Inputs and ground truth

Examples of script, photo, continuation and video tasks, their hidden ground-truth facts and generated proxy frames

Each column shows an agent input, selected hidden facts with evaluation-dimension tags, and a generated proxy frame. These figures illustrate the benchmark; complete third-party source images and videos are retrieved separately.

Quickstart

Browse the 242 tasks

from datasets import load_dataset

tasks = load_dataset("Kling-Keye/vibeprevisbench", "tasks", split="test")
print(tasks)
print(tasks[0])

Install the reader with pip install datasets. The tasks configuration exposes all 242 tasks in the test split. output_spec and entities in viewer rows are JSON-encoded strings; per-task JSON files contain the structured objects. Viewer rows do not include rubric files or full input media. Download the repository snapshot for evaluation.

If accessing a private copy, authenticate first with hf auth login using an account with access.

Download annotations and verify the release

The recommended path uses the verification tool in the code repository:

git clone https://github.com/vibeprevisbench/vibeprevisbench.git
cd vibeprevisbench
pip install huggingface_hub

# Download task briefs, annotations and manifests; verify SHA256SUMS.
python vpbench.py download --repo Kling-Keye/vibeprevisbench --data data --data-only

# Retrieve third-party input photos and videos from their sources.
pip install -r requirements.txt
python vpbench.py fetch-media --data data
python vpbench.py fetch-media --data data --check

Media reconstruction also requires ffmpeg and ffprobe (5.1 or newer) on your PATH; see the full setup instructions.

The 215 third-party input images/videos are not bundled as standalone media files in this repository. external.json records their sources, extraction recipes, fingerprints and hashes; NOTICE.md records attribution and third-party terms. Some scan-dataset photos require an authorized local copy; see external-input documentation.

To download the complete snapshot directly:

from huggingface_hub import snapshot_download

root = snapshot_download(
    repo_id="Kling-Keye/vibeprevisbench",
    repo_type="dataset",
    local_dir="data/vibeprevisbench",
)

Dataset structure

tasks/index.json             task IDs, input kinds, frame counts and paths
tasks/tasks.jsonl            viewer rows (tasks configuration, test split)
tasks/<ID>/prompt.txt        English generation brief
tasks/<ID>/task.json         structured brief, output specification and media references
tasks/<ID>/entities.json     places and subjects to tag in the scene
rubrics/<ID>/               evaluator annotations; never expose to the evaluated agent
  task.json                 original brief pinned by the scorer
  gt.json                   ground-truth scene facts
  claims.json               script-task claims (94 files)
  gt_facts.json             checkable facts, where applicable
  checklist.json            per-check dimensions and time windows, where applicable
  scene_frames.json         reference time alignment, where applicable
  scene_reconciled.json     per-frame visible subjects and relations (3 tasks); not used in scoring
  refs/*.npz                derived reference features, where applicable
categories.json             category IDs for scene annotation
agent_instructions.txt      instructions appended to generation briefs
external.json               source links, extraction recipes, fingerprints and hashes
assets/                     dataset-card figures and branding
NOTICE.md                   source attributions and third-party terms
LICENSE                     CC BY-NC-SA 4.0
SHA256SUMS                  checksums of distributed files

Task IDs are stable. Index paths are relative to the dataset root; media paths in each task.json are relative to tasks/<ID>/. See the dataset documentation for file conventions and the separation between agent inputs and evaluator annotations.

Evaluation

Agents deliver a standalone Blender Python program. Evaluation replays it in a fresh sandbox and checks the executable scene and rendered proxy video:

Score What it measures
Scene Check Rule-based checks against hidden ground-truth facts in nine dimensions: entities, layout, spaces, picture structure, routes, events, camera, cuts and physical validity.
Screen Check VLM-based checks of identity, action, visibility and story in the proxy video, conditioned on facts that passed Scene Check.
Overall The mean of Scene and Screen scores; both are on a 0–100 scale.

Failed or missing submissions count as zero in the 242-task average. Keep rubrics/ hidden from the evaluated agent. For the full setup and commands, follow Evaluate your agent on GitHub. Results and examples are on the interactive leaderboard and case-study gallery.

Photo-task protocol: the released generation briefs add an eight-second orbit to the photo views. rubrics/<ID>/task.json retains the original photo-only brief used in the paper, and evaluation remains pinned to that brief. Report whether you use the release or paper protocol; see GitHub for details.

License and responsible use

Task briefs, annotations and manifests are released under CC BY-NC-SA 4.0. Third-party input media remain with their owners and are not covered by this license. See NOTICE.md for source attribution and terms; do not redistribute third-party media without permission.

Some reference videos show real people. Use this benchmark for research on scene reconstruction, not to identify or profile the people shown. For removal requests, open a discussion here or contact the project through the project website.

Citation

arXiv:2611.15789 · Read the latest paper (PDF).

Please cite VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender (2026).

@misc{vibeprevisbench2026,
  title         = {VibePrevisBench: Evaluating Agentic Coding for 4D Scene Generation in Blender},
  author        = {Yuzhe Chen and Muxi Diao and Lele Yang and Wenbin Xu and Xuyang Liu and Qile Su and Changyi Liu and Jiankang Chen and Kaiyu Jiang and Haoyu Zhang and Hongyi Fu and Jianxiong Wang and Tianke Zhang and Haonan Fan and Yingxin Li and Bingyi Li and Xu Wang and Xuanyu Zheng and Bin Wen and Tingting Gao and Wentao Zhang and Jiale Cao},
  year          = {2026},
  eprint        = {2611.15789},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url           = {https://arxiv.org/abs/2611.15789}
}

Citation metadata is maintained in the code repository's CITATION.cff.

Downloads last month
2,041

Paper for Kling-Keye/vibeprevisbench