Dataset Viewer
Auto-converted to Parquet Duplicate
episode_id
large_stringlengths
9
9
split
large_stringclasses
6 values
scene
large_stringclasses
4 values
instruction
large_stringclasses
186 values
target_category
large_stringclasses
11 values
target_id
large_stringclasses
209 values
spawn_tier
large_stringclasses
3 values
spawn_dist_m
float64
1.51
10.4
num_steps
int64
15
845
trajectory_length_m
float64
0.54
10.2
final_ne_m
float64
0.67
1
collision_count
int64
0
0
template_id
large_stringclasses
24 values
ep_000001
train_id
office.usd
Move toward the monitor.
monitor
monitor_041
near
1.662
58
0.699
0.9639
0
train_04
ep_000002
test_id
office.usd
Move to the monitor and halt.
monitor
monitor_024
mid
5.578
294
4.632
0.9597
0
train_10
ep_000003
train_id
office.usd
Go to the monitor.
monitor
monitor_028
near
2.115
85
1.149
0.9677
0
train_01
ep_000004
train_id
office.usd
Stop next to the table.
table
table_054
mid
5.86
313
4.937
0.9624
0
train_18
ep_000005
test_ood_obj
office.usd
Locate the fire extinguisher and stop in front of it.
fire_extinguisher
fire_extinguisher_004
near
1.558
55
0.632
0.9685
0
train_15
ep_000006
train_id
office.usd
Drive to the monitor and stop.
monitor
monitor_021
mid
4.839
248
3.87
0.9696
0
train_02
ep_000007
train_id
office.usd
Move all the way to the monitor.
monitor
monitor_039
mid
5.728
302
4.768
0.9663
0
train_17
ep_000008
test_ood_lang
office.usd
Find the table and come to a stop.
table
table_046
mid
6.662
358
5.704
0.9649
0
ood_03
ep_000009
val_id
office.usd
Move toward the trash can.
trash_can
trash_can_016
mid
6.346
340
5.398
0.9635
0
train_04
ep_000010
train_id
office.usd
Go to the monitor.
monitor
monitor_036
mid
4.245
506
5.215
0.9645
0
train_01
ep_000011
test_ood_lang
office.usd
Proceed to the table.
table
table_013
near
1.93
195
0.96
0.9788
0
ood_02
ep_000012
train_id
office.usd
Move to the table and halt.
table
table_049
near
3.342
159
2.384
0.9601
0
train_10
ep_000013
train_id
office.usd
Approach the monitor.
monitor
monitor_016
near
1.847
70
0.898
0.964
0
train_03
ep_000014
train_id
office.usd
Your destination is the chair.
chair
chair_029
near
1.995
78
1.036
0.961
0
train_14
ep_000015
train_id
office.usd
Stop next to the trash can.
trash_can
trash_can_017
near
2.906
143
1.989
0.9646
0
train_18
ep_000016
train_id
office.usd
Move all the way to the monitor.
monitor
monitor_030
mid
4.293
218
3.357
0.9666
0
train_17
ep_000017
test_ood_obj
office.usd
Move to the fire extinguisher and halt.
fire_extinguisher
fire_extinguisher_003
near
2.504
118
1.592
0.9617
0
train_10
ep_000018
train_id
office.usd
Go to the monitor.
monitor
monitor_018
near
3.16
150
2.222
0.9632
0
train_01
ep_000019
test_id
office.usd
Locate the monitor and stop in front of it.
monitor
monitor_036
mid
4.981
258
4.032
0.9621
0
train_15
ep_000020
train_id
office.usd
Head to the trash can.
trash_can
trash_can_020
near
2.814
127
1.83
0.9859
0
train_09
ep_000021
train_id
office.usd
Approach the trash can.
trash_can
trash_can_017
near
2.592
116
1.655
0.9593
0
train_03
ep_000022
train_id
office.usd
Move all the way to the monitor.
monitor
monitor_026
near
1.584
53
0.618
0.967
0
train_17
ep_000023
train_id
office.usd
Go to the trash can in front of you.
trash_can
trash_can_016
near
3.007
139
2.05
0.9618
0
train_11
ep_000024
test_ood_obj
office.usd
Head to the fire extinguisher.
fire_extinguisher
fire_extinguisher_005
mid
3.769
184
2.796
0.9703
0
train_09
ep_000025
train_id
office.usd
Go to the monitor in front of you.
monitor
monitor_028
near
2.927
135
1.981
0.9693
0
train_11
ep_000026
train_id
office.usd
Your destination is the trash can.
trash_can
trash_can_017
mid
3.727
184
2.791
0.9677
0
train_14
ep_000027
test_id
office.usd
Move to the monitor and halt.
monitor
monitor_037
mid
3.606
175
2.646
0.9605
0
train_10
ep_000028
train_id
office.usd
Approach the table.
table
table_050
mid
5.195
270
4.234
0.9638
0
train_03
ep_000029
val_id
office.usd
Drive toward the monitor until you reach it.
monitor
monitor_044
mid
4.74
243
3.788
0.9643
0
train_12
ep_000030
train_id
office.usd
Locate the table and stop in front of it.
table
table_048
near
2.468
107
1.516
0.9639
0
train_15
ep_000031
val_id
office.usd
Approach the table.
table
table_046
near
2.601
114
1.636
0.9673
0
train_03
ep_000032
train_id
office.usd
Head to the monitor.
monitor
monitor_017
near
2.418
105
1.478
0.9637
0
train_09
ep_000033
train_id
office.usd
Navigate to the monitor.
monitor
monitor_024
mid
5.064
262
4.102
0.9677
0
train_05
ep_000034
train_id
office.usd
Go to the chair in front of you.
chair
chair_030
mid
4.766
246
3.828
0.9662
0
train_11
ep_000035
test_ood_obj
office.usd
Go to the fire extinguisher.
fire_extinguisher
fire_extinguisher_006
near
1.86
70
0.901
0.9629
0
train_01
ep_000036
train_id
office.usd
Navigate to the monitor.
monitor
monitor_025
mid
5.544
291
4.589
0.9653
0
train_05
ep_000037
train_id
office.usd
Move all the way to the table.
table
table_050
near
1.775
68
0.851
0.9675
0
train_17
ep_000038
test_ood_obj
office.usd
Go to the fire extinguisher in front of you.
fire_extinguisher
fire_extinguisher_001
near
2.564
119
1.619
0.9944
0
train_11
ep_000039
test_id
office.usd
Approach the chair.
chair
chair_030
near
2.74
122
1.774
0.9702
0
train_03
ep_000040
train_id
office.usd
Move all the way to the trash can.
trash_can
trash_can_002
near
2.311
144
1.334
0.9995
0
train_17
ep_000041
test_ood_lang
office.usd
Find the table and come to a stop.
table
table_044
near
3.191
150
2.232
0.9644
0
ood_03
ep_000042
train_id
office.usd
Drive to the table and stop.
table
table_051
near
2.26
96
1.325
0.9625
0
train_02
ep_000043
test_ood_lang
office.usd
Get closer to the monitor.
monitor
monitor_037
near
2.261
94
1.302
0.9681
0
ood_11
ep_000044
train_id
office.usd
Drive to the monitor and stop.
monitor
monitor_044
near
1.514
320
2.577
0.9974
0
train_02
ep_000045
test_id
office.usd
Navigate to the monitor.
monitor
monitor_041
near
2.817
128
1.868
0.9662
0
train_05
ep_000046
train_id
office.usd
Navigate to the monitor.
monitor
monitor_031
mid
6.245
333
5.288
0.961
0
train_05
ep_000047
train_id
office.usd
Approach the trash can.
trash_can
trash_can_016
mid
4.19
210
3.232
0.9635
0
train_03
ep_000048
train_id
office.usd
Navigate to the monitor.
monitor
monitor_035
near
2.415
104
1.468
0.9633
0
train_05
ep_000049
train_id
office.usd
Head to the monitor.
monitor
monitor_035
mid
5.72
301
4.759
0.9674
0
train_09
ep_000050
test_ood_lang
office.usd
Park next to the trash can.
trash_can
trash_can_020
mid
5.257
317
4.28
0.9975
0
ood_12
ep_000051
test_ood_lang
office.usd
Find your way to the monitor.
monitor
monitor_016
near
3.088
146
2.156
0.9684
0
ood_09
ep_000052
train_id
office.usd
Move all the way to the trash can.
trash_can
trash_can_016
near
2.265
94
1.299
0.9677
0
train_17
ep_000053
train_id
office.usd
Move toward the monitor.
monitor
monitor_034
near
3.133
149
2.209
0.9631
0
train_04
ep_000054
val_id
office.usd
Approach the trash can.
trash_can
trash_can_023
near
2.136
87
1.18
0.9622
0
train_03
ep_000055
train_id
office.usd
Go to the trash can.
trash_can
trash_can_017
near
2.673
120
1.727
0.962
0
train_01
ep_000056
train_id
office.usd
Move all the way to the monitor.
monitor
monitor_023
near
1.585
54
0.631
0.961
0
train_17
ep_000057
test_id
office.usd
Drive toward the rack until you reach it.
rack
rack_016
near
2.861
130
1.9
0.9637
0
train_12
ep_000058
train_id
office.usd
Stop next to the monitor.
monitor
monitor_005
mid
4.206
211
3.252
0.9617
0
train_18
ep_000059
test_ood_obj
office.usd
Navigate to the fire extinguisher.
fire_extinguisher
fire_extinguisher_006
near
1.6
57
0.669
0.9631
0
train_05
ep_000060
train_id
office.usd
Stop next to the monitor.
monitor
monitor_034
mid
5.331
280
4.393
0.9679
0
train_18
ep_000061
train_id
office.usd
Get to the trash can.
trash_can
trash_can_016
near
2.787
127
1.843
0.9686
0
train_13
ep_000062
train_id
office.usd
Head to the monitor.
monitor
monitor_024
mid
5.718
301
4.756
0.9634
0
train_09
ep_000063
train_id
office.usd
Move to the monitor and halt.
monitor
monitor_018
mid
6
560
5.243
0.9627
0
train_10
ep_000064
val_id
office.usd
Go to the table in front of you.
table
table_045
near
1.94
76
0.996
0.9683
0
train_11
ep_000065
train_id
office.usd
Go to the trash can in front of you.
trash_can
trash_can_015
mid
4.115
206
3.168
0.9679
0
train_11
ep_000066
test_ood_lang
office.usd
Park next to the monitor.
monitor
monitor_037
mid
6.887
371
5.926
0.9614
0
ood_12
ep_000067
train_id
office.usd
Your destination is the monitor.
monitor
monitor_010
mid
5.557
297
4.622
0.9647
0
train_14
ep_000068
train_id
office.usd
Get to the trash can.
trash_can
trash_can_023
mid
4.426
224
3.467
0.9654
0
train_13
ep_000069
test_ood_lang
office.usd
Make your way to the chair.
chair
chair_030
mid
4.219
211
3.254
0.9673
0
ood_01
ep_000070
train_id
office.usd
Your destination is the monitor.
monitor
monitor_010
near
1.682
127
0.734
0.9613
0
train_14
ep_000071
train_id
office.usd
Drive to the table and stop.
table
table_051
mid
6.041
321
5.084
0.9601
0
train_02
ep_000072
train_id
office.usd
Get to the chair.
chair
chair_028
near
3.214
152
2.269
0.9693
0
train_13
ep_000073
test_id
office.usd
Go to the chair in front of you.
chair
chair_030
mid
3.529
170
2.568
0.9645
0
train_11
ep_000074
test_ood_lang
office.usd
Close in on the trash can.
trash_can
trash_can_011
near
2.361
103
1.381
0.9949
0
ood_06
ep_000075
test_id
office.usd
Move toward the chair.
chair
chair_028
mid
4.591
228
3.586
0.9659
0
train_04
ep_000076
train_id
office.usd
Drive to the chair and stop.
chair
chair_030
near
2.777
127
1.842
0.9666
0
train_02
ep_000077
train_id
office.usd
Move toward the monitor.
monitor
monitor_026
near
1.883
71
0.921
0.9696
0
train_04
ep_000078
train_id
office.usd
Drive to the monitor and stop.
monitor
monitor_048
near
1.962
77
1.013
0.9702
0
train_02
ep_000079
train_id
office.usd
Drive to the trash can and stop.
trash_can
trash_can_015
mid
5.492
290
4.561
0.9633
0
train_02
ep_000080
train_id
office.usd
Drive to the table and stop.
table
table_048
mid
5.149
268
4.197
0.9627
0
train_02
ep_000081
train_id
office.usd
Go to the trash can.
trash_can
trash_can_021
near
3.267
158
2.35
0.9619
0
train_01
ep_000082
train_id
office.usd
Move to the rack and halt.
rack
rack_002
near
1.805
70
0.888
0.9612
0
train_10
ep_000083
test_ood_lang
office.usd
Close in on the monitor.
monitor
monitor_018
mid
6.645
369
5.728
0.9683
0
ood_06
ep_000084
val_id
office.usd
Head to the table.
table
table_045
near
3.06
144
2.123
0.9609
0
train_09
ep_000085
train_id
office.usd
Move toward the monitor.
monitor
monitor_022
near
3.029
141
2.081
0.9595
0
train_04
ep_000086
test_ood_obj
office.usd
Locate the fire extinguisher and stop in front of it.
fire_extinguisher
fire_extinguisher_004
mid
4.21
212
3.262
0.9672
0
train_15
ep_000087
test_id
office.usd
Locate the monitor and stop in front of it.
monitor
monitor_038
mid
5.088
274
4.147
0.9639
0
train_15
ep_000088
val_id
office.usd
Your destination is the monitor.
monitor
monitor_020
near
2.822
128
1.863
0.9609
0
train_14
ep_000089
train_id
office.usd
Locate the rack and stop in front of it.
rack
rack_016
near
1.616
67
0.694
0.9675
0
train_15
ep_000090
test_id
office.usd
Move all the way to the sofa.
sofa
sofa_018
near
2.872
131
1.918
0.9635
0
train_17
ep_000091
train_id
office.usd
Navigate to the table.
table
table_048
mid
4.099
206
3.16
0.9689
0
train_05
ep_000092
train_id
office.usd
Drive to the monitor and stop.
monitor
monitor_047
mid
6
318
5.035
0.9659
0
train_02
ep_000093
train_id
office.usd
Get to the chair.
chair
chair_030
near
2.445
109
1.536
0.9592
0
train_13
ep_000094
train_id
office.usd
Drive to the chair and stop.
chair
chair_030
mid
3.619
211
2.678
0.9597
0
train_02
ep_000095
train_id
office.usd
Locate the monitor and stop in front of it.
monitor
monitor_017
near
2.089
87
1.169
0.9612
0
train_15
ep_000096
test_ood_lang
office.usd
Find your way to the table.
table
table_046
mid
4.914
255
3.975
0.9597
0
ood_09
ep_000097
test_ood_obj
office.usd
Your destination is the fire extinguisher.
fire_extinguisher
fire_extinguisher_005
mid
3.693
180
2.738
0.9598
0
train_14
ep_000098
train_id
office.usd
Your destination is the trash can.
trash_can
trash_can_016
mid
5.369
280
4.403
0.9667
0
train_14
ep_000099
train_id
office.usd
Go to the monitor in front of you.
monitor
monitor_025
near
1.514
50
0.561
0.9626
0
train_11
ep_000100
test_id
office.usd
Navigate to the monitor.
monitor
monitor_041
mid
4.942
251
3.942
0.9695
0
train_05
End of preview. Expand in Data Studio

MiniVLA-Nav v1

A Multi-Scene Simulation Dataset for Language-Conditioned Robot Navigation


Demo

Nova Carter navigating to named objects across all four Isaac Sim environments.


Dataset Summary

MiniVLA-Nav v1 is a simulation dataset for the Language-Conditioned Object Approach (LCOA) task: given a short natural-language instruction, an NVIDIA Nova Carter differential-drive robot must navigate to the named object and stop within 1 m. Data were collected inside four photorealistic NVIDIA Isaac Sim 5.1 environments (Office, Hospital, Full Warehouse, Warehouse with Multiple Shelves).

Each of the 1,174 episodes pairs a language instruction with per-timestep, synchronized multimodal observations:

Modality Resolution / Shape Format
Front RGB 640 Γ— 640 Γ— 3, uint8 PNG
Metric depth 640 Γ— 640, float32 (metres) NumPy
Instance segmentation 640 Γ— 640, uint16 PNG
Continuous actions (v, Ο‰) T Γ— 2, float32 NumPy
Tokenized actions (7Γ—7) T Γ— 2, int16 NumPy
Robot poses (x,y,z,qw,qx,qy,qz) T Γ— 7, float32 NumPy

All sensors operate at 60 Hz (physics Ξ”t = 1/60 s).


Supported Tasks

  • Language-Conditioned Object Approach (LCOA) β€” given a natural-language goal and front RGB-D observations, predict continuous (v, Ο‰) or discrete 7Γ—7 action tokens to drive a differential-drive robot within 1 m of the named object.
  • Behaviour Cloning / Imitation Learning β€” dense per-step expert labels enable direct supervised training.
  • OOD Generalisation β€” structured evaluation splits test template-paraphrase and object-category out-of-distribution robustness.

Multimodal Observations

Each timestep provides synchronized RGB, metric depth (float32, metres), and instance segmentation. The composites below show RGB (left) and depth colormap (right) from a mid-episode step.

Office Hospital
RGB+D office RGB+D hospital
Full Warehouse Warehouse (Multi-Shelf)
RGB+D full warehouse RGB+D warehouse shelves

Depth strip β€” consecutive frames from an office episode, showing depth (metres) as the robot approaches the target:

Depth strip office


Scenes

Four photorealistic Isaac Sim environments, each with curated seen/held-out object categories:

Office

Contact sheet β€” Office

Hospital

Contact sheet β€” Hospital

Full Warehouse

Contact sheet β€” Full Warehouse

Warehouse (Multiple Shelves)

Contact sheet β€” Warehouse Multi-Shelf

Scene Episodes Seen Categories Held-out Categories
Office 281 chair, sofa, table, monitor, plant, trash_can fire_extinguisher, whiteboard
Hospital 22 chair, trash_can fire_extinguisher, whiteboard
Full Warehouse 54 shelf, rack barrel
Warehouse (Multi-Shelf) 68 shelf, rack barrel

Object Categories

12 categories total β€” 9 seen during training, 3 held out for OOD evaluation.

Seen categories:

chair monitor table trash can
chair monitor table trash can
rack crate shelf barrel (OOD)
rack crate shelf barrel

Held-out (OOD): fire_extinguisher, whiteboard, barrel β€” appear only in test_ood_obj split.


Object Category Demo

All object categories navigated to in the Office scene.


Dataset Structure

v1/
β”œβ”€β”€ dataset_meta.json            # Global metadata (scenes, camera, action space, splits)
β”œβ”€β”€ assets/                      # README visual assets
β”œβ”€β”€ splits/
β”‚   β”œβ”€β”€ train_id.txt             # 261 episode IDs
β”‚   β”œβ”€β”€ val_id.txt               #  41 episode IDs
β”‚   β”œβ”€β”€ test_id.txt              #  50 episode IDs
β”‚   β”œβ”€β”€ test_ood_obj.txt         #  37 episode IDs  (held-out object categories)
β”‚   └── test_ood_lang.txt        #  36 episode IDs  (paraphrase OOD templates)
β”œβ”€β”€ targets_office.yaml          # Per-scene object catalogs (3-D centroids)
β”œβ”€β”€ targets_hospital.yaml
β”œβ”€β”€ targets_full_warehouse.yaml
β”œβ”€β”€ targets_warehouse_multiple_shelves.yaml
└── episodes/
    └── ep_{N:06d}/
        β”œβ”€β”€ meta.json                 # Full episode metadata
        β”œβ”€β”€ rgb_front/{t}.png         # 640Γ—640 RGB frame at step t
        β”œβ”€β”€ depth_front/{t}.npy       # 640Γ—640 float32 depth (m) at step t
        β”œβ”€β”€ seg_front/{t}.png         # 640Γ—640 uint16 instance segmentation at step t
        β”œβ”€β”€ actions_continuous.npy    # (T, 2) float32 β€” (v_t, Ο‰_t)
        β”œβ”€β”€ actions_tokens.npy        # (T, 2) int16  β€” discretized 7Γ—7 tokens
        └── poses.npy                 # (T, 7) float32 β€” (x,y,z,qw,qx,qy,qz)

Episode Metadata (meta.json)

Each episode's sidecar JSON records the full configuration:

{
  "episode_id": "ep_000321",
  "scene_id": "full_warehouse.usd",
  "goal": {
    "target_category": "crate",
    "target_id": "crate_038",
    "goal_position_xyz_m": [-15.08, 10.77, 2.93]
  },
  "instruction": {
    "text": "Go to the crate.",
    "template_id": "train_01"
  },
  "spawn": { "tier": "mid", "spawn_to_target_dist_m": 3.574 },
  "rollout": {
    "num_steps": 219,
    "terminated_by": "success",
    "success": true,
    "collision_count": 0,
    "final_ne_m": 0.966,
    "trajectory_length_m": 2.61
  }
}

Splits

Split Episodes Description
train_id 261 Seen objects, seen instruction templates
val_id 41 Seen objects, seen templates (validation)
test_id 50 Seen objects, seen templates (held-out test)
test_ood_obj 37 Held-out object categories (fire extinguisher, whiteboard, barrel)
test_ood_lang 36 Paraphrase OOD instruction templates
Total 425 (current snapshot; full budget: 2,000)

Language Instructions

Instructions are generated from slot-fill templates with {object} and {color} placeholders.

18 training templates (T1–T18), examples:

  • "Go to the {object}."
  • "Drive to the {object} and stop."
  • "Approach the {object}."
  • "Navigate to the {object}."
  • "Your destination is the {object}."

12 paraphrase-OOD templates (O1–O12), examples:

  • "Make your way to the {object}."
  • "Proceed to the {object}."
  • "Find the {object} and come to a stop."
  • "Close in on the {object}."

Note: Color-slot templates are suppressed in v1 β€” all targets carry color=unknown because USD assets do not expose material-color attributes through a standard prim API. Active pool: 13 train + 10 paraphrase-OOD templates.


Task Definition

LCOA formulation: Given instruction $\ell$ and observations $o_t = (I_t^\text{RGB}, D_t)$, output actions $a_t = (v_t, \omega_t)$ such that the robot stops within $r_\text{success} = 1.0$ m of the target object centroid.

Action space:

  • Continuous: $(v, \omega) \in [0, 1]$ m/s Γ— $[-1.5, 1.5]$ rad/s
  • Tokenized: each dimension quantized to 7 uniform bins β†’ 49-token vocabulary

Episode termination:

  • Success β€” within 1 m and stationary for β‰₯ 5 consecutive steps
  • Collision β€” stall detected (no forward progress for β‰₯ 16 steps near obstacle)
  • Timeout β€” 1,000 steps reached without success

Only successful episodes are retained in the dataset.


Spawn Tiers

Trajectory diversity is ensured through three distance tiers:

Tier Weight Radius
Near 30% 1.5–3.5 m from target
Mid 40% 3.5–7.0 m from target
Far 30% Global curated floor points

Pearson correlation between spawn distance and trajectory length: r = 0.94.


Expert Controller

The data-collection expert is a proportional controller using pixel-level target visibility from the instance segmentation mask:

  • Target visible (β‰₯ 32 px): angular correction from mask centroid column + depth-based speed
  • Target not visible: bearing-only proportional law from known goal position
  • Obstacle avoidance: speed clamped when depth in central foreground crop < 0.25 m

Rollout Statistics

Split N Mean NE (m) Mean TL (m) Mean Steps
train_id 261 0.967 2.75 197.6
val_id 41 0.967 2.83 205.6
test_id 50 0.966 2.74 190.6
test_ood_obj 37 0.967 2.38 174.7
test_ood_lang 36 0.967 3.07 229.7

NE = final navigation error (distance to goal at termination). TL = trajectory length.


Collection Setup

Property Value
Simulator NVIDIA Isaac Sim 5.1.0-rc.19
Robot NVIDIA Nova Carter (differential-drive)
Camera front_hawk/right stereo camera
Physics rate 60 Hz (Ξ”t = 1/60 s)
Image resolution 640 Γ— 640 px
Random seed 42
Generation date 2026-04-22

Loading the Dataset

import json
import numpy as np
from pathlib import Path
from PIL import Image

root = Path("v1")

# Load split
with open(root / "splits" / "train_id.txt") as f:
    train_ids = [line.strip() for line in f]

# Load an episode
ep_dir = root / "episodes" / train_ids[0]
meta = json.loads((ep_dir / "meta.json").read_text())

instruction = meta["instruction"]["text"]             # "Go to the monitor."
actions = np.load(ep_dir / "actions_continuous.npy")  # (T, 2) float32
tokens  = np.load(ep_dir / "actions_tokens.npy")      # (T, 2) int16
poses   = np.load(ep_dir / "poses.npy")               # (T, 7) float32

# Load frame t=0
rgb   = np.array(Image.open(ep_dir / "rgb_front" / "0.png"))   # (640, 640, 3)
depth = np.load(ep_dir / "depth_front" / "0.npy")              # (640, 640) metres
seg   = np.array(Image.open(ep_dir / "seg_front" / "0.png"))   # (640, 640) instance IDs

Citation

If you use MiniVLA-Nav v1 in your research, please cite:

@misc{albustami2026minivlanavv1multiscenesimulation,
  title={MiniVLA-Nav v1: A Multi-Scene Simulation Dataset for Language-Conditioned Robot Navigation}, 
  author={Ali Al-Bustami and Jaerock Kwon},
  year={2026},
  eprint={2605.00397},
  archivePrefix={arXiv},
  primaryClass={cs.RO},
  url={https://arxiv.org/abs/2605.00397}, 
}

License

This dataset is released under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.


Contact

Ali Al-Bustami - abustami@umich.edu

Downloads last month
812

Paper for alibustami/miniVLA-Nav