Dataset Viewer
Auto-converted to Parquet Duplicate
ready_for_fresh_episode
bool
1 class
image
stringclasses
1 value
native_revision
stringclasses
1 value
passed
listlengths
4
4
swap_blocks_initialization
dict
button_measurements
listlengths
1
1
known_limitation
stringclasses
1 value
evidence_sha256
dict
ready_image
stringclasses
1 value
runtime_readiness
stringclasses
1 value
true
sha256:7e145476e697eef228b1304af8a9c395b41f862e6a6a744276258c9f8037d32c
c3af1ac6ad7302c0c09d17f4b050fd3a5c68c921
[ "initial RGB spectator rendering", "25 Hz control and joint mapping", "rendering leaves physics unchanged", "fresh swap_blocks layout 25 initialization and native button arming" ]
{ "requested_seed": 25, "effective_layout": 25, "layout_ids": [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, ...
[ { "name": "joint_ratio_above", "value": 1, "op": ">", "threshold": 0.95, "unit": "ratio", "inputs": { "joint": "PrismaticJoint_orangeswitch_up", "position": 0.004999999888241291, "lower": -0.005040000192821026, "upper": 0.004999999888241291 } } ]
A second reset in the same native process renames the layout articulation but retains the first object. This campaign creates a fresh simulator process per episode; no simulator is reused across episodes. The failing repeated-reset test is retained in native.log.
{ "native.log": "97563c51a64f2be4075b8b1b1d8f02e2d78f42bf6010d80eb6a4fb07afa59efd", "probe.log": "a19c1fe6bbbeb5c8dcde238f0fb52d62612f5052e1e6b3261224e520c7777f3a" }
sha256:2ae102a399aa889d6fafd8d6030e699add7bca35089754991831162fd71a597f
rosbridge doctor: all 54 catalog entries ready, no issues after upstream-checkout synchronization and ROS rebuild.
true
sha256:7e145476e697eef228b1304af8a9c395b41f862e6a6a744276258c9f8037d32c
c3af1ac6ad7302c0c09d17f4b050fd3a5c68c921
[ "initial RGB spectator rendering", "25 Hz control and joint mapping", "rendering leaves physics unchanged", "fresh swap_blocks layout 25 initialization and native button arming" ]
{ "requested_seed": 25, "effective_layout": 25, "layout_ids": [ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, ...
[ { "name": "joint_ratio_above", "value": 1, "op": ">", "threshold": 0.95, "unit": "ratio", "inputs": { "joint": "PrismaticJoint_orangeswitch_up", "position": 0.004999999888241291, "lower": -0.005040000192821026, "upper": 0.004999999888241291 } } ]
A second reset in the same native process renames the layout articulation but retains the first object. This campaign creates a fresh simulator process per episode; no simulator is reused across episodes. The failing repeated-reset test is retained in native.log.
{ "native.log": "97563c51a64f2be4075b8b1b1d8f02e2d78f42bf6010d80eb6a4fb07afa59efd", "probe.log": "a19c1fe6bbbeb5c8dcde238f0fb52d62612f5052e1e6b3261224e520c7777f3a" }
sha256:2ae102a399aa889d6fafd8d6030e699add7bca35089754991831162fd71a597f
rosbridge doctor: all 54 catalog entries ready, no issues after upstream-checkout synchronization and ROS rebuild.

Kinex benchmark on RoboDojo and RoboCasa

The current results cover the single-seed Kinex and rosbridge evaluation. Each row identifies its simulator, task, seed and evaluation status. Accepted rows include their native evidence.

Archive v1

The archived evaluation preserves the original multi-seed protocol, records and media.

Downloads last month
173

Space using RLE-Bench/kinex-benchmark-robodojo-robocasa 1