diff --git a/masters/roboolympics-fun/alpine_slalom/meta.json b/masters/roboolympics-fun/alpine_slalom/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..6a0add1928bfa338182b47145242d864a85df27e --- /dev/null +++ b/masters/roboolympics-fun/alpine_slalom/meta.json @@ -0,0 +1,140 @@ +{ + "suite": "roboolympics-fun", + "task": "alpine_slalom", + "family": "alpine-slalom", + "task_dir": "embodied-olympics-fun-alpine-slalom-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-alpine-slalom-i00-privileged/embodied-olympics-fun-alpine-sla__eV2gD7X", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; snow venue: a snow shader (bright blue-white, rough, slight subsurface, fine bump) on the course's snow geoms, snow-white valley floor; a chase camera: per-frame fit of the robot (height included, so jumps and spins stay framed), at least 6 m away, 24 degree elevation, smoothed in time", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 820, + "seconds": 27.33, + "speed": 3.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (d52b119c47b1deaf)", + "end_check": { + "success": true, + "stages": [ + "start", + "gate 1", + "gate 2", + "gate 3", + "gate 4", + "gate 5", + "gate 6", + "gate 7", + "gate 8", + "gate 9", + "gate 10", + "gate 11", + "gate 12", + "gate 13", + "gate 14", + "gate 15", + "gate 16", + "gate 17", + "gate 18", + "gate 19", + "gate 20", + "gate 21", + "gate 22", + "gate 23", + "gate 24", + "gate 25", + "gate 26", + "gate 27", + "gate 28", + "gate 29", + "gate 30", + "gate 31", + "gate 32", + "finish" + ], + "stages_total": 34, + "progress": 1.0, + "metrics": { + "speed_m_s": 5.35, + "gates": 32, + "time_s": 79.058 + }, + "time_s": 83.41, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "d52b119c47b1deaf514fcdd4c0b2ac5dfa81b8ff3ca0aa278a1bb90505da9309", + "steps": 8341, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-alpine-slalom-i00-privileged/embodied-olympics-fun-alpine-sla__eV2gD7X/artifacts/app/output/trajectory.npz", + "n_actions": 8341, + "control_steps_replayed": 8341, + "control_hz": 100.0, + "every_control_steps": 11, + "final_state_sha": "d52b119c47b1deaf514fcdd4c0b2ac5dfa81b8ff3ca0aa278a1bb90505da9309", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "textured snow drawn with a snow shader instead of its checker texture's mean colour", + "the start house, the start and finish arches, the finish posts and banner (drawn-only, no contact) not drawn: they hid the robot from the chase camera; the safety nets and the gates are drawn" + ], + "camera": { + "mode": "director", + "az_deg": -70.61, + "az_rule": "robot-travel-70", + "el_deg": 24.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 6.24, + 9.304 + ], + "robot_root": "pelvis", + "travel_m": 307.125, + "focus": { + "robot+object": 742, + "robot": 18 + }, + "objects": { + "gate1_turn": 0.0, + "gate2_turn": 0.0, + "gate3_turn": 0.0, + "gate8_turn": 0.0, + "gate9_turn": 0.0, + "gate17_turn": 0.0, + "gate19_turn": 0.0, + "gate23_turn": 0.0, + "gate24_turn": 0.0, + "gate26_turn": 0.0, + "gate27_turn": 0.01, + "gate29_turn": 0.02, + "gate30_turn": 0.04, + "gate31_turn": 0.03, + "left_ski": 307.43, + "right_ski": 307.29 + } + }, + "orbit_deg": 0.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 4.72, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:37-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/alpine_slalom/poster.png b/masters/roboolympics-fun/alpine_slalom/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..dd4e30c8895c94dbd6bc444f2b1777e6981a2783 --- /dev/null +++ b/masters/roboolympics-fun/alpine_slalom/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9eddb864bc3a96372cf583bf7e9fad3f1d3ffbd2889f72335a581166795cf5a8 +size 3036918 diff --git a/masters/roboolympics-fun/alpine_slalom/video.mp4 b/masters/roboolympics-fun/alpine_slalom/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..4085c473cab225f8c53535c4558d5e43c0c5f0bd --- /dev/null +++ b/masters/roboolympics-fun/alpine_slalom/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:43c60360d5509c0445d422f0ff9ae6618a2b2cabc9aa9dc28ab2903cd655c8f9 +size 12693590 diff --git a/masters/roboolympics-fun/backflip/meta.json b/masters/roboolympics-fun/backflip/meta.json index 2db9359db87ca84b6a22b4859e0196311d7cad61..5b585041434195724f427939531012538746a00a 100644 --- a/masters/roboolympics-fun/backflip/meta.json +++ b/masters/roboolympics-fun/backflip/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "backflip", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-backflip-i00-privileged/roboolympics-backflip-i00-privil__fuKrrEN", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "backflip", + "task_dir": "embodied-olympics-fun-backflip-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-backflip-i00-privileged/embodied-olympics-fun-backflip-i__L6om7HM", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 227, - "seconds": 7.57, + "frames": 237, + "seconds": 7.9, "speed": 0.6, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (f5c0b3c390fd3db4)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (c1ad751ccaaf63b9)", "end_check": { "success": true, "stages": [ @@ -23,22 +26,30 @@ "stages_total": 4, "progress": 1.0, "metrics": { - "turn_deg": 319.3, - "air_s": 0.673 + "turn_deg": 304.2, + "air_s": 0.436 }, - "time_s": 3.31 + "time_s": 3.52, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "c1ad751ccaaf63b91879e4b69137dcd58180235c8bd7ddf39a078bb84bdd4bf1", + "steps": 352, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 331, - "control_steps_replayed": 331, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-backflip-i00-privileged/embodied-olympics-fun-backflip-i__L6om7HM/artifacts/app/output/trajectory.npz", + "n_actions": 352, + "control_steps_replayed": 352, "control_hz": 100.0, "every_control_steps": 2, - "final_state_sha": "f5c0b3c390fd3db47e93a8f22414ca675082bb9b748945b35436320990730f4c", + "final_state_sha": "c1ad751ccaaf63b91879e4b69137dcd58180235c8bd7ddf39a078bb84bdd4bf1", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", @@ -48,30 +59,31 @@ "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.264, - 3.385 + 3.094, + 3.358 ], "robot_root": "pelvis", - "travel_m": 0.017, + "travel_m": 0.813, "focus": { - "robot": 167 + "robot": 177 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.8, - "image": "eai-robofunolympics:0.2.1", + "seconds_per_frame": 2.62, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:31-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:37-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/backflip/poster.png b/masters/roboolympics-fun/backflip/poster.png index a74e07d8d3fdf49ae918903f6d96dda3672b8723..605635917cdf8166cf2752adeb2e3d9fa5889df2 100644 --- a/masters/roboolympics-fun/backflip/poster.png +++ b/masters/roboolympics-fun/backflip/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b517c8f4ec244ebb4c38eec2366e8c0891ad479311359d69aca2b5a85c33c4d6 -size 3104830 +oid sha256:eb44a96f36ebbe06ea2963219cdea3fc168b5cae53d3ec30b62d4aa063c37fc8 +size 3111651 diff --git a/masters/roboolympics-fun/backflip/video.mp4 b/masters/roboolympics-fun/backflip/video.mp4 index c9d4ea832874644582f406306fe815c0f98a3a30..d5a04ca7e2a63eb2dcbc95010423f195300e5ef0 100644 --- a/masters/roboolympics-fun/backflip/video.mp4 +++ b/masters/roboolympics-fun/backflip/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a3695ad8018ec20e9298dda93e279bdf920ccfe997a75a631fc86bb8b1e49481 -size 4075675 +oid sha256:f9c39256f792c5d233bc2545f7807b7157727972fe35cb0cea667e6253e12067 +size 4084672 diff --git a/masters/roboolympics-fun/badminton_serve/meta.json b/masters/roboolympics-fun/badminton_serve/meta.json index 18bcf5471f69ac7f6eef5801787dada4a51e972e..4dd412922623363f8387ae3dcc83dcd15d5908a1 100644 --- a/masters/roboolympics-fun/badminton_serve/meta.json +++ b/masters/roboolympics-fun/badminton_serve/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "badminton_serve", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-fun-badminton-serve-i00-privileged/roboolympics-badminton-serve-i00__D5pDUDF", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "badminton-serve", + "task_dir": "embodied-olympics-fun-badminton-serve-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-badminton-serve-i00-privileged/embodied-olympics-fun-badminton__dU3SuCW", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 279, - "seconds": 9.3, + "frames": 359, + "seconds": 11.97, "speed": 0.3, "success_check": true, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (9afedd98535665a1)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (8ad0c87716c42927)", "end_check": { "success": true, "stages": [ @@ -23,81 +26,71 @@ "stages_total": 4, "progress": 1.0, "metrics": { - "hit_height_m": 0.4881, - "shuttle_speed_m_s": 11.9426, - "shuttle_angle_deg": 45.48, - "landing_x_m": 5.788, - "landing_y_m": 1.716 + "hit_height_m": 0.7396, + "shuttle_speed_m_s": 11.6684, + "shuttle_angle_deg": 38.16, + "landing_x_m": 5.933, + "landing_y_m": 1.554 }, - "time_s": 2.18 + "time_s": 2.98, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 1, - "n_actions": 218, - "progress": 1.0, - "stages_reached": 4, - "stages_total": 4, - "metric_hit_height_m": 0.4881, - "metric_landing_x_m": 5.788, - "metric_landing_y_m": 1.716, - "metric_shuttle_angle_deg": 45.48, - "metric_shuttle_speed_m_s": 11.9426 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "8ad0c87716c429273574da561b752e29645f02dedfe0e43cb6ae80d442ae5be9", + "steps": 298, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-badminton-serve-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-badminton-serve-i00-privileged/roboolympics-badminton-serve-i00__D5pDUDF/artifacts/app/output/trajectory.npz", - "n_actions": 218, - "control_steps_replayed": 218, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-badminton-serve-i00-privileged/embodied-olympics-fun-badminton__dU3SuCW/artifacts/app/output/trajectory.npz", + "n_actions": 298, + "control_steps_replayed": 298, "control_hz": 100.0, "every_control_steps": 1, - "final_state_sha": "9afedd98535665a1dee70b866bc1073569264d702c09337714ba6088628908f6", - "replays_agree": true, - "mujoco": "3.3.7", + "final_state_sha": "8ad0c87716c429273574da561b752e29645f02dedfe0e43cb6ae80d442ae5be9", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -99.04, + "az_deg": -100.81, "az_rule": "object-travel-115", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.504, - 7.156 + 3.576, + 6.469 ], "robot_root": "pelvis", - "travel_m": 0.006, + "travel_m": 0.242, "focus": { - "robot+object": 212, - "robot": 7 + "robot+object": 259, + "robot": 40 }, "objects": { - "shuttle": 5.62, - "racket": 0.31 + "shuttle": 5.81, + "racket": 0.3 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.43, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-badminton-serve-i00-privileged", + "seconds_per_frame": 3.03, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-badminton-serve-i00-privileged/roboolympics-badminton-serve-i00__D5pDUDF", - "trial_choice": "a counted GPT-6 Astra success in privileged mode", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "vast.ai RTX 4090 gpu0", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:14:57-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/badminton_serve/poster.png b/masters/roboolympics-fun/badminton_serve/poster.png index 0f653a5d3fc7596dd831595b9543c291790194c4..1ade24eed48b6ce743f7f6f244f1cca06d92a39d 100644 --- a/masters/roboolympics-fun/badminton_serve/poster.png +++ b/masters/roboolympics-fun/badminton_serve/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b9c8b0913908c85282c15c17efbc2162d8ad6e0dab8d84a98892d0ade2fb2d14 -size 3176992 +oid sha256:6318a8e21a3950506ca30ce5eccb3c30e39e2e24205317238b7cd3bf01012e28 +size 3209522 diff --git a/masters/roboolympics-fun/badminton_serve/video.mp4 b/masters/roboolympics-fun/badminton_serve/video.mp4 index 6b6957a27525336476d35f8882c6f55bac2a03ea..3f4a8fe796264bf193541014aade24df159a31e6 100644 --- a/masters/roboolympics-fun/badminton_serve/video.mp4 +++ b/masters/roboolympics-fun/badminton_serve/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:53d49e8d1404649fbd1331dffae3745cae61bbd1f82ff29e816261746aa4860a -size 5400934 +oid sha256:e139dff6733b24ac6bfc6a8baf4b85417431909909ba556a02ee8e95e90cb815 +size 6319444 diff --git a/masters/roboolympics-fun/balance_beam/meta.json b/masters/roboolympics-fun/balance_beam/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..bc26e87492f64b63904a6603d8ca5bdb631e9c1b --- /dev/null +++ b/masters/roboolympics-fun/balance_beam/meta.json @@ -0,0 +1,88 @@ +{ + "suite": "roboolympics-fun", + "task": "balance_beam", + "family": "balance-beam", + "task_dir": "embodied-olympics-fun-balance-beam-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-balance-beam-i00-privileged/embodied-olympics-fun-balance-be__FJ5XCVd", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 780, + "seconds": 26.0, + "speed": 3.0, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (1dbc37fa79cb9961)", + "end_check": { + "success": true, + "stages": [ + "half way", + "far end", + "stand" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "distance_m": 4.541, + "time_s": 71.89 + }, + "time_s": 71.89, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "1dbc37fa79cb99619dea2129218d3366ca0592f3dcdf9d80838046ae0072e3b0", + "steps": 7189, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-balance-beam-i00-privileged/embodied-olympics-fun-balance-be__FJ5XCVd/artifacts/app/output/trajectory.npz", + "n_actions": 7189, + "control_steps_replayed": 7189, + "control_hz": 100.0, + "every_control_steps": 10, + "final_state_sha": "1dbc37fa79cb99619dea2129218d3366ca0592f3dcdf9d80838046ae0072e3b0", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -70.02, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.087, + 3.22 + ], + "robot_root": "pelvis", + "travel_m": 4.451, + "focus": { + "robot": 720 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.68, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/balance_beam/poster.png b/masters/roboolympics-fun/balance_beam/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2061004a7859f1e21dfb33ac37daab4714a7d960 --- /dev/null +++ b/masters/roboolympics-fun/balance_beam/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:20b1f9384aea1bca2899f1731f88f40a4320b74ed93004d993ca23b4a110021b +size 3100072 diff --git a/masters/roboolympics-fun/balance_beam/video.mp4 b/masters/roboolympics-fun/balance_beam/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..011d7a2404bb8789f7f96a66269d160d8831f268 --- /dev/null +++ b/masters/roboolympics-fun/balance_beam/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b3cc3f0294c632d4838dc59f7d742a0875ebc3c2ed19d141e1a16a1aeb0ea08b +size 11572565 diff --git a/masters/roboolympics-fun/balance_board/meta.json b/masters/roboolympics-fun/balance_board/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..14e8785edd1ddf854d284dc968303d28d093e7d3 --- /dev/null +++ b/masters/roboolympics-fun/balance_board/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "balance_board", + "family": "balance-board", + "task_dir": "embodied-olympics-fun-balance-board-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-balance-board-i00-privileged/embodied-olympics-fun-balance-bo__RkHCQBb", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 811, + "seconds": 27.03, + "speed": 1.2, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (1d23a15ed5f9c597)", + "end_check": { + "success": true, + "stages": [ + "10 s", + "20 s", + "30 s" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "max_tilt_deg": 10.15, + "time_on_board_s": 30.0 + }, + "time_s": 30.0, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "1d23a15ed5f9c597c27a2c93f86851b09296a0c5f0c37554600b31af5019f914", + "steps": 3000, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-balance-board-i00-privileged/embodied-olympics-fun-balance-bo__RkHCQBb/artifacts/app/output/trajectory.npz", + "n_actions": 3000, + "control_steps_replayed": 3000, + "control_hz": 100.0, + "every_control_steps": 4, + "final_state_sha": "1d23a15ed5f9c597c27a2c93f86851b09296a0c5f0c37554600b31af5019f914", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.737, + 3.927 + ], + "robot_root": "pelvis", + "travel_m": 0.039, + "focus": { + "robot": 751 + }, + "objects": { + "roller": 0.12, + "deck": 0.23 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.58, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/balance_board/poster.png b/masters/roboolympics-fun/balance_board/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..ed8d60c480147f7db1cca3c228a11165f3383a11 --- /dev/null +++ b/masters/roboolympics-fun/balance_board/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:99ddf6692308f33f64b438c76ed742eb3e1f0228ddfe8f702add9cb0dcfb91b0 +size 3105285 diff --git a/masters/roboolympics-fun/balance_board/video.mp4 b/masters/roboolympics-fun/balance_board/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..c8589acd2dcfd45fa3de877374eb4c0e432ad52b --- /dev/null +++ b/masters/roboolympics-fun/balance_board/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b5671816567883b76efe463240635c7a0e9180c6da0177c90d62a908a532a637 +size 8246896 diff --git a/masters/roboolympics-fun/basketball_dribble/meta.json b/masters/roboolympics-fun/basketball_dribble/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..b71dfc687314e3cb5c8e2bc37ea573277827acb7 --- /dev/null +++ b/masters/roboolympics-fun/basketball_dribble/meta.json @@ -0,0 +1,93 @@ +{ + "suite": "roboolympics-fun", + "task": "basketball_dribble", + "family": "basketball-dribble", + "task_dir": "embodied-olympics-fun-basketball-dribble-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-basketball-dribble-i00-privileged/embodied-olympics-fun-basketball__yqWK2fP", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 565, + "seconds": 18.83, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (75298314bd477697)", + "end_check": { + "success": true, + "stages": [ + "dribble", + "past the free-throw line", + "in the frontcourt" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "distance_m": 14.358, + "ball_height_m": 0.146, + "dribbles": 28 + }, + "time_s": 15.11, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "75298314bd4776976ff97d2b893e81d3817bf85a109132d1b1c384e617e77909", + "steps": 1511, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-basketball-dribble-i00-privileged/embodied-olympics-fun-basketball__yqWK2fP/artifacts/app/output/trajectory.npz", + "n_actions": 1511, + "control_steps_replayed": 1511, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "75298314bd4776976ff97d2b893e81d3817bf85a109132d1b1c384e617e77909", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the ball drawn as a basketball (orange, black channels)" + ], + "camera": { + "mode": "director", + "az_deg": -67.0, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.298, + 4.388 + ], + "robot_root": "pelvis", + "travel_m": 13.877, + "focus": { + "robot+object": 503, + "robot": 2 + }, + "objects": { + "ball": 13.9 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.86, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/basketball_dribble/poster.png b/masters/roboolympics-fun/basketball_dribble/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..fab084b9155bd982b1603e66a5b58d95148ae078 --- /dev/null +++ b/masters/roboolympics-fun/basketball_dribble/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:391e7d5d80e1a7cb92dbe2a82e33618b997171e0a2febe691f13e1086a2d7439 +size 3144397 diff --git a/masters/roboolympics-fun/basketball_dribble/video.mp4 b/masters/roboolympics-fun/basketball_dribble/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..026e2cf8aac43aa969431617c0a35f3dffc7aa33 --- /dev/null +++ b/masters/roboolympics-fun/basketball_dribble/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ba3ff70303fe1825d1f8c4488897762a9a6789a7cc4399e4eff0cb596e96077 +size 11305885 diff --git a/masters/roboolympics-fun/bowling/meta.json b/masters/roboolympics-fun/bowling/meta.json index fc2069ff5ffcec4b5186ea54c00e1806bf1b0b32..c6a77b351fefafde6de65e46df9a9c77465bf069 100644 --- a/masters/roboolympics-fun/bowling/meta.json +++ b/masters/roboolympics-fun/bowling/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "bowling", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-bowling-i00-privileged/roboolympics-bowling-i00-privile__wK9Ujtv", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "bowling", + "task_dir": "embodied-olympics-fun-bowling-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-bowling-i00-privileged/embodied-olympics-fun-bowling-i0__nD6fgoT", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 427, - "seconds": 14.23, + "frames": 402, + "seconds": 13.4, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (369fbd0f61bc2aa9)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e0df2c83e0638eca)", "end_check": { "success": true, "stages": [ @@ -24,72 +27,81 @@ "metrics": { "pins_down": 10, "release_speed_m_s": 5.06, - "entry_speed_m_s": 3.0, - "entry_y_m": -0.058, - "entry_angle_deg": 2.25 + "entry_speed_m_s": 3.83, + "entry_y_m": -0.12, + "entry_angle_deg": -0.47 }, - "time_s": 10.98 + "time_s": 10.22, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "e0df2c83e0638ecad98881b4688e1c53d1f4714ece2174affdbfcefd26c5fcdd", + "steps": 1022, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 1098, - "control_steps_replayed": 1098, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-bowling-i00-privileged/embodied-olympics-fun-bowling-i0__nD6fgoT/artifacts/app/output/trajectory.npz", + "n_actions": 1022, + "control_steps_replayed": 1022, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "369fbd0f61bc2aa9631a38b6bd217be0232e741da17cb30936ce994358f45960", + "final_state_sha": "e0df2c83e0638ecad98881b4688e1c53d1f4714ece2174affdbfcefd26c5fcdd", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -115.86, + "az_deg": -116.4, "az_rule": "object-travel-115", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.599, - 35.626 + 4.697, + 36.677 ], "robot_root": "pelvis", - "travel_m": 0.128, + "travel_m": 0.086, "focus": { - "robot": 16, - "robot(incoming)": 3, - "object": 201, - "robot+object": 71, - "robot+landed": 76 + "robot+landed": 81, + "robot+object": 49, + "robot": 29, + "object": 173, + "robot(incoming)": 10 }, "objects": { - "pin1": 0.55, - "pin2": 0.52, - "pin3": 0.93, - "pin4": 0.87, - "pin5": 0.72, - "pin6": 0.32, - "pin7": 0.17, - "pin8": 0.87, - "pin9": 0.64, - "pin10": 0.84, - "ball": 19.71 + "pin1": 0.59, + "pin2": 0.28, + "pin3": 0.44, + "pin4": 0.56, + "pin5": 0.32, + "pin6": 0.55, + "pin7": 0.43, + "pin8": 0.32, + "pin9": 0.39, + "pin10": 0.48, + "ball": 20.16 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 4.45, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", + "seconds_per_frame": 2.97, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial_choice": "no counted Astra success; a counted success of another run", - "poster": "frame 46: just after the release / strike, robot and object in view (the last frame no longer shows the robot)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:32-0500" + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 38 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/bowling/poster.png b/masters/roboolympics-fun/bowling/poster.png index a67d1a0ef72296a09c807c723f5444b5e4de6ba7..d72608f26f4d5776a7c3da71c3be4d0da882d9b6 100644 --- a/masters/roboolympics-fun/bowling/poster.png +++ b/masters/roboolympics-fun/bowling/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:03d1740eb42f1f976be380f36d8806f4607b7abbb02fe25b7f1c4423a2d3e5df -size 3118879 +oid sha256:c9c1d981c877d1564e218c009658f8f5ae790bfd115551bc11ae9ad8bb2bfcbd +size 3132893 diff --git a/masters/roboolympics-fun/bowling/video.mp4 b/masters/roboolympics-fun/bowling/video.mp4 index d549ba16195dd0eb0cd51808ea1a330584c25dd2..cb03df1a2a1aad02745a2d570ad362e3265b8343 100644 --- a/masters/roboolympics-fun/bowling/video.mp4 +++ b/masters/roboolympics-fun/bowling/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e1ef37f2a4b8868940e9286310b7aac43301b9a63434d806e6d1070997a19461 -size 7391936 +oid sha256:d600112ba6c9fa9ff5651698830de3ad4142a4eda94e1ae2811c7972ad859a44 +size 7445595 diff --git a/masters/roboolympics-fun/box_jump/meta.json b/masters/roboolympics-fun/box_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..ab200a5ec0e38d1277e147ef72254ef1fd622280 --- /dev/null +++ b/masters/roboolympics-fun/box_jump/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "box_jump", + "family": "box-jump", + "task_dir": "embodied-olympics-fun-box-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-box-jump-i00-privileged/embodied-olympics-fun-box-jump-i__o9U84U8", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 244, + "seconds": 8.13, + "speed": 0.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7dba1a5cc42e13fe)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "landing", + "stand" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "air_s": 0.514, + "pelvis_rise_m": 0.531, + "pelvis_above_box_m": 1.027 + }, + "time_s": 3.65, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "7dba1a5cc42e13fe30558cf7684a49b635f8d495b2317dcff26d53cd3d7bee7e", + "steps": 365, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-box-jump-i00-privileged/embodied-olympics-fun-box-jump-i__o9U84U8/artifacts/app/output/trajectory.npz", + "n_actions": 365, + "control_steps_replayed": 365, + "control_hz": 100.0, + "every_control_steps": 2, + "final_state_sha": "7dba1a5cc42e13fe30558cf7684a49b635f8d495b2317dcff26d53cd3d7bee7e", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.354, + 5.638 + ], + "robot_root": "pelvis", + "travel_m": 0.958, + "focus": { + "robot": 184 + }, + "objects": { + "box": 0.0 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.68, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/box_jump/poster.png b/masters/roboolympics-fun/box_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..3f0318bc7e9b9172f62b9465d1efde205a0fe9e5 --- /dev/null +++ b/masters/roboolympics-fun/box_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e78d9e90ba155cd4f925170fa20b8feae37ecdff414eb42a67f06cf4fd4bd797 +size 3046876 diff --git a/masters/roboolympics-fun/box_jump/video.mp4 b/masters/roboolympics-fun/box_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..57e4399a65ea3788085922ddade847d021ace8e0 --- /dev/null +++ b/masters/roboolympics-fun/box_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5b723c54603a03565367ec78d3927231dd6646a899e4ce70bc415911ed4370c +size 3383986 diff --git a/masters/roboolympics-fun/boxing/meta.json b/masters/roboolympics-fun/boxing/meta.json index 1c9614d8a4ff3d52be5ee194bd298df57b49a126..aad1b9466e2d7e79cd927f1767796bd0e2643be4 100644 --- a/masters/roboolympics-fun/boxing/meta.json +++ b/masters/roboolympics-fun/boxing/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "boxing", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-boxing-i00-standard/roboolympics-boxing-i00-standard__a7gYtdP", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "boxing", + "task_dir": "embodied-olympics-fun-boxing-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-boxing-i00-privileged/embodied-olympics-fun-boxing-i00__pTjF4SX", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 302, - "seconds": 10.07, + "frames": 257, + "seconds": 8.57, "speed": 0.6, "success_check": true, - "success_check_how": "the task's own verdict (environment/sim/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7ac9e6b973578113)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (bee9e7f8378fcca5)", "end_check": { "success": true, "stages": [ @@ -21,24 +24,32 @@ "stages_total": 2, "progress": 1.0, "metrics": { - "hit_height_m": 1.231, - "impulse_n_s": 13.39, - "contact_impulse_n_s": 15.12, - "contact_s": 0.089 + "hit_height_m": 1.4, + "impulse_n_s": 20.64, + "contact_impulse_n_s": 26.35, + "contact_s": 0.059 }, - "time_s": 4.82 + "time_s": 3.92, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "bee9e7f8378fcca5da7f32d97f4377381ec226b82dd2e0695270f802ee6a5550", + "steps": 392, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 482, - "control_steps_replayed": 482, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-boxing-i00-privileged/embodied-olympics-fun-boxing-i00__pTjF4SX/artifacts/app/output/trajectory.npz", + "n_actions": 392, + "control_steps_replayed": 392, "control_hz": 100.0, "every_control_steps": 2, - "final_state_sha": "7ac9e6b973578113abfa0fd35a2cb6f66ad6d75ef834e83ec2ca8ada36fc449e", + "final_state_sha": "bee9e7f8378fcca5da7f32d97f4377381ec226b82dd2e0695270f802ee6a5550", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", @@ -48,32 +59,33 @@ "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.919, - 5.205 + 4.872, + 5.158 ], "robot_root": "pelvis", - "travel_m": 0.489, + "travel_m": 0.762, "focus": { - "robot": 242 + "robot": 197 }, "objects": { - "bag": 0.06 + "bag": 0.04 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.79, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", - "mode": "standard", - "trial_choice": "no counted Astra success; a counted success of another run", + "seconds_per_frame": 2.66, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu1 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:32-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:38-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/boxing/poster.png b/masters/roboolympics-fun/boxing/poster.png index 3fb228a523a26c23a5ff671f13c868fb658a3fcd..854ad0d2fcb691ca53d095f35233b12323ff232a 100644 --- a/masters/roboolympics-fun/boxing/poster.png +++ b/masters/roboolympics-fun/boxing/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:bca214eb8b8171b1e7c0eb4c95b45d5b1b1556d3d36973ef86b0bd5c68e51028 -size 3150505 +oid sha256:56ba6fec874e0f8cc9719b27b2ec8dbbc7f3592c3badbebb550969babaa8ca87 +size 3155145 diff --git a/masters/roboolympics-fun/boxing/video.mp4 b/masters/roboolympics-fun/boxing/video.mp4 index 6ae6ce228117f01f29b8bc1e980629fc4da6b21f..e25ae10e802796fbac0cc5a94f1e4a5f83e67bd1 100644 --- a/masters/roboolympics-fun/boxing/video.mp4 +++ b/masters/roboolympics-fun/boxing/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:f2d2fb7df606b44637eaaad02b05369aa8cbb35de0739cb61c008d3765d4728c -size 4264670 +oid sha256:ed2d114a79d4593712748a225191dec068cfbe47d9c4ff2cd0850c0c56015666 +size 4041069 diff --git a/masters/roboolympics-fun/breaking_windmill/meta.json b/masters/roboolympics-fun/breaking_windmill/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..cad8f4603a114481fd6a88b1234374d918433641 --- /dev/null +++ b/masters/roboolympics-fun/breaking_windmill/meta.json @@ -0,0 +1,89 @@ +{ + "suite": "roboolympics-fun", + "task": "breaking_windmill", + "family": "breaking-windmill", + "task_dir": "embodied-olympics-fun-breaking-windmill-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra_quick-high-embodied-olympics-fun-breaking-windmill-i00-privileged/embodied-olympics-fun-breaking-w__mxtoRx8", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 748, + "seconds": 24.93, + "speed": 0.9, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b35bdb9e0bd9d696)", + "end_check": { + "success": false, + "stages": [ + "down", + "1 windmill", + "2 windmills" + ], + "stages_total": 4, + "progress": 0.75, + "metrics": { + "windmills": 2 + }, + "time_s": 20.6, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "b35bdb9e0bd9d69658ac77d30d57ad97af8c5903ab8b883e8d1af4b46d77e9f2", + "steps": 2060, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-breaking-windmill-i00-privileged/embodied-olympics-fun-breaking-w__mxtoRx8/artifacts/app/output/trajectory.npz", + "n_actions": 2060, + "control_steps_replayed": 2060, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "b35bdb9e0bd9d69658ac77d30d57ad97af8c5903ab8b883e8d1af4b46d77e9f2", + "original_verdict": "counted failure (verifier reward 0)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 2.072, + 3.389 + ], + "robot_root": "pelvis", + "travel_m": 0.655, + "focus": { + "robot": 688 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:39-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/breaking_windmill/poster.png b/masters/roboolympics-fun/breaking_windmill/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..f4d2c1d69fd7a7cdc9eb8006174983d2d027dca1 --- /dev/null +++ b/masters/roboolympics-fun/breaking_windmill/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1118a0dfac387175aaf8c838d1efc2dca72a8a34ae0875c5f534b6d55f3c1499 +size 3052344 diff --git a/masters/roboolympics-fun/breaking_windmill/video.mp4 b/masters/roboolympics-fun/breaking_windmill/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..7adddc40a0a3a0c4ed09933cb0cb3e643a8c8a23 --- /dev/null +++ b/masters/roboolympics-fun/breaking_windmill/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d15b4138e5fae82de7c099ee1833b5ff6ead3771b547b78dffd8e394cc4d646f +size 15904376 diff --git a/masters/roboolympics-fun/catch/meta.json b/masters/roboolympics-fun/catch/meta.json index 9fd53f44b78b8e7c6280efc8697fdeb3d0cb02b3..d1631b9f8a3279347e1f9037d54e66a351ad6a22 100644 --- a/masters/roboolympics-fun/catch/meta.json +++ b/masters/roboolympics-fun/catch/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "catch", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-catch-i00-privileged/roboolympics-catch-i00-privilege__n9GbsX7", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "catch", + "task_dir": "embodied-olympics-fun-catch-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-catch-i00-privileged/embodied-olympics-fun-catch-i00__qBKtgfp", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 306, - "seconds": 10.2, - "speed": 0.6, + "frames": 227, + "seconds": 7.57, + "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (738f01722d4c5c4f)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7143bce9fa2d82ad)", "end_check": { "success": true, "stages": [ @@ -21,58 +24,68 @@ "stages_total": 2, "progress": 1.0, "metrics": { - "catch_height_m": 1.186 + "catch_height_m": 1.042, + "catch_speed_m_s": 9.938 }, - "time_s": 4.9 + "time_s": 4.96, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "7143bce9fa2d82ad900e6e5d780b651fc77a8b98cdc4c93ba9633157e3447da7", + "steps": 496, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 490, - "control_steps_replayed": 490, + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-catch-i00-privileged/embodied-olympics-fun-catch-i00__qBKtgfp/artifacts/app/output/trajectory.npz", + "n_actions": 496, + "control_steps_replayed": 496, "control_hz": 100.0, - "every_control_steps": 2, - "final_state_sha": "738f01722d4c5c4f0582593ef988a9d1960dd893a892e62ebb2a91046478b10f", + "every_control_steps": 3, + "final_state_sha": "7143bce9fa2d82ad900e6e5d780b651fc77a8b98cdc4c93ba9633157e3447da7", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -89.44, - "az_rule": "robot-travel-70", + "az_deg": -293.81, + "az_rule": "object-travel-115", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.446, - 17.158 + 4.356, + 16.306 ], "robot_root": "pelvis", - "travel_m": 2.245, + "travel_m": 1.66, "focus": { - "robot(incoming)": 124, - "robot+object": 71, - "robot": 51 + "robot+object": 48, + "robot": 33, + "robot(incoming)": 86 }, "objects": { - "ball": 37.58 + "ball": 38.23 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 8.15, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", + "seconds_per_frame": 3.92, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", "mode": "privileged", - "trial_choice": "no counted Astra success; a counted success of another run", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", "poster": "last frame (solved state)", - "host": "gpu2 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:32-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:39-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/catch/poster.png b/masters/roboolympics-fun/catch/poster.png index 976bb2b31934ed26841b9e1e2a756c18c85b7c76..e0c2128a34684173c75937960d9f79e72aa7ae9c 100644 --- a/masters/roboolympics-fun/catch/poster.png +++ b/masters/roboolympics-fun/catch/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:70e0c2b6ca8d3e403942cf9d9d577b43cd732c8f460993cc007d019cc9aec568 -size 3037721 +oid sha256:d4edd616756ca828a97062c9a7f9d14bac9671236591d4de49d11de8c5e6b59a +size 3049411 diff --git a/masters/roboolympics-fun/catch/video.mp4 b/masters/roboolympics-fun/catch/video.mp4 index 76675cf67d45247727671e12c390020db9b9c188..e748a2f0e2f7a3a7771e73ece4b0823a18a856f5 100644 --- a/masters/roboolympics-fun/catch/video.mp4 +++ b/masters/roboolympics-fun/catch/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:eed355e941a9d4720fa02f51106a13cb45d2b49c8e06ab3ebb9dff0408e6b8ca -size 4733797 +oid sha256:4fcf5e3a498b657b438460f17ac9e231f2076972e0ab9d577e309b8fcd5d895b +size 3564063 diff --git a/masters/roboolympics-fun/catch_and_shoot/meta.json b/masters/roboolympics-fun/catch_and_shoot/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..d34d00fffb4bef815599200a0a7cf9afa417f941 --- /dev/null +++ b/masters/roboolympics-fun/catch_and_shoot/meta.json @@ -0,0 +1,92 @@ +{ + "suite": "roboolympics-fun", + "task": "catch_and_shoot", + "family": "catch-and-shoot", + "task_dir": "embodied-olympics-fun-catch-and-shoot-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-catch-and-shoot-i00-privileged/embodied-olympics-fun-catch-and__GZgBxmp", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 282, + "seconds": 9.4, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (c4b69de29627ac7b)", + "end_check": { + "success": true, + "stages": [ + "catch", + "release", + "through the ring" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "ball_height_m": 2.775 + }, + "time_s": 6.61, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "c4b69de29627ac7bfb36bec68f4b47ea21a2275a4b0a6221f7bfdfd4579bdb91", + "steps": 661, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-catch-and-shoot-i00-privileged/embodied-olympics-fun-catch-and__GZgBxmp/artifacts/app/output/trajectory.npz", + "n_actions": 661, + "control_steps_replayed": 661, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "c4b69de29627ac7bfb36bec68f4b47ea21a2275a4b0a6221f7bfdfd4579bdb91", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the ball drawn as a basketball (orange, black channels)" + ], + "camera": { + "mode": "director", + "az_deg": -155.62, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.886, + 16.553 + ], + "robot_root": "pelvis", + "travel_m": 0.024, + "focus": { + "robot+object": 217, + "robot": 5 + }, + "objects": { + "passer": 0.03, + "ball": 6.5 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.91, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 162 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:39-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/catch_and_shoot/poster.png b/masters/roboolympics-fun/catch_and_shoot/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..b9c8d1b6bd5c36097e92dee6ee10f6184b6d6dcc --- /dev/null +++ b/masters/roboolympics-fun/catch_and_shoot/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5cb18a0b34045617126dbc9bf727d299d5013342193e0bb9bad65df92ed5c07c +size 3143831 diff --git a/masters/roboolympics-fun/catch_and_shoot/video.mp4 b/masters/roboolympics-fun/catch_and_shoot/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..c2e67a8154fad460cf5fa5dec6b9bb389112b89e --- /dev/null +++ b/masters/roboolympics-fun/catch_and_shoot/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:28001c0fb9bd2e9d63658dc845f8755d317c2451ceacb388f69b45f908a86cda +size 5684331 diff --git a/masters/roboolympics-fun/clean_and_jerk/meta.json b/masters/roboolympics-fun/clean_and_jerk/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..9125f384a7a69d1014543021143669adad742434 --- /dev/null +++ b/masters/roboolympics-fun/clean_and_jerk/meta.json @@ -0,0 +1,95 @@ +{ + "suite": "roboolympics-fun", + "task": "clean_and_jerk", + "family": "clean-and-jerk", + "task_dir": "embodied-olympics-fun-clean-and-jerk-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-clean-and-jerk-i00-privileged/embodied-olympics-fun-clean-and__HRohaub", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 387, + "seconds": 12.9, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e574484c3443a628)", + "end_check": { + "success": true, + "stages": [ + "lift-off", + "clean", + "overhead", + "down signal" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "bar_height_m": 2.122, + "bar_max_height_m": 2.123, + "elbow_flex_deg": 0.3, + "knee_flex_deg": 3.1, + "time_s": 9.76 + }, + "time_s": 9.76, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "e574484c3443a628a9ff27f8df93ad0fbb777b314b89318a1bc1430d6647c76f", + "steps": 976, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-clean-and-jerk-i00-privileged/embodied-olympics-fun-clean-and__HRohaub/artifacts/app/output/trajectory.npz", + "n_actions": 976, + "control_steps_replayed": 976, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "e574484c3443a628a9ff27f8df93ad0fbb777b314b89318a1bc1430d6647c76f", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.151, + 6.408 + ], + "robot_root": "pelvis", + "travel_m": 0.076, + "focus": { + "robot": 191, + "robot+object": 136 + }, + "objects": { + "bar": 0.21 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:39-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/clean_and_jerk/poster.png b/masters/roboolympics-fun/clean_and_jerk/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2a3c11f8da1edddffeaa73e07b3abaed4a7872f0 --- /dev/null +++ b/masters/roboolympics-fun/clean_and_jerk/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:18d98d68879eb9777501b8e346b1036397a08e7f540fcf2717d351b6996f041b +size 3088548 diff --git a/masters/roboolympics-fun/clean_and_jerk/video.mp4 b/masters/roboolympics-fun/clean_and_jerk/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..35ecaae0f22b7f0aef36a300162817d38cee6ad1 --- /dev/null +++ b/masters/roboolympics-fun/clean_and_jerk/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad7cdfce402d206c89478423922feb7322dbd479f5e3da34ebceb3b96caa2b52 +size 5720370 diff --git a/masters/roboolympics-fun/crawl/meta.json b/masters/roboolympics-fun/crawl/meta.json index 8ac87015c2ec0e9541be9cc6aa5d3d092e23d913..2db0bab191bd69b0edd5e8cfcc675663cdb6e791 100644 --- a/masters/roboolympics-fun/crawl/meta.json +++ b/masters/roboolympics-fun/crawl/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "crawl", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-crawl-i00-privileged/roboolympics-crawl-i00-privilege__Che2uUd", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "crawl", + "task_dir": "embodied-olympics-fun-crawl-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-crawl-i00-privileged/embodied-olympics-fun-crawl-i00__YDvTC8q", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 806, - "seconds": 26.87, - "speed": 2.7, + "frames": 815, + "seconds": 27.17, + "speed": 3.0, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e3c198584948a82e)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (faa48f5cc065243c)", "end_check": { "success": true, "stages": [ @@ -22,55 +25,64 @@ "stages_total": 3, "progress": 1.0, "metrics": { - "rear_x_m": 11.004, - "time_s": 67.01 + "rear_x_m": 11.029, + "time_s": 75.38 }, - "time_s": 67.01 + "time_s": 75.38, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "faa48f5cc065243cb786397eb654755d85ac321ded2ad43b197fc3e384130c93", + "steps": 7538, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 6701, - "control_steps_replayed": 6701, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-crawl-i00-privileged/embodied-olympics-fun-crawl-i00__YDvTC8q/artifacts/app/output/trajectory.npz", + "n_actions": 7538, + "control_steps_replayed": 7538, "control_hz": 100.0, - "every_control_steps": 9, - "final_state_sha": "e3c198584948a82ed464af7e6034f377f6d997d95e75dd3b591284913baa82b5", + "every_control_steps": 10, + "final_state_sha": "faa48f5cc065243cb786397eb654755d85ac321ded2ad43b197fc3e384130c93", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -68.56, + "az_deg": -70.09, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 1.64, - 3.345 + 2.141, + 3.303 ], "robot_root": "pelvis", - "travel_m": 11.186, + "travel_m": 11.431, "focus": { - "robot": 746 + "robot": 755 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 4.58, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", + "seconds_per_frame": 2.81, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial_choice": "no counted Astra success; a counted success of another run", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:32-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:39-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/crawl/poster.png b/masters/roboolympics-fun/crawl/poster.png index e0a8c50915d2204aa973359a5698325ec5ff00b7..67908f916f541b7b596cffdd108aa5eedbe95dca 100644 --- a/masters/roboolympics-fun/crawl/poster.png +++ b/masters/roboolympics-fun/crawl/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:b6a0569f768ccb6878ca7611b4e051a9fb557668f3f054698b711a310e1e4fbc -size 3305343 +oid sha256:01ce4dc4205fd6d5464e2b72b0204baf7255b1f77a7377bdc823973194ad4faa +size 3311344 diff --git a/masters/roboolympics-fun/crawl/video.mp4 b/masters/roboolympics-fun/crawl/video.mp4 index c17dc17d2e898bc35aaa8d6056a710ef64782cc1..f8b16300226f638567ffa1f2fbd1613434063733 100644 --- a/masters/roboolympics-fun/crawl/video.mp4 +++ b/masters/roboolympics-fun/crawl/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2c4a29b327feff7223f85ae06c8e6dcf23d418a9c17722cc6a525a39c1d4d9de -size 31976838 +oid sha256:fb0a674a9f857c644ae9c9df03b7d39e26d6e2baa09326728770e56e1417aa14 +size 28413111 diff --git a/masters/roboolympics-fun/discus/meta.json b/masters/roboolympics-fun/discus/meta.json index 9cc9ce27b1f8a85517d72c05bc4fd187d50f7277..66fabd07c3adc697bd44976a58870789521089f5 100644 --- a/masters/roboolympics-fun/discus/meta.json +++ b/masters/roboolympics-fun/discus/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "discus", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-discus-i00-privileged/roboolympics-discus-i00-privileg__rnFDUxc", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "discus", + "task_dir": "embodied-olympics-fun-discus-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-discus-i00-privileged/embodied-olympics-fun-discus-i00__Uhxr9X9", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 332, - "seconds": 11.07, + "frames": 289, + "seconds": 9.63, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (01d601fa5a0b19e1)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (fb017b6a0c005f22)", "end_check": { "success": true, "stages": [ @@ -22,62 +25,72 @@ "stages_total": 3, "progress": 1.0, "metrics": { - "release_speed_m_s": 15.58, - "release_height_m": 1.463, - "release_tilt_deg": 24.3, - "release_spin_rad_s": 27.8, - "distance_m": 21.73 + "release_time_s": 1.118, + "release_speed_m_s": 12.68, + "release_height_m": 2.087, + "release_tilt_deg": 71.7, + "release_spin_rad_s": 24.4, + "distance_m": 16.267 }, - "time_s": 8.11 + "time_s": 6.83, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "fb017b6a0c005f2289db7a3c21dd584ce7e7760fecfa933897e2c74709c462af", + "steps": 683, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 811, - "control_steps_replayed": 811, + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-discus-i00-privileged/embodied-olympics-fun-discus-i00__Uhxr9X9/artifacts/app/output/trajectory.npz", + "n_actions": 683, + "control_steps_replayed": 683, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "01d601fa5a0b19e1e9b352392b51dc840637256ae9ef2faf0d302bffdac3f53f", + "final_state_sha": "fb017b6a0c005f2289db7a3c21dd584ce7e7760fecfa933897e2c74709c462af", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -105.14, + "az_deg": -64.76, "az_rule": "object-travel-115", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.527, - 17.207 + 5.578, + 14.867 ], "robot_root": "pelvis", - "travel_m": 1.495, + "travel_m": 1.332, "focus": { - "robot": 8, - "robot+object": 103, - "object": 161 + "robot": 7, + "object": 171, + "robot+object": 51 }, "objects": { - "discus": 45.2 + "discus": 25.02 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 4.32, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Astra (high)", - "run": "codex-gpt6_astra-high", + "seconds_per_frame": 2.7, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", "mode": "privileged", - "trial_choice": "a counted GPT-6 Astra success", - "poster": "frame 97: just after the release / strike, robot and object in view (the last frame no longer shows the robot)", - "host": "gpu3 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:33-0500" + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 41 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:40-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/discus/poster.png b/masters/roboolympics-fun/discus/poster.png index 51e9969c6cdb22c6119e142cf8d077620d3d3bd4..1e5d40946130866fb63a00cc8a03abefce73ad85 100644 --- a/masters/roboolympics-fun/discus/poster.png +++ b/masters/roboolympics-fun/discus/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:dd97cce6bb1297d55de31859b398d4bc0709b63a4e0a940cbfb9e1d7c247c1f4 -size 2992058 +oid sha256:a2e9bbed96c92f81f5713939192d201936a0d65817d9fd6119d80f29401edcbd +size 3027679 diff --git a/masters/roboolympics-fun/discus/video.mp4 b/masters/roboolympics-fun/discus/video.mp4 index ac2737ec6fcc89791509df71f717172e71b1d116..b53ce90c0866d4442671cd5de4fa22dcd50839ba 100644 --- a/masters/roboolympics-fun/discus/video.mp4 +++ b/masters/roboolympics-fun/discus/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:caf6cca967fa537c1ef73468aaa419ff5225b5d19cd67ecf7fc39ead80519913 -size 6489823 +oid sha256:898a064a6519084d2ce9a32d2cda3f4779613d68916e3ecbd401592011041d09 +size 6559398 diff --git a/masters/roboolympics-fun/dodgeball/meta.json b/masters/roboolympics-fun/dodgeball/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..ff2f39281fd57edd2f5d55acf9ff21491d2b3310 --- /dev/null +++ b/masters/roboolympics-fun/dodgeball/meta.json @@ -0,0 +1,103 @@ +{ + "suite": "roboolympics-fun", + "task": "dodgeball", + "family": "dodgeball", + "task_dir": "embodied-olympics-fun-dodgeball-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-dodgeball-i00-privileged/embodied-olympics-fun-dodgeball__sZ5uVQa", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 497, + "seconds": 16.57, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (33db8589fff69e88)", + "end_check": { + "success": true, + "stages": [ + "dodge 1", + "dodge 2", + "dodge 3", + "dodge 4", + "dodge 5" + ], + "stages_total": 5, + "progress": 1.0, + "metrics": { + "dodged": 5 + }, + "time_s": 13.06, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "33db8589fff69e880eb1e6125ee601631f2b646ad4125da63a52e9fd967fcbe9", + "steps": 1306, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-dodgeball-i00-privileged/embodied-olympics-fun-dodgeball__sZ5uVQa/artifacts/app/output/trajectory.npz", + "n_actions": 1306, + "control_steps_replayed": 1306, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "33db8589fff69e880eb1e6125ee601631f2b646ad4125da63a52e9fd967fcbe9", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": 25.39, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 7.133, + 19.085 + ], + "robot_root": "pelvis", + "travel_m": 2.383, + "focus": { + "robot+object": 277, + "robot": 1, + "object": 106, + "robot(incoming)": 53 + }, + "objects": { + "ball1": 20.24, + "turret1_yaw": 0.03, + "ball2": 20.02, + "turret2_yaw": 0.03, + "ball3": 20.18, + "turret3_yaw": 0.03, + "ball4": 18.31, + "turret4_yaw": 0.03, + "ball5": 9.92, + "turret5_yaw": 0.03 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 84 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:40-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/dodgeball/poster.png b/masters/roboolympics-fun/dodgeball/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..f61fc7ebf9f6e7200409adef8141d02b47474c7b --- /dev/null +++ b/masters/roboolympics-fun/dodgeball/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4030ec93781a3e111ec8923c758390c44d9b284f21378b5b410180d5196c2cd5 +size 3081999 diff --git a/masters/roboolympics-fun/dodgeball/video.mp4 b/masters/roboolympics-fun/dodgeball/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..dab9cfa7143b2d11aa286c87df01ff042505628f --- /dev/null +++ b/masters/roboolympics-fun/dodgeball/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9bb5f06833231633b5804c87b60983589cb09ea9e4f35e4d92772aff601f8639 +size 9637449 diff --git a/masters/roboolympics-fun/double_dutch/meta.json b/masters/roboolympics-fun/double_dutch/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..9f14f8294957bccbdb021b40bfcb7d37540e4bb6 --- /dev/null +++ b/masters/roboolympics-fun/double_dutch/meta.json @@ -0,0 +1,102 @@ +{ + "suite": "roboolympics-fun", + "task": "double_dutch", + "family": "double-dutch", + "task_dir": "embodied-olympics-fun-double-dutch-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-double-dutch-i00-privileged/embodied-olympics-fun-double-dut__uSMXSEu", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 264, + "seconds": 8.8, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (ffb43c97d9893f26)", + "end_check": { + "success": true, + "stages": [ + "entry", + "jump 1", + "jump 2", + "jump 3", + "jump 4", + "jump 5", + "jump 6", + "jump 7", + "jump 8", + "jump 9", + "jump 10", + "landing" + ], + "stages_total": 12, + "progress": 1.0, + "metrics": { + "jumps": 10 + }, + "time_s": 6.07, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "ffb43c97d9893f263e1d5a5138d9fd154abe75f07feb95b43a1eebbf8cdb244d", + "steps": 607, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-double-dutch-i00-privileged/embodied-olympics-fun-double-dut__uSMXSEu/artifacts/app/output/trajectory.npz", + "n_actions": 607, + "control_steps_replayed": 607, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "ffb43c97d9893f263e1d5a5138d9fd154abe75f07feb95b43a1eebbf8cdb244d", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.949, + 6.393 + ], + "robot_root": "pelvis", + "travel_m": 1.448, + "focus": { + "robot": 1, + "robot+object": 203 + }, + "objects": { + "crank_a_s": 0.37, + "crank_a_n": 0.37, + "crank_b_s": 0.65, + "crank_b_n": 0.65 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.7, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:40-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/double_dutch/poster.png b/masters/roboolympics-fun/double_dutch/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2c08cac88c2025dcd413c80df31c9b33821c37ca --- /dev/null +++ b/masters/roboolympics-fun/double_dutch/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:912a9bf49848149d00bd9c9446f3cc386b3b47902e7ff21a184336df6abb3689 +size 3121155 diff --git a/masters/roboolympics-fun/double_dutch/video.mp4 b/masters/roboolympics-fun/double_dutch/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..6490d7d36c8374ba0422de64d3074b2303c7cecb --- /dev/null +++ b/masters/roboolympics-fun/double_dutch/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f168ac797460bce38ec7b1750f4ae53dd92e880d44a66db065998c525c1857f4 +size 6032125 diff --git a/masters/roboolympics-fun/dribble_slalom/meta.json b/masters/roboolympics-fun/dribble_slalom/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..fe218af9586c87b43c546221c2f53391baf721c2 --- /dev/null +++ b/masters/roboolympics-fun/dribble_slalom/meta.json @@ -0,0 +1,103 @@ +{ + "suite": "roboolympics-fun", + "task": "dribble_slalom", + "family": "dribble-slalom", + "task_dir": "embodied-olympics-fun-dribble-slalom-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-dribble-slalom-i00-privileged/embodied-olympics-fun-dribble-sl__rfFDxRp", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 791, + "seconds": 26.37, + "speed": 3.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e9cd787e07b81ef5)", + "end_check": { + "success": true, + "stages": [ + "out 1", + "out 2", + "out 3", + "out 4", + "out 5", + "out 6", + "back 6", + "back 5", + "back 4", + "back 3", + "back 2", + "back 1", + "finish" + ], + "stages_total": 13, + "progress": 1.0, + "metrics": { + "planes_ball": 12, + "planes_robot": 12, + "max_ball_distance_m": 1.05, + "touches": 21 + }, + "time_s": 80.25, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "e9cd787e07b81ef5af1a9b4aaa67bdc3547742416a5aac862bd92e5fc3288795", + "steps": 8025, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-dribble-slalom-i00-privileged/embodied-olympics-fun-dribble-sl__rfFDxRp/artifacts/app/output/trajectory.npz", + "n_actions": 8025, + "control_steps_replayed": 8025, + "control_hz": 100.0, + "every_control_steps": 11, + "final_state_sha": "e9cd787e07b81ef5af1a9b4aaa67bdc3547742416a5aac862bd92e5fc3288795", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.24, + 4.325 + ], + "robot_root": "pelvis", + "travel_m": 1.194, + "focus": { + "robot+object": 154, + "robot": 577 + }, + "objects": { + "ball": 0.98 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.61, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:40-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/dribble_slalom/poster.png b/masters/roboolympics-fun/dribble_slalom/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..623ff806941670c3551fb97b28562136b39ed9e8 --- /dev/null +++ b/masters/roboolympics-fun/dribble_slalom/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9040880a23184153d404d9cbd38158419b1cb06fe6c08a1a6411aab41a715f88 +size 3125980 diff --git a/masters/roboolympics-fun/dribble_slalom/video.mp4 b/masters/roboolympics-fun/dribble_slalom/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..9768f12392ad73124443b19528780bc6885ad1c1 --- /dev/null +++ b/masters/roboolympics-fun/dribble_slalom/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fe10e3fd016685d7f90b6689e72358130479ca5ea344bf2b708a810576e7e30e +size 19081995 diff --git a/masters/roboolympics-fun/first_time_shot/meta.json b/masters/roboolympics-fun/first_time_shot/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..92c85f594c980554a5341f7f0fb0c83a7a8c39ef --- /dev/null +++ b/masters/roboolympics-fun/first_time_shot/meta.json @@ -0,0 +1,96 @@ +{ + "suite": "roboolympics-fun", + "task": "first_time_shot", + "family": "first-time-shot", + "task_dir": "embodied-olympics-fun-first-time-shot-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-first-time-shot-i00-privileged/embodied-olympics-fun-first-time__PQr3kXP", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 262, + "seconds": 8.73, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (0ac5783e9b96e22c)", + "end_check": { + "success": true, + "stages": [ + "shot", + "goal" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_x_m": 11.525, + "ball_y_m": 0.007, + "gate_rad": -1.734, + "speed_in_m_s": 3.16, + "speed_out_m_s": 12.95 + }, + "time_s": 6.03, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "0ac5783e9b96e22c2126035694041656637513dced63224bc7f96a49e9e187df", + "steps": 603, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-first-time-shot-i00-privileged/embodied-olympics-fun-first-time__PQr3kXP/artifacts/app/output/trajectory.npz", + "n_actions": 603, + "control_steps_replayed": 603, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "0ac5783e9b96e22c2126035694041656637513dced63224bc7f96a49e9e187df", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -149.16, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.302, + 11.686 + ], + "robot_root": "pelvis", + "travel_m": 0.592, + "focus": { + "robot": 83, + "robot+object": 94, + "robot(incoming)": 11, + "object": 14 + }, + "objects": { + "gate": 0.08, + "ball": 13.58 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 3.14, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 171 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:40-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/first_time_shot/poster.png b/masters/roboolympics-fun/first_time_shot/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2e2d64481a88af7b16abaae42f220cd4a7caf039 --- /dev/null +++ b/masters/roboolympics-fun/first_time_shot/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ccb68fd9b4aa49e1d2b133b416b487a8233555cbb4ea0dd3f44f4157777eadd3 +size 3157732 diff --git a/masters/roboolympics-fun/first_time_shot/video.mp4 b/masters/roboolympics-fun/first_time_shot/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..22dcda68ae6b968d23e91f9fedd1e14022ac2649 --- /dev/null +++ b/masters/roboolympics-fun/first_time_shot/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f6d33ba8a2fe0f2633bcb42eb2021822f27ef130ba0f10c9b218d761f1f89b81 +size 5047706 diff --git a/masters/roboolympics-fun/free_kick_wall/meta.json b/masters/roboolympics-fun/free_kick_wall/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..c532c8794f2dab93aeebd0cbf44b08a224cf1f4d --- /dev/null +++ b/masters/roboolympics-fun/free_kick_wall/meta.json @@ -0,0 +1,93 @@ +{ + "suite": "roboolympics-fun", + "task": "free_kick_wall", + "family": "free-kick-wall", + "task_dir": "embodied-olympics-fun-free-kick-wall-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-free-kick-wall-i00-privileged/embodied-olympics-fun-free-kick__rReLRzw", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 241, + "seconds": 8.03, + "speed": 0.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4ae3234b6a775425)", + "end_check": { + "success": true, + "stages": [ + "kick", + "goal" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_x_m": 20.498, + "kick_speed_m_s": 10.48, + "cross_y_m": 1.343, + "cross_z_m": 0.207 + }, + "time_s": 3.59, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "4ae3234b6a77542559defb471eb76a83307d89a97b53f1b5b07b44533e2e3661", + "steps": 359, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-free-kick-wall-i00-privileged/embodied-olympics-fun-free-kick__rReLRzw/artifacts/app/output/trajectory.npz", + "n_actions": 359, + "control_steps_replayed": 359, + "control_hz": 100.0, + "every_control_steps": 2, + "final_state_sha": "4ae3234b6a77542559defb471eb76a83307d89a97b53f1b5b07b44533e2e3661", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -100.18, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.979, + 14.151 + ], + "robot_root": "pelvis", + "travel_m": 0.224, + "focus": { + "robot+object": 36, + "robot": 45, + "object": 100 + }, + "objects": { + "ball": 20.88 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.75, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 51 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:41-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/free_kick_wall/poster.png b/masters/roboolympics-fun/free_kick_wall/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..4226dc7126fc811d0a9e5b1e5d555765e3254550 --- /dev/null +++ b/masters/roboolympics-fun/free_kick_wall/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd28a6a1872c243a865d6d68e80c4dfcf7f8a3c22f95bff4e35d4f8680377e1e +size 3050730 diff --git a/masters/roboolympics-fun/free_kick_wall/video.mp4 b/masters/roboolympics-fun/free_kick_wall/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..32709677ad9c016ad7f406500ee542109d21c008 --- /dev/null +++ b/masters/roboolympics-fun/free_kick_wall/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4fbd5bc50fd2e104f117d944d01096d847da945693026fd5b75bcff2ee84f430 +size 4084171 diff --git a/masters/roboolympics-fun/free_throw/meta.json b/masters/roboolympics-fun/free_throw/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..168792b5b5148d16adfa70d4e1004708fa5eac31 --- /dev/null +++ b/masters/roboolympics-fun/free_throw/meta.json @@ -0,0 +1,90 @@ +{ + "suite": "roboolympics-fun", + "task": "free_throw", + "family": "free-throw", + "task_dir": "embodied-olympics-fun-free-throw-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-free-throw-i00-privileged/embodied-olympics-fun-free-throw__6uq7st2", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 241, + "seconds": 8.03, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (49b9a36faa29ac9a)", + "end_check": { + "success": true, + "stages": [ + "release", + "through the ring" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_height_m": 2.755 + }, + "time_s": 1.8, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "49b9a36faa29ac9a6580a56785f7e90c2d72c08a99c8d4d9abc7b2b69d4aadc1", + "steps": 180, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-free-throw-i00-privileged/embodied-olympics-fun-free-throw__6uq7st2/artifacts/app/output/trajectory.npz", + "n_actions": 180, + "control_steps_replayed": 180, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "49b9a36faa29ac9a6580a56785f7e90c2d72c08a99c8d4d9abc7b2b69d4aadc1", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the ball drawn as a basketball (orange, black channels)" + ], + "camera": { + "mode": "director", + "az_deg": -114.98, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 13.177, + 14.35 + ], + "robot_root": "pelvis", + "travel_m": 0.311, + "focus": { + "robot+object": 178, + "robot": 3 + }, + "objects": { + "ball": 4.28 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.84, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 23 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:41-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/free_throw/poster.png b/masters/roboolympics-fun/free_throw/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..5d552b135f863ff31d77232921555a3acf69b444 --- /dev/null +++ b/masters/roboolympics-fun/free_throw/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ab28719dfacc14410c0e14c71517fbdf1fb4682cc605cdc8b4f219716ffe5ce +size 3106870 diff --git a/masters/roboolympics-fun/free_throw/video.mp4 b/masters/roboolympics-fun/free_throw/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..d6e51ba9a0e90e7fe08e56766d5fb2c0c27c26f8 --- /dev/null +++ b/masters/roboolympics-fun/free_throw/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acb4cb377e1d02a6704334b35deeb254e3709e41220dba1c8d8639cb209b1919 +size 3735262 diff --git a/masters/roboolympics-fun/giant_slalom/meta.json b/masters/roboolympics-fun/giant_slalom/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..950549fc69aadf1b5f683a1efa49ea18b4aa91f1 --- /dev/null +++ b/masters/roboolympics-fun/giant_slalom/meta.json @@ -0,0 +1,120 @@ +{ + "suite": "roboolympics-fun", + "task": "giant_slalom", + "family": "giant-slalom", + "task_dir": "embodied-olympics-fun-giant-slalom-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-giant-slalom-i00-privileged/embodied-olympics-fun-giant-slal__rNyiV8M", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; snow venue: a snow shader (bright blue-white, rough, slight subsurface, fine bump) on the course's snow geoms, snow-white valley floor; a chase camera: per-frame fit of the robot (height included, so jumps and spins stay framed), at least 6 m away, 24 degree elevation, smoothed in time", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 841, + "seconds": 28.03, + "speed": 3.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (6ddc9991566aed5f)", + "end_check": { + "success": true, + "stages": [ + "start", + "gate 1", + "gate 2", + "gate 3", + "gate 4", + "gate 5", + "gate 6", + "gate 7", + "gate 8", + "gate 9", + "gate 10", + "gate 11", + "gate 12", + "gate 13", + "gate 14", + "gate 15", + "gate 16", + "gate 17", + "gate 18", + "gate 19", + "gate 20", + "gate 21", + "gate 22", + "gate 23", + "gate 24", + "gate 25", + "gate 26", + "finish" + ], + "stages_total": 28, + "progress": 1.0, + "metrics": { + "speed_m_s": 12.69, + "gates": 26, + "time_s": 89.0 + }, + "time_s": 93.53, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "6ddc9991566aed5fe5c4be6adfe40be210b7f157b9457cb04e3ea60a738d18e9", + "steps": 9353, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-giant-slalom-i00-privileged/embodied-olympics-fun-giant-slal__rNyiV8M/artifacts/app/output/trajectory.npz", + "n_actions": 9353, + "control_steps_replayed": 9353, + "control_hz": 100.0, + "every_control_steps": 12, + "final_state_sha": "6ddc9991566aed5fe5c4be6adfe40be210b7f157b9457cb04e3ea60a738d18e9", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "textured snow drawn with a snow shader instead of its checker texture's mean colour", + "the start house, the start and finish arches, the finish posts and banner (drawn-only, no contact) not drawn: they hid the robot from the chase camera; the safety nets and the gates are drawn" + ], + "camera": { + "mode": "director", + "az_deg": -70.31, + "az_rule": "robot-travel-70", + "el_deg": 24.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 6.24, + 8.667 + ], + "robot_root": "pelvis", + "travel_m": 611.005, + "focus": { + "robot": 35, + "robot+object": 746 + }, + "objects": { + "left_ski": 611.14, + "right_ski": 611.17 + } + }, + "orbit_deg": 0.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 5.06, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:41-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/giant_slalom/poster.png b/masters/roboolympics-fun/giant_slalom/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2604b7afea64c2d0920009a090ecb6f020c1ab5a --- /dev/null +++ b/masters/roboolympics-fun/giant_slalom/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c25e4ef2e97c01002caa58cad3371373ebe5fdb0da99cd2c4757b2169c0f5938 +size 2979842 diff --git a/masters/roboolympics-fun/giant_slalom/video.mp4 b/masters/roboolympics-fun/giant_slalom/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..acecd18cfdcf074f1597718c1d8318b3bff2dca3 --- /dev/null +++ b/masters/roboolympics-fun/giant_slalom/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0d1336d75ae24867d8020d805a213449d310e274ad1f67eee4a1408d774b1f06 +size 12654338 diff --git a/masters/roboolympics-fun/goalkeeper_save/meta.json b/masters/roboolympics-fun/goalkeeper_save/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..c4df413646350f71101ae9f12f34c2d272c1e14e --- /dev/null +++ b/masters/roboolympics-fun/goalkeeper_save/meta.json @@ -0,0 +1,97 @@ +{ + "suite": "roboolympics-fun", + "task": "goalkeeper_save", + "family": "goalkeeper-save", + "task_dir": "embodied-olympics-fun-goalkeeper-save-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-goalkeeper-save-i00-privileged/embodied-olympics-fun-goalkeeper__46wnCQn", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 729, + "seconds": 24.3, + "speed": 1.2, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (cf11b1261773597f)", + "end_check": { + "success": true, + "stages": [ + "save 1", + "save 2", + "save 3" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "saved": 3 + }, + "time_s": 26.7, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "cf11b1261773597ff4671b399c1b6ab0d8099cb314533409012daa673709f018", + "steps": 2670, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-goalkeeper-save-i00-privileged/embodied-olympics-fun-goalkeeper__46wnCQn/artifacts/app/output/trajectory.npz", + "n_actions": 2670, + "control_steps_replayed": 2670, + "control_hz": 100.0, + "every_control_steps": 4, + "final_state_sha": "cf11b1261773597ff4671b399c1b6ab0d8099cb314533409012daa673709f018", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -29.02, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.104, + 14.23 + ], + "robot_root": "pelvis", + "travel_m": 0.735, + "focus": { + "robot(outgoing)": 404, + "robot(incoming)": 22, + "robot": 152, + "robot+object": 91 + }, + "objects": { + "machine1": 0.02, + "ball1": 27.32, + "machine2": 0.02, + "ball2": 30.72, + "machine3": 0.02, + "ball3": 16.42 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 5.08, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:42-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/goalkeeper_save/poster.png b/masters/roboolympics-fun/goalkeeper_save/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..e914b15c8c5657d4895418ea23d185cdad2bb77a --- /dev/null +++ b/masters/roboolympics-fun/goalkeeper_save/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9872149837696a459769340efc91b441a200c1f281347d227efe4e7201790617 +size 3208403 diff --git a/masters/roboolympics-fun/goalkeeper_save/video.mp4 b/masters/roboolympics-fun/goalkeeper_save/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..f710bd2bfb32f4c33adc442f012f86dd2790da29 --- /dev/null +++ b/masters/roboolympics-fun/goalkeeper_save/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0c11f9453b3814162213c61a18bd99e60b61349a60522f2baa48fa2df7e7a096 +size 14645562 diff --git a/masters/roboolympics-fun/golf_putt/meta.json b/masters/roboolympics-fun/golf_putt/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..4714c722a70697bec3caa27ae2f10f1464407ab1 --- /dev/null +++ b/masters/roboolympics-fun/golf_putt/meta.json @@ -0,0 +1,95 @@ +{ + "suite": "roboolympics-fun", + "task": "golf_putt", + "family": "golf-putt", + "task_dir": "embodied-olympics-fun-golf-putt-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-golf-putt-i00-privileged/embodied-olympics-fun-golf-putt__eCpKzxA", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 239, + "seconds": 7.97, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (3e81e19dc0796ed5)", + "end_check": { + "success": true, + "stages": [ + "stroke", + "holed" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_to_hole_m": 0.02, + "ball_speed_m_s": 0.002, + "head_contact_s": 0.013, + "head_hits": 1, + "grip_break_max_s": 0.002 + }, + "time_s": 5.34, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "3e81e19dc0796ed5f6b65768e5569b9a0b5c28cdd28a66600dfc0ef84b2b1d3f", + "steps": 534, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-golf-putt-i00-privileged/embodied-olympics-fun-golf-putt__eCpKzxA/artifacts/app/output/trajectory.npz", + "n_actions": 534, + "control_steps_replayed": 534, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "3e81e19dc0796ed5f6b65768e5569b9a0b5c28cdd28a66600dfc0ef84b2b1d3f", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -25.12, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.455, + 3.587 + ], + "robot_root": "pelvis", + "travel_m": 0.003, + "focus": { + "robot+object": 141, + "robot": 19, + "robot+landed": 19 + }, + "objects": { + "putter": 0.4, + "ball": 3.02 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.67, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 73 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:42-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/golf_putt/poster.png b/masters/roboolympics-fun/golf_putt/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..ca719ecfc819f1a2a211a5e257ae7583743f2936 --- /dev/null +++ b/masters/roboolympics-fun/golf_putt/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8c0735ac30d8abe4a55d4189b3728882b49eaa4f68ed7f2c6bf0a74dda380e0 +size 3144933 diff --git a/masters/roboolympics-fun/golf_putt/video.mp4 b/masters/roboolympics-fun/golf_putt/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..71449b7ee42817b48182a94d6026a4bc420c0277 --- /dev/null +++ b/masters/roboolympics-fun/golf_putt/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c47f297c2a1372f1ea388e16fbb12bd9ebeb848cbca01a4a676e5e64faecbf76 +size 3984949 diff --git a/masters/roboolympics-fun/header/meta.json b/masters/roboolympics-fun/header/meta.json index e07aa8740e4bff193964146a7f9ae23f3502246b..cb4a76e2228d12ab949547dda95c301e71d52b7d 100644 --- a/masters/roboolympics-fun/header/meta.json +++ b/masters/roboolympics-fun/header/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "header", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-header-i00-privileged/roboolympics-header-i00-privileg__6BeSeLC", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "header", + "task_dir": "embodied-olympics-fun-header-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-header-i00-privileged/embodied-olympics-fun-header-i00__rStdfSM", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 236, - "seconds": 7.87, + "frames": 229, + "seconds": 7.63, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (83680837df080267)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (923a29d420403b34)", "end_check": { "success": true, "stages": [ @@ -21,61 +24,72 @@ "stages_total": 2, "progress": 1.0, "metrics": { - "ball_x_m": 5.698, - "speed_in_m_s": 11.72, - "speed_out_m_s": 11.74, - "contact_height_on_trunk_m": 0.703 + "ball_x_m": 5.804, + "speed_in_m_s": 11.59, + "speed_out_m_s": 17.32, + "contact_height_on_trunk_m": 0.725 }, - "time_s": 5.25 + "time_s": 5.02, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "923a29d420403b3463147673a20468ad87bdc30eab88bc59ede0859ff4832449", + "steps": 502, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 525, - "control_steps_replayed": 525, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-header-i00-privileged/embodied-olympics-fun-header-i00__rStdfSM/artifacts/app/output/trajectory.npz", + "n_actions": 502, + "control_steps_replayed": 502, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "83680837df08026708e8de6ff2bc71817e9ad3ecad4166d4cabf21c681c7d740", + "final_state_sha": "923a29d420403b3463147673a20468ad87bdc30eab88bc59ede0859ff4832449", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -198.32, + "az_deg": -195.82, "az_rule": "object-travel-115", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.443, - 13.455 + 8.312, + 35.181 ], "robot_root": "pelvis", - "travel_m": 1.006, + "travel_m": 0.28, "focus": { - "robot+object": 40, - "robot(incoming)": 35, - "robot": 101 + "robot(incoming)": 83, + "object": 51, + "robot": 1, + "robot+object": 34 }, "objects": { - "ball": 23.25 + "machine": 0.04, + "ball": 17.55 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 4.29, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", + "seconds_per_frame": 2.94, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial_choice": "no counted Astra success; a counted success of another run", - "poster": "frame 161: just after the release / strike, robot and object in view (the last frame no longer shows the robot)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:33-0500" + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:42-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/header/poster.png b/masters/roboolympics-fun/header/poster.png index 5116c045c8faa51aafb6aa01a264e20c4fdc85d0..8e7f306459e56f4a33d8db0db1f142fac44388ed 100644 --- a/masters/roboolympics-fun/header/poster.png +++ b/masters/roboolympics-fun/header/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:45cfb92c2b5f243ebbffba5566459f6048364f1640d0adbd4e9d2930cbdc94cb -size 3164804 +oid sha256:e2d3a242ec9aa42ec160ab1b56d401915f540817bcc4e1ef4900cfebce429020 +size 3147948 diff --git a/masters/roboolympics-fun/header/video.mp4 b/masters/roboolympics-fun/header/video.mp4 index a4476ff022bf178535c97c4bc6889debf0d3ecd1..cb93fffdaa6db22d323d47237a894484151d1218 100644 --- a/masters/roboolympics-fun/header/video.mp4 +++ b/masters/roboolympics-fun/header/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:48e7a624ee31da218c694cfad60986ea23bf11af3f70dfe84f2466c6458a6aa1 -size 4559684 +oid sha256:507e6dd285254f9a579440575b0ee5445d60109d0c56f0b77c6afd4780fa4ad3 +size 4520312 diff --git a/masters/roboolympics-fun/high_bar_swing/meta.json b/masters/roboolympics-fun/high_bar_swing/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..7c8bbb4b340fbb5f628df94772392b97d1a5d8eb --- /dev/null +++ b/masters/roboolympics-fun/high_bar_swing/meta.json @@ -0,0 +1,92 @@ +{ + "suite": "roboolympics-fun", + "task": "high_bar_swing", + "family": "high-bar-swing", + "task_dir": "embodied-olympics-fun-high-bar-swing-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-high-bar-swing-i00-privileged/embodied-olympics-fun-high-bar-s__QpTJC3u", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 279, + "seconds": 9.3, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (662b3d86be20cb20)", + "end_check": { + "success": true, + "stages": [ + "jumped", + "caught the bar", + "over the top", + "full circle" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "travel_deg": 360.2, + "max_forward_deg": 143.6, + "max_backward_deg": 464.0, + "grip_break_max_s": 0.001, + "time_s": 6.53 + }, + "time_s": 6.53, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "662b3d86be20cb209126aae2f22b506901cf6979f9a12963537d1937bfcdcc8b", + "steps": 653, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-high-bar-swing-i00-privileged/embodied-olympics-fun-high-bar-s__QpTJC3u/artifacts/app/output/trajectory.npz", + "n_actions": 653, + "control_steps_replayed": 653, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "662b3d86be20cb209126aae2f22b506901cf6979f9a12963537d1937bfcdcc8b", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.642, + 6.288 + ], + "robot_root": "pelvis", + "travel_m": 0.435, + "focus": { + "robot": 219 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.65, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:42-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/high_bar_swing/poster.png b/masters/roboolympics-fun/high_bar_swing/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..8483c66a2be8e5b85de11731c4e65efaea738c11 --- /dev/null +++ b/masters/roboolympics-fun/high_bar_swing/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a6a9d8b9ea29834194644bf184a7618c1bebce9673fbec51d4361e8d6cdd6e34 +size 3096126 diff --git a/masters/roboolympics-fun/high_bar_swing/video.mp4 b/masters/roboolympics-fun/high_bar_swing/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..f75f2049c2d2f19f7fbd6f7bf3281a9a44cffc8b --- /dev/null +++ b/masters/roboolympics-fun/high_bar_swing/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a2568041bf4bbe5728925a195f2c19a4db83050e58cc60123e1c71daa5ecd9ef +size 5421875 diff --git a/masters/roboolympics-fun/highlights.json b/masters/roboolympics-fun/highlights.json index 76f9920e697273f7dab6826a336c80f62dfe1f1a..887ce7a5f1ea0ba60ae2a76539b25b38e785a5cf 100644 --- a/masters/roboolympics-fun/highlights.json +++ b/masters/roboolympics-fun/highlights.json @@ -1,24 +1,22 @@ { - "teaser": "dunk", + "teaser": "clean_and_jerk", "highlights": [ - "dunk", + "clean_and_jerk", + "platform_10m", "backflip", - "high-bar-swing", + "high_bar_swing", "javelin", - "volleyball-spike", - "tennis-serve", - "juggle-three", - "box-jump" + "alpine_slalom", + "sumo" ], "why_fewer": "", "why": { - "dunk": "a Unitree H1-2 dunks a basketball through a regulation 3.05 m ring: jump, force the ball down from above with the hand on it, land on its feet (solved by GPT-6 Astra and Opus 5.5 only)", - "backflip": "a standing back salto: take-off, a full turn in the air, landing and standing (GPT-6 Astra)", - "high-bar-swing": "a full giant swing round the high bar from a dead hang, driven only by leg pumps (GPT-6 Luna)", - "javelin": "a standing javelin throw over the 20 m line, head first, then back behind the arc without a foul (GPT-6 Astra)", - "volleyball-spike": "a jump spike: the set comes in, the robot jumps and hits it down into the court over the net (GPT-6 Astra)", - "tennis-serve": "a legal tennis serve: toss, hit with the racket, into the diagonal service box (GPT-6 Astra)", - "juggle-three": "three-ball cascade with two hands: catches and throws in order, nothing dropped (GPT-6 Astra)", - "box-jump": "a two-foot jump onto a 0.61 m box, landing on the feet and standing (GPT-6 Astra)" + "backflip": "a standing backflip: jump, a full backward turn in the air, land on the feet and stand (GPT-6 Astra)", + "platform_10m": "a forward dive straight (101A) from the 10 m platform: take-off, half a somersault, a vertical head-first entry into the water (GPT-6 Astra)", + "high_bar_swing": "jump from the block, catch the high bar with both hands, then a full giant swing round the bar without letting go (GPT-6 Astra)", + "clean_and_jerk": "a Unitree H1 cleans the barbell to its shoulders, jerks it to locked arms overhead and holds it until the down signal (GPT-6 Astra)", + "javelin": "a Unitree H1 throws the javelin over the shoulder from behind the arc: head first, inside the sector, past 20 m, then leaves the runway and stands (GPT-6 Astra)", + "alpine_slalom": "one slalom run on skis: both tips and both feet through each of the 32 gates in order, then over the finish line (GPT-6 Astra)", + "sumo": "a sumo bout on a dohyo against a pushing opponent robot: force it out of the ring before being forced out (GPT-6 Astra)" } } \ No newline at end of file diff --git a/masters/roboolympics-fun/hurdles_freestanding/meta.json b/masters/roboolympics-fun/hurdles_freestanding/meta.json index a3cfb98854d6dd543128ff123f20e67d9d299fd0..03e2732bc12ed5d1ab13f2af533dd0bcb61a57f9 100644 --- a/masters/roboolympics-fun/hurdles_freestanding/meta.json +++ b/masters/roboolympics-fun/hurdles_freestanding/meta.json @@ -1,100 +1,97 @@ { "suite": "roboolympics-fun", "task": "hurdles_freestanding", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-fun-hurdles-freestanding-i00-privileged/roboolympics-fun-hurdles-freesta__HDyKuzX", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "hurdles-freestanding", + "task_dir": "embodied-olympics-fun-hurdles-freestanding-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-hurdles-freestanding-i00-privileged/embodied-olympics-fun-hurdles-fr__wqHuDBn", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 803, - "seconds": 26.77, - "speed": 4.5, - "success_check": true, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (d18f85870df6e886)", + "frames": 801, + "seconds": 26.7, + "speed": 2.7, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (188fb75cbb41f97a)", "end_check": { - "success": true, + "success": false, "stages": [ "hurdle 1", "hurdle 2", - "hurdle 3", - "hurdle 4", - "finish" + "hurdle 3" ], "stages_total": 5, - "progress": 1.0, + "progress": 0.6, "metrics": { - "hurdles_cleared": 4, - "distance_m": 49.36, - "hurdles_down": 1, - "time_s": 111.217 + "hurdles_cleared": 3, + "distance_m": 32.39, + "hurdles_down": 3 }, - "time_s": 111.22 + "time_s": 66.54, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 1, - "n_actions": 11122, - "progress": 1.0, - "stages_reached": 5, - "stages_total": 5, - "metric_distance_m": 49.36, - "metric_hurdles_cleared": 4.0, - "metric_hurdles_down": 1.0, - "metric_time_s": 111.217 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "188fb75cbb41f97a1a2c4c9ec1c9a9acdb939df7477124df214d58f754403729", + "steps": 6654, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-hurdles-freestanding-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-hurdles-freestanding-i00-privileged/roboolympics-fun-hurdles-freesta__HDyKuzX/artifacts/app/output/trajectory.npz", - "n_actions": 11122, - "control_steps_replayed": 11122, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-hurdles-freestanding-i00-privileged/embodied-olympics-fun-hurdles-fr__wqHuDBn/artifacts/app/output/trajectory.npz", + "n_actions": 6654, + "control_steps_replayed": 6654, "control_hz": 100.0, - "every_control_steps": 15, - "final_state_sha": "d18f85870df6e886cf46bdcfd1e512f6076e58c72cc56c4e33017a8e843309ef", - "replays_agree": true, - "mujoco": "3.3.7", - "original_verdict": "counted success (verifier reward 1)" + "every_control_steps": 9, + "final_state_sha": "188fb75cbb41f97a1a2c4c9ec1c9a9acdb939df7477124df214d58f754403729", + "original_verdict": "counted failure (verifier reward 0)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)", - "the finish posts (drawn-only, at the finish line) not drawn: they hid the runner at the end" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -69.99, + "az_deg": -69.69, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.99, - 5.644 + 4.308, + 9.634 ], "robot_root": "pelvis", - "travel_m": 49.583, + "travel_m": 32.557, "focus": { - "robot": 743 + "robot+object": 65, + "robot+landed": 228, + "robot": 448 }, - "objects": {}, - "note": "the hurdles are not followed by the camera (a knocked hurdle would otherwise pull the camera out to a 40 m wide shot)" + "objects": { + "hurdle1": 2.75, + "hurdle2": 2.52, + "hurdle3": 0.07 + } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.96, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-hurdles-freestanding-i00-privileged", + "seconds_per_frame": 2.65, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-hurdles-freestanding-i00-privileged/roboolympics-fun-hurdles-freesta__HDyKuzX", - "trial_choice": "a counted GPT-6 Astra success in privileged mode", - "poster": "frame 406 of the video's own frames: clearing hurdle 2 (the last frame shows the robot fallen past the finish line)", - "host": "vast.ai RTX 4090 gpu1", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:14:59-0500" + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/hurdles_freestanding/poster.png b/masters/roboolympics-fun/hurdles_freestanding/poster.png index b69462873a085ebb5136d2d9b25a3ad3028cc2b6..31cf417a2913949084fc067c16b06c105ceddfb5 100644 --- a/masters/roboolympics-fun/hurdles_freestanding/poster.png +++ b/masters/roboolympics-fun/hurdles_freestanding/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c4d7e74cf501298f3e023495b48ba94bbb41eb2b19b7ce6a86ca4a3be865bf5f -size 3186837 +oid sha256:ee56c1442f6010b75c84de54f802f6dfdc9ae81cf1cd9651e3fa6ed9e2cfd3a2 +size 3215729 diff --git a/masters/roboolympics-fun/hurdles_freestanding/video.mp4 b/masters/roboolympics-fun/hurdles_freestanding/video.mp4 index 4d807f70396460e8e501089e077da7e49e4a596a..b25f34c417f3825171ffe78d3a18f82db984b7c0 100644 --- a/masters/roboolympics-fun/hurdles_freestanding/video.mp4 +++ b/masters/roboolympics-fun/hurdles_freestanding/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8b6717d105928a0be34762cc0b8fb363b2e3b0a0fe3d93659cfe83694ad164bd -size 19964455 +oid sha256:c0c073fd3502353763eaea10bfc4ff49fca3572f19a54b7384657f6faefc7161 +size 18533123 diff --git a/masters/roboolympics-fun/inline_slalom/meta.json b/masters/roboolympics-fun/inline_slalom/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..3f34046f16ef862642a2ab7d51c1d4ea783fc37d --- /dev/null +++ b/masters/roboolympics-fun/inline_slalom/meta.json @@ -0,0 +1,99 @@ +{ + "suite": "roboolympics-fun", + "task": "inline_slalom", + "family": "inline-slalom", + "task_dir": "embodied-olympics-fun-inline-slalom-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-inline-slalom-i00-privileged/embodied-olympics-fun-inline-sla__azMgCQJ", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 824, + "seconds": 27.47, + "speed": 1.5, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (086a472071d0fa4f)", + "end_check": { + "success": false, + "stages": [ + "start", + "cone line", + "cone 5", + "cone 10", + "cone 15", + "cone 20" + ], + "stages_total": 7, + "progress": 0.8571, + "metrics": { + "raw_time_s": 36.624, + "penalties": 19, + "kicked": 0, + "missed_intervals": 19 + }, + "time_s": 38.13, + "failed": "more than 4 cone penalties: the run is void" + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "086a472071d0fa4f1f8618b258e821f1e48c7d8b8d31a71aef59c3b888af89f3", + "steps": 3813, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-inline-slalom-i00-privileged/embodied-olympics-fun-inline-sla__azMgCQJ/artifacts/app/output/trajectory.npz", + "n_actions": 3813, + "control_steps_replayed": 3813, + "control_hz": 100.0, + "every_control_steps": 5, + "final_state_sha": "086a472071d0fa4f1f8618b258e821f1e48c7d8b8d31a71aef59c3b888af89f3", + "original_verdict": "counted failure (verifier reward 0)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -71.35, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.224, + 3.65 + ], + "robot_root": "pelvis", + "travel_m": 28.001, + "focus": { + "robot+object": 63, + "robot": 701 + }, + "objects": { + "skate_left": 28.14, + "skate_right": 28.11 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.63, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/inline_slalom/poster.png b/masters/roboolympics-fun/inline_slalom/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..8c321501c33b318b93b30028e6627184b04eaa0e --- /dev/null +++ b/masters/roboolympics-fun/inline_slalom/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:169d61bcc477ec9533767a04bea54d2ca4ae16176ca2c1da2335277a7ec7c72e +size 3123107 diff --git a/masters/roboolympics-fun/inline_slalom/video.mp4 b/masters/roboolympics-fun/inline_slalom/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..f1535b6ae4d5a231f95878f566619c5cd9959a6e --- /dev/null +++ b/masters/roboolympics-fun/inline_slalom/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:713be793bc222e9ead1ea7d43784117078756e41311e11c5e1ea2729aafbf3a3 +size 10629146 diff --git a/masters/roboolympics-fun/javelin/meta.json b/masters/roboolympics-fun/javelin/meta.json index 539aae9e0b08159235d5ac3dfa233e5703f6c6c7..1e22795bf793b72afbc5a4c6e1ceb017a4741cc5 100644 --- a/masters/roboolympics-fun/javelin/meta.json +++ b/masters/roboolympics-fun/javelin/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "javelin", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-javelin-i00-privileged/roboolympics-javelin-i00-privile__r6nNZRF", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "javelin", + "task_dir": "embodied-olympics-fun-javelin-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-javelin-i00-privileged/embodied-olympics-fun-javelin-i0__cCBqY3s", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 344, - "seconds": 11.47, + "frames": 638, + "seconds": 21.27, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7dca85321b28cdb1)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b2d01dacb744669e)", "end_check": { "success": true, "stages": [ @@ -22,64 +25,74 @@ "stages_total": 3, "progress": 1.0, "metrics": { - "release_speed_m_s": 15.36, - "release_height_m": 1.915, - "release_elevation_deg": 34.1, - "release_attitude_deg": 50.7, - "distance_m": 24.43, + "release_speed_m_s": 14.0, + "release_height_m": 1.908, + "release_elevation_deg": 52.5, + "release_attitude_deg": 25.6, + "distance_m": 20.57, "head_first": 1.0 }, - "time_s": 8.48 + "time_s": 17.3, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "b2d01dacb744669e1345dc36f9a1475b2ed0c46168e33248b7ce6b3f961a134c", + "steps": 1730, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 848, - "control_steps_replayed": 848, + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-javelin-i00-privileged/embodied-olympics-fun-javelin-i0__cCBqY3s/artifacts/app/output/trajectory.npz", + "n_actions": 1730, + "control_steps_replayed": 1730, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "7dca85321b28cdb18baa8a75558063d1dee48e4375837cb34e5977a77eb49092", + "final_state_sha": "b2d01dacb744669e1345dc36f9a1475b2ed0c46168e33248b7ce6b3f961a134c", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": 105.42, + "az_deg": 109.17, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 6.257, - 47.401 + 4.484, + 34.973 ], "robot_root": "pelvis", - "travel_m": 2.765, + "travel_m": 2.686, "focus": { - "object": 109, - "robot+object": 74, - "robot+landed": 93, - "robot": 8 + "robot": 188, + "object": 88, + "robot+landed": 214, + "robot(incoming)": 5, + "robot+object": 83 }, "objects": { - "javelin": 29.17 + "javelin": 21.81 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.9, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Astra (high)", - "run": "codex-gpt6_astra-high", + "seconds_per_frame": 2.63, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", "mode": "privileged", - "trial_choice": "a counted GPT-6 Astra success", - "poster": "frame 69: just after the release / strike, robot and object in view (the last frame no longer shows the robot)", - "host": "gpu2 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:33-0500" + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 257 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/javelin/poster.png b/masters/roboolympics-fun/javelin/poster.png index eb3591279baafa525537b9f3b171af480163842a..5448397b80e7bb28da2f172919c52a35ebf7639c 100644 --- a/masters/roboolympics-fun/javelin/poster.png +++ b/masters/roboolympics-fun/javelin/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c9b6e9bfa985ec77373a923a6ca34f4ace97089d2ede2699199ea889a1d01cb5 -size 3031354 +oid sha256:216d8e24d991d0e437318ffba019b65f88cfdfe1778dcad5eaf11d02f8365251 +size 3020923 diff --git a/masters/roboolympics-fun/javelin/video.mp4 b/masters/roboolympics-fun/javelin/video.mp4 index 8b571163378552a20030ea33f19a094517a4befa..c054709f3d3feac445d10011b0fe4a189f663043 100644 --- a/masters/roboolympics-fun/javelin/video.mp4 +++ b/masters/roboolympics-fun/javelin/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:4c64e6277f001e9595dd254d51de24449ef3b3c94845f51716c1ec328d8d6417 -size 6862382 +oid sha256:404301378155861053661384d4633f78284488efb274ae111bd205a62b3f767d +size 10747105 diff --git a/masters/roboolympics-fun/juggle_three/meta.json b/masters/roboolympics-fun/juggle_three/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..feb558fbd966105f1b20676f23a5b47c55dbff7b --- /dev/null +++ b/masters/roboolympics-fun/juggle_three/meta.json @@ -0,0 +1,92 @@ +{ + "suite": "roboolympics-fun", + "task": "juggle_three", + "family": "juggle-three", + "task_dir": "embodied-olympics-fun-juggle-three-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:claude_code-opus_5_5-high-embodied-olympics-fun-juggle-three-i00-privileged/embodied-olympics-fun-juggle-thr__E3xTcmn", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 269, + "seconds": 8.97, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (d1566225b9ef4488)", + "end_check": { + "success": true, + "stages": [ + "first catch", + "half way", + "ten catches" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "catches": 10 + }, + "time_s": 6.23, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "d1566225b9ef44888203745bf374ffa45bef1882b1497fc34bd90fa07bc1a177", + "steps": 623, + "pass": true + }, + "replay": { + "trajectory": "claude_code-opus_5_5-high-embodied-olympics-fun-juggle-three-i00-privileged/embodied-olympics-fun-juggle-thr__E3xTcmn/artifacts/app/output/trajectory.npz", + "n_actions": 623, + "control_steps_replayed": 623, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "d1566225b9ef44888203745bf374ffa45bef1882b1497fc34bd90fa07bc1a177", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.216, + 4.722 + ], + "robot_root": "pelvis", + "travel_m": 0.008, + "focus": { + "robot+object": 144, + "robot": 65 + }, + "objects": { + "ball_a": 0.8, + "ball_b": 0.13, + "ball_c": 0.34 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.66, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "Claude Code + Opus 5.5 (high)", + "run": "claude_code-opus_5_5-high", + "mode": "privileged", + "trial_choice": "no GPT-6 Astra success; a counted success of the next run in order (Opus, Argon, ...)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/juggle_three/poster.png b/masters/roboolympics-fun/juggle_three/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..7ce5b20de6402ee06f2b092cd59b0e4a522d6a4c --- /dev/null +++ b/masters/roboolympics-fun/juggle_three/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4acbace4d334c2bff9f26c04740bb6cf8ea0a9b73ee824abac0d0f72baeb436d +size 3089447 diff --git a/masters/roboolympics-fun/juggle_three/video.mp4 b/masters/roboolympics-fun/juggle_three/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..1767944a6bb85010e49518495773fa25b764aced --- /dev/null +++ b/masters/roboolympics-fun/juggle_three/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:250c69c742aa2d0e14f9cb217991a20edb196aedc4523f202fa5bd59bc852d11 +size 4391512 diff --git a/masters/roboolympics-fun/juggling/meta.json b/masters/roboolympics-fun/juggling/meta.json index 8127b8e7859d6586adec312a16385497bf19e76a..b1b87d19696bf48aeb3b019188da8e9d1c00a7c8 100644 --- a/masters/roboolympics-fun/juggling/meta.json +++ b/masters/roboolympics-fun/juggling/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "juggling", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-juggling-i00-privileged/roboolympics-juggling-i00-privil__oLLX8k2", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "juggling", + "task_dir": "embodied-olympics-fun-juggling-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-juggling-i00-privileged/embodied-olympics-fun-juggling-i__d9MbrKA", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 369, - "seconds": 12.3, + "frames": 400, + "seconds": 13.33, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (184f602583dde2e6)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (41fb53007a2a88ba)", "end_check": { "success": true, "stages": [ @@ -30,57 +33,66 @@ "progress": 1.0, "metrics": { "touches": 10, - "ball_height_m": 1.568 + "ball_height_m": 1.609 }, - "time_s": 9.22 + "time_s": 10.15, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "41fb53007a2a88ba22da6b66553a126434d018106d3852f7aac6399867971653", + "steps": 1015, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 922, - "control_steps_replayed": 922, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-juggling-i00-privileged/embodied-olympics-fun-juggling-i__d9MbrKA/artifacts/app/output/trajectory.npz", + "n_actions": 1015, + "control_steps_replayed": 1015, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "184f602583dde2e6849da3f5b2f504903273c1e645e1f5399ddba091dcd2628c", + "final_state_sha": "41fb53007a2a88ba22da6b66553a126434d018106d3852f7aac6399867971653", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -50.0, + "az_deg": -50.08, "az_rule": "facing-50", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.395, - 4.752 + 3.27, + 5.265 ], "robot_root": "pelvis", - "travel_m": 0.006, + "travel_m": 0.075, "focus": { - "robot": 47, - "robot+object": 262 + "robot+object": 200, + "robot": 140 }, "objects": { - "ball": 0.28 + "ball": 0.38 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.77, - "image": "eai-robofunolympics:0.2.1", + "seconds_per_frame": 2.56, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu3 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:34-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/juggling/poster.png b/masters/roboolympics-fun/juggling/poster.png index 7850605dd2489ba5319ec1ad13c022a67ac355b2..3cb36c73d491a03c9d1df5a730a221b1aebb82f2 100644 --- a/masters/roboolympics-fun/juggling/poster.png +++ b/masters/roboolympics-fun/juggling/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:ff139b27e8ed6c257976223cbb8732f94694435462a42668bb5bbb83ec1ceb0d -size 3049303 +oid sha256:8fa82ecc99a44c6aa2afd07e016487556cdc136c9ae69a4c552ac7ea1fec8797 +size 3058387 diff --git a/masters/roboolympics-fun/juggling/video.mp4 b/masters/roboolympics-fun/juggling/video.mp4 index 77fa1a01eee19d3091fb2763489c5d19cc2073c3..e430626eb7de396d37b4180b3c026ed2525f1ce2 100644 --- a/masters/roboolympics-fun/juggling/video.mp4 +++ b/masters/roboolympics-fun/juggling/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:835628496cca87a9558ea4771acc10bd01def78a186d26009f87144ecf0805ab -size 5708853 +oid sha256:124ca24738e4db83b686db412254e40bb30933ef3c743a61aadf51d5137b4f64 +size 6213731 diff --git a/masters/roboolympics-fun/juggling_feet/meta.json b/masters/roboolympics-fun/juggling_feet/meta.json index e9d171db5b7b6bb1571e859359fd8143f2009577..20c89c2916662fa94c6d14173e6d80cc50c57f98 100644 --- a/masters/roboolympics-fun/juggling_feet/meta.json +++ b/masters/roboolympics-fun/juggling_feet/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "juggling_feet", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-fun-juggling-feet-i00-privileged/roboolympics-juggling-feet-i00-p__p6ohQgr", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "juggling-feet", + "task_dir": "embodied-olympics-fun-juggling-feet-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-juggling-feet-i00-privileged/embodied-olympics-fun-juggling-f__SnUL9zt", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 245, - "seconds": 8.17, + "frames": 318, + "seconds": 10.6, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (77de17d2e73b169a)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (09e01b816a4f00c7)", "end_check": { "success": true, "stages": [ @@ -31,74 +34,66 @@ "metrics": { "touches": 10, "bounces": 0, - "ball_height_m": 0.509 + "ball_height_m": 0.676 }, - "time_s": 5.52 + "time_s": 7.71, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 1, - "n_actions": 552, - "progress": 1.0, - "stages_reached": 10, - "stages_total": 10, - "metric_ball_height_m": 0.509, - "metric_bounces": 0.0, - "metric_touches": 10.0 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "09e01b816a4f00c77fe7eb824eee151a966c205b918546fcb7ed2c8061801f98", + "steps": 771, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-juggling-feet-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-juggling-feet-i00-privileged/roboolympics-juggling-feet-i00-p__p6ohQgr/artifacts/app/output/trajectory.npz", - "n_actions": 552, - "control_steps_replayed": 552, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-juggling-feet-i00-privileged/embodied-olympics-fun-juggling-f__SnUL9zt/artifacts/app/output/trajectory.npz", + "n_actions": 771, + "control_steps_replayed": 771, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "77de17d2e73b169ad6ef8ead57ad405b483971baac10f0d5352fdef7f00bf782", - "replays_agree": true, - "mujoco": "3.3.7", + "final_state_sha": "09e01b816a4f00c77fe7eb824eee151a966c205b918546fcb7ed2c8061801f98", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -50.0, + "az_deg": -50.08, "az_rule": "facing-50", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 2.993, - 3.327 + 3.098, + 3.264 ], "robot_root": "pelvis", - "travel_m": 0.158, + "travel_m": 0.172, "focus": { - "robot+object": 166, - "robot": 19 + "robot": 50, + "robot+object": 208 }, "objects": { - "ball": 0.53 + "ball": 0.19 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.97, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-juggling-feet-i00-privileged", + "seconds_per_frame": 2.61, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-juggling-feet-i00-privileged/roboolympics-juggling-feet-i00-p__p6ohQgr", - "trial_choice": "a counted GPT-6 Astra success in privileged mode", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "vast.ai RTX 4090 gpu4", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:20:23-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/juggling_feet/poster.png b/masters/roboolympics-fun/juggling_feet/poster.png index a43626412231dbc0f20d3a6ed07e6f51652b7252..747d1559d4cab1a2cc82390bcffb5bea166e3b93 100644 --- a/masters/roboolympics-fun/juggling_feet/poster.png +++ b/masters/roboolympics-fun/juggling_feet/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:cbd65f9f0a1dd4f7ff91fff452a8b6690ee267c8bba17eade07821986517e2aa -size 3119034 +oid sha256:d96bee5bc4cbd96d3f4b1a4ba5ac00206966f73ee563fa858dd4eab334ba6129 +size 3135823 diff --git a/masters/roboolympics-fun/juggling_feet/video.mp4 b/masters/roboolympics-fun/juggling_feet/video.mp4 index 459b7cd34aafea4afee5b6083e86139710012784..345a16f5dca498d2573677ce85d973fefc8eee99 100644 --- a/masters/roboolympics-fun/juggling_feet/video.mp4 +++ b/masters/roboolympics-fun/juggling_feet/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:13e6971e38dcb05dfe1370c681799e1b3cbe363e22ccbc65d4008cd8c2baf17d -size 4097903 +oid sha256:fa1edf3344b4e88e538bdba8ced7f0e0d12b4f6b790a032f2863adbf7e6372a3 +size 5523176 diff --git a/masters/roboolympics-fun/juggling_head/meta.json b/masters/roboolympics-fun/juggling_head/meta.json index c1cf79e50764114f699e17e35f84a56aed077069..34dfd74a7b05a210c907fa08011488b57ce52351 100644 --- a/masters/roboolympics-fun/juggling_head/meta.json +++ b/masters/roboolympics-fun/juggling_head/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "juggling_head", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-fun-juggling-head-i00-privileged/roboolympics-fun-juggling-head-i__Xvvrvjm", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "juggling-head", + "task_dir": "embodied-olympics-fun-juggling-head-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-juggling-head-i00-privileged/embodied-olympics-fun-juggling-h__YtL3oAJ", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 327, - "seconds": 10.9, + "frames": 364, + "seconds": 12.13, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (4e171afae3d0274d)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4578e30b1c6edb3d)", "end_check": { "success": true, "stages": [ @@ -30,35 +33,29 @@ "progress": 1.0, "metrics": { "headers": 10, - "ball_height_m": 1.62 + "ball_height_m": 1.608 }, - "time_s": 7.96 + "time_s": 9.08, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 1, - "n_actions": 796, - "progress": 1.0, - "stages_reached": 10, - "stages_total": 10, - "metric_ball_height_m": 1.62, - "metric_headers": 10.0 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "4578e30b1c6edb3d0b2891c7bc6949a6c605a043e4b8e967790794812cd6372a", + "steps": 908, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-juggling-head-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-juggling-head-i00-privileged/roboolympics-fun-juggling-head-i__Xvvrvjm/artifacts/app/output/trajectory.npz", - "n_actions": 796, - "control_steps_replayed": 796, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-juggling-head-i00-privileged/embodied-olympics-fun-juggling-h__YtL3oAJ/artifacts/app/output/trajectory.npz", + "n_actions": 908, + "control_steps_replayed": 908, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "4e171afae3d0274ddc9eb2656b7fbb8d13f353fd3a9c8567b2768d420c423aa1", - "replays_agree": true, - "mujoco": "3.3.7", + "final_state_sha": "4578e30b1c6edb3d0b2891c7bc6949a6c605a043e4b8e967790794812cd6372a", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", @@ -68,35 +65,35 @@ "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 5.142, - 5.158 + 4.67, + 6.438 ], "robot_root": "pelvis", - "travel_m": 0.045, + "travel_m": 0.028, "focus": { - "robot+object": 229, - "robot": 38 + "robot": 36, + "robot+object": 268 }, "objects": { - "ball": 0.0 + "machine": 0.02, + "ball": 1.08 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.93, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-juggling-head-i00-privileged", + "seconds_per_frame": 2.65, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-juggling-head-i00-privileged/roboolympics-fun-juggling-head-i__Xvvrvjm", - "trial_choice": "a counted GPT-6 Astra success in privileged mode", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "vast.ai RTX 4090 gpu4", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:15:00-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:43-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/juggling_head/poster.png b/masters/roboolympics-fun/juggling_head/poster.png index 56f56c7055d40ad96b5d426d5bfc872c447711e3..d3a4ceb0873d00fcb31e319fbebd6884447d2c39 100644 --- a/masters/roboolympics-fun/juggling_head/poster.png +++ b/masters/roboolympics-fun/juggling_head/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:df9ff321f8b0f3634b0fe4519f533b52fbb41eec9f3df78696c7b2f6201ac8cc -size 3104332 +oid sha256:6ef082de501b5d207e253890ef2ad3d5a384bec7c8421dc4c2151f356b33064e +size 3132918 diff --git a/masters/roboolympics-fun/juggling_head/video.mp4 b/masters/roboolympics-fun/juggling_head/video.mp4 index 53f3255b36e57f451424ac5fb3fac70931dc78ac..92519d1205e37d2865e8b6da12f5a9254ccb290b 100644 --- a/masters/roboolympics-fun/juggling_head/video.mp4 +++ b/masters/roboolympics-fun/juggling_head/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:659236fad7623e6c8325137ef53536072593cbe4fcc076324331a90cd9842dcc -size 5091620 +oid sha256:cb9241defef31711c935a277f6d6203de747abdb887559b4762b9e2904f0b814 +size 6180712 diff --git a/masters/roboolympics-fun/jump_rope/meta.json b/masters/roboolympics-fun/jump_rope/meta.json index 012c3b89c1db1a413db867fa6862f326dbb43d7b..e5ed4f91942554cb64dc7cee11d4cd5f1403e9d0 100644 --- a/masters/roboolympics-fun/jump_rope/meta.json +++ b/masters/roboolympics-fun/jump_rope/meta.json @@ -1,68 +1,57 @@ { "suite": "roboolympics-fun", "task": "jump_rope", + "family": "jump-rope", + "task_dir": "embodied-olympics-fun-jump-rope-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", "source": "scene-orbit", - "why_no_video": "No evaluated agent has solved this task (no counted success in any run when this was made). As for the other unsolved RoboOlympics-Fun tasks on the site (board 0161 rule), this video shows the best failed attempt instead, replayed exactly (not a solved run): the task's own check at the end says it failed, as the verifier did.", - "shows": "agent-attempt (failed):codex-gpt6_astra-high-roboolympics-fun-jump-rope-i00-privileged/roboolympics-fun-jump-rope-i00-p__5weSQuL", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-jump-rope-i00-privileged/embodied-olympics-fun-jump-rope__VhNzNrn", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 711, - "seconds": 23.7, + "frames": 347, + "seconds": 11.57, "speed": 0.9, "success_check": false, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (9c7a348dddae099e)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b7708509232220f7)", "end_check": { "success": false, "stages": [ "jump 1", "jump 2", "jump 3", - "jump 4", - "jump 5", - "jump 6", - "jump 7", - "jump 8", - "jump 9", - "jump 10", - "jump 11", - "jump 12" + "jump 4" ], "stages_total": 18, - "progress": 0.6667, + "progress": 0.2222, "metrics": { - "count": 12, - "rope_turns": 10.715 + "count": 4, + "rope_turns": 4.888 }, - "time_s": 19.5 + "time_s": 8.58, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 0, - "n_actions": 1950, - "progress": 0.6667, - "stages_reached": 12, - "stages_total": 18, - "metric_count": 12.0, - "metric_rope_turns": 10.715 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "b7708509232220f728a0b0657cdc97e0db96ed95a23493a9e99a105f2e104d86", + "steps": 858, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-jump-rope-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-jump-rope-i00-privileged/roboolympics-fun-jump-rope-i00-p__5weSQuL/artifacts/app/output/trajectory.npz", - "n_actions": 1950, - "control_steps_replayed": 1950, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-jump-rope-i00-privileged/embodied-olympics-fun-jump-rope__VhNzNrn/artifacts/app/output/trajectory.npz", + "n_actions": 858, + "control_steps_replayed": 858, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "9c7a348dddae099ee1b19c66e9b0146e3aaf79940cffa9b40d823803984b1a55", - "replays_agree": true, - "mujoco": "3.3.7", + "final_state_sha": "b7708509232220f728a0b0657cdc97e0db96ed95a23493a9e99a105f2e104d86", "original_verdict": "counted failure (verifier reward 0)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", @@ -72,64 +61,63 @@ "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.153, - 4.531 + 3.387, + 3.499 ], "robot_root": "pelvis", - "travel_m": 1.277, + "travel_m": 0.56, "focus": { - "robot+object": 645, - "robot": 6 + "robot+object": 286, + "robot": 1 }, "objects": { - "handle_left": 1.45, - "handle_right": 1.44, - "rope0": 1.41, - "rope1": 1.38, - "rope2": 1.33, - "rope3": 1.29, - "rope4": 1.25, - "rope5": 1.22, - "rope6": 1.14, - "rope7": 1.03, - "rope8": 0.92, - "rope9": 0.79, - "rope10": 0.66, - "rope11": 0.55, - "rope12": 0.44, - "rope13": 0.38, - "rope14": 0.36, - "rope15": 0.4, - "rope16": 0.52, - "rope17": 0.65, - "rope18": 0.77, - "rope19": 0.87, - "rope20": 0.98, - "rope21": 1.09, - "rope22": 1.2, - "rope23": 1.27, - "rope24": 1.3, - "rope25": 1.33, - "rope26": 1.36, - "rope27": 1.4 + "handle_left": 0.59, + "handle_right": 0.53, + "rope0": 0.6, + "rope1": 0.6, + "rope2": 0.6, + "rope3": 0.61, + "rope4": 0.59, + "rope5": 0.56, + "rope6": 0.51, + "rope7": 0.45, + "rope8": 0.41, + "rope9": 0.37, + "rope10": 0.33, + "rope11": 0.34, + "rope12": 0.42, + "rope13": 0.48, + "rope14": 0.49, + "rope15": 0.43, + "rope16": 0.38, + "rope17": 0.43, + "rope18": 0.49, + "rope19": 0.51, + "rope20": 0.53, + "rope21": 0.57, + "rope22": 0.58, + "rope23": 0.57, + "rope24": 0.57, + "rope25": 0.56, + "rope26": 0.56, + "rope27": 0.54 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.0, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-jump-rope-i00-privileged", + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-jump-rope-i00-privileged/roboolympics-fun-jump-rope-i00-p__5weSQuL", - "trial_choice": "no counted success in any run; GPT-6 Astra's privileged attempt with the most progress", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", "poster": "last frame (end of the failed attempt)", - "host": "vast.ai RTX 4090 gpu2", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:15:00-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/jump_rope/poster.png b/masters/roboolympics-fun/jump_rope/poster.png index 734e34f934b90eb34f9a0868580983ffe16ea777..91065dff3dfad494d62faf4f7495baf384c3510c 100644 --- a/masters/roboolympics-fun/jump_rope/poster.png +++ b/masters/roboolympics-fun/jump_rope/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:1026a4ffa273dbbdac6807f1d03d3092e9c372f6465c410548c51a58514dd97a -size 3116789 +oid sha256:52278e60056f74d6e8ab641efb393164feef98a1870539530c17e9b1daae5f5e +size 3117959 diff --git a/masters/roboolympics-fun/jump_rope/video.mp4 b/masters/roboolympics-fun/jump_rope/video.mp4 index 4fc347e504366dab2782ab8eae4fc2b4ab8134dd..62581354f184605b87a9f7748d463701285bd970 100644 --- a/masters/roboolympics-fun/jump_rope/video.mp4 +++ b/masters/roboolympics-fun/jump_rope/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:466feea0e154589680ef5e15fa2384e6299c2a5e47c88373cf6002c02a78d733 -size 13163764 +oid sha256:b0825c3e6b32989f8cfa709001b29b7f7c2cb7b9cd5df65dd5c0d7338a306a4f +size 6450165 diff --git a/masters/roboolympics-fun/kick_target/meta.json b/masters/roboolympics-fun/kick_target/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..46cae2f74f6e29ef60cea1999dd7b74c39ab7904 --- /dev/null +++ b/masters/roboolympics-fun/kick_target/meta.json @@ -0,0 +1,92 @@ +{ + "suite": "roboolympics-fun", + "task": "kick_target", + "family": "kick-target", + "task_dir": "embodied-olympics-fun-kick-target-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-kick-target-i00-privileged/embodied-olympics-fun-kick-targe__D9hQQeK", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 245, + "seconds": 8.17, + "speed": 0.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b1e4c8852510a551)", + "end_check": { + "success": true, + "stages": [ + "strike", + "stand" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "hit_height_m": 1.061, + "impulse_n_s": 32.35, + "contact_impulse_n_s": 37.89, + "contact_s": 0.091 + }, + "time_s": 3.67, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "b1e4c8852510a55118406050f885c50a11fa87c556ce5c8887a198ef646e793a", + "steps": 367, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-kick-target-i00-privileged/embodied-olympics-fun-kick-targe__D9hQQeK/artifacts/app/output/trajectory.npz", + "n_actions": 367, + "control_steps_replayed": 367, + "control_hz": 100.0, + "every_control_steps": 2, + "final_state_sha": "b1e4c8852510a55118406050f885c50a11fa87c556ce5c8887a198ef646e793a", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.142, + 5.167 + ], + "robot_root": "pelvis", + "travel_m": 0.111, + "focus": { + "robot": 59, + "robot+object": 126 + }, + "objects": { + "bag": 0.28 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.63, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 160 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/kick_target/poster.png b/masters/roboolympics-fun/kick_target/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..05a39a15394612ef46aab6554aade2cf9840be1a --- /dev/null +++ b/masters/roboolympics-fun/kick_target/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b0210231dc9eba4d503d6ab34aa1d3205ff85324f908b8b14ec1e039398ffe09 +size 3156743 diff --git a/masters/roboolympics-fun/kick_target/video.mp4 b/masters/roboolympics-fun/kick_target/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..e0679f04a8415f8408033ca09d956ee4b0729b6b --- /dev/null +++ b/masters/roboolympics-fun/kick_target/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:92fc646c990e315858f102f0bfa28ea6e500aae6d542d3304b4380b04bf422f2 +size 3884908 diff --git a/masters/roboolympics-fun/layup/meta.json b/masters/roboolympics-fun/layup/meta.json index fad0b3b3457d19b7c20d864d46ade3cd95039555..b9522c0f7c888e2e3d5d00f7f1958b1ee752bf1c 100644 --- a/masters/roboolympics-fun/layup/meta.json +++ b/masters/roboolympics-fun/layup/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "layup", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-layup-i00-privileged/roboolympics-layup-i00-privilege__fuzG3zE", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "layup", + "task_dir": "embodied-olympics-fun-layup-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-layup-i00-privileged/embodied-olympics-fun-layup-i00__MG8gHYw", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 249, - "seconds": 8.3, - "speed": 0.3, + "frames": 236, + "seconds": 7.87, + "speed": 0.6, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e44066c1f0b6286f)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b39a9d015080c249)", "end_check": { "success": true, "stages": [ @@ -22,60 +25,69 @@ "stages_total": 3, "progress": 1.0, "metrics": { - "pelvis_to_ring_m": 0.31, - "ball_height_m": 2.783, - "steps": 1 + "pelvis_to_ring_m": 1.881, + "ball_height_m": 2.78, + "steps": 2 }, - "time_s": 1.88 + "time_s": 3.49, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "b39a9d015080c24951eb6af19e049567a059664c3c8d3fe8f8679307ea8f0c91", + "steps": 349, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 188, - "control_steps_replayed": 188, + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-layup-i00-privileged/embodied-olympics-fun-layup-i00__MG8gHYw/artifacts/app/output/trajectory.npz", + "n_actions": 349, + "control_steps_replayed": 349, "control_hz": 100.0, - "every_control_steps": 1, - "final_state_sha": "e44066c1f0b6286f9dcf2ad28032b65c826d1e215fc6932da22fef62222a5ad6", + "every_control_steps": 2, + "final_state_sha": "b39a9d015080c24951eb6af19e049567a059664c3c8d3fe8f8679307ea8f0c91", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)", + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", "the ball drawn as a basketball (orange, black channels)" ], "camera": { "mode": "director", - "az_deg": -75.23, - "az_rule": "robot-travel-70", + "az_deg": -50.0, + "az_rule": "facing-50", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 5.227, - 7.886 + 4.25, + 8.211 ], "robot_root": "pelvis", - "travel_m": 2.842, + "travel_m": 1.147, "focus": { - "robot+object": 152, - "robot": 37 + "robot+object": 107, + "robot": 69 }, "objects": { - "ball": 2.52 + "ball": 2.7 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.02, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Astra (high)", - "run": "codex-gpt6_astra-high", + "seconds_per_frame": 2.9, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", "mode": "privileged", - "trial_choice": "a counted GPT-6 Astra success", - "poster": "last frame (solved state)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:34-0500" + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 133 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/layup/poster.png b/masters/roboolympics-fun/layup/poster.png index 803b7e9283f43bd921ac0c3f87ddc94cac85ccce..9b0d0eb28c3309a5b9e738a382a790d60980a0de 100644 --- a/masters/roboolympics-fun/layup/poster.png +++ b/masters/roboolympics-fun/layup/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:f7762e8ff9e3586a342affaed01010a5f0eec7c27cc948ba1446f91f54a330e5 -size 3224608 +oid sha256:e4774a2739f5745500d79e912003d3964cf717d6e102fc66e26b1d79927548d7 +size 3158165 diff --git a/masters/roboolympics-fun/layup/video.mp4 b/masters/roboolympics-fun/layup/video.mp4 index 99e7d160b756926558b9c414b231e6af2ef458ea..adb20663fb2f265eabd1b7d8a9ddc318ce89e66d 100644 --- a/masters/roboolympics-fun/layup/video.mp4 +++ b/masters/roboolympics-fun/layup/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7f8bafa9f73dffeda56529253bb43ee36a84f4ba2b416e4f41ba1946fd8b6bcf -size 4676662 +oid sha256:7c98b0f3d93beca87dcb619486efe9c613558f4e39e8a70a911ac92c291e7e3f +size 4788944 diff --git a/masters/roboolympics-fun/long_jump/meta.json b/masters/roboolympics-fun/long_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..c350b4a57cef7c7cb2a5792733077440224fbf8f --- /dev/null +++ b/masters/roboolympics-fun/long_jump/meta.json @@ -0,0 +1,89 @@ +{ + "suite": "roboolympics-fun", + "task": "long_jump", + "family": "long-jump", + "task_dir": "embodied-olympics-fun-long-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-long-jump-i00-privileged/embodied-olympics-fun-long-jump__mdnyE4R", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 369, + "seconds": 12.3, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b9bbe414cbb82521)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "landing" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "air_s": 0.469, + "rise_m": 0.167, + "takeoff_x_m": -0.033, + "mark_m": 2.519 + }, + "time_s": 9.22, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "b9bbe414cbb825218f00b2f71c68d931d3764b87b773ab85f4000b2181a2456b", + "steps": 922, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-long-jump-i00-privileged/embodied-olympics-fun-long-jump__mdnyE4R/artifacts/app/output/trajectory.npz", + "n_actions": 922, + "control_steps_replayed": 922, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "b9bbe414cbb825218f00b2f71c68d931d3764b87b773ab85f4000b2181a2456b", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -71.1, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 1.786, + 3.889 + ], + "robot_root": "pelvis", + "travel_m": 22.766, + "focus": { + "robot": 309 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.57, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/long_jump/poster.png b/masters/roboolympics-fun/long_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..b7b49b57d025fceecbc0c89cfc6440c2fd4dc2dd --- /dev/null +++ b/masters/roboolympics-fun/long_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad53969ea015869c7fb89491b65e67d100daf6a2d2a8a8019a23ba78bb01d572 +size 3164824 diff --git a/masters/roboolympics-fun/long_jump/video.mp4 b/masters/roboolympics-fun/long_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..cf0e2ca4859c620b20afcfce3b098403898aa4a2 --- /dev/null +++ b/masters/roboolympics-fun/long_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fda69ed651dd2c126f229cf6691c627b89654a8d228fb15ad4552862296e8a8e +size 8150586 diff --git a/masters/roboolympics-fun/long_rope/meta.json b/masters/roboolympics-fun/long_rope/meta.json index 3ba8f2a48bc67d9b09ee5046f898a70be9a51380..9c58f4fa5b3a3f23b54db2c1d644819474fdd802 100644 --- a/masters/roboolympics-fun/long_rope/meta.json +++ b/masters/roboolympics-fun/long_rope/meta.json @@ -1,8 +1,11 @@ { "suite": "roboolympics-fun", "task": "long_rope", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-fun-long-rope-i00-privileged/roboolympics-fun-long-rope-i00-p__4d9dxc6", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "long-rope", + "task_dir": "embodied-olympics-fun-long-rope-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-long-rope-i00-privileged/embodied-olympics-fun-long-rope__JkqfFEE", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, @@ -11,7 +14,7 @@ "seconds": 13.77, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (rcb_rfo.py Episode.verdict(), the referee, from the task dir's tests/ (privileged verifier copy)) at the end of this re-render's own replay of every recorded row from the task's frozen instance (t = 0 state hash checked); a second plain replay with the task's own replay() agrees and ends in the same final state hash (6bd8fc1468432496)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b5ae798b05e77ec4)", "end_check": { "success": true, "stages": [ @@ -33,32 +36,27 @@ "metrics": { "jumps": 10 }, - "time_s": 10.54 + "time_s": 10.55, + "failed": null }, - "recorded_result": { - "verifier_reward_json": { - "success": 1, - "n_actions": 1054, - "progress": 1.0, - "stages_reached": 12, - "stages_total": 12, - "metric_jumps": 10.0 - } + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "b5ae798b05e77ec458c04a835b5ac66e2b2c386ec8aa19d6fb629f6efedf19a1", + "steps": 1055, + "pass": true }, "replay": { - "task_dir": "roboolympics-fun-long-rope-i00-privileged", - "trajectory": "jobs/codex/gpt6_astra/high/codex-gpt6_astra-high-roboolympics-fun-long-rope-i00-privileged/roboolympics-fun-long-rope-i00-p__4d9dxc6/artifacts/app/output/trajectory.npz", - "n_actions": 1054, - "control_steps_replayed": 1054, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-long-rope-i00-privileged/embodied-olympics-fun-long-rope__JkqfFEE/artifacts/app/output/trajectory.npz", + "n_actions": 1055, + "control_steps_replayed": 1055, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "6bd8fc1468432496183321da325f06312c7482b58b9b961baa7b08330b172f49", - "replays_agree": true, - "mujoco": "3.3.7", + "final_state_sha": "b5ae798b05e77ec458c04a835b5ac66e2b2c386ec8aa19d6fb629f6efedf19a1", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", @@ -68,76 +66,75 @@ "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.736, - 4.816 + 4.706, + 4.801 ], "robot_root": "pelvis", - "travel_m": 0.031, + "travel_m": 0.041, "focus": { - "robot+object": 352, - "robot": 1 + "robot": 1, + "robot+object": 352 }, "objects": { - "hand_a": 0.59, - "hand_b": 0.59, - "rope0": 0.56, - "rope1": 0.58, - "rope2": 0.6, - "rope3": 0.63, - "rope4": 0.65, - "rope5": 0.67, - "rope6": 0.68, - "rope7": 0.7, - "rope8": 0.71, - "rope9": 0.72, - "rope10": 0.73, - "rope11": 0.74, - "rope12": 0.76, - "rope13": 0.77, - "rope14": 0.79, - "rope15": 0.8, - "rope16": 0.8, - "rope17": 0.8, - "rope18": 0.8, - "rope19": 0.79, - "rope20": 0.79, - "rope21": 0.79, - "rope22": 0.78, - "rope23": 0.77, - "rope24": 0.76, - "rope25": 0.74, - "rope26": 0.73, - "rope27": 0.71, - "rope28": 0.7, - "rope29": 0.68, - "rope30": 0.67, - "rope31": 0.66, - "rope32": 0.65, - "rope33": 0.64, - "rope34": 0.63, - "rope35": 0.61, - "rope36": 0.6, - "rope37": 0.6, - "rope38": 0.59, - "rope39": 0.57 + "crank_a": 0.3, + "crank_b": 0.3, + "rope0": 0.62, + "rope1": 0.67, + "rope2": 0.72, + "rope3": 0.76, + "rope4": 0.8, + "rope5": 0.83, + "rope6": 0.85, + "rope7": 0.86, + "rope8": 0.87, + "rope9": 0.87, + "rope10": 0.87, + "rope11": 0.87, + "rope12": 0.87, + "rope13": 0.89, + "rope14": 0.9, + "rope15": 0.92, + "rope16": 0.93, + "rope17": 0.95, + "rope18": 0.96, + "rope19": 0.97, + "rope20": 0.97, + "rope21": 0.97, + "rope22": 0.97, + "rope23": 0.95, + "rope24": 0.94, + "rope25": 0.93, + "rope26": 0.92, + "rope27": 0.91, + "rope28": 0.91, + "rope29": 0.91, + "rope30": 0.9, + "rope31": 0.9, + "rope32": 0.89, + "rope33": 0.87, + "rope34": 0.85, + "rope35": 0.83, + "rope36": 0.79, + "rope37": 0.74, + "rope38": 0.69, + "rope39": 0.63 } }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.12, - "image": "ghcr.io/mll-lab-nu/efe:robofunolympics-0.2.1 (simulator) + ghcr.io/mll-lab-nu/efe:roboolympics-fun-files-0.1.0@sha256:49a33342a72b57b56742cc754e427205a30f39dfcd69a99d1d7275c0eae4a619 (rcb_rfo.py, rfo/)", - "task_dir": "roboolympics-fun-long-rope-i00-privileged", + "seconds_per_frame": 2.66, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial": "codex-gpt6_astra-high-roboolympics-fun-long-rope-i00-privileged/roboolympics-fun-long-rope-i00-p__4d9dxc6", - "trial_choice": "a counted GPT-6 Astra success in privileged mode", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "vast.ai RTX 4090 gpu5", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709, faststart", - "created": "2026-10-19T06:15:01-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/long_rope/poster.png b/masters/roboolympics-fun/long_rope/poster.png index 9c77d92434a6af1709a571e51d9ed8d44904b98f..734330fbf1e702b082bbca27be386bae5c6c6943 100644 --- a/masters/roboolympics-fun/long_rope/poster.png +++ b/masters/roboolympics-fun/long_rope/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:c8298583b2ca4170ef8d6fbf5c7b92c5af7a67df72bd8cdbb9236c47f1371cc6 -size 3177442 +oid sha256:55aff98f35e85b94513bbfc72cbfd93f9b30a20d84ce86cc3a1c8049e43cab78 +size 3166255 diff --git a/masters/roboolympics-fun/long_rope/video.mp4 b/masters/roboolympics-fun/long_rope/video.mp4 index eb72eac3b4a21a427966445fa61d29675acb87b1..ae9eefb093de1badbc5d8574229a4963409fe9ab 100644 --- a/masters/roboolympics-fun/long_rope/video.mp4 +++ b/masters/roboolympics-fun/long_rope/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9ac8ee9076e610b8600e8ae5d76c0a91daf81125df069ac4670a27c5bc0c93d9 -size 7616865 +oid sha256:bbdb63537828bb64ba2b744edd331c03fd7c80d504125989f7f6c86570cdf390 +size 7526302 diff --git a/masters/roboolympics-fun/paddle_bounce/meta.json b/masters/roboolympics-fun/paddle_bounce/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..95f197e2ed27f9276f62bfa72fe733e0d16eddc4 --- /dev/null +++ b/masters/roboolympics-fun/paddle_bounce/meta.json @@ -0,0 +1,111 @@ +{ + "suite": "roboolympics-fun", + "task": "paddle_bounce", + "family": "paddle-bounce", + "task_dir": "embodied-olympics-fun-paddle-bounce-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-paddle-bounce-i00-privileged/embodied-olympics-fun-paddle-bou__yr6NTKi", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 544, + "seconds": 18.13, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (cdc57439332dfade)", + "end_check": { + "success": true, + "stages": [ + "bounce 1", + "bounce 2", + "bounce 3", + "bounce 4", + "bounce 5", + "bounce 6", + "bounce 7", + "bounce 8", + "bounce 9", + "bounce 10", + "bounce 11", + "bounce 12", + "bounce 13", + "bounce 14", + "bounce 15", + "bounce 16", + "bounce 17", + "bounce 18", + "bounce 19", + "bounce 20" + ], + "stages_total": 20, + "progress": 1.0, + "metrics": { + "ball_height_m": 1.217, + "grip_break_max_s": 0.003, + "bounces": 20 + }, + "time_s": 14.49, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "cdc57439332dfadea1c10757a1ca03c35ce2a715e8610ec886fd5e0d1c752039", + "steps": 1449, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-paddle-bounce-i00-privileged/embodied-olympics-fun-paddle-bou__yr6NTKi/artifacts/app/output/trajectory.npz", + "n_actions": 1449, + "control_steps_replayed": 1449, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "cdc57439332dfadea1c10757a1ca03c35ce2a715e8610ec886fd5e0d1c752039", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.714, + 3.852 + ], + "robot_root": "pelvis", + "travel_m": 0.002, + "focus": { + "robot": 35, + "robot+object": 449 + }, + "objects": { + "dropper": 0.3, + "ball": 0.01, + "paddle": 0.32 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.59, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:44-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/paddle_bounce/poster.png b/masters/roboolympics-fun/paddle_bounce/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..1755306563ee8086f8f16af91ec905a22e73e9c6 --- /dev/null +++ b/masters/roboolympics-fun/paddle_bounce/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:24c5c4eecacfc4aa53ec05ffdcd9b1de94c4dbba517688818ab5d22d88445780 +size 3127825 diff --git a/masters/roboolympics-fun/paddle_bounce/video.mp4 b/masters/roboolympics-fun/paddle_bounce/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..8c71f8f071a861b451ffb6504436092c2bb86a56 --- /dev/null +++ b/masters/roboolympics-fun/paddle_bounce/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8ce6f2fc2ced5746916f377d59e33c1274b4eb02cd873b700ad77418665d7544 +size 6422969 diff --git a/masters/roboolympics-fun/penalty_kick/meta.json b/masters/roboolympics-fun/penalty_kick/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..bce914507a369960e803d2a07070a090e9f8b46a --- /dev/null +++ b/masters/roboolympics-fun/penalty_kick/meta.json @@ -0,0 +1,93 @@ +{ + "suite": "roboolympics-fun", + "task": "penalty_kick", + "family": "penalty-kick", + "task_dir": "embodied-olympics-fun-penalty-kick-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-penalty-kick-i00-privileged/embodied-olympics-fun-penalty-ki__qWLCYv4", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 440, + "seconds": 14.67, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (6985609591a8407c)", + "end_check": { + "success": true, + "stages": [ + "kick", + "goal" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_x_m": 14.242, + "keeper_y_m": -0.941, + "kick_speed_m_s": 16.69 + }, + "time_s": 11.36, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "6985609591a8407c4d5e8fde7850af97883df2763479f98af009c9985e310d36", + "steps": 1136, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-penalty-kick-i00-privileged/embodied-olympics-fun-penalty-ki__qWLCYv4/artifacts/app/output/trajectory.npz", + "n_actions": 1136, + "control_steps_replayed": 1136, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "6985609591a8407c4d5e8fde7850af97883df2763479f98af009c9985e310d36", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -58.05, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.402, + 12.31 + ], + "robot_root": "pelvis", + "travel_m": 3.076, + "focus": { + "object": 11, + "robot": 355, + "robot+object": 14 + }, + "objects": { + "keeper": 0.94, + "ball": 11.52 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.6, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 357 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/penalty_kick/poster.png b/masters/roboolympics-fun/penalty_kick/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2a1023cf4d0fea639dca30d16cee42823bcfa870 --- /dev/null +++ b/masters/roboolympics-fun/penalty_kick/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:476922ac37df2a30aa761c88954c672e177e4c0fd43f780ac3dd0469c5dbee3b +size 3053194 diff --git a/masters/roboolympics-fun/penalty_kick/video.mp4 b/masters/roboolympics-fun/penalty_kick/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..41fc76244cfe31722ae4d2a1d011ac7f054535b2 --- /dev/null +++ b/masters/roboolympics-fun/penalty_kick/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8369e7518f1be245677de10c4b92f4545a077aa97bf1d9145958804cbfed478b +size 7597759 diff --git a/masters/roboolympics-fun/pitch_target/meta.json b/masters/roboolympics-fun/pitch_target/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..ee420f87ce518cb1d451696b4723dd04ab7bf4b0 --- /dev/null +++ b/masters/roboolympics-fun/pitch_target/meta.json @@ -0,0 +1,94 @@ +{ + "suite": "roboolympics-fun", + "task": "pitch_target", + "family": "pitch-target", + "task_dir": "embodied-olympics-fun-pitch-target-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-pitch-target-i00-privileged/embodied-olympics-fun-pitch-targ__C2UUpk9", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 324, + "seconds": 10.8, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4be8a19323b18236)", + "end_check": { + "success": true, + "stages": [ + "release", + "strike" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "release_speed_m_s": 16.58, + "release_height_m": 2.306, + "cross_y_m": -0.0, + "cross_z_m": 0.765, + "cross_speed_m_s": 15.83 + }, + "time_s": 2.63, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "4be8a19323b182362d35edced58e36393422790799b840896ee392bc149ae24d", + "steps": 263, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-pitch-target-i00-privileged/embodied-olympics-fun-pitch-targ__C2UUpk9/artifacts/app/output/trajectory.npz", + "n_actions": 263, + "control_steps_replayed": 263, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "4be8a19323b182362d35edced58e36393422790799b840896ee392bc149ae24d", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -114.54, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.933, + 17.033 + ], + "robot_root": "pelvis", + "travel_m": 0.082, + "focus": { + "robot+object": 176, + "object": 77, + "robot": 11 + }, + "objects": { + "ball": 17.72 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 152 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/pitch_target/poster.png b/masters/roboolympics-fun/pitch_target/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..1547735db0199feea7705ecdfae6f25aa4fbd6ca --- /dev/null +++ b/masters/roboolympics-fun/pitch_target/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:53628aa6a17b67e138e8e7f7319f1b862321140f2660a276ce616b05128605a0 +size 2966016 diff --git a/masters/roboolympics-fun/pitch_target/video.mp4 b/masters/roboolympics-fun/pitch_target/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..e56639dc6f128e8caad7dd528c8f1d5e90315ead --- /dev/null +++ b/masters/roboolympics-fun/pitch_target/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8558e1c7402b5098d771b6305252c0b25f663176e7b2c40dc0f4ec60a6261979 +size 5190187 diff --git a/masters/roboolympics-fun/platform_10m/meta.json b/masters/roboolympics-fun/platform_10m/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..35c2a068ac715e06c37777feb1b422c8740edcb7 --- /dev/null +++ b/masters/roboolympics-fun/platform_10m/meta.json @@ -0,0 +1,98 @@ +{ + "suite": "roboolympics-fun", + "task": "platform_10m", + "family": "platform-10m", + "task_dir": "embodied-olympics-fun-platform-10m-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-platform-10m-i00-privileged/embodied-olympics-fun-platform-1__xbuUTG2", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; the pool's water drawn as refractive water (IOR 1.33, gentle wave bump, blue-green absorption); a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 312, + "seconds": 10.4, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (0e79454172c3d073)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "entry", + "under water" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "somersault_deg": 175.3, + "entry_distance_m": 2.231, + "entry_tilt_deg": 0.1, + "entry_hip": 172.118, + "entry_knee": 170.0833, + "entry_feet": 0.0851, + "entry_toes": 0.5001, + "entry_beyond": 0.3713, + "entry_apart": 0.1339, + "entry_off": 0.067, + "entry_twist": 0.1703 + }, + "time_s": 2.51, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "0e79454172c3d073e29d1bd202caf5a63ccf4d7a1fc1a22d432fb9c5b343a9be", + "steps": 251, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-platform-10m-i00-privileged/embodied-olympics-fun-platform-1__xbuUTG2/artifacts/app/output/trajectory.npz", + "n_actions": 251, + "control_steps_replayed": 251, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "0e79454172c3d073e29d1bd202caf5a63ccf4d7a1fc1a22d432fb9c5b343a9be", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the pool's water box drawn as water instead of an opaque box" + ], + "camera": { + "mode": "director", + "az_deg": -69.99, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.94, + 13.034 + ], + "robot_root": "pelvis", + "travel_m": 2.43, + "focus": { + "robot": 252 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 3.41, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 80 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/platform_10m/poster.png b/masters/roboolympics-fun/platform_10m/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..da3a4ae52b22feba64f2b887c570e990c0f72d1f --- /dev/null +++ b/masters/roboolympics-fun/platform_10m/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:feb8217ccb98aa406c9c3efeb9c1fabf08760d8f26d00bf5ffc48e25d5d9f920 +size 3155599 diff --git a/masters/roboolympics-fun/platform_10m/video.mp4 b/masters/roboolympics-fun/platform_10m/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..18bc75d846a8c9f7a969d38eb71399e427fdd799 --- /dev/null +++ b/masters/roboolympics-fun/platform_10m/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d02ddb2d28ec40fd8d5f12a17a319f6bab270835391c7a142ecdd34fb7297f61 +size 5868567 diff --git a/masters/roboolympics-fun/pole_vault/meta.json b/masters/roboolympics-fun/pole_vault/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..76e9077b7bb6150d28f38a62cf8595644aad77b6 --- /dev/null +++ b/masters/roboolympics-fun/pole_vault/meta.json @@ -0,0 +1,90 @@ +{ + "suite": "roboolympics-fun", + "task": "pole_vault", + "family": "pole-vault", + "task_dir": "embodied-olympics-fun-pole-vault-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-pole-vault-i00-privileged/embodied-olympics-fun-pole-vault__C5Ja6o2", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 306, + "seconds": 10.2, + "speed": 0.9, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (a0ac85014e42cf86)", + "end_check": { + "success": false, + "stages": [ + "plant", + "take-off" + ], + "stages_total": 4, + "progress": 0.5, + "metrics": {}, + "time_s": 7.35, + "failed": "the bar did not stay on its pegs" + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "a0ac85014e42cf8687745b177fc0280ea624609f89f604bed6dfe7a23766d146", + "steps": 735, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-pole-vault-i00-privileged/embodied-olympics-fun-pole-vault__C5Ja6o2/artifacts/app/output/trajectory.npz", + "n_actions": 735, + "control_steps_replayed": 735, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "a0ac85014e42cf8687745b177fc0280ea624609f89f604bed6dfe7a23766d146", + "original_verdict": "counted failure (verifier reward 0)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -70.23, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.38, + 8.755 + ], + "robot_root": "pelvis", + "travel_m": 20.285, + "focus": { + "robot+object": 217, + "robot": 29 + }, + "objects": { + "bar": 0.28, + "pole": 19.11 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.58, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/pole_vault/poster.png b/masters/roboolympics-fun/pole_vault/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..e733a99fb307b1d19b8b9e2a72ff38ce078f6dea --- /dev/null +++ b/masters/roboolympics-fun/pole_vault/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8931a3c61c3e04ae7c0671d07ada051c3c8b38faf00eeae7c18da29a0f3d7773 +size 3078315 diff --git a/masters/roboolympics-fun/pole_vault/video.mp4 b/masters/roboolympics-fun/pole_vault/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..d913d51e7e317aec831dd60e9ca4fe5b4a9be042 --- /dev/null +++ b/masters/roboolympics-fun/pole_vault/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8a236cdc8bbfcd65239631c2bf41d996bcdf2eea4ea4ec5f02457de4c601a44 +size 6146674 diff --git a/masters/roboolympics-fun/round_off/meta.json b/masters/roboolympics-fun/round_off/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..8ea6c8663e536682d663e556c7c201298ddaa77f --- /dev/null +++ b/masters/roboolympics-fun/round_off/meta.json @@ -0,0 +1,88 @@ +{ + "suite": "roboolympics-fun", + "task": "round_off", + "family": "round-off", + "task_dir": "embodied-olympics-fun-round-off-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-round-off-i00-privileged/embodied-olympics-fun-round-off__9HYxiBi", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 339, + "seconds": 11.3, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (027cacf6006c3bce)", + "end_check": { + "success": true, + "stages": [ + "hand support", + "landing", + "stand" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "hand_support_s": 0.473, + "landing_facing_dev_deg": 9.1 + }, + "time_s": 2.78, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "027cacf6006c3bce3d60d40314472c2d162e394073eae8c36e9e610498ae782a", + "steps": 278, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-round-off-i00-privileged/embodied-olympics-fun-round-off__9HYxiBi/artifacts/app/output/trajectory.npz", + "n_actions": 278, + "control_steps_replayed": 278, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "027cacf6006c3bce3d60d40314472c2d162e394073eae8c36e9e610498ae782a", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -140.13, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.271, + 3.704 + ], + "robot_root": "pelvis", + "travel_m": 2.099, + "focus": { + "robot": 279 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.61, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/round_off/poster.png b/masters/roboolympics-fun/round_off/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..3aeabec65cec7fb3dbdd133b0014c9447df75731 --- /dev/null +++ b/masters/roboolympics-fun/round_off/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a2b7455f6b306b0f4bd32f1257b5775d7712551da2df94b41179c4c28f62524a +size 3088622 diff --git a/masters/roboolympics-fun/round_off/video.mp4 b/masters/roboolympics-fun/round_off/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..c377698b4fde15a6e06682dbb7a821b4a1e9b935 --- /dev/null +++ b/masters/roboolympics-fun/round_off/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8d905605fccdf82a13c007b6cf20a5f1148efd789daa80e960dba1c170cf863a +size 5492038 diff --git a/masters/roboolympics-fun/rugby_conversion/meta.json b/masters/roboolympics-fun/rugby_conversion/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..338ba4256f2bed9232584dca24b61c74429e9ccd --- /dev/null +++ b/masters/roboolympics-fun/rugby_conversion/meta.json @@ -0,0 +1,93 @@ +{ + "suite": "roboolympics-fun", + "task": "rugby_conversion", + "family": "rugby-conversion", + "task_dir": "embodied-olympics-fun-rugby-conversion-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-rugby-conversion-i00-privileged/embodied-olympics-fun-rugby-conv__WgWaLHT", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 185, + "seconds": 6.17, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (676e06934e772e34)", + "end_check": { + "success": true, + "stages": [ + "kick", + "goal" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_max_z_m": 6.079, + "plane_z_m": 6.073, + "plane_y_m": -0.206 + }, + "time_s": 1.24, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "676e06934e772e3459691634a223bcd5e7db1c983bfe6011a377330ebf20301e", + "steps": 124, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-rugby-conversion-i00-privileged/embodied-olympics-fun-rugby-conv__WgWaLHT/artifacts/app/output/trajectory.npz", + "n_actions": 124, + "control_steps_replayed": 124, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "676e06934e772e3459691634a223bcd5e7db1c983bfe6011a377330ebf20301e", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -115.41, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 13.458, + 15.563 + ], + "robot_root": "pelvis", + "travel_m": 0.002, + "focus": { + "robot+object": 89, + "robot": 7, + "object": 29 + }, + "objects": { + "tee": 2.09, + "ball": 5.44 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.58, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 20 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:45-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/rugby_conversion/poster.png b/masters/roboolympics-fun/rugby_conversion/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..bea1135b22195b8801fbca68b93333a9ed3c76cc --- /dev/null +++ b/masters/roboolympics-fun/rugby_conversion/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85a6e306e658ab24711ef93b40f383b95b398b261e2f1a94bde273db0dc0b788 +size 3125207 diff --git a/masters/roboolympics-fun/rugby_conversion/video.mp4 b/masters/roboolympics-fun/rugby_conversion/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..5bad9f554da6ac744793f32337c2e04ccf79e133 --- /dev/null +++ b/masters/roboolympics-fun/rugby_conversion/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0999289140897d28f7892f59fb3d1dc0e9f9b8f6f14cdf1cdb3e59db15915a44 +size 3019427 diff --git a/masters/roboolympics-fun/shot_put/meta.json b/masters/roboolympics-fun/shot_put/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..6d0a5b50461d471d7275e0ba922334675033a05a --- /dev/null +++ b/masters/roboolympics-fun/shot_put/meta.json @@ -0,0 +1,97 @@ +{ + "suite": "roboolympics-fun", + "task": "shot_put", + "family": "shot-put", + "task_dir": "embodied-olympics-fun-shot-put-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-shot-put-i00-privileged/embodied-olympics-fun-shot-put-i__WRpuCTL", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 598, + "seconds": 19.93, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4a99925fb38a90ea)", + "end_check": { + "success": true, + "stages": [ + "release", + "landing", + "exit" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "stance_neck_distance_m": 0.159, + "release_time_s": 10.212, + "release_speed_m_s": 6.99, + "release_height_m": 1.952, + "release_push_fraction": 0.926, + "distance_m": 6.664 + }, + "time_s": 16.09, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "4a99925fb38a90ea8c160612184dd9a64e31589f99ad91d6ff509f0d8cb4e051", + "steps": 1609, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-shot-put-i00-privileged/embodied-olympics-fun-shot-put-i__WRpuCTL/artifacts/app/output/trajectory.npz", + "n_actions": 1609, + "control_steps_replayed": 1609, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "4a99925fb38a90ea8c160612184dd9a64e31589f99ad91d6ff509f0d8cb4e051", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -120.29, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.377, + 19.644 + ], + "robot_root": "pelvis", + "travel_m": 0.816, + "focus": { + "robot+landed": 61, + "robot+object": 41, + "object": 101, + "robot": 335 + }, + "objects": { + "shot": 12.58 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.73, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 344 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/shot_put/poster.png b/masters/roboolympics-fun/shot_put/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..b594dad4edf79001733740c52071e5a91047ed82 --- /dev/null +++ b/masters/roboolympics-fun/shot_put/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3429c0b215f0a28896c66c06e465605c4e3c586a0be8dda3cca22cc374dd9072 +size 3105325 diff --git a/masters/roboolympics-fun/shot_put/video.mp4 b/masters/roboolympics-fun/shot_put/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..b04868f275a814c525a9dfbc8759a3618a3bb0dc --- /dev/null +++ b/masters/roboolympics-fun/shot_put/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e4f1af2a11520aaa13df22a4cdc103b929bc21fa7f2ecd39d0631c10fe1ea58 +size 9773666 diff --git a/masters/roboolympics-fun/shuttle_run/meta.json b/masters/roboolympics-fun/shuttle_run/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..062683655791f85fdb12ff07b5dd4ea58d15d245 --- /dev/null +++ b/masters/roboolympics-fun/shuttle_run/meta.json @@ -0,0 +1,89 @@ +{ + "suite": "roboolympics-fun", + "task": "shuttle_run", + "family": "shuttle-run", + "task_dir": "embodied-olympics-fun-shuttle-run-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-shuttle-run-i00-privileged/embodied-olympics-fun-shuttle-ru__HTxEsWc", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 684, + "seconds": 22.8, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (f389114b1225d8bd)", + "end_check": { + "success": true, + "stages": [ + "turn 1", + "turn 2", + "turn 3", + "finish" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "turns": 3, + "time_s": 18.683 + }, + "time_s": 18.69, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "f389114b1225d8bd582d969bcef4dbdba4eb2c8553117e62e6bcc8d935b65673", + "steps": 1869, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-shuttle-run-i00-privileged/embodied-olympics-fun-shuttle-ru__HTxEsWc/artifacts/app/output/trajectory.npz", + "n_actions": 1869, + "control_steps_replayed": 1869, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "f389114b1225d8bd582d969bcef4dbdba4eb2c8553117e62e6bcc8d935b65673", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.107, + 3.939 + ], + "robot_root": "pelvis", + "travel_m": 0.812, + "focus": { + "robot": 624 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.62, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/shuttle_run/poster.png b/masters/roboolympics-fun/shuttle_run/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..fba4775e779c641e96153ab1b512210a3b2b6888 --- /dev/null +++ b/masters/roboolympics-fun/shuttle_run/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f10d3a22768383984c44e13fa432d4f9d6690cd6c4d6eef879b55cfe4a249be4 +size 3078484 diff --git a/masters/roboolympics-fun/shuttle_run/video.mp4 b/masters/roboolympics-fun/shuttle_run/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..b2ecdb6658daeae05bf8668d5d45ecf79c34ec3f --- /dev/null +++ b/masters/roboolympics-fun/shuttle_run/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e7c752ada66caa2040cd2ba73a6b855e3c2a008f820d2653d9601bdb2075390 +size 15003552 diff --git a/masters/roboolympics-fun/skateboard_ollie/meta.json b/masters/roboolympics-fun/skateboard_ollie/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..3086bdad7c32099d4407fade318c148d460d5527 --- /dev/null +++ b/masters/roboolympics-fun/skateboard_ollie/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "skateboard_ollie", + "family": "skateboard-ollie", + "task_dir": "embodied-olympics-fun-skateboard-ollie-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-skateboard-ollie-i00-privileged/embodied-olympics-fun-skateboard__chkrGqp", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 299, + "seconds": 9.97, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4ce428034656acf8)", + "end_check": { + "success": true, + "stages": [ + "pop", + "clear", + "landed", + "stand" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "board_clearance_m": 0.0834 + }, + "time_s": 2.38, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "4ce428034656acf8cfbb034e6ddef9ff04605d1be4644a626f3d68302e1e3360", + "steps": 238, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-skateboard-ollie-i00-privileged/embodied-olympics-fun-skateboard__chkrGqp/artifacts/app/output/trajectory.npz", + "n_actions": 238, + "control_steps_replayed": 238, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "4ce428034656acf8cfbb034e6ddef9ff04605d1be4644a626f3d68302e1e3360", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -140.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.29, + 3.578 + ], + "robot_root": "pelvis", + "travel_m": 0.048, + "focus": { + "robot+object": 134, + "robot": 105 + }, + "objects": { + "board": 0.05 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.57, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/skateboard_ollie/poster.png b/masters/roboolympics-fun/skateboard_ollie/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..a8f3135e3709572774f9ff4f38b5bf5ef72a17fe --- /dev/null +++ b/masters/roboolympics-fun/skateboard_ollie/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bbfe0fb8eaeaf8b1b9ac7ec9956a60b0c91a85eeffa0d6c053902b0fc92aea2d +size 3075737 diff --git a/masters/roboolympics-fun/skateboard_ollie/video.mp4 b/masters/roboolympics-fun/skateboard_ollie/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..2e2b442eac494c42d56dced713632bc96d00bc7c --- /dev/null +++ b/masters/roboolympics-fun/skateboard_ollie/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:be688d49ab0ae7a7733f5fc15801aa52958cc8626f15d36a5480eb885afe336a +size 4132561 diff --git a/masters/roboolympics-fun/slalom/meta.json b/masters/roboolympics-fun/slalom/meta.json index b212c886d9fbecc81b54f207af8b6917a21aff18..340fdd14c333cbad3d1ae63b9e38413536dbc37d 100644 --- a/masters/roboolympics-fun/slalom/meta.json +++ b/masters/roboolympics-fun/slalom/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "slalom", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-slalom-i00-privileged/roboolympics-slalom-i00-privileg__6LuNqSG", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "slalom", + "task_dir": "embodied-olympics-fun-slalom-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-slalom-i00-privileged/embodied-olympics-fun-slalom-i00__mvdtZvN", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 799, - "seconds": 26.63, - "speed": 1.5, + "frames": 770, + "seconds": 25.67, + "speed": 2.1, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (518fca824284cc74)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (df7415308111f196)", "end_check": { "success": true, "stages": [ @@ -32,54 +35,63 @@ "progress": 1.0, "metrics": { "gates": 12, - "time_s": 36.883 + "time_s": 49.595 }, - "time_s": 36.89 + "time_s": 49.6, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "df7415308111f196472e1b43d461139ee949197255b6905ae6b14a93e1feee6d", + "steps": 4960, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 3689, - "control_steps_replayed": 3689, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-slalom-i00-privileged/embodied-olympics-fun-slalom-i00__mvdtZvN/artifacts/app/output/trajectory.npz", + "n_actions": 4960, + "control_steps_replayed": 4960, "control_hz": 100.0, - "every_control_steps": 5, - "final_state_sha": "518fca824284cc7451f9a55c97e16240820eb70cf99f5ef3fbb238d8db2ce85f", + "every_control_steps": 7, + "final_state_sha": "df7415308111f196472e1b43d461139ee949197255b6905ae6b14a93e1feee6d", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -157.29, + "az_deg": -156.75, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.174, - 5.304 + 3.154, + 3.935 ], "robot_root": "pelvis", - "travel_m": 5.686, + "travel_m": 3.697, "focus": { - "robot": 739 + "robot": 710 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.87, - "image": "eai-robofunolympics:0.2.1", + "seconds_per_frame": 2.6, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu1 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:34-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/slalom/poster.png b/masters/roboolympics-fun/slalom/poster.png index b262d4be4633ca6c7cf158314c763b650d9ed45c..a60fbb08d9642040afaf8c238be7fe185b4a3bb8 100644 --- a/masters/roboolympics-fun/slalom/poster.png +++ b/masters/roboolympics-fun/slalom/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:e58010ab5978c5d10fc2187d1a35c011622d0a3ec3567517f28be01f8e926841 -size 3076565 +oid sha256:5337fa82e50cb9f23c3445baab2c8f819d81d1451b3b2401bb5ade71c449df83 +size 3204974 diff --git a/masters/roboolympics-fun/slalom/video.mp4 b/masters/roboolympics-fun/slalom/video.mp4 index 99e4a9f80219efb2ac8e84d81e7bf355bbb2aedc..b7f53b9572889551c52533180384ff13caa84fbf 100644 --- a/masters/roboolympics-fun/slalom/video.mp4 +++ b/masters/roboolympics-fun/slalom/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:2803ce79f2f74ceb1e6cde42f4732f8774c3971b267cc6b55ab05893b9e94ae3 -size 18370653 +oid sha256:dff2a4ca8745baa3be8269db96133531d774948a44f12e85dd8e2cbc3a34fd09 +size 17859480 diff --git a/masters/roboolympics-fun/sled_push/meta.json b/masters/roboolympics-fun/sled_push/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..0b944a5059978307cce4c771eb7892f82004a143 --- /dev/null +++ b/masters/roboolympics-fun/sled_push/meta.json @@ -0,0 +1,90 @@ +{ + "suite": "roboolympics-fun", + "task": "sled_push", + "family": "sled-push", + "task_dir": "embodied-olympics-fun-sled-push-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-sled-push-i00-privileged/embodied-olympics-fun-sled-push__6xkVG2A", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 477, + "seconds": 15.9, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e9a55efaa8257525)", + "end_check": { + "success": true, + "stages": [ + "moved", + "line" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "sled_rear_x_m": 13.933, + "sled_distance_m": 13.433 + }, + "time_s": 12.47, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "e9a55efaa8257525b2d78b27430174ca2bdab66723d30dbda35d4359cda4d0b1", + "steps": 1247, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-sled-push-i00-privileged/embodied-olympics-fun-sled-push__6xkVG2A/artifacts/app/output/trajectory.npz", + "n_actions": 1247, + "control_steps_replayed": 1247, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "e9a55efaa8257525b2d78b27430174ca2bdab66723d30dbda35d4359cda4d0b1", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -69.9, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.449, + 5.666 + ], + "robot_root": "pelvis", + "travel_m": 13.438, + "focus": { + "robot": 128, + "robot+object": 289 + }, + "objects": { + "sled": 13.44 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.59, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/sled_push/poster.png b/masters/roboolympics-fun/sled_push/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..250521a62d0e95653c853310b021efbc2c1af705 --- /dev/null +++ b/masters/roboolympics-fun/sled_push/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:202457efc08658f2f8ef98031170a30aa94cf5904e3e7203332ab39e7a653d9e +size 3169197 diff --git a/masters/roboolympics-fun/sled_push/video.mp4 b/masters/roboolympics-fun/sled_push/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..d0f8af494a7b83214c5324f277495aa784d2619f --- /dev/null +++ b/masters/roboolympics-fun/sled_push/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc5683530a63507634ae196ed859fd3ed8ef6d3726ee74dc7c4e02033f00cbf8 +size 7355645 diff --git a/masters/roboolympics-fun/snowboard/meta.json b/masters/roboolympics-fun/snowboard/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..76d324bf42cbae1f6d1daef8b7d11f752b95b09e --- /dev/null +++ b/masters/roboolympics-fun/snowboard/meta.json @@ -0,0 +1,114 @@ +{ + "suite": "roboolympics-fun", + "task": "snowboard", + "family": "snowboard", + "task_dir": "embodied-olympics-fun-snowboard-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra_quick-high-embodied-olympics-fun-snowboard-i00-privileged/embodied-olympics-fun-snowboard__sKGCCXG", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; snow venue: a snow shader (bright blue-white, rough, slight subsurface, fine bump) on the course's snow geoms, snow-white valley floor; a chase camera: per-frame fit of the robot (height included, so jumps and spins stay framed), at least 6 m away, 24 degree elevation, smoothed in time", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 839, + "seconds": 27.97, + "speed": 1.5, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4147534a1b59fee6)", + "end_check": { + "success": false, + "stages": [ + "start", + "gate 1", + "gate 2", + "gate 3", + "gate 4", + "gate 5", + "gate 6", + "gate 7", + "gate 8", + "gate 9", + "gate 10", + "gate 11", + "gate 12", + "gate 13" + ], + "stages_total": 20, + "progress": 0.7, + "metrics": { + "speed_m_s": 7.126, + "gates": 13 + }, + "time_s": 38.86, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "4147534a1b59fee6e16941a75e18dcb9f97830da697efc96e5ba1743972075d4", + "steps": 3886, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-snowboard-i00-privileged/embodied-olympics-fun-snowboard__sKGCCXG/artifacts/app/output/trajectory.npz", + "n_actions": 3886, + "control_steps_replayed": 3886, + "control_hz": 100.0, + "every_control_steps": 5, + "final_state_sha": "4147534a1b59fee6e16941a75e18dcb9f97830da697efc96e5ba1743972075d4", + "original_verdict": "counted failure (verifier reward 0)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "textured snow drawn with a snow shader instead of its checker texture's mean colour", + "the start house, the start and finish arches, the finish posts and banner (drawn-only, no contact) not drawn: they hid the robot from the chase camera; the safety nets and the gates are drawn" + ], + "camera": { + "mode": "director", + "az_deg": -69.4, + "az_rule": "robot-travel-70", + "el_deg": 24.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 6.24, + 11.451 + ], + "robot_root": "pelvis", + "travel_m": 172.758, + "focus": { + "robot": 234, + "robot+object": 545 + }, + "objects": { + "gate0_turn": 0.0, + "gate1_turn": 0.0, + "gate4_turn": 0.0, + "gate5_turn": 0.0, + "gate8_turn": 0.04, + "gate9_turn": 0.08, + "gate11_turn": 0.44, + "gate12_turn": 0.62, + "board": 172.79 + } + }, + "orbit_deg": 0.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 4.56, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:46-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/snowboard/poster.png b/masters/roboolympics-fun/snowboard/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..75a04c471c513c2260f5cc0bdd913cb830abcbb1 --- /dev/null +++ b/masters/roboolympics-fun/snowboard/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:89496f810fbaa964d2dd8f6465d27b034c5eac0d43db4bab4f234b8fe5459248 +size 2914692 diff --git a/masters/roboolympics-fun/snowboard/video.mp4 b/masters/roboolympics-fun/snowboard/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..2f100c84df9652cbedfc7cb5ae97fa5a87fab08e --- /dev/null +++ b/masters/roboolympics-fun/snowboard/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b15d7e1c6e3ec18df853c2890048e719685aa4d05945e06f28377c1818fa63e +size 9803920 diff --git a/masters/roboolympics-fun/springboard_3m/meta.json b/masters/roboolympics-fun/springboard_3m/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..b8ceb21f688189e3611452d6d37b2f6fc24ec110 --- /dev/null +++ b/masters/roboolympics-fun/springboard_3m/meta.json @@ -0,0 +1,102 @@ +{ + "suite": "roboolympics-fun", + "task": "springboard_3m", + "family": "springboard-3m", + "task_dir": "embodied-olympics-fun-springboard-3m-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-springboard-3m-i00-privileged/embodied-olympics-fun-springboar__vjAXTWo", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; the pool's water drawn as refractive water (IOR 1.33, gentle wave bump, blue-green absorption); a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 245, + "seconds": 8.17, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (26597142afef38a8)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "entry", + "under water" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "board_tip_lowest_m": 0.1758, + "somersault_deg": 166.9, + "entry_distance_m": 1.739, + "entry_tilt_deg": 0.0, + "entry_hip": 171.6261, + "entry_knee": 170.1238, + "entry_feet": 0.0964, + "entry_toes": 0.4995, + "entry_beyond": 0.3688, + "entry_apart": 0.1099, + "entry_off": 0.0557, + "entry_twist": 0.0572 + }, + "time_s": 1.84, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "26597142afef38a81b13c4cc6e5180091e4619acac91fad70d95af4b96293d7d", + "steps": 184, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-springboard-3m-i00-privileged/embodied-olympics-fun-springboar__vjAXTWo/artifacts/app/output/trajectory.npz", + "n_actions": 184, + "control_steps_replayed": 184, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "26597142afef38a81b13c4cc6e5180091e4619acac91fad70d95af4b96293d7d", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the pool's water box drawn as water instead of an opaque box" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.09, + 12.505 + ], + "robot_root": "pelvis", + "travel_m": 1.919, + "focus": { + "robot+object": 130, + "robot": 55 + }, + "objects": { + "board_seg0": 0.0 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 4.27, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 59 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/springboard_3m/poster.png b/masters/roboolympics-fun/springboard_3m/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..a3cdce502ea9cd53c2c7601199c7751f92a2fb16 --- /dev/null +++ b/masters/roboolympics-fun/springboard_3m/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:be93182847eeba94e1eda052d5b69b2ec023a8acdfb20d7dacf0a798a2cfc9f6 +size 3189888 diff --git a/masters/roboolympics-fun/springboard_3m/video.mp4 b/masters/roboolympics-fun/springboard_3m/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..ef180af93e051d84426fd5fa5ce6aa65eafbb67f --- /dev/null +++ b/masters/roboolympics-fun/springboard_3m/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1dd678d92e38406bef3872bc9b48a766f3d12cd9fe843de44a69c67cb534f594 +size 4296160 diff --git a/masters/roboolympics-fun/sprint/meta.json b/masters/roboolympics-fun/sprint/meta.json index 6dfcfc298d442f569fb9fecd67db78b075ddfa35..c37526783239b711d9b59ebce1df0b7c66f532ef 100644 --- a/masters/roboolympics-fun/sprint/meta.json +++ b/masters/roboolympics-fun/sprint/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "sprint", - "source": "agent-success:codex-gpt6_luna-xhigh-roboolympics-sprint-i00-privileged/roboolympics-sprint-i00-privileg__L22fzrw", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "sprint", + "task_dir": "embodied-olympics-fun-sprint-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-sprint-i00-privileged/embodied-olympics-fun-sprint-i00__pp2ccSJ", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 621, - "seconds": 20.7, + "frames": 584, + "seconds": 19.47, "speed": 0.9, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (82c06e8346e98bc6)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (bd7d8fa1b4352883)", "end_check": { "success": true, "stages": [ @@ -22,56 +25,65 @@ "progress": 1.0, "metrics": { "distance_m": 50.0, - "reaction_s": 0.33, - "time_s": 13.057 + "reaction_s": 0.25, + "time_s": 11.946 }, - "time_s": 16.8 + "time_s": 15.69, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "bd7d8fa1b4352883da8bf63d67dc31e3259dfd692258ae0704f10f2a58af06ba", + "steps": 1569, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 1680, - "control_steps_replayed": 1680, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-sprint-i00-privileged/embodied-olympics-fun-sprint-i00__pp2ccSJ/artifacts/app/output/trajectory.npz", + "n_actions": 1569, + "control_steps_replayed": 1569, "control_hz": 100.0, "every_control_steps": 3, - "final_state_sha": "82c06e8346e98bc655d6e4a659a90390c43429d573806a8c49d55772d9674e21", + "final_state_sha": "bd7d8fa1b4352883da8bf63d67dc31e3259dfd692258ae0704f10f2a58af06ba", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)", - "the finish posts and a lamp post (drawn-only, at the finish line) not drawn: they hid the runner" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)", + "the finish posts and a lamp post (drawn-only) not drawn: they hid the runner" ], "camera": { "mode": "director", - "az_deg": -70.4, + "az_deg": -70.26, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 3.049, - 4.272 + 2.999, + 4.902 ], "robot_root": "pelvis", - "travel_m": 49.992, + "travel_m": 49.883, "focus": { - "robot": 561 + "robot": 524 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.88, - "image": "eai-robofunolympics:0.2.1", - "agent": "codex + GPT-6 Luna (xhigh)", - "run": "codex-gpt6_luna-xhigh", + "seconds_per_frame": 2.6, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", "mode": "privileged", - "trial_choice": "no counted Astra success; a counted success of another run", + "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu2 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:35-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/sprint/poster.png b/masters/roboolympics-fun/sprint/poster.png index b216caf957a569f09fc9cb69977f7a0b833ada36..621c0f7b3fbd52cd6d0c4c981545b47aeda3ad4d 100644 --- a/masters/roboolympics-fun/sprint/poster.png +++ b/masters/roboolympics-fun/sprint/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9600a905139b497731a39a6cc95c8e4d9855f39722e4eb554610ced7334ea53f -size 3173456 +oid sha256:0b98587029f7346f203eb54691181c3c1f7643c3d001a27d619cd0adeb05c52a +size 3131054 diff --git a/masters/roboolympics-fun/sprint/video.mp4 b/masters/roboolympics-fun/sprint/video.mp4 index f552f43ca729be8c23bb4bbd2cb572e23681c4cc..cbb4e5ca49a2050c3558a3217ead75eb4417f5ba 100644 --- a/masters/roboolympics-fun/sprint/video.mp4 +++ b/masters/roboolympics-fun/sprint/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:69639cd60f47614808b8cdbc4a4e695fb0f90d96479eb98742a2b8b599696a74 -size 14615332 +oid sha256:61c652651b88052ff0191047a0e978786dc8bc9ca7fd2c08d712599a0e602933 +size 14024444 diff --git a/masters/roboolympics-fun/standing_long_jump/meta.json b/masters/roboolympics-fun/standing_long_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..d9232e4811cf0c7c68ca68465a565551f9803112 --- /dev/null +++ b/masters/roboolympics-fun/standing_long_jump/meta.json @@ -0,0 +1,89 @@ +{ + "suite": "roboolympics-fun", + "task": "standing_long_jump", + "family": "standing-long-jump", + "task_dir": "embodied-olympics-fun-standing-long-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-standing-long-jump-i00-privileged/embodied-olympics-fun-standing-l__6XUKDGd", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 358, + "seconds": 11.93, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (82054e3b5694f846)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "landing", + "stand" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "air_s": 0.589, + "rise_m": 0.437, + "mark_m": 1.6 + }, + "time_s": 2.97, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "82054e3b5694f8468f71bb2ea450087163fc143f843b61d7a34b1f33744c70bf", + "steps": 297, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-standing-long-jump-i00-privileged/embodied-olympics-fun-standing-l__6XUKDGd/artifacts/app/output/trajectory.npz", + "n_actions": 297, + "control_steps_replayed": 297, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "82054e3b5694f8468f71bb2ea450087163fc143f843b61d7a34b1f33744c70bf", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.293, + 4.363 + ], + "robot_root": "pelvis", + "travel_m": 1.865, + "focus": { + "robot": 298 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.61, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/standing_long_jump/poster.png b/masters/roboolympics-fun/standing_long_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..f5dc48cb7de8183bc9d29f9e812c986d74037a92 --- /dev/null +++ b/masters/roboolympics-fun/standing_long_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:998a3cb8be611b87a9aed94afeca6cb3512ee8d8a3ed0028177d77a6e1683daf +size 3174821 diff --git a/masters/roboolympics-fun/standing_long_jump/video.mp4 b/masters/roboolympics-fun/standing_long_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..d04723bc1f31884a120258b28720c9c72855fb2f --- /dev/null +++ b/masters/roboolympics-fun/standing_long_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5ad7fd202cc50cb5c2c2edce20080ab8cd9aa2861769cf6b664dbaec522f3a9f +size 6523515 diff --git a/masters/roboolympics-fun/steeplechase/meta.json b/masters/roboolympics-fun/steeplechase/meta.json index ba7d435323366eb544bee761c9dd3ff88fdb15fe..b2f79caffe2ac76a3c75a779cf752433c230b473 100644 --- a/masters/roboolympics-fun/steeplechase/meta.json +++ b/masters/roboolympics-fun/steeplechase/meta.json @@ -1,17 +1,20 @@ { "suite": "roboolympics-fun", "task": "steeplechase", - "source": "agent-success:codex-gpt6_astra-high-roboolympics-steeplechase-i00-privileged/roboolympics-steeplechase-i00-pr__4Db4Amv", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "family": "steeplechase", + "task_dir": "embodied-olympics-fun-steeplechase-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-steeplechase-i00-privileged/embodied-olympics-fun-steeplecha__q4qT2kF", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 778, - "seconds": 25.93, + "frames": 735, + "seconds": 24.5, "speed": 1.2, "success_check": true, - "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (38dfb1923c458a0c)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (414adc5fa4c3c129)", "end_check": { "success": true, "stages": [ @@ -22,55 +25,64 @@ "stages_total": 3, "progress": 1.0, "metrics": { - "rear_x_m": 14.486, - "time_s": 28.68 + "rear_x_m": 14.301, + "time_s": 26.94 }, - "time_s": 28.68 + "time_s": 26.94, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "414adc5fa4c3c1295ba4f9df10d9018ed2ed52a3efde8df5ffe74097a5a04d53", + "steps": 2694, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 2868, - "control_steps_replayed": 2868, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-steeplechase-i00-privileged/embodied-olympics-fun-steeplecha__q4qT2kF/artifacts/app/output/trajectory.npz", + "n_actions": 2694, + "control_steps_replayed": 2694, "control_hz": 100.0, "every_control_steps": 4, - "final_state_sha": "38dfb1923c458a0c832738e21479956912751749deae014ede48486e33d68c5d", + "final_state_sha": "414adc5fa4c3c1295ba4f9df10d9018ed2ed52a3efde8df5ffe74097a5a04d53", "original_verdict": "counted success (verifier reward 1)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -69.75, + "az_deg": -70.7, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.328, - 5.105 + 3.886, + 4.41 ], "robot_root": "pelvis", - "travel_m": 14.599, + "travel_m": 14.413, "focus": { - "robot": 718 + "robot": 675 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 3.4, - "image": "eai-robofunolympics:0.2.1", + "seconds_per_frame": 3.14, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", "mode": "privileged", "trial_choice": "a counted GPT-6 Astra success", "poster": "last frame (solved state)", - "host": "gpu0 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:35-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/steeplechase/poster.png b/masters/roboolympics-fun/steeplechase/poster.png index e841bbf90498c859b484e00d68d79a400d43c863..62df563901ef63133c03b79764ded61ab414ed1a 100644 --- a/masters/roboolympics-fun/steeplechase/poster.png +++ b/masters/roboolympics-fun/steeplechase/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:6a4111b8e54c41fe200683996476f45cef2f096069e39c850fe377cf1d275845 -size 3166342 +oid sha256:48e604b05083fccd63db1247f9ca48b418e4437a1626f25f8011362b4be1c65a +size 3170891 diff --git a/masters/roboolympics-fun/steeplechase/video.mp4 b/masters/roboolympics-fun/steeplechase/video.mp4 index 214fa4aa632e0bda685a560e33ed730ff9902bc3..52d3ced4b5e07681114bbf39a7d0b0951d1b3b3a 100644 --- a/masters/roboolympics-fun/steeplechase/video.mp4 +++ b/masters/roboolympics-fun/steeplechase/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:70016976b45b8a5afc8eb55b52479580469ec970f92b38921d9096d4b86bb5c9 -size 14481088 +oid sha256:0fc5a1e8e9e5eaded93214ab4c7bdf5ead064013275421bf96f251f903764cc2 +size 14309105 diff --git a/masters/roboolympics-fun/sumo/meta.json b/masters/roboolympics-fun/sumo/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..4e5d7027c83c85e4832baed92e2331785e5a27ad --- /dev/null +++ b/masters/roboolympics-fun/sumo/meta.json @@ -0,0 +1,88 @@ +{ + "suite": "roboolympics-fun", + "task": "sumo", + "family": "sumo", + "task_dir": "embodied-olympics-fun-sumo-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-sumo-i00-privileged/embodied-olympics-fun-sumo-i00-p__HgxJJxC", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 265, + "seconds": 8.83, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (230ef78948b765a1)", + "end_check": { + "success": true, + "stages": [ + "win" + ], + "stages_total": 1, + "progress": 1.0, + "metrics": { + "win_t": 2.031 + }, + "time_s": 2.04, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "230ef78948b765a1d178e21f056b6d743df2715388607a1691ca4ee25f4353f5", + "steps": 204, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-sumo-i00-privileged/embodied-olympics-fun-sumo-i00-p__HgxJJxC/artifacts/app/output/trajectory.npz", + "n_actions": 204, + "control_steps_replayed": 204, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "230ef78948b765a1d178e21f056b6d743df2715388607a1691ca4ee25f4353f5", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.741, + 4.117 + ], + "robot_root": "pelvis", + "travel_m": 1.88, + "focus": { + "robot": 67, + "robot+object": 138 + }, + "objects": { + "opp_pelvis": 0.55 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.73, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/sumo/poster.png b/masters/roboolympics-fun/sumo/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..dcdffa83d9c312b0b7013ba9ca976920ecd010c7 --- /dev/null +++ b/masters/roboolympics-fun/sumo/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2d6f3051374550973ce50bb9e16f4d392b150c5c3bab435f74a4a48632fa6a69 +size 3260155 diff --git a/masters/roboolympics-fun/sumo/video.mp4 b/masters/roboolympics-fun/sumo/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..8ded366e13ac1b30f3d2bbafb18eb2065f5836d3 --- /dev/null +++ b/masters/roboolympics-fun/sumo/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e4474d8179a8a7a14360dd29bfc3c625db305c23b9f0c98cc483c3a3da41a13d +size 5550747 diff --git a/masters/roboolympics-fun/sweeper_jump/meta.json b/masters/roboolympics-fun/sweeper_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..9a955d8e3b11f101e2d5868fdb30a1d1bbb6b608 --- /dev/null +++ b/masters/roboolympics-fun/sweeper_jump/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "sweeper_jump", + "family": "sweeper-jump", + "task_dir": "embodied-olympics-fun-sweeper-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-sweeper-jump-i00-privileged/embodied-olympics-fun-sweeper-ju__TqwYdsR", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 775, + "seconds": 25.83, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (e4cef589c9a5ac6b)", + "end_check": { + "success": true, + "stages": [ + "jump 1", + "jump 2", + "jump 3", + "landing", + "stand" + ], + "stages_total": 5, + "progress": 1.0, + "metrics": { + "jumps": 3 + }, + "time_s": 21.42, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "e4cef589c9a5ac6b73d001f8474852d9f5c91883c0789da1b2ccbe888a4cdb8d", + "steps": 2142, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-sweeper-jump-i00-privileged/embodied-olympics-fun-sweeper-ju__TqwYdsR/artifacts/app/output/trajectory.npz", + "n_actions": 2142, + "control_steps_replayed": 2142, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "e4cef589c9a5ac6b73d001f8474852d9f5c91883c0789da1b2ccbe888a4cdb8d", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": 130.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.316, + 8.914 + ], + "robot_root": "pelvis", + "travel_m": 0.126, + "focus": { + "robot": 715 + }, + "objects": { + "boom": 1.03 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.71, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:47-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/sweeper_jump/poster.png b/masters/roboolympics-fun/sweeper_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..5c9709449f4fd5efa6ce0c8cb59241fbefd15b90 --- /dev/null +++ b/masters/roboolympics-fun/sweeper_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:76528b2d6d70ca162949c0efa6d4b7ffd32fbcb72d266b0bd0ff68555f007d38 +size 3132496 diff --git a/masters/roboolympics-fun/sweeper_jump/video.mp4 b/masters/roboolympics-fun/sweeper_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..7305917592238be367c609050eb9f06ca65b8a04 --- /dev/null +++ b/masters/roboolympics-fun/sweeper_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e0c172ebe278846669a774dda5f13c3315f6636a9762e54c5a4d63b13259cba1 +size 14891831 diff --git a/masters/roboolympics-fun/table_tennis_return/meta.json b/masters/roboolympics-fun/table_tennis_return/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..2227516f1f45dcfbbef1d6fbecea314aaea785fd --- /dev/null +++ b/masters/roboolympics-fun/table_tennis_return/meta.json @@ -0,0 +1,94 @@ +{ + "suite": "roboolympics-fun", + "task": "table_tennis_return", + "family": "table-tennis-return", + "task_dir": "embodied-olympics-fun-table-tennis-return-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-table-tennis-return-i00-privileged/embodied-olympics-fun-table-tenn__uMUjtpN", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 301, + "seconds": 10.03, + "speed": 0.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (2cb9a380c5f7d00d)", + "end_check": { + "success": true, + "stages": [ + "bounce", + "strike", + "opponent's court" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "grip_break_max_s": 0.006, + "landing_x_m": 2.573, + "landing_y_m": 0.018, + "net_touch": 0.0 + }, + "time_s": 4.79, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "2cb9a380c5f7d00d8dbb8c0278e046eb81717d6378f53a5189315f4b81871e22", + "steps": 479, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-table-tennis-return-i00-privileged/embodied-olympics-fun-table-tenn__uMUjtpN/artifacts/app/output/trajectory.npz", + "n_actions": 479, + "control_steps_replayed": 479, + "control_hz": 100.0, + "every_control_steps": 2, + "final_state_sha": "2cb9a380c5f7d00d8dbb8c0278e046eb81717d6378f53a5189315f4b81871e22", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.033, + 5.167 + ], + "robot_root": "pelvis", + "travel_m": 0.453, + "focus": { + "robot+object": 184, + "robot": 57 + }, + "objects": { + "ball": 0.76, + "paddle": 0.98 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.79, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 240 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/table_tennis_return/poster.png b/masters/roboolympics-fun/table_tennis_return/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..2cfe545a2e7bc0c152f16e86e6bea9ebf6ed5aa3 --- /dev/null +++ b/masters/roboolympics-fun/table_tennis_return/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10d3d560bd278ad69ebfc68886b525007542c192136e18dbf2c5964a581a6143 +size 3142889 diff --git a/masters/roboolympics-fun/table_tennis_return/video.mp4 b/masters/roboolympics-fun/table_tennis_return/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..e003339e3cbe8f578691e92ba47893ae38b4adcb --- /dev/null +++ b/masters/roboolympics-fun/table_tennis_return/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:864120cc60f42b8b8b1afabc01178e5c2297944aedd72b10802b98a00023856d +size 5719878 diff --git a/masters/roboolympics-fun/taekwondo_kick/meta.json b/masters/roboolympics-fun/taekwondo_kick/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..fff5dc6445994836a10afa0bdda424510842bcf2 --- /dev/null +++ b/masters/roboolympics-fun/taekwondo_kick/meta.json @@ -0,0 +1,88 @@ +{ + "suite": "roboolympics-fun", + "task": "taekwondo_kick", + "family": "taekwondo-kick", + "task_dir": "embodied-olympics-fun-taekwondo-kick-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-taekwondo-kick-i00-privileged/embodied-olympics-fun-taekwondo__Sqd8zUs", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 267, + "seconds": 8.9, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (ac7fda0b27d9581a)", + "end_check": { + "success": true, + "stages": [ + "kick", + "stand" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "impulse_n_s": 13.27, + "peak_force_n": 966.0, + "contact_s": 0.034 + }, + "time_s": 6.18, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "ac7fda0b27d9581a2b4de94eebe4a8e60b4c35a7b67c862d6638947b4e1a63d0", + "steps": 618, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-taekwondo-kick-i00-privileged/embodied-olympics-fun-taekwondo__Sqd8zUs/artifacts/app/output/trajectory.npz", + "n_actions": 618, + "control_steps_replayed": 618, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "ac7fda0b27d9581a2b4de94eebe4a8e60b4c35a7b67c862d6638947b4e1a63d0", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 3.196, + 3.341 + ], + "robot_root": "pelvis", + "travel_m": 0.099, + "focus": { + "robot": 207 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.64, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/taekwondo_kick/poster.png b/masters/roboolympics-fun/taekwondo_kick/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..907d191e2f6662221386f77635304ebb0eedafdc --- /dev/null +++ b/masters/roboolympics-fun/taekwondo_kick/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9797f4f48b3aeb2633b30dd9923846b7035a0da4167313603af8a3b2c18844f5 +size 3238496 diff --git a/masters/roboolympics-fun/taekwondo_kick/video.mp4 b/masters/roboolympics-fun/taekwondo_kick/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..9e0f3f956842796469cf1f9a6db01f8d9f22607e --- /dev/null +++ b/masters/roboolympics-fun/taekwondo_kick/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9b5be1df24548ecfe0ea77394f460c6ef04b3bec9c2b6b28e46ca4ba23d59e1b +size 5085340 diff --git a/masters/roboolympics-fun/teaser.mp4 b/masters/roboolympics-fun/teaser.mp4 index 7040173cdfb97da84d0205d2bec532ca3f178a79..35ecaae0f22b7f0aef36a300162817d38cee6ad1 100644 --- a/masters/roboolympics-fun/teaser.mp4 +++ b/masters/roboolympics-fun/teaser.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:7cec832633f667b9d1db4d995ac256a3d2dcb08377e95346860cfffb273f96c9 -size 8713925 +oid sha256:ad7cdfce402d206c89478423922feb7322dbd479f5e3da34ebceb3b96caa2b52 +size 5720370 diff --git a/masters/roboolympics-fun/teaser_poster.png b/masters/roboolympics-fun/teaser_poster.png index 9917dc7666d82814e54f1f8ee8fa06359e112595..2a3c11f8da1edddffeaa73e07b3abaed4a7872f0 100644 --- a/masters/roboolympics-fun/teaser_poster.png +++ b/masters/roboolympics-fun/teaser_poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:8ff334034fd3cbe1c54064a69195c76a680d9aae195b1fd080f4b592045f2d94 -size 3237854 +oid sha256:18d98d68879eb9777501b8e346b1036397a08e7f540fcf2717d351b6996f041b +size 3088548 diff --git a/masters/roboolympics-fun/tennis_return/meta.json b/masters/roboolympics-fun/tennis_return/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..3e5847b7e223d3289e0700ba566cdcad57f0ecdf --- /dev/null +++ b/masters/roboolympics-fun/tennis_return/meta.json @@ -0,0 +1,96 @@ +{ + "suite": "roboolympics-fun", + "task": "tennis_return", + "family": "tennis-return", + "task_dir": "embodied-olympics-fun-tennis-return-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-tennis-return-i00-privileged/embodied-olympics-fun-tennis-ret__hbqEGUK", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 250, + "seconds": 8.33, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (094e5b44625c333f)", + "end_check": { + "success": true, + "stages": [ + "bounce", + "racket", + "over the net", + "in" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "landing_x_m": 18.313, + "landing_y_m": 2.757 + }, + "time_s": 5.66, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "094e5b44625c333f972f5c6178eda9b83622e8c21d6733d2644d26e235e45e41", + "steps": 566, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-tennis-return-i00-privileged/embodied-olympics-fun-tennis-ret__hbqEGUK/artifacts/app/output/trajectory.npz", + "n_actions": 566, + "control_steps_replayed": 566, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "094e5b44625c333f972f5c6178eda9b83622e8c21d6733d2644d26e235e45e41", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -287.7, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.318, + 13.752 + ], + "robot_root": "pelvis", + "travel_m": 1.507, + "focus": { + "robot": 59, + "object": 38, + "robot(incoming)": 24, + "robot+object": 69 + }, + "objects": { + "server": 0.01, + "ball": 6.29, + "racket": 2.03 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.71, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 135 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/tennis_return/poster.png b/masters/roboolympics-fun/tennis_return/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..7e362343337d1068679162522e79336db5dc9451 --- /dev/null +++ b/masters/roboolympics-fun/tennis_return/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8042632030f9903ef86aca8bded922d9f7f1a9df2be88142138cb5cb26bc3fef +size 3052457 diff --git a/masters/roboolympics-fun/tennis_return/video.mp4 b/masters/roboolympics-fun/tennis_return/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..8eac598d60cd0bfa6f1e182870fe29b8ac3420a3 --- /dev/null +++ b/masters/roboolympics-fun/tennis_return/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8dcfc162125af8db457fde0ef3781ac41bec42483eae981c8c0170087b7257eb +size 4012382 diff --git a/masters/roboolympics-fun/tennis_serve/meta.json b/masters/roboolympics-fun/tennis_serve/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..3358d8f84b3b308b86bc391499de5772950c2b84 --- /dev/null +++ b/masters/roboolympics-fun/tennis_serve/meta.json @@ -0,0 +1,95 @@ +{ + "suite": "roboolympics-fun", + "task": "tennis_serve", + "family": "tennis-serve", + "task_dir": "embodied-olympics-fun-tennis-serve-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra_quick-high-embodied-olympics-fun-tennis-serve-i00-privileged/embodied-olympics-fun-tennis-ser__mTzrBbF", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 276, + "seconds": 9.2, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (b33a8195925f69f8)", + "end_check": { + "success": true, + "stages": [ + "toss", + "strike", + "over the net", + "in" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "strike_speed_m_s": 15.197, + "landing_x_m": 15.59, + "landing_y_m": 3.036 + }, + "time_s": 2.15, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "0dca23c35f (files 0.3.3)", + "final_state_sha": "b33a8195925f69f88ae327af37dfa8d8013a1ee09a5d09b16f9603046fd1284b", + "steps": 215, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra_quick-high-embodied-olympics-fun-tennis-serve-i00-privileged/embodied-olympics-fun-tennis-ser__mTzrBbF/artifacts/app/output/trajectory.npz", + "n_actions": 215, + "control_steps_replayed": 215, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "b33a8195925f69f88ae327af37dfa8d8013a1ee09a5d09b16f9603046fd1284b", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -104.21, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.664, + 15.68 + ], + "robot_root": "pelvis", + "travel_m": 0.012, + "focus": { + "robot+object": 131, + "robot": 8, + "object": 77 + }, + "objects": { + "ball": 15.46, + "racket": 0.3 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.87, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high), quick-check run", + "run": "codex-gpt6_astra_quick-high", + "mode": "privileged", + "trial_choice": "GPT-6 Astra's privileged success in the quick check of this task version (no formal run on it yet)", + "poster": "frame 96 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/tennis_serve/poster.png b/masters/roboolympics-fun/tennis_serve/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..9b5ca7db77c560241782772bc46f585ec1487dac --- /dev/null +++ b/masters/roboolympics-fun/tennis_serve/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9651838eed909bda6bb4f26b121525d0d04fbb9133aa6df16fec335647e5fe38 +size 3134247 diff --git a/masters/roboolympics-fun/tennis_serve/video.mp4 b/masters/roboolympics-fun/tennis_serve/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..908d1799f910c8588640bf72d5ae9692a84df8fd --- /dev/null +++ b/masters/roboolympics-fun/tennis_serve/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1ab327c1ea09b9ec906b39d0a3ca0224a0b6657ef9f44af83aa4b08b14a10594 +size 5328967 diff --git a/masters/roboolympics-fun/toes_to_bar/meta.json b/masters/roboolympics-fun/toes_to_bar/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..6d6c6fdc6f76a0524b2839bd1a1c034879d52233 --- /dev/null +++ b/masters/roboolympics-fun/toes_to_bar/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "toes_to_bar", + "family": "toes-to-bar", + "task_dir": "embodied-olympics-fun-toes-to-bar-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-toes-to-bar-i00-privileged/embodied-olympics-fun-toes-to-ba__x8iZgcy", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 797, + "seconds": 26.57, + "speed": 1.8, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (82d634c2693cd127)", + "end_check": { + "success": true, + "stages": [ + "rep 1", + "rep 2", + "rep 3", + "rep 4", + "rep 5" + ], + "stages_total": 5, + "progress": 1.0, + "metrics": { + "touches": 5, + "reps": 5, + "time_s": 44.14 + }, + "time_s": 44.14, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "82d634c2693cd127b69563d23ad95b71fa990f11ad0cbfa9308302693ed1c1e8", + "steps": 4414, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-toes-to-bar-i00-privileged/embodied-olympics-fun-toes-to-ba__x8iZgcy/artifacts/app/output/trajectory.npz", + "n_actions": 4414, + "control_steps_replayed": 4414, + "control_hz": 100.0, + "every_control_steps": 6, + "final_state_sha": "82d634c2693cd127b69563d23ad95b71fa990f11ad0cbfa9308302693ed1c1e8", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 2.267, + 3.92 + ], + "robot_root": "pelvis", + "travel_m": 0.049, + "focus": { + "robot": 737 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.6, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/toes_to_bar/poster.png b/masters/roboolympics-fun/toes_to_bar/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..fe3d87465c9161bb85d2f3ffde56dcafb5e32446 --- /dev/null +++ b/masters/roboolympics-fun/toes_to_bar/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b37209e0517bdbf22e36cc4cdba9d1323359f43062492987e1f352bd3082291d +size 3160660 diff --git a/masters/roboolympics-fun/toes_to_bar/video.mp4 b/masters/roboolympics-fun/toes_to_bar/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..75d3d190705dc5f84b192805d159265865528fc7 --- /dev/null +++ b/masters/roboolympics-fun/toes_to_bar/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0ac10aff752e22d8e22bff26d20b054deeb61f2a18093965df9f96f18fefbaae +size 13523955 diff --git a/masters/roboolympics-fun/trampoline/meta.json b/masters/roboolympics-fun/trampoline/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..daa2b0bd631c8b0d1ddb982f77df1cc93e4d7f6f --- /dev/null +++ b/masters/roboolympics-fun/trampoline/meta.json @@ -0,0 +1,92 @@ +{ + "suite": "roboolympics-fun", + "task": "trampoline", + "family": "trampoline", + "task_dir": "embodied-olympics-fun-trampoline-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-trampoline-i00-privileged/embodied-olympics-fun-trampoline__THfW4gu", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 268, + "seconds": 8.93, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (6c270741a83ee4fc)", + "end_check": { + "success": true, + "stages": [ + "tuck jump", + "landing", + "stop" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "flights": 1, + "tuck_jump_air_s": 0.46, + "tuck_jump_com_rise_m": 0.3265 + }, + "time_s": 6.21, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "6c270741a83ee4fc905a1552cb21aaf1b10430ca7def814042bec65302226c7f", + "steps": 621, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-trampoline-i00-privileged/embodied-olympics-fun-trampoline__THfW4gu/artifacts/app/output/trajectory.npz", + "n_actions": 621, + "control_steps_replayed": 621, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "6c270741a83ee4fc905a1552cb21aaf1b10430ca7def814042bec65302226c7f", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 6.415, + 6.541 + ], + "robot_root": "pelvis", + "travel_m": 0.059, + "focus": { + "robot+object": 61, + "robot": 147 + }, + "objects": { + "bed": 0.0 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.89, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:48-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/trampoline/poster.png b/masters/roboolympics-fun/trampoline/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..b4107e841a0b3b710c94be233c0cdf455269e163 --- /dev/null +++ b/masters/roboolympics-fun/trampoline/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d6624f5e7e3b5f41dd5f333c110b631a29137322998693041eb03773aea02ad3 +size 3266836 diff --git a/masters/roboolympics-fun/trampoline/video.mp4 b/masters/roboolympics-fun/trampoline/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..afd1921c68251d4864559dda8986cbfa7330af69 --- /dev/null +++ b/masters/roboolympics-fun/trampoline/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7a148f335f3ec351586f8916f1339c328a2244a008587949c500a33b467a3a33 +size 5188501 diff --git a/masters/roboolympics-fun/trap_ball/meta.json b/masters/roboolympics-fun/trap_ball/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..fc9f4ac6e7a3213f7d733a962476d11b7b67a315 --- /dev/null +++ b/masters/roboolympics-fun/trap_ball/meta.json @@ -0,0 +1,95 @@ +{ + "suite": "roboolympics-fun", + "task": "trap_ball", + "family": "trap-ball", + "task_dir": "embodied-olympics-fun-trap-ball-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-trap-ball-i00-privileged/embodied-olympics-fun-trap-ball__PAXYgiG", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 261, + "seconds": 8.7, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (21eef66fd644f88a)", + "end_check": { + "success": true, + "stages": [ + "first touch", + "control" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_speed_m_s": 0.193, + "ball_from_pelvis_m": 0.363, + "touches": 2, + "control_after_s": 0.84 + }, + "time_s": 5.98, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "21eef66fd644f88a9876284956b174bc7376ba5aff7a23dd73c5f0cbcdf4f740", + "steps": 598, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-trap-ball-i00-privileged/embodied-olympics-fun-trap-ball__PAXYgiG/artifacts/app/output/trajectory.npz", + "n_actions": 598, + "control_steps_replayed": 598, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "21eef66fd644f88a9876284956b174bc7376ba5aff7a23dd73c5f0cbcdf4f740", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -294.58, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 10.927, + 32.692 + ], + "robot_root": "pelvis", + "travel_m": 0.328, + "focus": { + "object": 61, + "robot(incoming)": 99, + "robot": 3, + "robot+object": 38 + }, + "objects": { + "machine": 0.03, + "ball": 14.95 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.56, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 169 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/trap_ball/poster.png b/masters/roboolympics-fun/trap_ball/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..a131d079c2abe02ce1b22aed3880449e777fc390 --- /dev/null +++ b/masters/roboolympics-fun/trap_ball/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:575cadfc3dd00ec1fab8f73f9b7c7ddc9a5a8f1de9fe5ea4ba43b5f1b43c35d8 +size 2957243 diff --git a/masters/roboolympics-fun/trap_ball/video.mp4 b/masters/roboolympics-fun/trap_ball/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..94934475eed8f429b8fc5a751a30883ac008db4a --- /dev/null +++ b/masters/roboolympics-fun/trap_ball/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:160e790a2f9f93d1184247ebf2a946a55db4fb64fe1177968ed10a7b510194e3 +size 3942959 diff --git a/masters/roboolympics-fun/triple_jump/meta.json b/masters/roboolympics-fun/triple_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..e6bd1b2ffb5592adc7ea43cbd0fe0d1b80ba92ea --- /dev/null +++ b/masters/roboolympics-fun/triple_jump/meta.json @@ -0,0 +1,93 @@ +{ + "suite": "roboolympics-fun", + "task": "triple_jump", + "family": "triple-jump", + "task_dir": "embodied-olympics-fun-triple-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "scene-orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-triple-jump-i00-privileged/embodied-olympics-fun-triple-jum__QzLruGy", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 330, + "seconds": 11.0, + "speed": 0.9, + "success_check": false, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7e329d5295a252a3)", + "end_check": { + "success": false, + "stages": [ + "hop", + "step" + ], + "stages_total": 4, + "progress": 0.5, + "metrics": { + "hop_air_s": 0.275, + "hop_rise_m": 0.14, + "hop_m": 1.6, + "step_air_s": 0.176, + "step_m": 3.305, + "jump_air_s": 0.526 + }, + "time_s": 8.06, + "failed": "the jump did not come down in the sand" + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "7e329d5295a252a36c9ab9854820b6ad6882392b552ce47d9d0be245e9cf41fb", + "steps": 806, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-triple-jump-i00-privileged/embodied-olympics-fun-triple-jum__QzLruGy/artifacts/app/output/trajectory.npz", + "n_actions": 806, + "control_steps_replayed": 806, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "7e329d5295a252a36c9ab9854820b6ad6882392b552ce47d9d0be245e9cf41fb", + "original_verdict": "counted failure (verifier reward 0)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -68.84, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 2.97, + 3.883 + ], + "robot_root": "pelvis", + "travel_m": 26.423, + "focus": { + "robot": 270 + }, + "objects": {} + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.54, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", + "poster": "last frame (end of the failed attempt)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/triple_jump/poster.png b/masters/roboolympics-fun/triple_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..07cc5eccce5e0708781baf325064451fc5e06b06 --- /dev/null +++ b/masters/roboolympics-fun/triple_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b6e486cd631be3ea688a3189559dc16d02e8698c26daf3cac31c1854092264ae +size 3037777 diff --git a/masters/roboolympics-fun/triple_jump/video.mp4 b/masters/roboolympics-fun/triple_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..5a8dad6c43780de838251370e92c1253034f900c --- /dev/null +++ b/masters/roboolympics-fun/triple_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88df6da3d9ad1cac9591c3374ca1e2abe73e5c3cca1b29d6973e2cd0ab1df95d +size 7828365 diff --git a/masters/roboolympics-fun/vault/meta.json b/masters/roboolympics-fun/vault/meta.json index f52d2f0fb997a07c2e217fdd0b74bf547c7163b2..323f4b818cc742424f86619e97d7b663c9e56123 100644 --- a/masters/roboolympics-fun/vault/meta.json +++ b/masters/roboolympics-fun/vault/meta.json @@ -1,73 +1,87 @@ { "suite": "roboolympics-fun", "task": "vault", + "family": "vault", + "task_dir": "embodied-olympics-fun-vault-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", "source": "scene-orbit", - "why_no_video": "No evaluated agent has solved this task (no counted success in any run). Owner's rule for the site (board 0161): show GPT-6 Astra's best failed attempt instead. This video is that attempt replayed exactly (not a solved run): the task's own check at the end says it failed, as the verifier did.", - "shows": "agent-attempt (failed):codex-gpt6_astra-high-roboolympics-vault-i00-standard/roboolympics-vault-i00-standard__Qf7h8z8", - "renderer": "Blender 4.2.23 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (fixed 22 degree elevation, per-frame fit of the robot and the object in play, smoothed in time) with a slow eased 10-degree orbit", + "why_no_video": "No evaluated agent has a success on this task version yet. Site rule (boards 0161 / 0252): show GPT-6 Astra's failed attempt with the most progress instead. This video is that attempt replayed exactly (not a solved run): the task's own referee at the end says it failed, as the verifier did.", + "shows": "agent-attempt (failed):codex-gpt6_astra-high-embodied-olympics-fun-vault-i00-privileged/embodied-olympics-fun-vault-i00__ojeVhV4", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", "rendering": "path-tracing", "width": 2560, "height": 1440, "fps": 30, - "frames": 806, - "seconds": 26.87, - "speed": 1.5, + "frames": 761, + "seconds": 25.37, + "speed": 2.4, "success_check": false, - "success_check_how": "the task's own verdict (environment/sim/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance; a second plain replay with the task's own replay() agrees and ends in the same final state hash (8b5b5f301acd5130)", + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (dd8d92e140ed39e7)", "end_check": { "success": false, "stages": [ - "hands on the table" + "hands on the table", + "over the table", + "landing" ], "stages_total": 4, - "progress": 0.25, + "progress": 0.75, "metrics": {}, - "time_s": 37.22 + "time_s": 55.95, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "dd8d92e140ed39e7ca5da619fa80966abba802421c1bd002689a87ad759680df", + "steps": 5595, + "pass": true }, "replay": { - "trajectory": "a run of the exam", - "n_actions": 3722, - "control_steps_replayed": 3722, + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-vault-i00-privileged/embodied-olympics-fun-vault-i00__ojeVhV4/artifacts/app/output/trajectory.npz", + "n_actions": 5595, + "control_steps_replayed": 5595, "control_hz": 100.0, - "every_control_steps": 5, - "final_state_sha": "8b5b5f301acd51303e1f2ccc092c0fedf6a8942a4d5e073bd53b3e581a8e99c3", + "every_control_steps": 8, + "final_state_sha": "dd8d92e140ed39e7ca5da619fa80966abba802421c1bd002689a87ad759680df", "original_verdict": "counted failure (verifier reward 0)" }, "visual_changes": [ - "daylight sky and a plain ground plane in the venue's ground colour instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery)" + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" ], "camera": { "mode": "director", - "az_deg": -69.86, + "az_deg": -71.56, "az_rule": "robot-travel-70", "el_deg": 22.0, "fovy_deg": 34.0, "fill": 0.72, "distance_m": [ - 4.206, - 4.405 + 3.782, + 4.862 ], "robot_root": "pelvis", - "travel_m": 24.398, + "travel_m": 27.132, "focus": { - "robot": 746 + "robot": 701 }, "objects": {} }, "orbit_deg": 10.0, "holds": { + "start_frames": 15, + "end_frames": 45, "start_s": 0.5, "end_s": 1.5, "note": "first/last frame held, no extra simulation" }, - "seconds_per_frame": 2.86, - "image": "eai-robofunolympics:0.2.1", + "seconds_per_frame": 2.63, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", "agent": "codex + GPT-6 Astra (high)", "run": "codex-gpt6_astra-high", - "mode": "standard", - "trial_choice": "no counted success in any run; GPT-6 Astra's failed attempt (most progress)", + "mode": "privileged", + "trial_choice": "no success on this task version; GPT-6 Astra's failed attempt (most progress)", "poster": "last frame (end of the failed attempt)", - "host": "gpu3 062523e6f067", - "encode": "libx264 preset slow CRF 16, yuv420p, bt709", - "created": "2026-10-18T07:23:35-0500" + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" } \ No newline at end of file diff --git a/masters/roboolympics-fun/vault/poster.png b/masters/roboolympics-fun/vault/poster.png index 3b3a0d4a883b5fc6b3d527ab2d644956f0678e6d..eac3c1017b7caa6dc47e65ca890409a2c269dc58 100644 --- a/masters/roboolympics-fun/vault/poster.png +++ b/masters/roboolympics-fun/vault/poster.png @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:9e896ffb3861b8626941767bae387cc1ab13d1846bb2b65825fa872911df30c5 -size 3161484 +oid sha256:8909a4f9413d26e3e9e3ce74b19d6850f8f3c50db146ce3a86bfb422c73a65d9 +size 3152122 diff --git a/masters/roboolympics-fun/vault/video.mp4 b/masters/roboolympics-fun/vault/video.mp4 index 4c91e14ee3d8a71fc51843e5d3ae41bd7ca8d67a..150830f88e95577f71efbc3ef40ac9066f4888da 100644 --- a/masters/roboolympics-fun/vault/video.mp4 +++ b/masters/roboolympics-fun/vault/video.mp4 @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:a399be3904ed8fe8e9d015f6947cc39467716cdd3a7619b9c22712b06dc32dee -size 14732707 +oid sha256:d81ab3ece855a80e5f5d1c7dea1dfbfebac3951a23f976736c48736e251676ac +size 14721898 diff --git a/masters/roboolympics-fun/vertical_jump/meta.json b/masters/roboolympics-fun/vertical_jump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..64df8e8ab9fc0ab6b3129bc7069bd91c2fdc7e6e --- /dev/null +++ b/masters/roboolympics-fun/vertical_jump/meta.json @@ -0,0 +1,91 @@ +{ + "suite": "roboolympics-fun", + "task": "vertical_jump", + "family": "vertical-jump", + "task_dir": "embodied-olympics-fun-vertical-jump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-vertical-jump-i00-privileged/embodied-olympics-fun-vertical-j__kjbpYBB", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 254, + "seconds": 8.47, + "speed": 0.6, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (42e788ae54ecefb8)", + "end_check": { + "success": true, + "stages": [ + "take-off", + "touch", + "landing", + "stand" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "air_s": 0.502, + "rise_m": 0.331 + }, + "time_s": 3.86, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "42e788ae54ecefb89ab303c1098e042f807f429c1d218c92c4be963fc2cbe4e5", + "steps": 386, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-vertical-jump-i00-privileged/embodied-olympics-fun-vertical-j__kjbpYBB/artifacts/app/output/trajectory.npz", + "n_actions": 386, + "control_steps_replayed": 386, + "control_hz": 100.0, + "every_control_steps": 2, + "final_state_sha": "42e788ae54ecefb89ab303c1098e042f807f429c1d218c92c4be963fc2cbe4e5", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -50.0, + "az_rule": "facing-50", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 5.819, + 5.952 + ], + "robot_root": "pelvis", + "travel_m": 0.215, + "focus": { + "robot": 194 + }, + "objects": { + "vane": 0.4 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.61, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/vertical_jump/poster.png b/masters/roboolympics-fun/vertical_jump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..39897e7e415f1be87a396c856ac0f6686d2f17b6 --- /dev/null +++ b/masters/roboolympics-fun/vertical_jump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3485234a85dc901902cc5f1d852aeb9c4c507ac44064a84f6a374b33d4abd8b +size 3046555 diff --git a/masters/roboolympics-fun/vertical_jump/video.mp4 b/masters/roboolympics-fun/vertical_jump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..19874f056fa6a86a359d7cc5e90aa82721f063c0 --- /dev/null +++ b/masters/roboolympics-fun/vertical_jump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e098f1d1c355554a05e71aaef83344f608b5f8ffc7b30c93079d93e55f64ada3 +size 3723221 diff --git a/masters/roboolympics-fun/volley_shot/meta.json b/masters/roboolympics-fun/volley_shot/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..cc7164ca4d4db176c5cc9d19834c09c9321319e8 --- /dev/null +++ b/masters/roboolympics-fun/volley_shot/meta.json @@ -0,0 +1,94 @@ +{ + "suite": "roboolympics-fun", + "task": "volley_shot", + "family": "volley-shot", + "task_dir": "embodied-olympics-fun-volley-shot-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-volley-shot-i00-privileged/embodied-olympics-fun-volley-sho__7Kg3EQH", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 246, + "seconds": 8.2, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (4c29d605ec7f041d)", + "end_check": { + "success": true, + "stages": [ + "volley", + "goal" + ], + "stages_total": 2, + "progress": 1.0, + "metrics": { + "ball_x_m": 11.23, + "speed_in_m_s": 11.71, + "speed_out_m_s": 19.62 + }, + "time_s": 5.55, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "4c29d605ec7f041da029dd17a3a75904edca5923c63568a7ccb737189953d0de", + "steps": 555, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-volley-shot-i00-privileged/embodied-olympics-fun-volley-sho__7Kg3EQH/artifacts/app/output/trajectory.npz", + "n_actions": 555, + "control_steps_replayed": 555, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "4c29d605ec7f041da029dd17a3a75904edca5923c63568a7ccb737189953d0de", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -190.83, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 10.23, + 32.454 + ], + "robot_root": "pelvis", + "travel_m": 0.102, + "focus": { + "object": 57, + "robot": 1, + "robot+object": 35, + "robot(incoming)": 93 + }, + "objects": { + "machine": 0.04, + "ball": 21.19 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.82, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 163 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/volley_shot/poster.png b/masters/roboolympics-fun/volley_shot/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..e5724bb2a94da5c9bdf12973d2bea85671d22f11 --- /dev/null +++ b/masters/roboolympics-fun/volley_shot/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:62db62c2088f999e0c47f2412143f62ae51f62d6326e45d45c7439f3795a10cc +size 3071094 diff --git a/masters/roboolympics-fun/volley_shot/video.mp4 b/masters/roboolympics-fun/volley_shot/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..d72e6a3144eaa747245824d8c2c2b3cbc3c08dfd --- /dev/null +++ b/masters/roboolympics-fun/volley_shot/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2236829e9dbfbb0f94321ac6ec97e7cd78eea0c1480fee6f899e0d76c1aec2cf +size 4782639 diff --git a/masters/roboolympics-fun/volleyball_bump/meta.json b/masters/roboolympics-fun/volleyball_bump/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..6619e1c044bf63ffed6335ccb49b3906b8addc68 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_bump/meta.json @@ -0,0 +1,99 @@ +{ + "suite": "roboolympics-fun", + "task": "volleyball_bump", + "family": "volleyball-bump", + "task_dir": "embodied-olympics-fun-volleyball-bump-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-volleyball-bump-i00-privileged/embodied-olympics-fun-volleyball__hL2NK3F", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 259, + "seconds": 8.63, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (6eb2a1c340f76811)", + "end_check": { + "success": true, + "stages": [ + "pass", + "above the net height", + "in the target" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "ball_height_m": 0.102, + "pass_apex_m": 3.444, + "speed_in_m_s": 11.4, + "speed_out_m_s": 7.68, + "landing_x_m": 5.0, + "landing_y_m": -1.5, + "from_target_centre_m": 0.0 + }, + "time_s": 5.93, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "6eb2a1c340f76811bf5280f9c437ca309352bdf382746fdd917e27a30ee17b71", + "steps": 593, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-volleyball-bump-i00-privileged/embodied-olympics-fun-volleyball__hL2NK3F/artifacts/app/output/trajectory.npz", + "n_actions": 593, + "control_steps_replayed": 593, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "6eb2a1c340f76811bf5280f9c437ca309352bdf382746fdd917e27a30ee17b71", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -279.46, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 8.373, + 33.383 + ], + "robot_root": "pelvis", + "travel_m": 0.932, + "focus": { + "robot+object": 75, + "robot(incoming)": 91, + "robot": 1, + "object": 32 + }, + "objects": { + "machine": 0.03, + "ball": 11.24 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.7, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 145 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:49-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/volleyball_bump/poster.png b/masters/roboolympics-fun/volleyball_bump/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..990914a66645b25fe7b0b3c3a2733c6a54a179b1 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_bump/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff59bf2f925452b1c397e8b514b1a95cd0ce9c95e088068fa5998c5086384da5 +size 3098821 diff --git a/masters/roboolympics-fun/volleyball_bump/video.mp4 b/masters/roboolympics-fun/volleyball_bump/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..b343b322581e14ed3c7dbdd47420bb4641d62d68 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_bump/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dfc5b2b3488e33ef48ca92bd27fd9a1b4761629b6ede1d5e740a304cfcd11d8b +size 4871630 diff --git a/masters/roboolympics-fun/volleyball_serve/meta.json b/masters/roboolympics-fun/volleyball_serve/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..2a8fa8a3628aaac7e890a750bc8d465ec9c10284 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_serve/meta.json @@ -0,0 +1,96 @@ +{ + "suite": "roboolympics-fun", + "task": "volleyball_serve", + "family": "volleyball-serve", + "task_dir": "embodied-olympics-fun-volleyball-serve-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-volleyball-serve-i00-privileged/embodied-olympics-fun-volleyball__kiujwRr", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 270, + "seconds": 9.0, + "speed": 0.3, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (40c648e737ed8159)", + "end_check": { + "success": true, + "stages": [ + "toss", + "hit", + "over the net", + "in the court" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "ball_height_m": 0.081, + "speed_in_m_s": 1.63, + "speed_out_m_s": 12.94, + "landing_x_m": 14.279, + "landing_y_m": 0.112 + }, + "time_s": 2.09, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "40c648e737ed8159d28c138f152bb65af2c6b8425c78cec93da664962b229fbb", + "steps": 209, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-volleyball-serve-i00-privileged/embodied-olympics-fun-volleyball__kiujwRr/artifacts/app/output/trajectory.npz", + "n_actions": 209, + "control_steps_replayed": 209, + "control_hz": 100.0, + "every_control_steps": 1, + "final_state_sha": "40c648e737ed8159d28c138f152bb65af2c6b8425c78cec93da664962b229fbb", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -115.42, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 10.736, + 15.145 + ], + "robot_root": "pelvis", + "travel_m": 0.009, + "focus": { + "robot+object": 75, + "robot": 15, + "object": 120 + }, + "objects": { + "ball": 13.84 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.75, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "frame 31 (same scene, camera and settings as the video)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:50-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/volleyball_serve/poster.png b/masters/roboolympics-fun/volleyball_serve/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..64e5bb0bff3d6176b92eb0db4582a51a36a4c2b0 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_serve/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d2c69e65fd70e70fb6e8ff6dca2fcd0b36435e4cdf8c9e186d135fa77b1c231e +size 3062444 diff --git a/masters/roboolympics-fun/volleyball_serve/video.mp4 b/masters/roboolympics-fun/volleyball_serve/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..45ae12216168cf5390d20e698c5f0557150ebcc3 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_serve/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bc8ef0132f2a45b07fd25144a3b135b28b105105c41558ab5554b3a7b3c3a526 +size 5267889 diff --git a/masters/roboolympics-fun/volleyball_set/meta.json b/masters/roboolympics-fun/volleyball_set/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..b0323b0d523b41ede96bd50deec8da9c0abd20bb --- /dev/null +++ b/masters/roboolympics-fun/volleyball_set/meta.json @@ -0,0 +1,97 @@ +{ + "suite": "roboolympics-fun", + "task": "volleyball_set", + "family": "volleyball-set", + "task_dir": "embodied-olympics-fun-volleyball-set-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-volleyball-set-i00-privileged/embodied-olympics-fun-volleyball__X8pPtPq", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 264, + "seconds": 8.8, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (13403adda32efd70)", + "end_check": { + "success": true, + "stages": [ + "set", + "above the net", + "in the target" + ], + "stages_total": 3, + "progress": 1.0, + "metrics": { + "ball_height_m": 0.074, + "set_apex_m": 4.816, + "speed_in_m_s": 6.85, + "speed_out_m_s": 8.19, + "landing_x_m": 5.001, + "landing_y_m": -0.0, + "from_target_centre_m": 0.001 + }, + "time_s": 6.09, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "13403adda32efd70ac218f2c0f23ea91775f84301a351a84e1aa7fc38afc19fe", + "steps": 609, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-volleyball-set-i00-privileged/embodied-olympics-fun-volleyball__X8pPtPq/artifacts/app/output/trajectory.npz", + "n_actions": 609, + "control_steps_replayed": 609, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "13403adda32efd70ac218f2c0f23ea91775f84301a351a84e1aa7fc38afc19fe", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -173.32, + "az_rule": "object-travel-115", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.442, + 12.387 + ], + "robot_root": "pelvis", + "travel_m": 0.014, + "focus": { + "robot": 96, + "robot+object": 108 + }, + "objects": { + "machine": 0.02, + "ball": 6.61 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.96, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:50-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/volleyball_set/poster.png b/masters/roboolympics-fun/volleyball_set/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..e545239f6a3a04bb62fce475d983bb31a7fd6738 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_set/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:af67f6d0d426cd79d5e34ac6f8ff400808d351b3bbbab9556ec3e6d9700d94c4 +size 3217910 diff --git a/masters/roboolympics-fun/volleyball_set/video.mp4 b/masters/roboolympics-fun/volleyball_set/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..a8a5c75b56ed8ee1603f88f906d0f3d4b8ac0292 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_set/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:207095f1fe554fabd983516eac52eed8fc6c151c39159b6246d110cc326d5738 +size 5237453 diff --git a/masters/roboolympics-fun/volleyball_spike/meta.json b/masters/roboolympics-fun/volleyball_spike/meta.json new file mode 100644 index 0000000000000000000000000000000000000000..4f27e74f9e044413696859299759992d086a3f58 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_spike/meta.json @@ -0,0 +1,100 @@ +{ + "suite": "roboolympics-fun", + "task": "volleyball_spike", + "family": "volleyball-spike", + "task_dir": "embodied-olympics-fun-volleyball-spike-i00-privileged", + "task_version": "JamesKrW/embodied-first-exam@4fde6ba91c", + "source": "agent-success:codex-gpt6_astra-high-embodied-olympics-fun-volleyball-spike-i00-privileged/embodied-olympics-fun-volleyball__gAT4uFP", + "renderer": "Blender 4.5.3 LTS Cycles path tracing (GPU OptiX), adaptive sampling (threshold 0.01, 64-128 spp), OpenImageDenoise with albedo + normal passes; kinematic playback of the exact MuJoCo replay poses (MuJoCo's own compiled meshes, primitives and colours); daylight: a sun lamp (2 degree disc) and a Nishita sky (no sun disc), the venue's own ground colour on a large ground plane with a faint world-space variation; AgX view transform, Medium High Contrast look; a tracking camera (per-frame fit of the robot and the object in play, smoothed in time, zooming out early) with a slow eased 10.0-degree orbit", + "rendering": "path-tracing", + "width": 2560, + "height": 1440, + "fps": 30, + "frames": 283, + "seconds": 9.43, + "speed": 0.9, + "success_check": true, + "success_check_how": "the task's own verdict (tests/rcb_rfo.py Episode.verdict(), the referee) at the end of this re-render's own replay of every recorded row from the task's frozen instance on the new task version; a second plain replay with the task's own replay() agrees and ends in the same final state hash (7a71d895784e3faf)", + "end_check": { + "success": true, + "stages": [ + "spike", + "over the net", + "in the court", + "landing" + ], + "stages_total": 4, + "progress": 1.0, + "metrics": { + "ball_height_m": 1.096, + "speed_in_m_s": 7.01, + "speed_out_m_s": 7.62, + "contact_height_m": 2.479, + "jump_rise_m": 0.3761, + "air_s": 0.575, + "landing_x_m": 7.58, + "landing_y_m": -0.168 + }, + "time_s": 6.66, + "failed": null + }, + "replay_hash_check": { + "how": "the trial's trajectory replayed with the task's own `rcb_rfo.py replay`, twice in the new version's verifier image and once in the image it was graded on; all three end-state hashes equal, and the verdict, steps, progress, stages and metrics equal the recorded reward.json", + "graded_on": "6109cac419 (files 0.3.2)", + "final_state_sha": "7a71d895784e3faf942dfeb33af2a54991952025aa49b33d93508ad8493c088e", + "steps": 666, + "pass": true + }, + "replay": { + "trajectory": "codex-gpt6_astra-high-embodied-olympics-fun-volleyball-spike-i00-privileged/embodied-olympics-fun-volleyball__gAT4uFP/artifacts/app/output/trajectory.npz", + "n_actions": 666, + "control_steps_replayed": 666, + "control_hz": 100.0, + "every_control_steps": 3, + "final_state_sha": "7a71d895784e3faf942dfeb33af2a54991952025aa49b33d93508ad8493c088e", + "original_verdict": "counted success (verifier reward 1)" + }, + "visual_changes": [ + "daylight sky and a plain ground plane instead of the venue's painted sky, hills, trees, grandstands and perimeter boards (drawn-only scenery, geoms named rfo_*)" + ], + "camera": { + "mode": "director", + "az_deg": -69.4, + "az_rule": "robot-travel-70", + "el_deg": 22.0, + "fovy_deg": 34.0, + "fill": 0.72, + "distance_m": [ + 4.548, + 10.605 + ], + "robot_root": "pelvis", + "travel_m": 2.43, + "focus": { + "object": 4, + "robot+object": 122, + "robot": 97 + }, + "objects": { + "machine": 0.02, + "ball": 9.29 + } + }, + "orbit_deg": 10.0, + "holds": { + "start_frames": 15, + "end_frames": 45, + "start_s": 0.5, + "end_s": 1.5, + "note": "first/last frame held, no extra simulation" + }, + "seconds_per_frame": 2.87, + "image": "ghcr.io/mll-lab-nu/efe:embodied-olympics-fun-0.2.1 + embodied-olympics-fun-files-0.3.3 (the task's tests/ verifier image)", + "agent": "codex + GPT-6 Astra (high)", + "run": "codex-gpt6_astra-high", + "mode": "privileged", + "trial_choice": "a counted GPT-6 Astra success", + "poster": "last frame (solved state)", + "encode": "libx264 preset slow CRF 16, yuv420p, bt709, +faststart", + "created": "2026-10-24T13:13:50-0500" +} \ No newline at end of file diff --git a/masters/roboolympics-fun/volleyball_spike/poster.png b/masters/roboolympics-fun/volleyball_spike/poster.png new file mode 100644 index 0000000000000000000000000000000000000000..07874d378e7a53e09b20ce3fee0b458bcc911b64 --- /dev/null +++ b/masters/roboolympics-fun/volleyball_spike/poster.png @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:855e26a6e35eb7419b872952e70698920e39f55e98a26ea78d6edd867456dbc8 +size 3113351 diff --git a/masters/roboolympics-fun/volleyball_spike/video.mp4 b/masters/roboolympics-fun/volleyball_spike/video.mp4 new file mode 100644 index 0000000000000000000000000000000000000000..feb92f4f9f9e8b9cc80b53b22aafec52c3eda6bf --- /dev/null +++ b/masters/roboolympics-fun/volleyball_spike/video.mp4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fb5b98c53f39748d3f5051935e7e9c4ab674df6ea209f61a7b3825e67f347b7e +size 6214779