Add GigaWorld-Policy-0.6 evaluation logs

#8
by UniCat - opened
This view is limited to 50 files because it contains too many changes. See the raw diff here.
Files changed (50) hide show
  1. gigaworld-policy-06/README.md +5 -0
  2. gigaworld-policy-06/leaderboard_gigaworld-policy-06.json +2626 -0
  3. gigaworld-policy-06/logo.png +3 -0
  4. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/cov_only/per_episode.jsonl +0 -0
  5. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/cov_only/summary.json +224 -0
  6. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_cov/per_episode.jsonl +0 -0
  7. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_cov/summary.json +223 -0
  8. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_only/per_episode.jsonl +0 -0
  9. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_only/summary.json +224 -0
  10. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/none/per_episode.jsonl +0 -0
  11. gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/none/summary.json +209 -0
  12. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/cov_only/per_episode.jsonl +0 -0
  13. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/cov_only/summary.json +266 -0
  14. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_cov/per_episode.jsonl +0 -0
  15. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_cov/summary.json +262 -0
  16. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_only/per_episode.jsonl +0 -0
  17. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_only/summary.json +257 -0
  18. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/none/per_episode.jsonl +0 -0
  19. gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/none/summary.json +261 -0
  20. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/cov_only/per_episode.jsonl +0 -0
  21. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/cov_only/summary.json +271 -0
  22. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_cov/per_episode.jsonl +0 -0
  23. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_cov/summary.json +284 -0
  24. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_only/per_episode.jsonl +0 -0
  25. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_only/summary.json +293 -0
  26. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/none/per_episode.jsonl +0 -0
  27. gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/none/summary.json +284 -0
  28. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/cov_only/per_episode.jsonl +0 -0
  29. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/cov_only/summary.json +236 -0
  30. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_cov/per_episode.jsonl +0 -0
  31. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_cov/summary.json +239 -0
  32. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_only/per_episode.jsonl +0 -0
  33. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_only/summary.json +234 -0
  34. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/none/per_episode.jsonl +0 -0
  35. gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/none/summary.json +237 -0
  36. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/per_episode.jsonl +0 -0
  37. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/summary.json +229 -0
  38. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/per_episode.jsonl +0 -0
  39. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/summary.json +239 -0
  40. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/per_episode.jsonl +0 -0
  41. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/summary.json +225 -0
  42. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/none/per_episode.jsonl +0 -0
  43. gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/none/summary.json +228 -0
  44. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/cov_only/per_episode.jsonl +0 -0
  45. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/cov_only/summary.json +266 -0
  46. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_cov/per_episode.jsonl +0 -0
  47. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_cov/summary.json +265 -0
  48. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_only/per_episode.jsonl +0 -0
  49. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_only/summary.json +259 -0
  50. gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/none/per_episode.jsonl +0 -0
gigaworld-policy-06/README.md ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ # GigaWorld-Policy-0.6
2
+
3
+ **Contributor:** [GigaAI](https://gigaai.cc/)
4
+
5
+ **Project:** https://open-gigaai.github.io/giga-world-policy/
gigaworld-policy-06/leaderboard_gigaworld-policy-06.json ADDED
@@ -0,0 +1,2626 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": {
3
+ "id": "gigaworld-policy-06",
4
+ "label": "GigaWorld-Policy-0.6",
5
+ "variantLabel": "",
6
+ "family": "GigaWorld-Policy",
7
+ "modelVersion": "GigaWorld-Policy-0.6",
8
+ "checkpoint": "epoch5 / step12870; SHA256 afed58631710411829692e0f9d5ce618a5f2866c2d239482c7315eb1113be56b",
9
+ "inputs": "RGB + proprioception + task instruction",
10
+ "protocolId": "standard",
11
+ "displayOrder": 140,
12
+ "trainingSetting": "Multi-task fine-tuned checkpoint, epoch 5 / step 12870",
13
+ "evaluationSetting": "26 tasks, four channels, 50 episodes per task/channel; fixed checkpoint across all tasks",
14
+ "resultVersion": "2026-10-28 evaluation",
15
+ "updatedAt": "2026-11-17",
16
+ "sourceLabel": "GigaAI Bench2Dex evaluation",
17
+ "sourceUrl": "https://open-gigaai.github.io/giga-world-policy/",
18
+ "resultDetailsUrl": null,
19
+ "contributor": "GigaAI",
20
+ "contributorUrl": "https://gigaai.cc/",
21
+ "type": "WAM",
22
+ "modelUrl": "https://open-gigaai.github.io/giga-world-policy/",
23
+ "logoPath": "leaderboard/gigaworld-policy-06/logo.png",
24
+ "logoAlt": "GigaAI",
25
+ "weightsUrl": null,
26
+ "codeUrl": null
27
+ },
28
+ "overall": {
29
+ "status": "reported",
30
+ "taskCount": 26,
31
+ "scope": "task-macro",
32
+ "sr": 51.36538461538461,
33
+ "lscr": 70.9099358974359,
34
+ "efficiency": 0.961797457000364,
35
+ "efficiencyStatus": "reported",
36
+ "efficiencyTaskCount": 26,
37
+ "channelSr": {
38
+ "None": 55.38461538461539,
39
+ "Equi.": 48.23076923076923,
40
+ "Inv.": 51.38461538461539,
41
+ "Full": 50.46153846153846
42
+ },
43
+ "channelLscr": {
44
+ "None": 73.81794871794872,
45
+ "Equi.": 68.9448717948718,
46
+ "Inv.": 71.1,
47
+ "Full": 69.77692307692308
48
+ },
49
+ "channelTaskCounts": {
50
+ "None": 26,
51
+ "Equi.": 26,
52
+ "Inv.": 26,
53
+ "Full": 26
54
+ }
55
+ },
56
+ "per_task": {
57
+ "03": {
58
+ "channels": {
59
+ "Equi.": {
60
+ "status": "reported",
61
+ "episodes": 50,
62
+ "successes": 30,
63
+ "lscr": 88.66666666666666,
64
+ "source": {
65
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
66
+ "perEpisodeFile": "per_task_eval/03_wine_glass_plate_balance/cov_only/per_episode.jsonl",
67
+ "perEpisodeSha256": "d6fed1a145a56acec35383f9368c308967d94f28144cec274ed8acfac1ed5935",
68
+ "summaryFile": "per_task_eval/03_wine_glass_plate_balance/cov_only/summary.json",
69
+ "summarySha256": "f22c4f8e5c8e091778568b15607e85e85aef5f7ce6a8c338ee5bef7ee9be4072"
70
+ }
71
+ },
72
+ "Full": {
73
+ "status": "reported",
74
+ "episodes": 50,
75
+ "successes": 27,
76
+ "lscr": 89.33333333333333,
77
+ "source": {
78
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
79
+ "perEpisodeFile": "per_task_eval/03_wine_glass_plate_balance/inv_cov/per_episode.jsonl",
80
+ "perEpisodeSha256": "0f86425b83ad558eaba050e47fbafe98425c16c848390108f4c55cb874ea707e",
81
+ "summaryFile": "per_task_eval/03_wine_glass_plate_balance/inv_cov/summary.json",
82
+ "summarySha256": "2a14fc8fc7cc04778d250943c02ba74d09e103d9e4eb586d006d649691711326"
83
+ }
84
+ },
85
+ "Inv.": {
86
+ "status": "reported",
87
+ "episodes": 50,
88
+ "successes": 31,
89
+ "lscr": 95.33333333333334,
90
+ "source": {
91
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
92
+ "perEpisodeFile": "per_task_eval/03_wine_glass_plate_balance/inv_only/per_episode.jsonl",
93
+ "perEpisodeSha256": "df12b982cc17762e6f2054d08bec3dac6f223cf60b665970afb976c9fd9a53b6",
94
+ "summaryFile": "per_task_eval/03_wine_glass_plate_balance/inv_only/summary.json",
95
+ "summarySha256": "572b19c68c2ff1b3908d085982f1a714567a823cc459f4d117ff01eedfec2f36"
96
+ }
97
+ },
98
+ "None": {
99
+ "status": "reported",
100
+ "episodes": 50,
101
+ "successes": 31,
102
+ "lscr": 91.33333333333333,
103
+ "source": {
104
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
105
+ "perEpisodeFile": "per_task_eval/03_wine_glass_plate_balance/none/per_episode.jsonl",
106
+ "perEpisodeSha256": "4f4a2be5c63f37630d59a05907de982a90cf53255a304fa5ed331e721d59df03",
107
+ "summaryFile": "per_task_eval/03_wine_glass_plate_balance/none/summary.json",
108
+ "summarySha256": "7db6af9bd4eadcf8e73a10d1949ad592e92cef471dd3f773d76bb213adfe5789"
109
+ }
110
+ }
111
+ },
112
+ "mean": {
113
+ "status": "reported",
114
+ "srScope": "four-channel-mean",
115
+ "lscrScope": "four-channel-mean",
116
+ "sr": 59.5,
117
+ "lscr": 91.16666666666666,
118
+ "efficiency": 1.0797859850390796,
119
+ "efficiencyStatus": "reported",
120
+ "srEpisodes": 200,
121
+ "lscrEpisodes": 200
122
+ }
123
+ },
124
+ "06": {
125
+ "channels": {
126
+ "Equi.": {
127
+ "status": "reported",
128
+ "episodes": 50,
129
+ "successes": 38,
130
+ "lscr": 91.0,
131
+ "source": {
132
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
133
+ "perEpisodeFile": "per_task_eval/06_fruit_bowl_loading/cov_only/per_episode.jsonl",
134
+ "perEpisodeSha256": "a1528d7c53dc05819f680e6707e60fe2d95b0f348f3577887197715c21b2b507",
135
+ "summaryFile": "per_task_eval/06_fruit_bowl_loading/cov_only/summary.json",
136
+ "summarySha256": "07e8bbdde4e64002f4855b9ff4b359496ce08bb03f0bf31a17016ffc0a0bd3ee"
137
+ }
138
+ },
139
+ "Full": {
140
+ "status": "reported",
141
+ "episodes": 50,
142
+ "successes": 35,
143
+ "lscr": 96.0,
144
+ "source": {
145
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
146
+ "perEpisodeFile": "per_task_eval/06_fruit_bowl_loading/inv_cov/per_episode.jsonl",
147
+ "perEpisodeSha256": "60f9f60bb33d96128284ac40f8337f6ec3e451a4bb24b8fd717b1d59d8b4e213",
148
+ "summaryFile": "per_task_eval/06_fruit_bowl_loading/inv_cov/summary.json",
149
+ "summarySha256": "e0711b090438219ef961f10ea15c537f1625f28d8d91ebb118dab70faf2b62ef"
150
+ }
151
+ },
152
+ "Inv.": {
153
+ "status": "reported",
154
+ "episodes": 50,
155
+ "successes": 41,
156
+ "lscr": 93.0,
157
+ "source": {
158
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
159
+ "perEpisodeFile": "per_task_eval/06_fruit_bowl_loading/inv_only/per_episode.jsonl",
160
+ "perEpisodeSha256": "432dbe1740fddd1e786ab1fc220de89aadb72ee7880b17b8f12a8307168173ff",
161
+ "summaryFile": "per_task_eval/06_fruit_bowl_loading/inv_only/summary.json",
162
+ "summarySha256": "8ec1d20333b562121aad84c48c900d31a80257409bd97d0af35563b529e1929e"
163
+ }
164
+ },
165
+ "None": {
166
+ "status": "reported",
167
+ "episodes": 50,
168
+ "successes": 44,
169
+ "lscr": 95.5,
170
+ "source": {
171
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
172
+ "perEpisodeFile": "per_task_eval/06_fruit_bowl_loading/none/per_episode.jsonl",
173
+ "perEpisodeSha256": "cdbb56c2cc5cfcdbcc507de52efe0b61a2f84256ccb989ba9635bf529f878da0",
174
+ "summaryFile": "per_task_eval/06_fruit_bowl_loading/none/summary.json",
175
+ "summarySha256": "f3e43d7e8d17095c6f220bda2dd89f98c4358f8788e7c5a33c256e87a744e299"
176
+ }
177
+ }
178
+ },
179
+ "mean": {
180
+ "status": "reported",
181
+ "srScope": "four-channel-mean",
182
+ "lscrScope": "four-channel-mean",
183
+ "sr": 79.0,
184
+ "lscr": 93.875,
185
+ "efficiency": 0.9312434383806384,
186
+ "efficiencyStatus": "reported",
187
+ "srEpisodes": 200,
188
+ "lscrEpisodes": 200
189
+ }
190
+ },
191
+ "07": {
192
+ "channels": {
193
+ "Equi.": {
194
+ "status": "reported",
195
+ "episodes": 50,
196
+ "successes": 43,
197
+ "lscr": 98.0,
198
+ "source": {
199
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
200
+ "perEpisodeFile": "per_task_eval/07_citrus_plate_loading/cov_only/per_episode.jsonl",
201
+ "perEpisodeSha256": "35bbb7e0e5ff4b05ec9d60af2b4ed59d6bfcd62d3681d98be439a157a0fe3c1a",
202
+ "summaryFile": "per_task_eval/07_citrus_plate_loading/cov_only/summary.json",
203
+ "summarySha256": "9d8d9b1f638315c49972b08d0928da875053b941f31380b00866c1fcd7b33738"
204
+ }
205
+ },
206
+ "Full": {
207
+ "status": "reported",
208
+ "episodes": 50,
209
+ "successes": 43,
210
+ "lscr": 92.0,
211
+ "source": {
212
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
213
+ "perEpisodeFile": "per_task_eval/07_citrus_plate_loading/inv_cov/per_episode.jsonl",
214
+ "perEpisodeSha256": "a35c1c904fe9edeabce230e63bf02b1b58d261b6560aad389fce19a4ff399e1a",
215
+ "summaryFile": "per_task_eval/07_citrus_plate_loading/inv_cov/summary.json",
216
+ "summarySha256": "de2e9bac49ea3b291d955f86c6354898bce5dfd2db4899284e17866ef46958b4"
217
+ }
218
+ },
219
+ "Inv.": {
220
+ "status": "reported",
221
+ "episodes": 50,
222
+ "successes": 44,
223
+ "lscr": 94.5,
224
+ "source": {
225
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
226
+ "perEpisodeFile": "per_task_eval/07_citrus_plate_loading/inv_only/per_episode.jsonl",
227
+ "perEpisodeSha256": "a316f73cb351fbdee5fef2f81a7984e98483b917fef83865bb59cc06a210cad9",
228
+ "summaryFile": "per_task_eval/07_citrus_plate_loading/inv_only/summary.json",
229
+ "summarySha256": "28491ce0a26bac7f852c3fcfcd4c6026295a7c8c53636b85658a400898ddd441"
230
+ }
231
+ },
232
+ "None": {
233
+ "status": "reported",
234
+ "episodes": 50,
235
+ "successes": 45,
236
+ "lscr": 98.0,
237
+ "source": {
238
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
239
+ "perEpisodeFile": "per_task_eval/07_citrus_plate_loading/none/per_episode.jsonl",
240
+ "perEpisodeSha256": "8cc687916a4c203e14ee45082302684f0f7b1d913f184d1c97dfe4c2fb279c86",
241
+ "summaryFile": "per_task_eval/07_citrus_plate_loading/none/summary.json",
242
+ "summarySha256": "be473140eb01d7f2af1648649d23f1fdade796bfce1b343710be7053fd0e9bd4"
243
+ }
244
+ }
245
+ },
246
+ "mean": {
247
+ "status": "reported",
248
+ "srScope": "four-channel-mean",
249
+ "lscrScope": "four-channel-mean",
250
+ "sr": 87.5,
251
+ "lscr": 95.625,
252
+ "efficiency": 0.8419888574708139,
253
+ "efficiencyStatus": "reported",
254
+ "srEpisodes": 200,
255
+ "lscrEpisodes": 200
256
+ }
257
+ },
258
+ "08": {
259
+ "channels": {
260
+ "Equi.": {
261
+ "status": "reported",
262
+ "episodes": 50,
263
+ "successes": 46,
264
+ "lscr": 94.0,
265
+ "source": {
266
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
267
+ "perEpisodeFile": "per_task_eval/08_frypan_stand_pour/cov_only/per_episode.jsonl",
268
+ "perEpisodeSha256": "ec084d07e414d05bbf5dae3ea88bff588debc19d2fcb684fab2262b2dccf44ea",
269
+ "summaryFile": "per_task_eval/08_frypan_stand_pour/cov_only/summary.json",
270
+ "summarySha256": "1a55b9a44ac882fae78484059764a54fea6b089ab3e32941ee2d5c4e80353185"
271
+ }
272
+ },
273
+ "Full": {
274
+ "status": "reported",
275
+ "episodes": 50,
276
+ "successes": 42,
277
+ "lscr": 90.0,
278
+ "source": {
279
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
280
+ "perEpisodeFile": "per_task_eval/08_frypan_stand_pour/inv_cov/per_episode.jsonl",
281
+ "perEpisodeSha256": "c0463790147a6c4ff670f23d6ef556a33cefa7a3e534dce635b2fcea92bfb06f",
282
+ "summaryFile": "per_task_eval/08_frypan_stand_pour/inv_cov/summary.json",
283
+ "summarySha256": "755524d5d5333649dc6ce4c965d96bc710840f0cbf7dd5452451140eb0eedac5"
284
+ }
285
+ },
286
+ "Inv.": {
287
+ "status": "reported",
288
+ "episodes": 50,
289
+ "successes": 46,
290
+ "lscr": 97.0,
291
+ "source": {
292
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
293
+ "perEpisodeFile": "per_task_eval/08_frypan_stand_pour/inv_only/per_episode.jsonl",
294
+ "perEpisodeSha256": "f11db4ed6407af3d610d0092f312b269cdec53db9dd3e72f1dbce1ec490facea",
295
+ "summaryFile": "per_task_eval/08_frypan_stand_pour/inv_only/summary.json",
296
+ "summarySha256": "9c4eeed5294a2e8c2a4d7b35e76c85b96138e353d141a93888e1c72bce59fd8c"
297
+ }
298
+ },
299
+ "None": {
300
+ "status": "reported",
301
+ "episodes": 50,
302
+ "successes": 46,
303
+ "lscr": 95.0,
304
+ "source": {
305
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
306
+ "perEpisodeFile": "per_task_eval/08_frypan_stand_pour/none/per_episode.jsonl",
307
+ "perEpisodeSha256": "44aad2322ae487de7a233841f5429a25093c214b45162c25886232f081257df5",
308
+ "summaryFile": "per_task_eval/08_frypan_stand_pour/none/summary.json",
309
+ "summarySha256": "e29eff97524218ee63f627353596a25db07a4fc8bc1f4a3180eefdf8a1106cb2"
310
+ }
311
+ }
312
+ },
313
+ "mean": {
314
+ "status": "reported",
315
+ "srScope": "four-channel-mean",
316
+ "lscrScope": "four-channel-mean",
317
+ "sr": 90.0,
318
+ "lscr": 94.0,
319
+ "efficiency": 1.2919700472756133,
320
+ "efficiencyStatus": "reported",
321
+ "srEpisodes": 200,
322
+ "lscrEpisodes": 200
323
+ }
324
+ },
325
+ "09": {
326
+ "channels": {
327
+ "Equi.": {
328
+ "status": "reported",
329
+ "episodes": 50,
330
+ "successes": 39,
331
+ "lscr": 98.0,
332
+ "source": {
333
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
334
+ "perEpisodeFile": "per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/per_episode.jsonl",
335
+ "perEpisodeSha256": "295ac5f7d7a08689639202cc1d10b7aea171650ae56f40dec6276b0993044008",
336
+ "summaryFile": "per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/summary.json",
337
+ "summarySha256": "c35c1506148ed35c9cbe1f5332b21597982e804bbc17be24285f1b5dc2320336"
338
+ }
339
+ },
340
+ "Full": {
341
+ "status": "reported",
342
+ "episodes": 50,
343
+ "successes": 37,
344
+ "lscr": 96.0,
345
+ "source": {
346
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
347
+ "perEpisodeFile": "per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/per_episode.jsonl",
348
+ "perEpisodeSha256": "b1644c3d2d02927ff959162096cab165516f15d4d398078bdba4cf849b608289",
349
+ "summaryFile": "per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/summary.json",
350
+ "summarySha256": "cd694428dae64e857943b8c43db60ed2747b9a8b172bfab8841e446cd11c001e"
351
+ }
352
+ },
353
+ "Inv.": {
354
+ "status": "reported",
355
+ "episodes": 50,
356
+ "successes": 36,
357
+ "lscr": 98.0,
358
+ "source": {
359
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
360
+ "perEpisodeFile": "per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/per_episode.jsonl",
361
+ "perEpisodeSha256": "5c400781e3efb77258e7d18a18920af8befbe08f2939146a431bed92e39fb215",
362
+ "summaryFile": "per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/summary.json",
363
+ "summarySha256": "69633aa6214598fbdcaf734efc9ee93fb792f58074156cd68d3fdb375ebd39fe"
364
+ }
365
+ },
366
+ "None": {
367
+ "status": "reported",
368
+ "episodes": 50,
369
+ "successes": 42,
370
+ "lscr": 98.0,
371
+ "source": {
372
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
373
+ "perEpisodeFile": "per_task_eval/09_cleaner_moisturizer_box_loading/none/per_episode.jsonl",
374
+ "perEpisodeSha256": "c4e28d99a963ca069857611dff690f723832023f16fb7b218bcde3ec410b6ad8",
375
+ "summaryFile": "per_task_eval/09_cleaner_moisturizer_box_loading/none/summary.json",
376
+ "summarySha256": "3d6ed7656e381d63c77f38760fad615655a1d9715e0ae142c094638060fb7d68"
377
+ }
378
+ }
379
+ },
380
+ "mean": {
381
+ "status": "reported",
382
+ "srScope": "four-channel-mean",
383
+ "lscrScope": "four-channel-mean",
384
+ "sr": 77.0,
385
+ "lscr": 97.5,
386
+ "efficiency": 0.9249617294989491,
387
+ "efficiencyStatus": "reported",
388
+ "srEpisodes": 200,
389
+ "lscrEpisodes": 200
390
+ }
391
+ },
392
+ "12": {
393
+ "channels": {
394
+ "Equi.": {
395
+ "status": "reported",
396
+ "episodes": 50,
397
+ "successes": 26,
398
+ "lscr": 85.0,
399
+ "source": {
400
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
401
+ "perEpisodeFile": "per_task_eval/12_screwdriver_box_and_hammer/cov_only/per_episode.jsonl",
402
+ "perEpisodeSha256": "74ef81f67213282710a64c6439e253ccf4e37e90a5a5b8e98f2489dc9a1ba1f1",
403
+ "summaryFile": "per_task_eval/12_screwdriver_box_and_hammer/cov_only/summary.json",
404
+ "summarySha256": "db3db531404074f1de8f68a9aaa19158f8a1e8d039a4069c7c907127a3105331"
405
+ }
406
+ },
407
+ "Full": {
408
+ "status": "reported",
409
+ "episodes": 50,
410
+ "successes": 30,
411
+ "lscr": 83.0,
412
+ "source": {
413
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
414
+ "perEpisodeFile": "per_task_eval/12_screwdriver_box_and_hammer/inv_cov/per_episode.jsonl",
415
+ "perEpisodeSha256": "67c03d7845ce035b5a076daca538f389e21fd01dc449eb512744f23495dcaefd",
416
+ "summaryFile": "per_task_eval/12_screwdriver_box_and_hammer/inv_cov/summary.json",
417
+ "summarySha256": "b4b741b3c147c6a3ab9b8ca5b25e5763fd61f52daa68480c6fac82b5ec7e9f5d"
418
+ }
419
+ },
420
+ "Inv.": {
421
+ "status": "reported",
422
+ "episodes": 50,
423
+ "successes": 40,
424
+ "lscr": 90.5,
425
+ "source": {
426
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
427
+ "perEpisodeFile": "per_task_eval/12_screwdriver_box_and_hammer/inv_only/per_episode.jsonl",
428
+ "perEpisodeSha256": "836540412ca739eee839a6a646302f493946c2c14b3bf78bf6585b9fda301231",
429
+ "summaryFile": "per_task_eval/12_screwdriver_box_and_hammer/inv_only/summary.json",
430
+ "summarySha256": "7954672a0433f1013cd6e755b2baf95776e82b79e3a16133a26e74487f5bf693"
431
+ }
432
+ },
433
+ "None": {
434
+ "status": "reported",
435
+ "episodes": 50,
436
+ "successes": 35,
437
+ "lscr": 91.0,
438
+ "source": {
439
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
440
+ "perEpisodeFile": "per_task_eval/12_screwdriver_box_and_hammer/none/per_episode.jsonl",
441
+ "perEpisodeSha256": "e6a1dbdb11429d66fd2fc4a244739778094fc9dccd71f7ae6fbcc8dd85c5c3f6",
442
+ "summaryFile": "per_task_eval/12_screwdriver_box_and_hammer/none/summary.json",
443
+ "summarySha256": "ecec0842be75cfbd5be42980a26d89339b0e720e513faae2c9aabc3e294b745d"
444
+ }
445
+ }
446
+ },
447
+ "mean": {
448
+ "status": "reported",
449
+ "srScope": "four-channel-mean",
450
+ "lscrScope": "four-channel-mean",
451
+ "sr": 65.5,
452
+ "lscr": 87.375,
453
+ "efficiency": 0.9687895402783636,
454
+ "efficiencyStatus": "reported",
455
+ "srEpisodes": 200,
456
+ "lscrEpisodes": 200
457
+ }
458
+ },
459
+ "21": {
460
+ "channels": {
461
+ "Equi.": {
462
+ "status": "reported",
463
+ "episodes": 50,
464
+ "successes": 36,
465
+ "lscr": 97.33333333333334,
466
+ "source": {
467
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
468
+ "perEpisodeFile": "per_task_eval/21_condiment_box_loading/cov_only/per_episode.jsonl",
469
+ "perEpisodeSha256": "44b40ece67cf36899ccd74e78b3f0b5e5642f5f0c1e1e51050061e855d30dee3",
470
+ "summaryFile": "per_task_eval/21_condiment_box_loading/cov_only/summary.json",
471
+ "summarySha256": "80314b57774a54a7f8064149d4ffd72c0ffb29e5cf127144158e46f102d2c502"
472
+ }
473
+ },
474
+ "Full": {
475
+ "status": "reported",
476
+ "episodes": 50,
477
+ "successes": 36,
478
+ "lscr": 94.0,
479
+ "source": {
480
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
481
+ "perEpisodeFile": "per_task_eval/21_condiment_box_loading/inv_cov/per_episode.jsonl",
482
+ "perEpisodeSha256": "20469b08eb0eaca46889d63d4b4f2f7de176e79735faa920332b6f1dfa36c47d",
483
+ "summaryFile": "per_task_eval/21_condiment_box_loading/inv_cov/summary.json",
484
+ "summarySha256": "4691f5b354fd1c6aebf36dec5e0c9860ff896daf19acb7d57e30dd7e952e8b36"
485
+ }
486
+ },
487
+ "Inv.": {
488
+ "status": "reported",
489
+ "episodes": 50,
490
+ "successes": 33,
491
+ "lscr": 90.66666666666666,
492
+ "source": {
493
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
494
+ "perEpisodeFile": "per_task_eval/21_condiment_box_loading/inv_only/per_episode.jsonl",
495
+ "perEpisodeSha256": "d6659d5c0cc9a9e9d847e1a44ae65137f0bb79b6f0ff3a3c1ea4eca6ffda7679",
496
+ "summaryFile": "per_task_eval/21_condiment_box_loading/inv_only/summary.json",
497
+ "summarySha256": "f0fa007fe15bc27766d00daea39f993079e244e5f49b2c29526e020e6bc354d6"
498
+ }
499
+ },
500
+ "None": {
501
+ "status": "reported",
502
+ "episodes": 50,
503
+ "successes": 38,
504
+ "lscr": 97.33333333333334,
505
+ "source": {
506
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
507
+ "perEpisodeFile": "per_task_eval/21_condiment_box_loading/none/per_episode.jsonl",
508
+ "perEpisodeSha256": "3b8eea84a9aa96621c98c4b9d7647e07a040d02ffb58278c849640521298fb92",
509
+ "summaryFile": "per_task_eval/21_condiment_box_loading/none/summary.json",
510
+ "summarySha256": "06b1df715e5c8eef6d4f61a93797381b4cdea2b7314c4ee06a6016fc1b3b61eb"
511
+ }
512
+ }
513
+ },
514
+ "mean": {
515
+ "status": "reported",
516
+ "srScope": "four-channel-mean",
517
+ "lscrScope": "four-channel-mean",
518
+ "sr": 71.5,
519
+ "lscr": 94.83333333333334,
520
+ "efficiency": 0.9274499654934437,
521
+ "efficiencyStatus": "reported",
522
+ "srEpisodes": 200,
523
+ "lscrEpisodes": 200
524
+ }
525
+ },
526
+ "22": {
527
+ "channels": {
528
+ "Equi.": {
529
+ "status": "reported",
530
+ "episodes": 50,
531
+ "successes": 8,
532
+ "lscr": 39.0,
533
+ "source": {
534
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
535
+ "perEpisodeFile": "per_task_eval/22_tool_box_loading/cov_only/per_episode.jsonl",
536
+ "perEpisodeSha256": "62eb49bcb7a3e1fa2d719b5e8db43d76e99d7364a04090da2b6ba8dcd13f07fa",
537
+ "summaryFile": "per_task_eval/22_tool_box_loading/cov_only/summary.json",
538
+ "summarySha256": "1239495b8c0f04f876b37bec1d71695161e9235d9137a58f6870634fd6493c66"
539
+ }
540
+ },
541
+ "Full": {
542
+ "status": "reported",
543
+ "episodes": 50,
544
+ "successes": 13,
545
+ "lscr": 43.0,
546
+ "source": {
547
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
548
+ "perEpisodeFile": "per_task_eval/22_tool_box_loading/inv_cov/per_episode.jsonl",
549
+ "perEpisodeSha256": "186ba40ea3f17b63eef080d4111ac8dff9e2743b98a80ce036c72463f6977355",
550
+ "summaryFile": "per_task_eval/22_tool_box_loading/inv_cov/summary.json",
551
+ "summarySha256": "3c76347ededfb2626cf48243c0e99e9081a91819aca58d9177b009e549c7fa20"
552
+ }
553
+ },
554
+ "Inv.": {
555
+ "status": "reported",
556
+ "episodes": 50,
557
+ "successes": 11,
558
+ "lscr": 38.5,
559
+ "source": {
560
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
561
+ "perEpisodeFile": "per_task_eval/22_tool_box_loading/inv_only/per_episode.jsonl",
562
+ "perEpisodeSha256": "667b4100e21697e83fb371d75d34cb6f5ccaddd0204f2b8b121b50b4261326ea",
563
+ "summaryFile": "per_task_eval/22_tool_box_loading/inv_only/summary.json",
564
+ "summarySha256": "fc454c8a93e4208afd3a8f21a931c717d41d31fbfbb14092177208fffa01a870"
565
+ }
566
+ },
567
+ "None": {
568
+ "status": "reported",
569
+ "episodes": 50,
570
+ "successes": 13,
571
+ "lscr": 42.0,
572
+ "source": {
573
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
574
+ "perEpisodeFile": "per_task_eval/22_tool_box_loading/none/per_episode.jsonl",
575
+ "perEpisodeSha256": "4df94bb6f70a91ec9302be8d98c6cad52bc353fd7377af27ea8b4c4f994467e1",
576
+ "summaryFile": "per_task_eval/22_tool_box_loading/none/summary.json",
577
+ "summarySha256": "35c9525dde92c7eef0d9a7b8f4f33794d3064410a2f18e6da94ef87faf583f79"
578
+ }
579
+ }
580
+ },
581
+ "mean": {
582
+ "status": "reported",
583
+ "srScope": "four-channel-mean",
584
+ "lscrScope": "four-channel-mean",
585
+ "sr": 22.5,
586
+ "lscr": 40.625,
587
+ "efficiency": 0.8941789671937278,
588
+ "efficiencyStatus": "reported",
589
+ "srEpisodes": 200,
590
+ "lscrEpisodes": 200
591
+ }
592
+ },
593
+ "24": {
594
+ "channels": {
595
+ "Equi.": {
596
+ "status": "reported",
597
+ "episodes": 50,
598
+ "successes": 17,
599
+ "lscr": 68.0,
600
+ "source": {
601
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
602
+ "perEpisodeFile": "per_task_eval/24_stationery_category_sorting/cov_only/per_episode.jsonl",
603
+ "perEpisodeSha256": "1ed7eccbd2b9fc5883402117282bcb623cc697d317a4114abe9b8614452cd867",
604
+ "summaryFile": "per_task_eval/24_stationery_category_sorting/cov_only/summary.json",
605
+ "summarySha256": "7264e270719491d491a7ecd1306b5740effee9147a9cf24bbd542d65922d4f69"
606
+ }
607
+ },
608
+ "Full": {
609
+ "status": "reported",
610
+ "episodes": 50,
611
+ "successes": 28,
612
+ "lscr": 72.5,
613
+ "source": {
614
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
615
+ "perEpisodeFile": "per_task_eval/24_stationery_category_sorting/inv_cov/per_episode.jsonl",
616
+ "perEpisodeSha256": "84f9bd3665efad2eaa632c4127e36bbab715cf8e10d1c1122c3292b5b8bf4933",
617
+ "summaryFile": "per_task_eval/24_stationery_category_sorting/inv_cov/summary.json",
618
+ "summarySha256": "7e09728dc15081510d66f5d34d829f6b85ad8d4100bb8c2126f2b4df6ca05351"
619
+ }
620
+ },
621
+ "Inv.": {
622
+ "status": "reported",
623
+ "episodes": 50,
624
+ "successes": 20,
625
+ "lscr": 62.5,
626
+ "source": {
627
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
628
+ "perEpisodeFile": "per_task_eval/24_stationery_category_sorting/inv_only/per_episode.jsonl",
629
+ "perEpisodeSha256": "6e4c3657070b5f1bba56cd1862c29b89fe851e8e2a4872b9b34de79bb8de6f0a",
630
+ "summaryFile": "per_task_eval/24_stationery_category_sorting/inv_only/summary.json",
631
+ "summarySha256": "7ea8fca41a5f815f67e7510906274174c90102b889c9a1214c344bb68827debe"
632
+ }
633
+ },
634
+ "None": {
635
+ "status": "reported",
636
+ "episodes": 50,
637
+ "successes": 27,
638
+ "lscr": 82.0,
639
+ "source": {
640
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
641
+ "perEpisodeFile": "per_task_eval/24_stationery_category_sorting/none/per_episode.jsonl",
642
+ "perEpisodeSha256": "949a6708bfbb9e6f5c39929f42ae9957850b1eceb5769754eb46e55ec0bfc320",
643
+ "summaryFile": "per_task_eval/24_stationery_category_sorting/none/summary.json",
644
+ "summarySha256": "1347b5bd04171be4b24f40d06c90944548fee0ea9352f7496c54fa9a906a3155"
645
+ }
646
+ }
647
+ },
648
+ "mean": {
649
+ "status": "reported",
650
+ "srScope": "four-channel-mean",
651
+ "lscrScope": "four-channel-mean",
652
+ "sr": 46.0,
653
+ "lscr": 71.25,
654
+ "efficiency": 1.047794959696784,
655
+ "efficiencyStatus": "reported",
656
+ "srEpisodes": 200,
657
+ "lscrEpisodes": 200
658
+ }
659
+ },
660
+ "26": {
661
+ "channels": {
662
+ "Equi.": {
663
+ "status": "reported",
664
+ "episodes": 50,
665
+ "successes": 18,
666
+ "lscr": 64.5,
667
+ "source": {
668
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
669
+ "perEpisodeFile": "per_task_eval/26_canned_food_tray_line_arrangement/cov_only/per_episode.jsonl",
670
+ "perEpisodeSha256": "f14ad81eef62cec48f445b809671039ea7d6e72d538c6efc9daf3321e29b3760",
671
+ "summaryFile": "per_task_eval/26_canned_food_tray_line_arrangement/cov_only/summary.json",
672
+ "summarySha256": "55fb2d263ea6bb75a4ed21fc30e80c4a72704d52c7c2b04548dfd06b2b520e9f"
673
+ }
674
+ },
675
+ "Full": {
676
+ "status": "reported",
677
+ "episodes": 50,
678
+ "successes": 24,
679
+ "lscr": 71.5,
680
+ "source": {
681
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
682
+ "perEpisodeFile": "per_task_eval/26_canned_food_tray_line_arrangement/inv_cov/per_episode.jsonl",
683
+ "perEpisodeSha256": "0bb0a52759fa793131405e71e51c6cf5d68b908d49fa9a7bf3069dc6b7c56558",
684
+ "summaryFile": "per_task_eval/26_canned_food_tray_line_arrangement/inv_cov/summary.json",
685
+ "summarySha256": "43b16e3652d6fd2c32d16832e8eb448b7fd100cef58ef8f235081897dc10c3ef"
686
+ }
687
+ },
688
+ "Inv.": {
689
+ "status": "reported",
690
+ "episodes": 50,
691
+ "successes": 21,
692
+ "lscr": 72.5,
693
+ "source": {
694
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
695
+ "perEpisodeFile": "per_task_eval/26_canned_food_tray_line_arrangement/inv_only/per_episode.jsonl",
696
+ "perEpisodeSha256": "60c9763e139e329b6f0872c5ebcd7e55db7d43a18b952e74e1838f4186c182c7",
697
+ "summaryFile": "per_task_eval/26_canned_food_tray_line_arrangement/inv_only/summary.json",
698
+ "summarySha256": "7810b52d485c1c40f89eed4baacc4034919210ea6c4398b90141bec9e7d0028a"
699
+ }
700
+ },
701
+ "None": {
702
+ "status": "reported",
703
+ "episodes": 50,
704
+ "successes": 17,
705
+ "lscr": 62.5,
706
+ "source": {
707
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
708
+ "perEpisodeFile": "per_task_eval/26_canned_food_tray_line_arrangement/none/per_episode.jsonl",
709
+ "perEpisodeSha256": "c468fc7896c1858c814c512e26f58b82ef9c333cde38b657acc45aa2f70c42b8",
710
+ "summaryFile": "per_task_eval/26_canned_food_tray_line_arrangement/none/summary.json",
711
+ "summarySha256": "5c31ca39b1c942e52d19cd665d8e5a087cf530bf7f23af88e9c3516644e79bcf"
712
+ }
713
+ }
714
+ },
715
+ "mean": {
716
+ "status": "reported",
717
+ "srScope": "four-channel-mean",
718
+ "lscrScope": "four-channel-mean",
719
+ "sr": 40.0,
720
+ "lscr": 67.75,
721
+ "efficiency": 0.9435469674606292,
722
+ "efficiencyStatus": "reported",
723
+ "srEpisodes": 200,
724
+ "lscrEpisodes": 200
725
+ }
726
+ },
727
+ "27": {
728
+ "channels": {
729
+ "Equi.": {
730
+ "status": "reported",
731
+ "episodes": 50,
732
+ "successes": 30,
733
+ "lscr": 80.0,
734
+ "source": {
735
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
736
+ "perEpisodeFile": "per_task_eval/27_ball_box_loading/cov_only/per_episode.jsonl",
737
+ "perEpisodeSha256": "7aed108637e4aa1fb7d69dd873c65a241ccc9f379425982d6009d862155b68f5",
738
+ "summaryFile": "per_task_eval/27_ball_box_loading/cov_only/summary.json",
739
+ "summarySha256": "7d5db98065424c62602ad7b73227ef63b682598ec5ddbea0db4e5a686db2d00c"
740
+ }
741
+ },
742
+ "Full": {
743
+ "status": "reported",
744
+ "episodes": 50,
745
+ "successes": 30,
746
+ "lscr": 83.0,
747
+ "source": {
748
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
749
+ "perEpisodeFile": "per_task_eval/27_ball_box_loading/inv_cov/per_episode.jsonl",
750
+ "perEpisodeSha256": "acaa6cd38dad2c4cdac9b01a01a8a765fa1990a1f4e54c109a0e74ec5490fb43",
751
+ "summaryFile": "per_task_eval/27_ball_box_loading/inv_cov/summary.json",
752
+ "summarySha256": "3065d3d608eb399aab98fb963c898548defae9741ad5838b81dbf5887cbd1a74"
753
+ }
754
+ },
755
+ "Inv.": {
756
+ "status": "reported",
757
+ "episodes": 50,
758
+ "successes": 33,
759
+ "lscr": 85.0,
760
+ "source": {
761
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
762
+ "perEpisodeFile": "per_task_eval/27_ball_box_loading/inv_only/per_episode.jsonl",
763
+ "perEpisodeSha256": "aafd978c5033634d96b3bce82e7ca7315587649192e83949678764f6c557ac31",
764
+ "summaryFile": "per_task_eval/27_ball_box_loading/inv_only/summary.json",
765
+ "summarySha256": "ae355019ab7ccabc52063b1e95dbb4ff91b55f6539bec345d86adc9f4c2052c0"
766
+ }
767
+ },
768
+ "None": {
769
+ "status": "reported",
770
+ "episodes": 50,
771
+ "successes": 39,
772
+ "lscr": 86.5,
773
+ "source": {
774
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_wuji",
775
+ "perEpisodeFile": "per_task_eval/27_ball_box_loading/none/per_episode.jsonl",
776
+ "perEpisodeSha256": "fafcdf7ed23ed6ea49dd2886bc67d35b7d7998c4ae87ed5e3e526c2df3921f8a",
777
+ "summaryFile": "per_task_eval/27_ball_box_loading/none/summary.json",
778
+ "summarySha256": "6fd93a21eae83a801f2d5043700add56f1c3d0a620d66c54a0a191fc3c1c8528"
779
+ }
780
+ }
781
+ },
782
+ "mean": {
783
+ "status": "reported",
784
+ "srScope": "four-channel-mean",
785
+ "lscrScope": "four-channel-mean",
786
+ "sr": 66.0,
787
+ "lscr": 83.625,
788
+ "efficiency": 0.9758127890615537,
789
+ "efficiencyStatus": "reported",
790
+ "srEpisodes": 200,
791
+ "lscrEpisodes": 200
792
+ }
793
+ },
794
+ "32": {
795
+ "channels": {
796
+ "Equi.": {
797
+ "status": "reported",
798
+ "episodes": 50,
799
+ "successes": 8,
800
+ "lscr": 55.00000000000001,
801
+ "source": {
802
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
803
+ "perEpisodeFile": "per_task_eval/32_baking_tray_prep_with_tools/cov_only/per_episode.jsonl",
804
+ "perEpisodeSha256": "15d557ac5af327f269aef0b556aab9ac31c35eaa83453b8067813de9bc932555",
805
+ "summaryFile": "per_task_eval/32_baking_tray_prep_with_tools/cov_only/summary.json",
806
+ "summarySha256": "7810e5240e89e4f0f5668e6a9db3068b5e4e6ad860c9062bc46e002992e547ee"
807
+ }
808
+ },
809
+ "Full": {
810
+ "status": "reported",
811
+ "episodes": 50,
812
+ "successes": 7,
813
+ "lscr": 57.99999999999999,
814
+ "source": {
815
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
816
+ "perEpisodeFile": "per_task_eval/32_baking_tray_prep_with_tools/inv_cov/per_episode.jsonl",
817
+ "perEpisodeSha256": "7b627bc63946e353a4f0942a044925f62a08f59b85a6a52860e480d905660e30",
818
+ "summaryFile": "per_task_eval/32_baking_tray_prep_with_tools/inv_cov/summary.json",
819
+ "summarySha256": "38009ca9a8406993dad302e9bce4ba49f30290af5bf02dc4a8e4bfa96c8fe40a"
820
+ }
821
+ },
822
+ "Inv.": {
823
+ "status": "reported",
824
+ "episodes": 50,
825
+ "successes": 14,
826
+ "lscr": 72.0,
827
+ "source": {
828
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
829
+ "perEpisodeFile": "per_task_eval/32_baking_tray_prep_with_tools/inv_only/per_episode.jsonl",
830
+ "perEpisodeSha256": "f5668c4cb8bf3038e186f848b324229fe065c49d53ce00c6dfff74a034865e8f",
831
+ "summaryFile": "per_task_eval/32_baking_tray_prep_with_tools/inv_only/summary.json",
832
+ "summarySha256": "c6a531d02ad10ab45b5a402d2945c4e6f37415ddc0cd4c321c4f20a2391e88f8"
833
+ }
834
+ },
835
+ "None": {
836
+ "status": "reported",
837
+ "episodes": 50,
838
+ "successes": 13,
839
+ "lscr": 67.5,
840
+ "source": {
841
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
842
+ "perEpisodeFile": "per_task_eval/32_baking_tray_prep_with_tools/none/per_episode.jsonl",
843
+ "perEpisodeSha256": "0d4d247f2172afb6deedbb93fe39030e2596a570d6ccff8ec8c22b807c8ce9e1",
844
+ "summaryFile": "per_task_eval/32_baking_tray_prep_with_tools/none/summary.json",
845
+ "summarySha256": "677b00c6de4093fd3d26b026f5e09c9062b748df337d2f6f88b6075cb5498e17"
846
+ }
847
+ }
848
+ },
849
+ "mean": {
850
+ "status": "reported",
851
+ "srScope": "four-channel-mean",
852
+ "lscrScope": "four-channel-mean",
853
+ "sr": 21.0,
854
+ "lscr": 63.125,
855
+ "efficiency": 0.8586039253346137,
856
+ "efficiencyStatus": "reported",
857
+ "srEpisodes": 200,
858
+ "lscrEpisodes": 200
859
+ }
860
+ },
861
+ "34": {
862
+ "channels": {
863
+ "Equi.": {
864
+ "status": "reported",
865
+ "episodes": 50,
866
+ "successes": 31,
867
+ "lscr": 80.4,
868
+ "source": {
869
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
870
+ "perEpisodeFile": "per_task_eval/34_fridge_wine_interhand_pour/cov_only/per_episode.jsonl",
871
+ "perEpisodeSha256": "d66862ea6fe68303863050ed959badd97d005bea7e901e1421bbf7b7fdaaae42",
872
+ "summaryFile": "per_task_eval/34_fridge_wine_interhand_pour/cov_only/summary.json",
873
+ "summarySha256": "13bf8d7689f51d6d3200e0d84890d29bcef1e79f7c6bb154dae52d311d9034ff"
874
+ }
875
+ },
876
+ "Full": {
877
+ "status": "reported",
878
+ "episodes": 50,
879
+ "successes": 36,
880
+ "lscr": 85.2,
881
+ "source": {
882
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
883
+ "perEpisodeFile": "per_task_eval/34_fridge_wine_interhand_pour/inv_cov/per_episode.jsonl",
884
+ "perEpisodeSha256": "f3ae033404f0402e587c944d33b655a07c8369a6ee73da8286103da71230cb60",
885
+ "summaryFile": "per_task_eval/34_fridge_wine_interhand_pour/inv_cov/summary.json",
886
+ "summarySha256": "35d81e38f244efe45be3f7b74059384da264c676fc664d466a3e732acef0a2fb"
887
+ }
888
+ },
889
+ "Inv.": {
890
+ "status": "reported",
891
+ "episodes": 50,
892
+ "successes": 34,
893
+ "lscr": 83.6,
894
+ "source": {
895
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
896
+ "perEpisodeFile": "per_task_eval/34_fridge_wine_interhand_pour/inv_only/per_episode.jsonl",
897
+ "perEpisodeSha256": "d28b3ba37c3386851c66eac6535841419b97ffdea9053eea3bd5ee9d330bc1e8",
898
+ "summaryFile": "per_task_eval/34_fridge_wine_interhand_pour/inv_only/summary.json",
899
+ "summarySha256": "027c2d171ea610659200b98c4aa0dc944cfc020998a4fd182e125ff627f0cce2"
900
+ }
901
+ },
902
+ "None": {
903
+ "status": "reported",
904
+ "episodes": 50,
905
+ "successes": 32,
906
+ "lscr": 83.6,
907
+ "source": {
908
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
909
+ "perEpisodeFile": "per_task_eval/34_fridge_wine_interhand_pour/none/per_episode.jsonl",
910
+ "perEpisodeSha256": "9eeb5d48a2afd179c20b4b96d1ef4f8284da58e1abf31507f599cb15456773b3",
911
+ "summaryFile": "per_task_eval/34_fridge_wine_interhand_pour/none/summary.json",
912
+ "summarySha256": "6fa5ea4d0b37b36f7fdd980742f770c2d2cf5e6a025975acc75021846c065db4"
913
+ }
914
+ }
915
+ },
916
+ "mean": {
917
+ "status": "reported",
918
+ "srScope": "four-channel-mean",
919
+ "lscrScope": "four-channel-mean",
920
+ "sr": 66.5,
921
+ "lscr": 83.2,
922
+ "efficiency": 1.0229749170310654,
923
+ "efficiencyStatus": "reported",
924
+ "srEpisodes": 200,
925
+ "lscrEpisodes": 200
926
+ }
927
+ },
928
+ "42": {
929
+ "channels": {
930
+ "Equi.": {
931
+ "status": "reported",
932
+ "episodes": 50,
933
+ "successes": 27,
934
+ "lscr": 90.0,
935
+ "source": {
936
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
937
+ "perEpisodeFile": "per_task_eval/42_trash_disposal/cov_only/per_episode.jsonl",
938
+ "perEpisodeSha256": "5ba0b5069f0a45c0db0c45527830c0b966e304ff47a622ca353983e9882de382",
939
+ "summaryFile": "per_task_eval/42_trash_disposal/cov_only/summary.json",
940
+ "summarySha256": "b0fc2300aef0df19fd85f0946833504b739996608aabaf1d308f16666eb75f18"
941
+ }
942
+ },
943
+ "Full": {
944
+ "status": "reported",
945
+ "episodes": 50,
946
+ "successes": 28,
947
+ "lscr": 88.66666666666667,
948
+ "source": {
949
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
950
+ "perEpisodeFile": "per_task_eval/42_trash_disposal/inv_cov/per_episode.jsonl",
951
+ "perEpisodeSha256": "efb3234f609005b95ca2a6164cd795788ca689281d51dc022004a24eef2dd790",
952
+ "summaryFile": "per_task_eval/42_trash_disposal/inv_cov/summary.json",
953
+ "summarySha256": "b7217c8a4de5420781a5053b424e49b5511426995b77c294698167283856b1ca"
954
+ }
955
+ },
956
+ "Inv.": {
957
+ "status": "reported",
958
+ "episodes": 50,
959
+ "successes": 32,
960
+ "lscr": 86.66666666666667,
961
+ "source": {
962
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
963
+ "perEpisodeFile": "per_task_eval/42_trash_disposal/inv_only/per_episode.jsonl",
964
+ "perEpisodeSha256": "2a9a1194859b8ffc9935466d7b4495ffd71224080c820ab0ebb0361b2ffe8b36",
965
+ "summaryFile": "per_task_eval/42_trash_disposal/inv_only/summary.json",
966
+ "summarySha256": "3d3ed27fcd3b82b65dd0cba38c91f1d57f4ce01ab81884a47526d5efd792a786"
967
+ }
968
+ },
969
+ "None": {
970
+ "status": "reported",
971
+ "episodes": 50,
972
+ "successes": 30,
973
+ "lscr": 87.33333333333333,
974
+ "source": {
975
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh56dfx",
976
+ "perEpisodeFile": "per_task_eval/42_trash_disposal/none/per_episode.jsonl",
977
+ "perEpisodeSha256": "cfff3f6189a5bdc1ba7e9bc40a624dd36048b7dbe4764b84f3686c7045f11453",
978
+ "summaryFile": "per_task_eval/42_trash_disposal/none/summary.json",
979
+ "summarySha256": "08ebe6b1fd74ef1758cfec6487142c1cedd0d151e597786ec037b1b322b3d3c5"
980
+ }
981
+ }
982
+ },
983
+ "mean": {
984
+ "status": "reported",
985
+ "srScope": "four-channel-mean",
986
+ "lscrScope": "four-channel-mean",
987
+ "sr": 58.5,
988
+ "lscr": 88.16666666666667,
989
+ "efficiency": 0.918423464711274,
990
+ "efficiencyStatus": "reported",
991
+ "srEpisodes": 200,
992
+ "lscrEpisodes": 200
993
+ }
994
+ },
995
+ "43": {
996
+ "channels": {
997
+ "Equi.": {
998
+ "status": "reported",
999
+ "episodes": 50,
1000
+ "successes": 17,
1001
+ "lscr": 56.99999999999999,
1002
+ "source": {
1003
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1004
+ "perEpisodeFile": "per_task_eval/43_fridge_fruit_shelf_sorting/cov_only/per_episode.jsonl",
1005
+ "perEpisodeSha256": "3181526abfe9ea3645a0f8a1d1e7a857af66887ac353654492df60257f031100",
1006
+ "summaryFile": "per_task_eval/43_fridge_fruit_shelf_sorting/cov_only/summary.json",
1007
+ "summarySha256": "f65f5f75d1971f948353961c7aab34030656371cba9406369c63a982ee427098"
1008
+ }
1009
+ },
1010
+ "Full": {
1011
+ "status": "reported",
1012
+ "episodes": 50,
1013
+ "successes": 5,
1014
+ "lscr": 41.5,
1015
+ "source": {
1016
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1017
+ "perEpisodeFile": "per_task_eval/43_fridge_fruit_shelf_sorting/inv_cov/per_episode.jsonl",
1018
+ "perEpisodeSha256": "be6bd84da21a35a3e3188c6ac9c088a3782e90de831e58668b6e8816b03a6284",
1019
+ "summaryFile": "per_task_eval/43_fridge_fruit_shelf_sorting/inv_cov/summary.json",
1020
+ "summarySha256": "186f91060a4990e2282156de9ed1312c5be316119c68af79081a99a8e0d493b3"
1021
+ }
1022
+ },
1023
+ "Inv.": {
1024
+ "status": "reported",
1025
+ "episodes": 50,
1026
+ "successes": 6,
1027
+ "lscr": 39.5,
1028
+ "source": {
1029
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1030
+ "perEpisodeFile": "per_task_eval/43_fridge_fruit_shelf_sorting/inv_only/per_episode.jsonl",
1031
+ "perEpisodeSha256": "dff42e1632b518189bf4b580544aae6da73a0858af11c9c3ec0dbcd17d4e471f",
1032
+ "summaryFile": "per_task_eval/43_fridge_fruit_shelf_sorting/inv_only/summary.json",
1033
+ "summarySha256": "b8aee49f4b697d6abe0cbc2e47299f6571abd513cc3db010b149de123c0ea2d4"
1034
+ }
1035
+ },
1036
+ "None": {
1037
+ "status": "reported",
1038
+ "episodes": 50,
1039
+ "successes": 18,
1040
+ "lscr": 60.5,
1041
+ "source": {
1042
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1043
+ "perEpisodeFile": "per_task_eval/43_fridge_fruit_shelf_sorting/none/per_episode.jsonl",
1044
+ "perEpisodeSha256": "7bb03cbd644710751bd0cbdd536fcc80a96c615b5777bd5ddfa0c4d2f222f211",
1045
+ "summaryFile": "per_task_eval/43_fridge_fruit_shelf_sorting/none/summary.json",
1046
+ "summarySha256": "dead0fa82f87516a76404a18c055622c019b91a188dcad00dedd6477b4a694a2"
1047
+ }
1048
+ }
1049
+ },
1050
+ "mean": {
1051
+ "status": "reported",
1052
+ "srScope": "four-channel-mean",
1053
+ "lscrScope": "four-channel-mean",
1054
+ "sr": 23.0,
1055
+ "lscr": 49.625,
1056
+ "efficiency": 0.9464241626203604,
1057
+ "efficiencyStatus": "reported",
1058
+ "srEpisodes": 200,
1059
+ "lscrEpisodes": 200
1060
+ }
1061
+ },
1062
+ "44": {
1063
+ "channels": {
1064
+ "Equi.": {
1065
+ "status": "reported",
1066
+ "episodes": 50,
1067
+ "successes": 48,
1068
+ "lscr": 97.5,
1069
+ "source": {
1070
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
1071
+ "perEpisodeFile": "per_task_eval/44_microwave_bowl_loading/cov_only/per_episode.jsonl",
1072
+ "perEpisodeSha256": "3269fe348b7320dee36cb2cc9d661dca23ab492d16d9bab0ca5f71a2f5ff3c26",
1073
+ "summaryFile": "per_task_eval/44_microwave_bowl_loading/cov_only/summary.json",
1074
+ "summarySha256": "ffef5b8b07facb1f0aeddd8282e8fa0b6b043670efbe673d873511be6988c719"
1075
+ }
1076
+ },
1077
+ "Full": {
1078
+ "status": "reported",
1079
+ "episodes": 50,
1080
+ "successes": 47,
1081
+ "lscr": 96.0,
1082
+ "source": {
1083
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
1084
+ "perEpisodeFile": "per_task_eval/44_microwave_bowl_loading/inv_cov/per_episode.jsonl",
1085
+ "perEpisodeSha256": "7e0a07a8be79cfaae91491a7272de94d3b3ff114354c9a2873c9ef090fcf0f08",
1086
+ "summaryFile": "per_task_eval/44_microwave_bowl_loading/inv_cov/summary.json",
1087
+ "summarySha256": "8fdfa7064059756444c85c817c5db9ff01dd4f53d308c63b1c2d953fe27bb724"
1088
+ }
1089
+ },
1090
+ "Inv.": {
1091
+ "status": "reported",
1092
+ "episodes": 50,
1093
+ "successes": 48,
1094
+ "lscr": 98.0,
1095
+ "source": {
1096
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
1097
+ "perEpisodeFile": "per_task_eval/44_microwave_bowl_loading/inv_only/per_episode.jsonl",
1098
+ "perEpisodeSha256": "ebbe6c27b40cdca31da9f84c1b8e2abe3de896a04295da07ae165965aa4a2456",
1099
+ "summaryFile": "per_task_eval/44_microwave_bowl_loading/inv_only/summary.json",
1100
+ "summarySha256": "9a344c0ce091efdfb7bd676d01ee7ae9daf6421fd7867f3f8bba3fd79bdc591f"
1101
+ }
1102
+ },
1103
+ "None": {
1104
+ "status": "reported",
1105
+ "episodes": 50,
1106
+ "successes": 48,
1107
+ "lscr": 98.0,
1108
+ "source": {
1109
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_schunk_hand",
1110
+ "perEpisodeFile": "per_task_eval/44_microwave_bowl_loading/none/per_episode.jsonl",
1111
+ "perEpisodeSha256": "dc464e365d4b0a61093bdf6dbebcb6b26587dbd341e2450abfd4dcbe03baaad4",
1112
+ "summaryFile": "per_task_eval/44_microwave_bowl_loading/none/summary.json",
1113
+ "summarySha256": "aacf4f227f0667a579f825e6cbcd2a1f0c5e89c179b49d4601cebfcb4abe9e95"
1114
+ }
1115
+ }
1116
+ },
1117
+ "mean": {
1118
+ "status": "reported",
1119
+ "srScope": "four-channel-mean",
1120
+ "lscrScope": "four-channel-mean",
1121
+ "sr": 95.5,
1122
+ "lscr": 97.375,
1123
+ "efficiency": 1.0128473289500977,
1124
+ "efficiencyStatus": "reported",
1125
+ "srEpisodes": 200,
1126
+ "lscrEpisodes": 200
1127
+ }
1128
+ },
1129
+ "51": {
1130
+ "channels": {
1131
+ "Equi.": {
1132
+ "status": "reported",
1133
+ "episodes": 50,
1134
+ "successes": 14,
1135
+ "lscr": 55.333333333333336,
1136
+ "source": {
1137
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1138
+ "perEpisodeFile": "per_task_eval/51_toilet_lid_cleaner_pour/cov_only/per_episode.jsonl",
1139
+ "perEpisodeSha256": "2d5c187f15ddf001d81279c5ef201869b4456aa47314bd3a298a623dc04fa927",
1140
+ "summaryFile": "per_task_eval/51_toilet_lid_cleaner_pour/cov_only/summary.json",
1141
+ "summarySha256": "b270a62e134e68b8d331040085f403109feb4038137e9706eff10f13b03df1cb"
1142
+ }
1143
+ },
1144
+ "Full": {
1145
+ "status": "reported",
1146
+ "episodes": 50,
1147
+ "successes": 12,
1148
+ "lscr": 44.666666666666664,
1149
+ "source": {
1150
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1151
+ "perEpisodeFile": "per_task_eval/51_toilet_lid_cleaner_pour/inv_cov/per_episode.jsonl",
1152
+ "perEpisodeSha256": "fc29958d04158f1a68fdc082ecc99087ac8bfaa183e10d378271fcbc6eda1190",
1153
+ "summaryFile": "per_task_eval/51_toilet_lid_cleaner_pour/inv_cov/summary.json",
1154
+ "summarySha256": "74cd67fe8aba65f20030302c3edc97f565bb233546cf4aded0fe21cce3cb38d6"
1155
+ }
1156
+ },
1157
+ "Inv.": {
1158
+ "status": "reported",
1159
+ "episodes": 50,
1160
+ "successes": 8,
1161
+ "lscr": 44.666666666666664,
1162
+ "source": {
1163
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1164
+ "perEpisodeFile": "per_task_eval/51_toilet_lid_cleaner_pour/inv_only/per_episode.jsonl",
1165
+ "perEpisodeSha256": "053e646acebf544c18aeadaf85f7909930939380bb166c3c43979f7581f2bbd1",
1166
+ "summaryFile": "per_task_eval/51_toilet_lid_cleaner_pour/inv_only/summary.json",
1167
+ "summarySha256": "69dce6ab0439c4257dab7bf777086aeaaa06e30accad982cc5305a7af11f1cdb"
1168
+ }
1169
+ },
1170
+ "None": {
1171
+ "status": "reported",
1172
+ "episodes": 50,
1173
+ "successes": 14,
1174
+ "lscr": 58.666666666666664,
1175
+ "source": {
1176
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1177
+ "perEpisodeFile": "per_task_eval/51_toilet_lid_cleaner_pour/none/per_episode.jsonl",
1178
+ "perEpisodeSha256": "0841964692b46f10460646032fe0a3ced5e76345c012d739a0ed2e34784c2d0e",
1179
+ "summaryFile": "per_task_eval/51_toilet_lid_cleaner_pour/none/summary.json",
1180
+ "summarySha256": "70b5fbeac88632bb40ab67956c2809ff9b5962b52a60136d3cca83335a615e1c"
1181
+ }
1182
+ }
1183
+ },
1184
+ "mean": {
1185
+ "status": "reported",
1186
+ "srScope": "four-channel-mean",
1187
+ "lscrScope": "four-channel-mean",
1188
+ "sr": 24.0,
1189
+ "lscr": 50.83333333333333,
1190
+ "efficiency": 0.9908390169094798,
1191
+ "efficiencyStatus": "reported",
1192
+ "srEpisodes": 200,
1193
+ "lscrEpisodes": 200
1194
+ }
1195
+ },
1196
+ "60": {
1197
+ "channels": {
1198
+ "Equi.": {
1199
+ "status": "reported",
1200
+ "episodes": 50,
1201
+ "successes": 29,
1202
+ "lscr": 66.0,
1203
+ "source": {
1204
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
1205
+ "perEpisodeFile": "per_task_eval/60_breadbasket_fast_food_loading/cov_only/per_episode.jsonl",
1206
+ "perEpisodeSha256": "bca012e15a414c42445b4f01e3dc0606fc62ac00ce6721b25c4f2e23ee6158ed",
1207
+ "summaryFile": "per_task_eval/60_breadbasket_fast_food_loading/cov_only/summary.json",
1208
+ "summarySha256": "f27ed8e5021b0b19bebcec478332b88ef8f6dbaa2554324260d6aac313457181"
1209
+ }
1210
+ },
1211
+ "Full": {
1212
+ "status": "reported",
1213
+ "episodes": 50,
1214
+ "successes": 32,
1215
+ "lscr": 68.5,
1216
+ "source": {
1217
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
1218
+ "perEpisodeFile": "per_task_eval/60_breadbasket_fast_food_loading/inv_cov/per_episode.jsonl",
1219
+ "perEpisodeSha256": "767a6a7c5c85df2b0fd5543550f0e6372e9b1c10a095fb25bff255c83d4de730",
1220
+ "summaryFile": "per_task_eval/60_breadbasket_fast_food_loading/inv_cov/summary.json",
1221
+ "summarySha256": "4bb0c0d204e767d12e21d07245da0f0aa854f536ca50d1507a7046384d47fc3e"
1222
+ }
1223
+ },
1224
+ "Inv.": {
1225
+ "status": "reported",
1226
+ "episodes": 50,
1227
+ "successes": 39,
1228
+ "lscr": 84.5,
1229
+ "source": {
1230
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
1231
+ "perEpisodeFile": "per_task_eval/60_breadbasket_fast_food_loading/inv_only/per_episode.jsonl",
1232
+ "perEpisodeSha256": "5a0a9afe0a2a2b0c911ce77387a1a72b2ebd3ecd3d152ecdd365b81a5cbde117",
1233
+ "summaryFile": "per_task_eval/60_breadbasket_fast_food_loading/inv_only/summary.json",
1234
+ "summarySha256": "82a17c79b3aad4fe4037891f0483b588396787042af57e96bd32dcd001b3e96b"
1235
+ }
1236
+ },
1237
+ "None": {
1238
+ "status": "reported",
1239
+ "episodes": 50,
1240
+ "successes": 32,
1241
+ "lscr": 67.5,
1242
+ "source": {
1243
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rh5dg2",
1244
+ "perEpisodeFile": "per_task_eval/60_breadbasket_fast_food_loading/none/per_episode.jsonl",
1245
+ "perEpisodeSha256": "5a867cafc4a1cce1d486bbb63bb0242bad705db979c462e22466ed622689423f",
1246
+ "summaryFile": "per_task_eval/60_breadbasket_fast_food_loading/none/summary.json",
1247
+ "summarySha256": "afa0facfbc0d7ebae62c5329cba1dcd10dddfa86fad6f9cb7212cdff25749e4b"
1248
+ }
1249
+ }
1250
+ },
1251
+ "mean": {
1252
+ "status": "reported",
1253
+ "srScope": "four-channel-mean",
1254
+ "lscrScope": "four-channel-mean",
1255
+ "sr": 66.0,
1256
+ "lscr": 71.625,
1257
+ "efficiency": 0.9715292219566166,
1258
+ "efficiencyStatus": "reported",
1259
+ "srEpisodes": 200,
1260
+ "lscrEpisodes": 200
1261
+ }
1262
+ },
1263
+ "61": {
1264
+ "channels": {
1265
+ "Equi.": {
1266
+ "status": "reported",
1267
+ "episodes": 50,
1268
+ "successes": 16,
1269
+ "lscr": 60.0,
1270
+ "source": {
1271
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
1272
+ "perEpisodeFile": "per_task_eval/61_medicine_shoebox_pack/cov_only/per_episode.jsonl",
1273
+ "perEpisodeSha256": "423b1bd3dc09674407dd2d52fbde81a807460228ebafa1b536919c8b99e8e996",
1274
+ "summaryFile": "per_task_eval/61_medicine_shoebox_pack/cov_only/summary.json",
1275
+ "summarySha256": "e2a4b020106e305bfddf06580eed9109ce0843bbee290a22ee3887f3c675b9bf"
1276
+ }
1277
+ },
1278
+ "Full": {
1279
+ "status": "reported",
1280
+ "episodes": 50,
1281
+ "successes": 28,
1282
+ "lscr": 75.33333333333333,
1283
+ "source": {
1284
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
1285
+ "perEpisodeFile": "per_task_eval/61_medicine_shoebox_pack/inv_cov/per_episode.jsonl",
1286
+ "perEpisodeSha256": "54a8126e8c08782054aa3a6b88c35ea6315cca2eb819cd5aeb56c20483d8ab8c",
1287
+ "summaryFile": "per_task_eval/61_medicine_shoebox_pack/inv_cov/summary.json",
1288
+ "summarySha256": "a30e73eafecee62a0ea62fc1eeaec0eafd6dd51f26fc8bd22d99532c77dabf69"
1289
+ }
1290
+ },
1291
+ "Inv.": {
1292
+ "status": "reported",
1293
+ "episodes": 50,
1294
+ "successes": 20,
1295
+ "lscr": 68.66666666666667,
1296
+ "source": {
1297
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
1298
+ "perEpisodeFile": "per_task_eval/61_medicine_shoebox_pack/inv_only/per_episode.jsonl",
1299
+ "perEpisodeSha256": "d07ab99192b22913df93709f65137db7cad0303c221912b4153bf6e4a42a46e2",
1300
+ "summaryFile": "per_task_eval/61_medicine_shoebox_pack/inv_only/summary.json",
1301
+ "summarySha256": "7c674141240b77bc9053687fd6092c5fbf67b3af18df1b4941e9042af45f3531"
1302
+ }
1303
+ },
1304
+ "None": {
1305
+ "status": "reported",
1306
+ "episodes": 50,
1307
+ "successes": 31,
1308
+ "lscr": 78.0,
1309
+ "source": {
1310
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_orca",
1311
+ "perEpisodeFile": "per_task_eval/61_medicine_shoebox_pack/none/per_episode.jsonl",
1312
+ "perEpisodeSha256": "95ef8132de5aaca64d5071cbaaeaaafac0a295fe8122266d41cb1c9247f0a8c9",
1313
+ "summaryFile": "per_task_eval/61_medicine_shoebox_pack/none/summary.json",
1314
+ "summarySha256": "f18f74f3e806d620c0620813ebf738fac85470ac9fcb54fa8e353bf360058d0c"
1315
+ }
1316
+ }
1317
+ },
1318
+ "mean": {
1319
+ "status": "reported",
1320
+ "srScope": "four-channel-mean",
1321
+ "lscrScope": "four-channel-mean",
1322
+ "sr": 47.5,
1323
+ "lscr": 70.5,
1324
+ "efficiency": 0.9606287271357496,
1325
+ "efficiencyStatus": "reported",
1326
+ "srEpisodes": 200,
1327
+ "lscrEpisodes": 200
1328
+ }
1329
+ },
1330
+ "62": {
1331
+ "channels": {
1332
+ "Equi.": {
1333
+ "status": "reported",
1334
+ "episodes": 50,
1335
+ "successes": 43,
1336
+ "lscr": 91.33333333333333,
1337
+ "source": {
1338
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1339
+ "perEpisodeFile": "per_task_eval/62_shoebox_accessory_pack/cov_only/per_episode.jsonl",
1340
+ "perEpisodeSha256": "3453467722f1831ff749c3b78a810ec2e44ca29b88e60750fff82d5ff9198cd5",
1341
+ "summaryFile": "per_task_eval/62_shoebox_accessory_pack/cov_only/summary.json",
1342
+ "summarySha256": "7d396986372b9502076c53bf811356d5939f332b412e4fd4a2b1b3cf081b2eed"
1343
+ }
1344
+ },
1345
+ "Full": {
1346
+ "status": "reported",
1347
+ "episodes": 50,
1348
+ "successes": 43,
1349
+ "lscr": 91.33333333333333,
1350
+ "source": {
1351
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1352
+ "perEpisodeFile": "per_task_eval/62_shoebox_accessory_pack/inv_cov/per_episode.jsonl",
1353
+ "perEpisodeSha256": "e5df1f7f3560ff3d412c97991a4a15955c59ac0be3c103f6ed63d4d99c5d5e79",
1354
+ "summaryFile": "per_task_eval/62_shoebox_accessory_pack/inv_cov/summary.json",
1355
+ "summarySha256": "06b3c7daacd9b900e798a6faf1556848414dec82a12104ed361054f48ca76dd0"
1356
+ }
1357
+ },
1358
+ "Inv.": {
1359
+ "status": "reported",
1360
+ "episodes": 50,
1361
+ "successes": 43,
1362
+ "lscr": 90.66666666666666,
1363
+ "source": {
1364
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1365
+ "perEpisodeFile": "per_task_eval/62_shoebox_accessory_pack/inv_only/per_episode.jsonl",
1366
+ "perEpisodeSha256": "047c555d3bdb639f9e32916dadbbfce122abb1f7ce70122bdbd50f55b35bc523",
1367
+ "summaryFile": "per_task_eval/62_shoebox_accessory_pack/inv_only/summary.json",
1368
+ "summarySha256": "da60737a2f09fc1981e7c01b4ba96b9349806a147b742488f839153e0c7eb799"
1369
+ }
1370
+ },
1371
+ "None": {
1372
+ "status": "reported",
1373
+ "episodes": 50,
1374
+ "successes": 45,
1375
+ "lscr": 92.0,
1376
+ "source": {
1377
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7",
1378
+ "perEpisodeFile": "per_task_eval/62_shoebox_accessory_pack/none/per_episode.jsonl",
1379
+ "perEpisodeSha256": "ad953ac83b7921193a7496ad0c2d08f8977a32e50ba9c05d19079a20329eaa79",
1380
+ "summaryFile": "per_task_eval/62_shoebox_accessory_pack/none/summary.json",
1381
+ "summarySha256": "64a7ddbc789e20683ca35996537a60d3aa853c4fc0d202fd12dd14b19e176f5d"
1382
+ }
1383
+ }
1384
+ },
1385
+ "mean": {
1386
+ "status": "reported",
1387
+ "srScope": "four-channel-mean",
1388
+ "lscrScope": "four-channel-mean",
1389
+ "sr": 87.0,
1390
+ "lscr": 91.33333333333333,
1391
+ "efficiency": 0.9812184202947876,
1392
+ "efficiencyStatus": "reported",
1393
+ "srEpisodes": 200,
1394
+ "lscrEpisodes": 200
1395
+ }
1396
+ },
1397
+ "64": {
1398
+ "channels": {
1399
+ "Equi.": {
1400
+ "status": "reported",
1401
+ "episodes": 50,
1402
+ "successes": 14,
1403
+ "lscr": 49.0,
1404
+ "source": {
1405
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_allegro",
1406
+ "perEpisodeFile": "per_task_eval/64_sports_ball_cup_sort/cov_only/per_episode.jsonl",
1407
+ "perEpisodeSha256": "5d6589292776d64f1b0bf7787fa71e6219ef75dd02520772ce413ec13b7b645a",
1408
+ "summaryFile": "per_task_eval/64_sports_ball_cup_sort/cov_only/summary.json",
1409
+ "summarySha256": "f822feaa6c6901e0441b7c9eed5437848e2fee2451152e4a47024e342c0c4369"
1410
+ }
1411
+ },
1412
+ "Full": {
1413
+ "status": "reported",
1414
+ "episodes": 50,
1415
+ "successes": 17,
1416
+ "lscr": 49.5,
1417
+ "source": {
1418
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_allegro",
1419
+ "perEpisodeFile": "per_task_eval/64_sports_ball_cup_sort/inv_cov/per_episode.jsonl",
1420
+ "perEpisodeSha256": "f5c68ca55c35c6f18357a29a4f86d1e04eabc55537b01b51ca746d1fb66a34e6",
1421
+ "summaryFile": "per_task_eval/64_sports_ball_cup_sort/inv_cov/summary.json",
1422
+ "summarySha256": "15725735ecc274f1822f1b82ed0b3c35450ba99eca95e1c563841eadbfb4ddb1"
1423
+ }
1424
+ },
1425
+ "Inv.": {
1426
+ "status": "reported",
1427
+ "episodes": 50,
1428
+ "successes": 14,
1429
+ "lscr": 48.0,
1430
+ "source": {
1431
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_allegro",
1432
+ "perEpisodeFile": "per_task_eval/64_sports_ball_cup_sort/inv_only/per_episode.jsonl",
1433
+ "perEpisodeSha256": "934efd61803f01bf32a807715e28f12703336cd781b76ede3614d6d35054e3b0",
1434
+ "summaryFile": "per_task_eval/64_sports_ball_cup_sort/inv_only/summary.json",
1435
+ "summarySha256": "28cc147c03e310f7eca4d321779949b33fe5f93fe6315fdc43189f98200af7ba"
1436
+ }
1437
+ },
1438
+ "None": {
1439
+ "status": "reported",
1440
+ "episodes": 50,
1441
+ "successes": 15,
1442
+ "lscr": 48.5,
1443
+ "source": {
1444
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_panda_with_allegro",
1445
+ "perEpisodeFile": "per_task_eval/64_sports_ball_cup_sort/none/per_episode.jsonl",
1446
+ "perEpisodeSha256": "57dcacd5b73301362a8f80ff114f93c126443f60ba6a19c6209ad8b3dc8cbfa0",
1447
+ "summaryFile": "per_task_eval/64_sports_ball_cup_sort/none/summary.json",
1448
+ "summarySha256": "3d0142b7283cd83db629c2fd2a21d8304791d8b82518ca2bbc006cd31ecf8f7f"
1449
+ }
1450
+ }
1451
+ },
1452
+ "mean": {
1453
+ "status": "reported",
1454
+ "srScope": "four-channel-mean",
1455
+ "lscrScope": "four-channel-mean",
1456
+ "sr": 30.0,
1457
+ "lscr": 48.75,
1458
+ "efficiency": 0.772247886341846,
1459
+ "efficiencyStatus": "reported",
1460
+ "srEpisodes": 200,
1461
+ "lscrEpisodes": 200
1462
+ }
1463
+ },
1464
+ "67": {
1465
+ "channels": {
1466
+ "Equi.": {
1467
+ "status": "reported",
1468
+ "episodes": 50,
1469
+ "successes": 26,
1470
+ "lscr": 65.5,
1471
+ "source": {
1472
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
1473
+ "perEpisodeFile": "per_task_eval/67_faucet_cup_water_fill/cov_only/per_episode.jsonl",
1474
+ "perEpisodeSha256": "aebde177450ec358896a6d11f84e2639246f0fa1563c63dc7fc1607adc2ce423",
1475
+ "summaryFile": "per_task_eval/67_faucet_cup_water_fill/cov_only/summary.json",
1476
+ "summarySha256": "9aca09644d997fe1a7fdef0c130d32eed718a32f95f2c7b67c0ce7c4e49a8639"
1477
+ }
1478
+ },
1479
+ "Full": {
1480
+ "status": "reported",
1481
+ "episodes": 50,
1482
+ "successes": 29,
1483
+ "lscr": 78.0,
1484
+ "source": {
1485
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
1486
+ "perEpisodeFile": "per_task_eval/67_faucet_cup_water_fill/inv_cov/per_episode.jsonl",
1487
+ "perEpisodeSha256": "bc76e24a3c9442a09d5c1c361b3cf63013309aa634fc2187748fc6a1b1e3e29a",
1488
+ "summaryFile": "per_task_eval/67_faucet_cup_water_fill/inv_cov/summary.json",
1489
+ "summarySha256": "58092bc194495b3d9d86e7ac15376fee6429e2329a1590b7d13d0738aabc0e70"
1490
+ }
1491
+ },
1492
+ "Inv.": {
1493
+ "status": "reported",
1494
+ "episodes": 50,
1495
+ "successes": 26,
1496
+ "lscr": 74.5,
1497
+ "source": {
1498
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
1499
+ "perEpisodeFile": "per_task_eval/67_faucet_cup_water_fill/inv_only/per_episode.jsonl",
1500
+ "perEpisodeSha256": "1c2a8e167764cf6e4acf1c8ad0d85a0cee348d9c821ea3c30c77d19b6bf470e8",
1501
+ "summaryFile": "per_task_eval/67_faucet_cup_water_fill/inv_only/summary.json",
1502
+ "summarySha256": "83a5e1a1d20c917a0ad4baf571d24b814d624b634a9237eb1f68e087975c272a"
1503
+ }
1504
+ },
1505
+ "None": {
1506
+ "status": "reported",
1507
+ "episodes": 50,
1508
+ "successes": 29,
1509
+ "lscr": 74.5,
1510
+ "source": {
1511
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_rm_65_with_revo2",
1512
+ "perEpisodeFile": "per_task_eval/67_faucet_cup_water_fill/none/per_episode.jsonl",
1513
+ "perEpisodeSha256": "81d94128f3c1f7988dd4ef11ecdc01a4bc6ada9f83c7c216605e7d141f1ede2f",
1514
+ "summaryFile": "per_task_eval/67_faucet_cup_water_fill/none/summary.json",
1515
+ "summarySha256": "fced09500df25dabfaa112cf55b7be19c19427cb7bde14681709a58825b8f387"
1516
+ }
1517
+ }
1518
+ },
1519
+ "mean": {
1520
+ "status": "reported",
1521
+ "srScope": "four-channel-mean",
1522
+ "lscrScope": "four-channel-mean",
1523
+ "sr": 55.0,
1524
+ "lscr": 73.125,
1525
+ "efficiency": 0.9688353692519941,
1526
+ "efficiencyStatus": "reported",
1527
+ "srEpisodes": 200,
1528
+ "lscrEpisodes": 200
1529
+ }
1530
+ },
1531
+ "73": {
1532
+ "channels": {
1533
+ "Equi.": {
1534
+ "status": "reported",
1535
+ "episodes": 50,
1536
+ "successes": 1,
1537
+ "lscr": 36.0,
1538
+ "source": {
1539
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
1540
+ "perEpisodeFile": "per_task_eval/73_jigsaw_puzzle_assembly/cov_only/per_episode.jsonl",
1541
+ "perEpisodeSha256": "07111590f588f7f5f6226648cbcd68be997c1b9b3c2dae141f0a5d14e6115415",
1542
+ "summaryFile": "per_task_eval/73_jigsaw_puzzle_assembly/cov_only/summary.json",
1543
+ "summarySha256": "a02550d1c0305659baaa49151530aa7482407e761b69265573a839db3a31e49e"
1544
+ }
1545
+ },
1546
+ "Full": {
1547
+ "status": "reported",
1548
+ "episodes": 50,
1549
+ "successes": 1,
1550
+ "lscr": 39.0,
1551
+ "source": {
1552
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
1553
+ "perEpisodeFile": "per_task_eval/73_jigsaw_puzzle_assembly/inv_cov/per_episode.jsonl",
1554
+ "perEpisodeSha256": "3419b15230dcbbc030728bd19105a15e1b28bdd1c191397a09e496dead90f271",
1555
+ "summaryFile": "per_task_eval/73_jigsaw_puzzle_assembly/inv_cov/summary.json",
1556
+ "summarySha256": "5d0a0208813ef5816b6b1bcf09a318dbd84964b895bfbdb9e73308bb0a604500"
1557
+ }
1558
+ },
1559
+ "Inv.": {
1560
+ "status": "reported",
1561
+ "episodes": 50,
1562
+ "successes": 2,
1563
+ "lscr": 50.5,
1564
+ "source": {
1565
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
1566
+ "perEpisodeFile": "per_task_eval/73_jigsaw_puzzle_assembly/inv_only/per_episode.jsonl",
1567
+ "perEpisodeSha256": "d405d05b3b14a6bd0d195bc450c660e3cf4896b0d328c61314c88df4ea775c6b",
1568
+ "summaryFile": "per_task_eval/73_jigsaw_puzzle_assembly/inv_only/summary.json",
1569
+ "summarySha256": "11ce4a1db18502b451b09f4ed36dec7a8a7815d26ffbf11f3cff206a67bb2548"
1570
+ }
1571
+ },
1572
+ "None": {
1573
+ "status": "reported",
1574
+ "episodes": 50,
1575
+ "successes": 1,
1576
+ "lscr": 62.0,
1577
+ "source": {
1578
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_iiwa7_with_sharpa",
1579
+ "perEpisodeFile": "per_task_eval/73_jigsaw_puzzle_assembly/none/per_episode.jsonl",
1580
+ "perEpisodeSha256": "188697847762a7660f7fdb398b77341c91c6180e1a9d6e1eb9107a8a6060c83c",
1581
+ "summaryFile": "per_task_eval/73_jigsaw_puzzle_assembly/none/summary.json",
1582
+ "summarySha256": "dd4b1157ffc84215c70777d6b6fd13484d4ccde236b4bb8d9446d5e87200d311"
1583
+ }
1584
+ }
1585
+ },
1586
+ "mean": {
1587
+ "status": "reported",
1588
+ "srScope": "four-channel-mean",
1589
+ "lscrScope": "four-channel-mean",
1590
+ "sr": 2.5,
1591
+ "lscr": 46.875,
1592
+ "efficiency": 0.9288613216938509,
1593
+ "efficiencyStatus": "reported",
1594
+ "srEpisodes": 200,
1595
+ "lscrEpisodes": 200
1596
+ }
1597
+ },
1598
+ "76": {
1599
+ "channels": {
1600
+ "Equi.": {
1601
+ "status": "reported",
1602
+ "episodes": 50,
1603
+ "successes": 3,
1604
+ "lscr": 35.0,
1605
+ "source": {
1606
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1607
+ "perEpisodeFile": "per_task_eval/76_soup_serving/cov_only/per_episode.jsonl",
1608
+ "perEpisodeSha256": "efc8631e3f6949337f9b797c9d7eba796ef343ec53f704a003e1ac7819987453",
1609
+ "summaryFile": "per_task_eval/76_soup_serving/cov_only/summary.json",
1610
+ "summarySha256": "1cd60c2fd6770c59a9ac17b7c90190aa4ff26364e9098876d4fdbb0669da494e"
1611
+ }
1612
+ },
1613
+ "Full": {
1614
+ "status": "reported",
1615
+ "episodes": 50,
1616
+ "successes": 2,
1617
+ "lscr": 28.499999999999996,
1618
+ "source": {
1619
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1620
+ "perEpisodeFile": "per_task_eval/76_soup_serving/inv_cov/per_episode.jsonl",
1621
+ "perEpisodeSha256": "e4c14a84716f172a0f905ac78dc2ac3759f87ce6366953d41516d85f796177bc",
1622
+ "summaryFile": "per_task_eval/76_soup_serving/inv_cov/summary.json",
1623
+ "summarySha256": "e5f0847c0d63c6c00c815dbda8aeb8196d716cbb5c35eb721c3e7072d14bbcea"
1624
+ }
1625
+ },
1626
+ "Inv.": {
1627
+ "status": "reported",
1628
+ "episodes": 50,
1629
+ "successes": 4,
1630
+ "lscr": 36.5,
1631
+ "source": {
1632
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1633
+ "perEpisodeFile": "per_task_eval/76_soup_serving/inv_only/per_episode.jsonl",
1634
+ "perEpisodeSha256": "d7b36366a571f2c54fc5a279bba2cf895da4a26d1b7818417724511403dd71f0",
1635
+ "summaryFile": "per_task_eval/76_soup_serving/inv_only/summary.json",
1636
+ "summarySha256": "90edf666561835ea4bcb701480c1651f30c5cdb9302f0b1f0efd2b36e524c8c1"
1637
+ }
1638
+ },
1639
+ "None": {
1640
+ "status": "reported",
1641
+ "episodes": 50,
1642
+ "successes": 3,
1643
+ "lscr": 28.999999999999996,
1644
+ "source": {
1645
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_shadow_hand",
1646
+ "perEpisodeFile": "per_task_eval/76_soup_serving/none/per_episode.jsonl",
1647
+ "perEpisodeSha256": "66f95c5a070248e9794e1358ae7c5b24f543c69db39a00651c18749758ed3a75",
1648
+ "summaryFile": "per_task_eval/76_soup_serving/none/summary.json",
1649
+ "summarySha256": "b327ddbab6ec3c1677a15a41e43dd9f8721b5475c5254ad9a2569a44d93df4aa"
1650
+ }
1651
+ }
1652
+ },
1653
+ "mean": {
1654
+ "status": "reported",
1655
+ "srScope": "four-channel-mean",
1656
+ "lscrScope": "four-channel-mean",
1657
+ "sr": 6.0,
1658
+ "lscr": 32.25,
1659
+ "efficiency": 0.9847280595896967,
1660
+ "efficiencyStatus": "reported",
1661
+ "srEpisodes": 200,
1662
+ "lscrEpisodes": 200
1663
+ }
1664
+ },
1665
+ "79": {
1666
+ "channels": {
1667
+ "Equi.": {
1668
+ "status": "reported",
1669
+ "episodes": 50,
1670
+ "successes": 2,
1671
+ "lscr": 7.000000000000001,
1672
+ "source": {
1673
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
1674
+ "perEpisodeFile": "per_task_eval/79_bimanual_piano_melody/cov_only/per_episode.jsonl",
1675
+ "perEpisodeSha256": "5257fa2e5bae19c4cd5f261175f46cfa058c63c2628d16499808dd63f1a941a0",
1676
+ "summaryFile": "per_task_eval/79_bimanual_piano_melody/cov_only/summary.json",
1677
+ "summarySha256": "70a7972772e32c18594d54572e10ef0dd5aec3063c4e2c99187d4574edea07d5"
1678
+ }
1679
+ },
1680
+ "Full": {
1681
+ "status": "reported",
1682
+ "episodes": 50,
1683
+ "successes": 4,
1684
+ "lscr": 15.0,
1685
+ "source": {
1686
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
1687
+ "perEpisodeFile": "per_task_eval/79_bimanual_piano_melody/inv_cov/per_episode.jsonl",
1688
+ "perEpisodeSha256": "598e811874a0c32c092706016515a5eca4eaa76670d983ef4012c24767dff06a",
1689
+ "summaryFile": "per_task_eval/79_bimanual_piano_melody/inv_cov/summary.json",
1690
+ "summarySha256": "68284bd4b93242e83b5ba4da4755484db86fb507643510b7d3c3d0249e2e8eec"
1691
+ }
1692
+ },
1693
+ "Inv.": {
1694
+ "status": "reported",
1695
+ "episodes": 50,
1696
+ "successes": 5,
1697
+ "lscr": 12.0,
1698
+ "source": {
1699
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
1700
+ "perEpisodeFile": "per_task_eval/79_bimanual_piano_melody/inv_only/per_episode.jsonl",
1701
+ "perEpisodeSha256": "64d2ad27ea87efa9ddf6f61aad02c022ad8eb750ec930223e046b6d59b37c0df",
1702
+ "summaryFile": "per_task_eval/79_bimanual_piano_melody/inv_only/summary.json",
1703
+ "summarySha256": "c926c08a439c453efe7c2844fc983a716c7fcdc6017180ada62d2d7e55369573"
1704
+ }
1705
+ },
1706
+ "None": {
1707
+ "status": "reported",
1708
+ "episodes": 50,
1709
+ "successes": 5,
1710
+ "lscr": 15.0,
1711
+ "source": {
1712
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_xarm7_with_ability",
1713
+ "perEpisodeFile": "per_task_eval/79_bimanual_piano_melody/none/per_episode.jsonl",
1714
+ "perEpisodeSha256": "d03aed23e960ee858e3e70c3ec7de141597b23767f68c5066fbc7e3ef56d792a",
1715
+ "summaryFile": "per_task_eval/79_bimanual_piano_melody/none/summary.json",
1716
+ "summarySha256": "4fc0b76179370220e8c96ac0eafc2bdb8aa10c78cee4ab73ffcc297a0686d8ea"
1717
+ }
1718
+ }
1719
+ },
1720
+ "mean": {
1721
+ "status": "reported",
1722
+ "srScope": "four-channel-mean",
1723
+ "lscrScope": "four-channel-mean",
1724
+ "sr": 8.0,
1725
+ "lscr": 12.25,
1726
+ "efficiency": 0.8285301497984944,
1727
+ "efficiencyStatus": "reported",
1728
+ "srEpisodes": 200,
1729
+ "lscrEpisodes": 200
1730
+ }
1731
+ },
1732
+ "80": {
1733
+ "channels": {
1734
+ "Equi.": {
1735
+ "status": "reported",
1736
+ "episodes": 50,
1737
+ "successes": 17,
1738
+ "lscr": 44.0,
1739
+ "source": {
1740
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_jaka_zu7_dexhand021",
1741
+ "perEpisodeFile": "per_task_eval/80_gaming_desk_setup/cov_only/per_episode.jsonl",
1742
+ "perEpisodeSha256": "33b407307d2d0918f733e3c76682ba9957676c766fd75497e195fb725a951bda",
1743
+ "summaryFile": "per_task_eval/80_gaming_desk_setup/cov_only/summary.json",
1744
+ "summarySha256": "26f154002fac61984b1750f6f8676dec084bf5f134868248f7ab019359b5742e"
1745
+ }
1746
+ },
1747
+ "Full": {
1748
+ "status": "reported",
1749
+ "episodes": 50,
1750
+ "successes": 20,
1751
+ "lscr": 44.666666666666664,
1752
+ "source": {
1753
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_jaka_zu7_dexhand021",
1754
+ "perEpisodeFile": "per_task_eval/80_gaming_desk_setup/inv_cov/per_episode.jsonl",
1755
+ "perEpisodeSha256": "779f99fa5b2f9a4ccc1293cab6064366ce16c7e9299460724e1bb84cb6884292",
1756
+ "summaryFile": "per_task_eval/80_gaming_desk_setup/inv_cov/summary.json",
1757
+ "summarySha256": "b28a8ad57966c2818c5d5eeade034d25bd1e856b06a6fbd6a2c487bff09b9e1f"
1758
+ }
1759
+ },
1760
+ "Inv.": {
1761
+ "status": "reported",
1762
+ "episodes": 50,
1763
+ "successes": 17,
1764
+ "lscr": 41.333333333333336,
1765
+ "source": {
1766
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_jaka_zu7_dexhand021",
1767
+ "perEpisodeFile": "per_task_eval/80_gaming_desk_setup/inv_only/per_episode.jsonl",
1768
+ "perEpisodeSha256": "3eb58c46d8e9ff52b81001533a1b213a152578a849d026fe7530fed3f8346254",
1769
+ "summaryFile": "per_task_eval/80_gaming_desk_setup/inv_only/summary.json",
1770
+ "summarySha256": "ff5f521cc0b6cae1924b849cf219ce86293428c3f2e63990737c2ca7ab229b88"
1771
+ }
1772
+ },
1773
+ "None": {
1774
+ "status": "reported",
1775
+ "episodes": 50,
1776
+ "successes": 27,
1777
+ "lscr": 57.99999999999999,
1778
+ "source": {
1779
+ "runId": "GigaWorld-Policy-0.6_all_1028_0600_jaka_zu7_dexhand021",
1780
+ "perEpisodeFile": "per_task_eval/80_gaming_desk_setup/none/per_episode.jsonl",
1781
+ "perEpisodeSha256": "83bea52330c68699b387c308a63cabec3078c8fe76d223e8c6ac1a10ca03aad9",
1782
+ "summaryFile": "per_task_eval/80_gaming_desk_setup/none/summary.json",
1783
+ "summarySha256": "fc40a4571fef3d94544314e9881f39c7f98dba298eed2c3eb532933b0a8d5efb"
1784
+ }
1785
+ }
1786
+ },
1787
+ "mean": {
1788
+ "status": "reported",
1789
+ "srScope": "four-channel-mean",
1790
+ "lscrScope": "four-channel-mean",
1791
+ "sr": 40.5,
1792
+ "lscr": 47.0,
1793
+ "efficiency": 1.032518663539941,
1794
+ "efficiencyStatus": "reported",
1795
+ "srEpisodes": 200,
1796
+ "lscrEpisodes": 200
1797
+ }
1798
+ }
1799
+ },
1800
+ "evaluation": {
1801
+ "schemaVersion": "1.0",
1802
+ "isTemplate": false,
1803
+ "protocol": {
1804
+ "benchmarkCommit": null,
1805
+ "taskCount": 26,
1806
+ "channels": [
1807
+ "None",
1808
+ "Equi.",
1809
+ "Inv.",
1810
+ "Full"
1811
+ ],
1812
+ "episodesPerChannel": 50,
1813
+ "metricSchemaVersion": "core-v1.8-policy-step-query-split",
1814
+ "seeds": {
1815
+ "03": {
1816
+ "Equi.": {
1817
+ "base": 100300000,
1818
+ "first": 1,
1819
+ "last": 50
1820
+ },
1821
+ "Full": {
1822
+ "base": 200300000,
1823
+ "first": 1,
1824
+ "last": 50
1825
+ },
1826
+ "Inv.": {
1827
+ "base": 100300000,
1828
+ "first": 1,
1829
+ "last": 50
1830
+ },
1831
+ "None": {
1832
+ "base": 100300000,
1833
+ "first": 1,
1834
+ "last": 50
1835
+ }
1836
+ },
1837
+ "06": {
1838
+ "Equi.": {
1839
+ "base": 100600000,
1840
+ "first": 1,
1841
+ "last": 50
1842
+ },
1843
+ "Full": {
1844
+ "base": 200600000,
1845
+ "first": 1,
1846
+ "last": 50
1847
+ },
1848
+ "Inv.": {
1849
+ "base": 100600000,
1850
+ "first": 1,
1851
+ "last": 50
1852
+ },
1853
+ "None": {
1854
+ "base": 100600000,
1855
+ "first": 1,
1856
+ "last": 50
1857
+ }
1858
+ },
1859
+ "07": {
1860
+ "Equi.": {
1861
+ "base": 100700000,
1862
+ "first": 1,
1863
+ "last": 50
1864
+ },
1865
+ "Full": {
1866
+ "base": 200700000,
1867
+ "first": 1,
1868
+ "last": 50
1869
+ },
1870
+ "Inv.": {
1871
+ "base": 100700000,
1872
+ "first": 1,
1873
+ "last": 50
1874
+ },
1875
+ "None": {
1876
+ "base": 100700000,
1877
+ "first": 1,
1878
+ "last": 50
1879
+ }
1880
+ },
1881
+ "08": {
1882
+ "Equi.": {
1883
+ "base": 100800000,
1884
+ "first": 1,
1885
+ "last": 50
1886
+ },
1887
+ "Full": {
1888
+ "base": 200800000,
1889
+ "first": 1,
1890
+ "last": 50
1891
+ },
1892
+ "Inv.": {
1893
+ "base": 100800000,
1894
+ "first": 1,
1895
+ "last": 50
1896
+ },
1897
+ "None": {
1898
+ "base": 100800000,
1899
+ "first": 1,
1900
+ "last": 50
1901
+ }
1902
+ },
1903
+ "09": {
1904
+ "Equi.": {
1905
+ "base": 100900000,
1906
+ "first": 1,
1907
+ "last": 50
1908
+ },
1909
+ "Full": {
1910
+ "base": 200900000,
1911
+ "first": 1,
1912
+ "last": 50
1913
+ },
1914
+ "Inv.": {
1915
+ "base": 100900000,
1916
+ "first": 1,
1917
+ "last": 50
1918
+ },
1919
+ "None": {
1920
+ "base": 100900000,
1921
+ "first": 1,
1922
+ "last": 50
1923
+ }
1924
+ },
1925
+ "12": {
1926
+ "Equi.": {
1927
+ "base": 101200000,
1928
+ "first": 1,
1929
+ "last": 50
1930
+ },
1931
+ "Full": {
1932
+ "base": 201200000,
1933
+ "first": 1,
1934
+ "last": 50
1935
+ },
1936
+ "Inv.": {
1937
+ "base": 101200000,
1938
+ "first": 1,
1939
+ "last": 50
1940
+ },
1941
+ "None": {
1942
+ "base": 101200000,
1943
+ "first": 1,
1944
+ "last": 50
1945
+ }
1946
+ },
1947
+ "21": {
1948
+ "Equi.": {
1949
+ "base": 102100000,
1950
+ "first": 1,
1951
+ "last": 50
1952
+ },
1953
+ "Full": {
1954
+ "base": 202100000,
1955
+ "first": 1,
1956
+ "last": 50
1957
+ },
1958
+ "Inv.": {
1959
+ "base": 102100000,
1960
+ "first": 1,
1961
+ "last": 50
1962
+ },
1963
+ "None": {
1964
+ "base": 102100000,
1965
+ "first": 1,
1966
+ "last": 50
1967
+ }
1968
+ },
1969
+ "22": {
1970
+ "Equi.": {
1971
+ "base": 102200000,
1972
+ "first": 1,
1973
+ "last": 50
1974
+ },
1975
+ "Full": {
1976
+ "base": 202200000,
1977
+ "first": 1,
1978
+ "last": 50
1979
+ },
1980
+ "Inv.": {
1981
+ "base": 102200000,
1982
+ "first": 1,
1983
+ "last": 50
1984
+ },
1985
+ "None": {
1986
+ "base": 102200000,
1987
+ "first": 1,
1988
+ "last": 50
1989
+ }
1990
+ },
1991
+ "24": {
1992
+ "Equi.": {
1993
+ "base": 102400000,
1994
+ "first": 1,
1995
+ "last": 50
1996
+ },
1997
+ "Full": {
1998
+ "base": 202400000,
1999
+ "first": 1,
2000
+ "last": 50
2001
+ },
2002
+ "Inv.": {
2003
+ "base": 102400000,
2004
+ "first": 1,
2005
+ "last": 50
2006
+ },
2007
+ "None": {
2008
+ "base": 102400000,
2009
+ "first": 1,
2010
+ "last": 50
2011
+ }
2012
+ },
2013
+ "26": {
2014
+ "Equi.": {
2015
+ "base": 102600000,
2016
+ "first": 1,
2017
+ "last": 50
2018
+ },
2019
+ "Full": {
2020
+ "base": 202600000,
2021
+ "first": 1,
2022
+ "last": 50
2023
+ },
2024
+ "Inv.": {
2025
+ "base": 102600000,
2026
+ "first": 1,
2027
+ "last": 50
2028
+ },
2029
+ "None": {
2030
+ "base": 102600000,
2031
+ "first": 1,
2032
+ "last": 50
2033
+ }
2034
+ },
2035
+ "27": {
2036
+ "Equi.": {
2037
+ "base": 102700000,
2038
+ "first": 1,
2039
+ "last": 50
2040
+ },
2041
+ "Full": {
2042
+ "base": 202700000,
2043
+ "first": 1,
2044
+ "last": 50
2045
+ },
2046
+ "Inv.": {
2047
+ "base": 102700000,
2048
+ "first": 1,
2049
+ "last": 50
2050
+ },
2051
+ "None": {
2052
+ "base": 102700000,
2053
+ "first": 1,
2054
+ "last": 50
2055
+ }
2056
+ },
2057
+ "32": {
2058
+ "Equi.": {
2059
+ "base": 103200000,
2060
+ "first": 1,
2061
+ "last": 50
2062
+ },
2063
+ "Full": {
2064
+ "base": 203200000,
2065
+ "first": 1,
2066
+ "last": 50
2067
+ },
2068
+ "Inv.": {
2069
+ "base": 103200000,
2070
+ "first": 1,
2071
+ "last": 50
2072
+ },
2073
+ "None": {
2074
+ "base": 103200000,
2075
+ "first": 1,
2076
+ "last": 50
2077
+ }
2078
+ },
2079
+ "34": {
2080
+ "Equi.": {
2081
+ "base": 103400000,
2082
+ "first": 1,
2083
+ "last": 50
2084
+ },
2085
+ "Full": {
2086
+ "base": 203400000,
2087
+ "first": 1,
2088
+ "last": 50
2089
+ },
2090
+ "Inv.": {
2091
+ "base": 103400000,
2092
+ "first": 1,
2093
+ "last": 50
2094
+ },
2095
+ "None": {
2096
+ "base": 103400000,
2097
+ "first": 1,
2098
+ "last": 50
2099
+ }
2100
+ },
2101
+ "42": {
2102
+ "Equi.": {
2103
+ "base": 104200000,
2104
+ "first": 1,
2105
+ "last": 50
2106
+ },
2107
+ "Full": {
2108
+ "base": 204200000,
2109
+ "first": 1,
2110
+ "last": 50
2111
+ },
2112
+ "Inv.": {
2113
+ "base": 104200000,
2114
+ "first": 1,
2115
+ "last": 50
2116
+ },
2117
+ "None": {
2118
+ "base": 104200000,
2119
+ "first": 1,
2120
+ "last": 50
2121
+ }
2122
+ },
2123
+ "43": {
2124
+ "Equi.": {
2125
+ "base": 104300000,
2126
+ "first": 1,
2127
+ "last": 50
2128
+ },
2129
+ "Full": {
2130
+ "base": 204300000,
2131
+ "first": 1,
2132
+ "last": 50
2133
+ },
2134
+ "Inv.": {
2135
+ "base": 104300000,
2136
+ "first": 1,
2137
+ "last": 50
2138
+ },
2139
+ "None": {
2140
+ "base": 104300000,
2141
+ "first": 1,
2142
+ "last": 50
2143
+ }
2144
+ },
2145
+ "44": {
2146
+ "Equi.": {
2147
+ "base": 104400000,
2148
+ "first": 1,
2149
+ "last": 50
2150
+ },
2151
+ "Full": {
2152
+ "base": 204400000,
2153
+ "first": 1,
2154
+ "last": 50
2155
+ },
2156
+ "Inv.": {
2157
+ "base": 104400000,
2158
+ "first": 1,
2159
+ "last": 50
2160
+ },
2161
+ "None": {
2162
+ "base": 104400000,
2163
+ "first": 1,
2164
+ "last": 50
2165
+ }
2166
+ },
2167
+ "51": {
2168
+ "Equi.": {
2169
+ "base": 105100000,
2170
+ "first": 1,
2171
+ "last": 50
2172
+ },
2173
+ "Full": {
2174
+ "base": 205100000,
2175
+ "first": 1,
2176
+ "last": 50
2177
+ },
2178
+ "Inv.": {
2179
+ "base": 105100000,
2180
+ "first": 1,
2181
+ "last": 50
2182
+ },
2183
+ "None": {
2184
+ "base": 105100000,
2185
+ "first": 1,
2186
+ "last": 50
2187
+ }
2188
+ },
2189
+ "60": {
2190
+ "Equi.": {
2191
+ "base": 106000000,
2192
+ "first": 1,
2193
+ "last": 50
2194
+ },
2195
+ "Full": {
2196
+ "base": 206000000,
2197
+ "first": 1,
2198
+ "last": 50
2199
+ },
2200
+ "Inv.": {
2201
+ "base": 106000000,
2202
+ "first": 1,
2203
+ "last": 50
2204
+ },
2205
+ "None": {
2206
+ "base": 106000000,
2207
+ "first": 1,
2208
+ "last": 50
2209
+ }
2210
+ },
2211
+ "61": {
2212
+ "Equi.": {
2213
+ "base": 106100000,
2214
+ "first": 1,
2215
+ "last": 50
2216
+ },
2217
+ "Full": {
2218
+ "base": 206100000,
2219
+ "first": 1,
2220
+ "last": 50
2221
+ },
2222
+ "Inv.": {
2223
+ "base": 106100000,
2224
+ "first": 1,
2225
+ "last": 50
2226
+ },
2227
+ "None": {
2228
+ "base": 106100000,
2229
+ "first": 1,
2230
+ "last": 50
2231
+ }
2232
+ },
2233
+ "62": {
2234
+ "Equi.": {
2235
+ "base": 106200000,
2236
+ "first": 1,
2237
+ "last": 50
2238
+ },
2239
+ "Full": {
2240
+ "base": 206200000,
2241
+ "first": 1,
2242
+ "last": 50
2243
+ },
2244
+ "Inv.": {
2245
+ "base": 106200000,
2246
+ "first": 1,
2247
+ "last": 50
2248
+ },
2249
+ "None": {
2250
+ "base": 106200000,
2251
+ "first": 1,
2252
+ "last": 50
2253
+ }
2254
+ },
2255
+ "64": {
2256
+ "Equi.": {
2257
+ "base": 106400000,
2258
+ "first": 1,
2259
+ "last": 50
2260
+ },
2261
+ "Full": {
2262
+ "base": 206400000,
2263
+ "first": 1,
2264
+ "last": 50
2265
+ },
2266
+ "Inv.": {
2267
+ "base": 106400000,
2268
+ "first": 1,
2269
+ "last": 50
2270
+ },
2271
+ "None": {
2272
+ "base": 106400000,
2273
+ "first": 1,
2274
+ "last": 50
2275
+ }
2276
+ },
2277
+ "67": {
2278
+ "Equi.": {
2279
+ "base": 106700000,
2280
+ "first": 1,
2281
+ "last": 50
2282
+ },
2283
+ "Full": {
2284
+ "base": 206700000,
2285
+ "first": 1,
2286
+ "last": 50
2287
+ },
2288
+ "Inv.": {
2289
+ "base": 106700000,
2290
+ "first": 1,
2291
+ "last": 50
2292
+ },
2293
+ "None": {
2294
+ "base": 106700000,
2295
+ "first": 1,
2296
+ "last": 50
2297
+ }
2298
+ },
2299
+ "73": {
2300
+ "Equi.": {
2301
+ "base": 107300000,
2302
+ "first": 1,
2303
+ "last": 50
2304
+ },
2305
+ "Full": {
2306
+ "base": 207300000,
2307
+ "first": 1,
2308
+ "last": 50
2309
+ },
2310
+ "Inv.": {
2311
+ "base": 107300000,
2312
+ "first": 1,
2313
+ "last": 50
2314
+ },
2315
+ "None": {
2316
+ "base": 107300000,
2317
+ "first": 1,
2318
+ "last": 50
2319
+ }
2320
+ },
2321
+ "76": {
2322
+ "Equi.": {
2323
+ "base": 107600000,
2324
+ "first": 1,
2325
+ "last": 50
2326
+ },
2327
+ "Full": {
2328
+ "base": 207600000,
2329
+ "first": 1,
2330
+ "last": 50
2331
+ },
2332
+ "Inv.": {
2333
+ "base": 107600000,
2334
+ "first": 1,
2335
+ "last": 50
2336
+ },
2337
+ "None": {
2338
+ "base": 107600000,
2339
+ "first": 1,
2340
+ "last": 50
2341
+ }
2342
+ },
2343
+ "79": {
2344
+ "Equi.": {
2345
+ "base": 107900000,
2346
+ "first": 1,
2347
+ "last": 50
2348
+ },
2349
+ "Full": {
2350
+ "base": 207900000,
2351
+ "first": 1,
2352
+ "last": 50
2353
+ },
2354
+ "Inv.": {
2355
+ "base": 107900000,
2356
+ "first": 1,
2357
+ "last": 50
2358
+ },
2359
+ "None": {
2360
+ "base": 107900000,
2361
+ "first": 1,
2362
+ "last": 50
2363
+ }
2364
+ },
2365
+ "80": {
2366
+ "Equi.": {
2367
+ "base": 108000000,
2368
+ "first": 1,
2369
+ "last": 50
2370
+ },
2371
+ "Full": {
2372
+ "base": 208000000,
2373
+ "first": 1,
2374
+ "last": 50
2375
+ },
2376
+ "Inv.": {
2377
+ "base": 108000000,
2378
+ "first": 1,
2379
+ "last": 50
2380
+ },
2381
+ "None": {
2382
+ "base": 108000000,
2383
+ "first": 1,
2384
+ "last": 50
2385
+ }
2386
+ }
2387
+ },
2388
+ "generalizationSplit": "unseen",
2389
+ "stepUnit": "physics_steps",
2390
+ "physicsHz": 60,
2391
+ "controlHz": 20,
2392
+ "physicsStepsPerPolicyAction": 3,
2393
+ "executionStepBudget": {
2394
+ "03": {
2395
+ "Equi.": 3246,
2396
+ "Full": 3246,
2397
+ "Inv.": 3246,
2398
+ "None": 3246
2399
+ },
2400
+ "06": {
2401
+ "Equi.": 1920,
2402
+ "Full": 1920,
2403
+ "Inv.": 1920,
2404
+ "None": 1920
2405
+ },
2406
+ "07": {
2407
+ "Equi.": 1806,
2408
+ "Full": 1806,
2409
+ "Inv.": 1806,
2410
+ "None": 1806
2411
+ },
2412
+ "08": {
2413
+ "Equi.": 3432,
2414
+ "Full": 3432,
2415
+ "Inv.": 3432,
2416
+ "None": 3432
2417
+ },
2418
+ "09": {
2419
+ "Equi.": 1926,
2420
+ "Full": 1926,
2421
+ "Inv.": 1926,
2422
+ "None": 1926
2423
+ },
2424
+ "12": {
2425
+ "Equi.": 3447,
2426
+ "Full": 3447,
2427
+ "Inv.": 3447,
2428
+ "None": 3447
2429
+ },
2430
+ "21": {
2431
+ "Equi.": 2481,
2432
+ "Full": 2481,
2433
+ "Inv.": 2481,
2434
+ "None": 2481
2435
+ },
2436
+ "22": {
2437
+ "Equi.": 2043,
2438
+ "Full": 2043,
2439
+ "Inv.": 2043,
2440
+ "None": 2043
2441
+ },
2442
+ "24": {
2443
+ "Equi.": 4626,
2444
+ "Full": 4626,
2445
+ "Inv.": 4626,
2446
+ "None": 4626
2447
+ },
2448
+ "26": {
2449
+ "Equi.": 2613,
2450
+ "Full": 2613,
2451
+ "Inv.": 2613,
2452
+ "None": 2613
2453
+ },
2454
+ "27": {
2455
+ "Equi.": 3570,
2456
+ "Full": 3570,
2457
+ "Inv.": 3570,
2458
+ "None": 3570
2459
+ },
2460
+ "32": {
2461
+ "Equi.": 2556,
2462
+ "Full": 2556,
2463
+ "Inv.": 2556,
2464
+ "None": 2556
2465
+ },
2466
+ "34": {
2467
+ "Equi.": 3240,
2468
+ "Full": 3240,
2469
+ "Inv.": 3240,
2470
+ "None": 3240
2471
+ },
2472
+ "42": {
2473
+ "Equi.": 2004,
2474
+ "Full": 2004,
2475
+ "Inv.": 2004,
2476
+ "None": 2004
2477
+ },
2478
+ "43": {
2479
+ "Equi.": 4125,
2480
+ "Full": 4125,
2481
+ "Inv.": 4125,
2482
+ "None": 4125
2483
+ },
2484
+ "44": {
2485
+ "Equi.": 3183,
2486
+ "Full": 3183,
2487
+ "Inv.": 3183,
2488
+ "None": 3183
2489
+ },
2490
+ "51": {
2491
+ "Equi.": 3069,
2492
+ "Full": 3069,
2493
+ "Inv.": 3069,
2494
+ "None": 3069
2495
+ },
2496
+ "60": {
2497
+ "Equi.": 2985,
2498
+ "Full": 2985,
2499
+ "Inv.": 2985,
2500
+ "None": 2985
2501
+ },
2502
+ "61": {
2503
+ "Equi.": 3102,
2504
+ "Full": 3102,
2505
+ "Inv.": 3102,
2506
+ "None": 3102
2507
+ },
2508
+ "62": {
2509
+ "Equi.": 3231,
2510
+ "Full": 3231,
2511
+ "Inv.": 3231,
2512
+ "None": 3231
2513
+ },
2514
+ "64": {
2515
+ "Equi.": 1995,
2516
+ "Full": 1995,
2517
+ "Inv.": 1995,
2518
+ "None": 1995
2519
+ },
2520
+ "67": {
2521
+ "Equi.": 2892,
2522
+ "Full": 2892,
2523
+ "Inv.": 2892,
2524
+ "None": 2892
2525
+ },
2526
+ "73": {
2527
+ "Equi.": 3639,
2528
+ "Full": 3639,
2529
+ "Inv.": 3639,
2530
+ "None": 3639
2531
+ },
2532
+ "76": {
2533
+ "Equi.": 2958,
2534
+ "Full": 2958,
2535
+ "Inv.": 2958,
2536
+ "None": 2958
2537
+ },
2538
+ "79": {
2539
+ "Equi.": 2043,
2540
+ "Full": 2043,
2541
+ "Inv.": 2043,
2542
+ "None": 2043
2543
+ },
2544
+ "80": {
2545
+ "Equi.": 1779,
2546
+ "Full": 1779,
2547
+ "Inv.": 1779,
2548
+ "None": 1779
2549
+ }
2550
+ },
2551
+ "successDwellTimeSByTask": {
2552
+ "03": 0.5,
2553
+ "06": 0.5,
2554
+ "07": 0.5,
2555
+ "08": 0.5,
2556
+ "09": 0.5,
2557
+ "12": 0.5,
2558
+ "21": 0.5,
2559
+ "22": 0.5,
2560
+ "24": 0.5,
2561
+ "26": 0.5,
2562
+ "27": 0.5,
2563
+ "32": 0.5,
2564
+ "34": 0.5,
2565
+ "42": 0.5,
2566
+ "43": 0.5,
2567
+ "44": 0.5,
2568
+ "51": 0.5,
2569
+ "60": 0.5,
2570
+ "61": 0.5,
2571
+ "62": 0.5,
2572
+ "64": 0.5,
2573
+ "67": 3.0,
2574
+ "73": 0.5,
2575
+ "76": 0.5,
2576
+ "79": 0.5,
2577
+ "80": 0.5
2578
+ },
2579
+ "successProtocol": "reach_and_stop",
2580
+ "predictionHorizon": 32,
2581
+ "executionHorizon": 32,
2582
+ "denoiseSteps": 10,
2583
+ "flowSteps": 5,
2584
+ "cfg": 1,
2585
+ "expertReference": "Native per-episode expert_time_step, in physics steps.",
2586
+ "protocolChanges": "Custom GWP model integration and Drive v5 IK/execution adapter; robot_base EEF80 actions; prediction/execution horizon 32. Native physics budgets, seeds and stable-success dwell recorded in logs. All 60 scored IK failures remain in denominators; no scored episodes replaced. Native shards were merged into 50-episode task/channel outputs for the original submission export. Exact historical worker benchmark commits are not asserted."
2587
+ },
2588
+ "coverage": {
2589
+ "evaluatedTaskCount": 26,
2590
+ "evaluatedEpisodeCount": 5200,
2591
+ "completedTaskCount": {
2592
+ "None": 26,
2593
+ "Equi.": 26,
2594
+ "Inv.": 26,
2595
+ "Full": 26,
2596
+ "mean": 26
2597
+ },
2598
+ "expectedFullEpisodeCount": 1300,
2599
+ "expectedFourChannelEpisodeCount": 5200,
2600
+ "missingTaskChannels": {}
2601
+ },
2602
+ "aggregation": {
2603
+ "units": {
2604
+ "sr": "percent",
2605
+ "lscr": "percent",
2606
+ "efficiency": "ratio",
2607
+ "steps": "physics_steps"
2608
+ },
2609
+ "sr": "100 * successes / episodes per task/channel; equal channel mean per task; equal task mean over all 26 tasks.",
2610
+ "lscr": "100 * mean native latched_stage_completion_rate including failures; equal channel and task means.",
2611
+ "efficiency": "Per task, expert_time_step divided by mean steps_to_stable_success pooled over successful episodes from all four channels. Overall is the mean of valid task efficiencies.",
2612
+ "errors": "All 60 scored IK errors retained; 5200 episodes, 2671 successes."
2613
+ },
2614
+ "logs": {
2615
+ "repository": "Bench2Dex/Leaderboard",
2616
+ "policyPath": "gigaworld-policy-06",
2617
+ "revision": null,
2618
+ "status": "pending-review"
2619
+ },
2620
+ "checkpoint": {
2621
+ "epoch": 5,
2622
+ "step": 12870,
2623
+ "sha256": "afed58631710411829692e0f9d5ce618a5f2866c2d239482c7315eb1113be56b"
2624
+ }
2625
+ }
2626
+ }
gigaworld-policy-06/logo.png ADDED

Git LFS Details

  • SHA256: fed7675b19c1eee5e639cba1c4dc5113bf43d7c4ec16b2308f0869692e5d3759
  • Pointer size: 130 Bytes
  • Size of remote file: 31.8 kB
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/cov_only/summary.json ADDED
@@ -0,0 +1,224 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100300000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 30,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3246,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.6,
50
+ "success_rate_ci": {
51
+ "mean": 0.6,
52
+ "ci_low": 0.46,
53
+ "ci_high": 0.74
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.8866666666666666
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 33.581666666666045
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.5,
64
+ "hard_violation_rate": 0.48,
65
+ "drop_rate": 0.48,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.00026322716504343247
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.66,
74
+ "at_end_success_rate": 0.6,
75
+ "task_family_success_rate": {
76
+ "table_setting": 0.6
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.7933333333333333,
81
+ "current_chain_depth_progress_score": 0.7933333333333333,
82
+ "latched_chain_depth_progress_score": 0.8866666666666666,
83
+ "avg_current_chain_depth": 2.38,
84
+ "avg_latched_chain_depth": 2.66
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.081580259627804,
88
+ "expert_time_s": 36.06666666666666,
89
+ "expert_time_step": 2164,
90
+ "avg_steps_to_success": 2014.9,
91
+ "avg_policy_steps_to_success": 672.0333333333333,
92
+ "avg_policy_queries_to_success": 672.0333333333333
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00026322716504343247,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_194_wineglass_1": 8,
98
+ "drop_to_table_obj_194_wineglass_2": 9,
99
+ "drop_to_table_obj_194_wineglass_3": 16
100
+ },
101
+ "robot_constraint_violation_rate": 0.3,
102
+ "finger_joint_limit_saturation_rate": 0.02,
103
+ "joint_limit_active_step_rate": 0.0034299297263235143,
104
+ "finger_joint_limit_saturation_step_rate": 0.0020340280935174327,
105
+ "max_joint_limit_excess_rad": 0.6842067241668701,
106
+ "robot_constraint_diagnostics": {
107
+ "joint_limit_soft_counts": {
108
+ "index_q1": 261,
109
+ "index_q2": 3,
110
+ "middle_q1": 86,
111
+ "middle_q2": 68,
112
+ "multi_ring_q2": 20,
113
+ "pinky_q1": 80,
114
+ "pinky_q2": 49,
115
+ "ring_q1": 47,
116
+ "ring_q2": 196,
117
+ "thumb_q2": 229
118
+ },
119
+ "joint_limit_hard_counts": {
120
+ "index_q1": 4,
121
+ "index_q2": 7,
122
+ "middle_q1": 1,
123
+ "middle_q2": 13,
124
+ "multi_pinky_q1": 4,
125
+ "multi_ring_q2": 1,
126
+ "pinky_q1": 12,
127
+ "pinky_q2": 16,
128
+ "ring_q1": 4,
129
+ "ring_q2": 28,
130
+ "thumb_q2": 4
131
+ },
132
+ "joint_limit_hard_episode_events": 32,
133
+ "finger_joint_limit_saturation_counts": {
134
+ "index_q1": 194,
135
+ "middle_q1": 70,
136
+ "middle_q2": 1,
137
+ "ring_q1": 4,
138
+ "ring_q2": 73,
139
+ "thumb_q2": 197
140
+ },
141
+ "joint_limit_max_excess_by_joint": {
142
+ "index_q1": 0.6036486625671387,
143
+ "index_q2": 0.03606659173965454,
144
+ "middle_q1": 0.3010816276073456,
145
+ "middle_q2": 0.14975129067897797,
146
+ "multi_index_q1": 0.0015880088321864605,
147
+ "multi_index_q2": 1.1394365451167232e-08,
148
+ "multi_pinky_q1": 0.007418627385050058,
149
+ "multi_pinky_q2": 0.0006152152782306075,
150
+ "multi_ring_q1": 0.0033840443938970566,
151
+ "multi_ring_q2": 0.10467394441366196,
152
+ "pinky_q1": 0.1348232477903366,
153
+ "pinky_q2": 0.16952918469905853,
154
+ "ring_q1": 0.2046889066696167,
155
+ "ring_q2": 0.47269898653030396,
156
+ "thumb_q2": 0.6842067241668701
157
+ },
158
+ "coverage": {
159
+ "hard_limit_source": {
160
+ "available": 50,
161
+ "unavailable": 0,
162
+ "unknown": 0
163
+ },
164
+ "joint_role_manifest": {
165
+ "available": 50,
166
+ "unavailable": 0,
167
+ "unknown": 0
168
+ }
169
+ },
170
+ "hard_limit_sources": {
171
+ "runtime.joint_pos_limits": 50
172
+ },
173
+ "soft_limit_sources": {
174
+ "runtime.soft_joint_pos_limits": 50
175
+ },
176
+ "unknown_runtime_joint_names": [],
177
+ "baseline_checks": {
178
+ "post_home": {
179
+ "checked": 50,
180
+ "hard_violation_episodes": 0
181
+ },
182
+ "post_reset": {
183
+ "checked": 50,
184
+ "hard_violation_episodes": 0
185
+ }
186
+ }
187
+ }
188
+ },
189
+ "robot_motion": {
190
+ "joint_acceleration_rms": 11.988408964423835,
191
+ "joint_effort_rms": 107.05907427330409,
192
+ "joint_jerk_rms": 1190.1634973192324,
193
+ "joint_velocity_rms": 0.8839222127994585
194
+ },
195
+ "grasp": {
196
+ "kinematic_grasp_stability_index": 0.8729366078696843,
197
+ "mean_kinematic_grasp_stability": 0.5707345486036994,
198
+ "per_object_best_kinematic_grasp_stability": {
199
+ "obj_194_wineglass_1": 0.8034512635366662,
200
+ "obj_194_wineglass_2": 0.7957597441361669,
201
+ "obj_194_wineglass_3": 0.7141078225852985
202
+ },
203
+ "grasp_gsi_diagnostics": {
204
+ "mean_anti_slip_score": 0.9260494115837264,
205
+ "mean_duration_score": 0.53228202459471,
206
+ "mean_hold_score": 0.16642107840672848,
207
+ "mean_kinematic_stability_score": 0.09384709577451433,
208
+ "mean_legacy_gsi": 0.38632284733774647,
209
+ "mean_lift_score": 0.18967222294600977,
210
+ "mean_motion_stability_score": 0.14265571229471205
211
+ }
212
+ },
213
+ "tool": {
214
+ "tool_selection_accuracy": null,
215
+ "tool_switch_success_rate": null,
216
+ "tool_switch_total": 0,
217
+ "tool_switch_successful": 0
218
+ },
219
+ "runtime": {
220
+ "avg_steps": 2507.34,
221
+ "error_count": 0
222
+ }
223
+ }
224
+ }
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_cov/summary.json ADDED
@@ -0,0 +1,223 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 200300000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 27,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3246,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.54,
50
+ "success_rate_ci": {
51
+ "mean": 0.54,
52
+ "ci_low": 0.4,
53
+ "ci_high": 0.68
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.8933333333333333
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 33.21975308641915
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.42,
64
+ "hard_violation_rate": 0.54,
65
+ "drop_rate": 0.54,
66
+ "high_speed_violation_rate": 0.04,
67
+ "safety_violation_step_rate": 0.003572707318212245
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.56,
74
+ "at_end_success_rate": 0.54,
75
+ "task_family_success_rate": {
76
+ "table_setting": 0.54
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.8133333333333334,
81
+ "current_chain_depth_progress_score": 0.8133333333333334,
82
+ "latched_chain_depth_progress_score": 0.8933333333333333,
83
+ "avg_current_chain_depth": 2.44,
84
+ "avg_latched_chain_depth": 2.68
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.0866195836572154,
88
+ "expert_time_s": 36.06666666666666,
89
+ "expert_time_step": 2164,
90
+ "avg_steps_to_success": 1993.1851851851852,
91
+ "avg_policy_steps_to_success": 664.8148148148148,
92
+ "avg_policy_queries_to_success": 664.8148148148148
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00036583277550321467,
96
+ "violation_counts": {
97
+ "drop_obj_194_wineglass_3": 1,
98
+ "drop_to_table_obj_194_wineglass_1": 18,
99
+ "drop_to_table_obj_194_wineglass_2": 12,
100
+ "drop_to_table_obj_194_wineglass_3": 13,
101
+ "high_speed_obj_029_plate_1": 1,
102
+ "high_speed_obj_194_wineglass_3": 2
103
+ },
104
+ "robot_constraint_violation_rate": 0.36,
105
+ "finger_joint_limit_saturation_rate": 0.06,
106
+ "joint_limit_active_step_rate": 0.0024129395831063096,
107
+ "finger_joint_limit_saturation_step_rate": 0.0003269143951305322,
108
+ "max_joint_limit_excess_rad": 0.3934663236141205,
109
+ "robot_constraint_diagnostics": {
110
+ "joint_limit_soft_counts": {
111
+ "index_q1": 100,
112
+ "index_q2": 2,
113
+ "middle_q1": 90,
114
+ "multi_ring_q2": 26,
115
+ "pinky_q1": 41,
116
+ "pinky_q2": 12,
117
+ "ring_q1": 129,
118
+ "ring_q2": 30,
119
+ "thumb_q2": 67
120
+ },
121
+ "joint_limit_hard_counts": {
122
+ "index_q1": 4,
123
+ "index_q2": 1,
124
+ "middle_q1": 23,
125
+ "middle_q2": 7,
126
+ "multi_ring_q2": 1,
127
+ "pinky_q1": 27,
128
+ "pinky_q2": 4,
129
+ "ring_q1": 38,
130
+ "ring_q2": 26,
131
+ "thumb_q2": 3
132
+ },
133
+ "joint_limit_hard_episode_events": 68,
134
+ "finger_joint_limit_saturation_counts": {
135
+ "multi_ring_q2": 6,
136
+ "ring_q1": 8,
137
+ "thumb_q2": 28
138
+ },
139
+ "joint_limit_max_excess_by_joint": {
140
+ "index_q1": 0.14101532101631165,
141
+ "index_q2": 0.019567573443055153,
142
+ "middle_q1": 0.09257026761770248,
143
+ "middle_q2": 0.009539523161947727,
144
+ "multi_index_q1": 0.0028177194762974977,
145
+ "multi_index_q2": 8.882607538396314e-09,
146
+ "multi_pinky_q1": 0.003529840847477317,
147
+ "multi_pinky_q2": 0.0010383010376244783,
148
+ "multi_ring_q1": 0.004025915637612343,
149
+ "multi_ring_q2": 0.22264625132083893,
150
+ "pinky_q1": 0.056955430656671524,
151
+ "pinky_q2": 0.021508364006876945,
152
+ "ring_q1": 0.13023464381694794,
153
+ "ring_q2": 0.12480971962213516,
154
+ "thumb_q2": 0.3934663236141205,
155
+ "xarm_l_joint2": 1.1444091796875e-05
156
+ },
157
+ "coverage": {
158
+ "hard_limit_source": {
159
+ "available": 50,
160
+ "unavailable": 0,
161
+ "unknown": 0
162
+ },
163
+ "joint_role_manifest": {
164
+ "available": 50,
165
+ "unavailable": 0,
166
+ "unknown": 0
167
+ }
168
+ },
169
+ "hard_limit_sources": {
170
+ "runtime.joint_pos_limits": 50
171
+ },
172
+ "soft_limit_sources": {
173
+ "runtime.soft_joint_pos_limits": 50
174
+ },
175
+ "unknown_runtime_joint_names": [],
176
+ "baseline_checks": {
177
+ "post_home": {
178
+ "checked": 50,
179
+ "hard_violation_episodes": 0
180
+ },
181
+ "post_reset": {
182
+ "checked": 50,
183
+ "hard_violation_episodes": 0
184
+ }
185
+ }
186
+ }
187
+ },
188
+ "robot_motion": {
189
+ "joint_acceleration_rms": 11.063890979274499,
190
+ "joint_effort_rms": 106.55893736221267,
191
+ "joint_jerk_rms": 1091.406658126809,
192
+ "joint_velocity_rms": 0.8755614417444691
193
+ },
194
+ "grasp": {
195
+ "kinematic_grasp_stability_index": 0.8851189470624374,
196
+ "mean_kinematic_grasp_stability": 0.6127045533015832,
197
+ "per_object_best_kinematic_grasp_stability": {
198
+ "obj_194_wineglass_1": 0.8005190457570085,
199
+ "obj_194_wineglass_2": 0.8235188840956895,
200
+ "obj_194_wineglass_3": 0.6892694686722516
201
+ },
202
+ "grasp_gsi_diagnostics": {
203
+ "mean_anti_slip_score": 0.9105631498614818,
204
+ "mean_duration_score": 0.5515618717316947,
205
+ "mean_hold_score": 0.18985781315255643,
206
+ "mean_kinematic_stability_score": 0.11314032774507539,
207
+ "mean_legacy_gsi": 0.4003596182332944,
208
+ "mean_lift_score": 0.21146222477206225,
209
+ "mean_motion_stability_score": 0.16499641775232607
210
+ }
211
+ },
212
+ "tool": {
213
+ "tool_selection_accuracy": null,
214
+ "tool_switch_success_rate": null,
215
+ "tool_switch_total": 0,
216
+ "tool_switch_successful": 0
217
+ },
218
+ "runtime": {
219
+ "avg_steps": 2569.48,
220
+ "error_count": 0
221
+ }
222
+ }
223
+ }
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/inv_only/summary.json ADDED
@@ -0,0 +1,224 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100300000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 31,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3246,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.62,
50
+ "success_rate_ci": {
51
+ "mean": 0.62,
52
+ "ci_low": 0.48,
53
+ "ci_high": 0.76
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.9533333333333334
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 33.060215053762846
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.54,
64
+ "hard_violation_rate": 0.44,
65
+ "drop_rate": 0.44,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.0003004075800139649
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.68,
74
+ "at_end_success_rate": 0.62,
75
+ "task_family_success_rate": {
76
+ "table_setting": 0.62
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.8733333333333333,
81
+ "current_chain_depth_progress_score": 0.8733333333333333,
82
+ "latched_chain_depth_progress_score": 0.9533333333333334,
83
+ "avg_current_chain_depth": 2.62,
84
+ "avg_latched_chain_depth": 2.86
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.0950744704899888,
88
+ "expert_time_s": 36.06666666666666,
89
+ "expert_time_step": 2164,
90
+ "avg_steps_to_success": 1983.6129032258063,
91
+ "avg_policy_steps_to_success": 661.516129032258,
92
+ "avg_policy_queries_to_success": 661.516129032258
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0003004075800139649,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_194_wineglass_1": 10,
98
+ "drop_to_table_obj_194_wineglass_2": 12,
99
+ "drop_to_table_obj_194_wineglass_3": 15
100
+ },
101
+ "robot_constraint_violation_rate": 0.3,
102
+ "finger_joint_limit_saturation_rate": 0.1,
103
+ "joint_limit_active_step_rate": 0.0039458941591023495,
104
+ "finger_joint_limit_saturation_step_rate": 0.0012422259389766655,
105
+ "max_joint_limit_excess_rad": 0.6985994577407837,
106
+ "robot_constraint_diagnostics": {
107
+ "joint_limit_soft_counts": {
108
+ "index_q1": 97,
109
+ "index_q2": 7,
110
+ "middle_q1": 68,
111
+ "multi_middle_q2": 25,
112
+ "pinky_q1": 9,
113
+ "pinky_q2": 99,
114
+ "ring_q1": 4,
115
+ "ring_q2": 157,
116
+ "thumb_q2": 172
117
+ },
118
+ "joint_limit_hard_counts": {
119
+ "index_q1": 5,
120
+ "index_q2": 4,
121
+ "middle_q1": 15,
122
+ "multi_index_q1": 3,
123
+ "multi_middle_q2": 1,
124
+ "multi_pinky_q1": 1,
125
+ "multi_ring_q1": 1,
126
+ "pinky_q1": 5,
127
+ "pinky_q2": 5,
128
+ "ring_q1": 7,
129
+ "ring_q2": 26,
130
+ "thumb_q2": 6
131
+ },
132
+ "joint_limit_hard_episode_events": 49,
133
+ "finger_joint_limit_saturation_counts": {
134
+ "index_q1": 8,
135
+ "pinky_q2": 10,
136
+ "ring_q2": 52,
137
+ "thumb_q2": 93
138
+ },
139
+ "joint_limit_max_excess_by_joint": {
140
+ "index_q1": 0.1195741519331932,
141
+ "index_q2": 0.090233713388443,
142
+ "middle_q1": 0.0434245802462101,
143
+ "middle_q2": 0.0040107904933393,
144
+ "multi_index_q1": 0.0051511493511497974,
145
+ "multi_index_q2": 1.3135795562391195e-08,
146
+ "multi_middle_q1": 0.0006872953963465989,
147
+ "multi_middle_q2": 0.15990972518920898,
148
+ "multi_pinky_q1": 0.005502480547875166,
149
+ "multi_pinky_q2": 0.0011734918225556612,
150
+ "multi_ring_q1": 0.00500290235504508,
151
+ "multi_ring_q2": 0.0014249583473429084,
152
+ "pinky_q1": 0.0487305223941803,
153
+ "pinky_q2": 0.15635517239570618,
154
+ "ring_q1": 0.023890726268291473,
155
+ "ring_q2": 0.3032042980194092,
156
+ "thumb_q2": 0.6985994577407837
157
+ },
158
+ "coverage": {
159
+ "hard_limit_source": {
160
+ "available": 50,
161
+ "unavailable": 0,
162
+ "unknown": 0
163
+ },
164
+ "joint_role_manifest": {
165
+ "available": 50,
166
+ "unavailable": 0,
167
+ "unknown": 0
168
+ }
169
+ },
170
+ "hard_limit_sources": {
171
+ "runtime.joint_pos_limits": 50
172
+ },
173
+ "soft_limit_sources": {
174
+ "runtime.soft_joint_pos_limits": 50
175
+ },
176
+ "unknown_runtime_joint_names": [],
177
+ "baseline_checks": {
178
+ "post_home": {
179
+ "checked": 50,
180
+ "hard_violation_episodes": 0
181
+ },
182
+ "post_reset": {
183
+ "checked": 50,
184
+ "hard_violation_episodes": 0
185
+ }
186
+ }
187
+ }
188
+ },
189
+ "robot_motion": {
190
+ "joint_acceleration_rms": 11.16098033145452,
191
+ "joint_effort_rms": 108.08355145190777,
192
+ "joint_jerk_rms": 1105.198228857092,
193
+ "joint_velocity_rms": 0.8793243130092036
194
+ },
195
+ "grasp": {
196
+ "kinematic_grasp_stability_index": 0.8730928809084616,
197
+ "mean_kinematic_grasp_stability": 0.6129008210476189,
198
+ "per_object_best_kinematic_grasp_stability": {
199
+ "obj_194_wineglass_1": 0.8140910259363742,
200
+ "obj_194_wineglass_2": 0.8400803236223158,
201
+ "obj_194_wineglass_3": 0.7282407084725866
202
+ },
203
+ "grasp_gsi_diagnostics": {
204
+ "mean_anti_slip_score": 0.9501684749213619,
205
+ "mean_duration_score": 0.5446313939056353,
206
+ "mean_hold_score": 0.16281375688463065,
207
+ "mean_kinematic_stability_score": 0.1147987962165137,
208
+ "mean_legacy_gsi": 0.39178121858565756,
209
+ "mean_lift_score": 0.18171392594882277,
210
+ "mean_motion_stability_score": 0.13987992623596854
211
+ }
212
+ },
213
+ "tool": {
214
+ "tool_selection_accuracy": null,
215
+ "tool_switch_success_rate": null,
216
+ "tool_switch_total": 0,
217
+ "tool_switch_successful": 0
218
+ },
219
+ "runtime": {
220
+ "avg_steps": 2463.32,
221
+ "error_count": 0
222
+ }
223
+ }
224
+ }
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/03_wine_glass_plate_balance/none/summary.json ADDED
@@ -0,0 +1,209 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100300000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 31,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3246,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.62,
50
+ "success_rate_ci": {
51
+ "mean": 0.62,
52
+ "ci_low": 0.48,
53
+ "ci_high": 0.76
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.9133333333333333
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 33.72741935483808
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.46,
64
+ "hard_violation_rate": 0.5,
65
+ "drop_rate": 0.5,
66
+ "high_speed_violation_rate": 0.02,
67
+ "safety_violation_step_rate": 0.00036975411351451286
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.68,
74
+ "at_end_success_rate": 0.62,
75
+ "task_family_success_rate": {
76
+ "table_setting": 0.62
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.8466666666666667,
81
+ "current_chain_depth_progress_score": 0.8466666666666667,
82
+ "latched_chain_depth_progress_score": 0.9133333333333333,
83
+ "avg_current_chain_depth": 2.54,
84
+ "avg_latched_chain_depth": 2.74
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.0750447023590612,
88
+ "expert_time_s": 36.06666666666666,
89
+ "expert_time_step": 2164,
90
+ "avg_steps_to_success": 2023.6451612903227,
91
+ "avg_policy_steps_to_success": 674.9032258064516,
92
+ "avg_policy_queries_to_success": 674.9032258064516
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00029741091739210815,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_194_wineglass_1": 10,
98
+ "drop_to_table_obj_194_wineglass_2": 8,
99
+ "drop_to_table_obj_194_wineglass_3": 18,
100
+ "high_speed_distractor_00_clutter_008_pudding_box": 1
101
+ },
102
+ "robot_constraint_violation_rate": 0.3,
103
+ "finger_joint_limit_saturation_rate": 0.0,
104
+ "joint_limit_active_step_rate": 0.0005144405057593222,
105
+ "finger_joint_limit_saturation_step_rate": 0.0,
106
+ "max_joint_limit_excess_rad": 0.04555869847536087,
107
+ "robot_constraint_diagnostics": {
108
+ "joint_limit_soft_counts": {
109
+ "multi_ring_q1": 6,
110
+ "pinky_q1": 39,
111
+ "ring_q1": 11,
112
+ "ring_q2": 14
113
+ },
114
+ "joint_limit_hard_counts": {
115
+ "index_q1": 2,
116
+ "multi_pinky_q1": 3,
117
+ "multi_ring_q1": 3,
118
+ "pinky_q1": 10,
119
+ "ring_q1": 8,
120
+ "ring_q2": 14
121
+ },
122
+ "joint_limit_hard_episode_events": 32,
123
+ "finger_joint_limit_saturation_counts": {},
124
+ "joint_limit_max_excess_by_joint": {
125
+ "index_q1": 0.006058328319340944,
126
+ "index_q2": 1.0279842754812307e-08,
127
+ "middle_q1": 0.004658669698983431,
128
+ "middle_q2": 0.0031762155704200268,
129
+ "multi_index_q1": 0.0014145530294626951,
130
+ "multi_index_q2": 7.355778652851086e-09,
131
+ "multi_middle_q1": 0.00041098546353168786,
132
+ "multi_middle_q2": 0.0001388878154102713,
133
+ "multi_pinky_q1": 0.006252799183130264,
134
+ "multi_pinky_q2": 0.0006881741574034095,
135
+ "multi_ring_q1": 0.030369585379958153,
136
+ "multi_ring_q2": 0.0036631354596465826,
137
+ "pinky_q1": 0.04555869847536087,
138
+ "pinky_q2": 0.004730451852083206,
139
+ "ring_q1": 0.03248172998428345,
140
+ "ring_q2": 0.03466816991567612,
141
+ "thumb_q2": 0.0024808247108012438
142
+ },
143
+ "coverage": {
144
+ "hard_limit_source": {
145
+ "available": 50,
146
+ "unavailable": 0,
147
+ "unknown": 0
148
+ },
149
+ "joint_role_manifest": {
150
+ "available": 50,
151
+ "unavailable": 0,
152
+ "unknown": 0
153
+ }
154
+ },
155
+ "hard_limit_sources": {
156
+ "runtime.joint_pos_limits": 50
157
+ },
158
+ "soft_limit_sources": {
159
+ "runtime.soft_joint_pos_limits": 50
160
+ },
161
+ "unknown_runtime_joint_names": [],
162
+ "baseline_checks": {
163
+ "post_home": {
164
+ "checked": 50,
165
+ "hard_violation_episodes": 0
166
+ },
167
+ "post_reset": {
168
+ "checked": 50,
169
+ "hard_violation_episodes": 0
170
+ }
171
+ }
172
+ }
173
+ },
174
+ "robot_motion": {
175
+ "joint_acceleration_rms": 10.15736915562279,
176
+ "joint_effort_rms": 104.44700685542102,
177
+ "joint_jerk_rms": 992.8956161119934,
178
+ "joint_velocity_rms": 0.8634219137791083
179
+ },
180
+ "grasp": {
181
+ "kinematic_grasp_stability_index": 0.8687979906549208,
182
+ "mean_kinematic_grasp_stability": 0.5687891008820829,
183
+ "per_object_best_kinematic_grasp_stability": {
184
+ "obj_194_wineglass_1": 0.7947338077534061,
185
+ "obj_194_wineglass_2": 0.8257102318198812,
186
+ "obj_194_wineglass_3": 0.6851774397705891
187
+ },
188
+ "grasp_gsi_diagnostics": {
189
+ "mean_anti_slip_score": 0.9102969459465445,
190
+ "mean_duration_score": 0.5415311707183771,
191
+ "mean_hold_score": 0.16135363306066944,
192
+ "mean_kinematic_stability_score": 0.0910349135240453,
193
+ "mean_legacy_gsi": 0.38184927141361313,
194
+ "mean_lift_score": 0.18186692946244498,
195
+ "mean_motion_stability_score": 0.13799338735361935
196
+ }
197
+ },
198
+ "tool": {
199
+ "tool_selection_accuracy": null,
200
+ "tool_switch_success_rate": null,
201
+ "tool_switch_total": 0,
202
+ "tool_switch_successful": 0
203
+ },
204
+ "runtime": {
205
+ "avg_steps": 2488.14,
206
+ "error_count": 0
207
+ }
208
+ }
209
+ }
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/cov_only/summary.json ADDED
@@ -0,0 +1,266 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100600000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 38,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1920,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.76,
50
+ "success_rate_ci": {
51
+ "mean": 0.76,
52
+ "ci_low": 0.64,
53
+ "ci_high": 0.88
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.91
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.552631578947313
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.5,
64
+ "hard_violation_rate": 0.44,
65
+ "drop_rate": 0.2,
66
+ "high_speed_violation_rate": 0.34,
67
+ "safety_violation_step_rate": 0.000690643732082356
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.76,
74
+ "at_end_success_rate": 0.76,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.76
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.89,
81
+ "current_chain_depth_progress_score": 0.92,
82
+ "latched_chain_depth_progress_score": 0.9266666666666666,
83
+ "avg_current_chain_depth": 2.76,
84
+ "avg_latched_chain_depth": 2.78
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9158344194605128,
88
+ "expert_time_s": 21.333333333333332,
89
+ "expert_time_step": 1280,
90
+ "avg_steps_to_success": 1413.157894736842,
91
+ "avg_policy_steps_to_success": 471.42105263157896,
92
+ "avg_policy_queries_to_success": 471.42105263157896
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0004821475110763617,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_011_banana_2": 1,
98
+ "drop_to_table_obj_013_apple_3": 8,
99
+ "drop_to_table_obj_013_apple_4": 7,
100
+ "high_speed_obj_013_apple_3": 11,
101
+ "high_speed_obj_013_apple_4": 10
102
+ },
103
+ "robot_constraint_violation_rate": 1.0,
104
+ "finger_joint_limit_saturation_rate": 0.96,
105
+ "joint_limit_active_step_rate": 0.13105290591608026,
106
+ "finger_joint_limit_saturation_step_rate": 0.06512900703674745,
107
+ "max_joint_limit_excess_rad": 1.412924885749817,
108
+ "robot_constraint_diagnostics": {
109
+ "joint_limit_soft_counts": {
110
+ "left_index_1_joint": 1049,
111
+ "left_index_2_joint": 32,
112
+ "left_little_1_joint": 1057,
113
+ "left_little_2_joint": 125,
114
+ "left_middle_1_joint": 1397,
115
+ "left_middle_2_joint": 153,
116
+ "left_ring_1_joint": 1432,
117
+ "left_ring_2_joint": 201,
118
+ "left_thumb_2_joint": 8,
119
+ "left_thumb_3_joint": 18,
120
+ "left_thumb_4_joint": 66,
121
+ "right_index_1_joint": 695,
122
+ "right_index_2_joint": 13,
123
+ "right_little_1_joint": 270,
124
+ "right_little_2_joint": 80,
125
+ "right_middle_1_joint": 864,
126
+ "right_ring_1_joint": 452,
127
+ "right_ring_2_joint": 32,
128
+ "right_thumb_1_joint": 37,
129
+ "right_thumb_2_joint": 7207,
130
+ "right_thumb_3_joint": 37,
131
+ "right_thumb_4_joint": 161
132
+ },
133
+ "joint_limit_hard_counts": {
134
+ "left_index_1_joint": 58,
135
+ "left_index_2_joint": 47,
136
+ "left_little_1_joint": 46,
137
+ "left_little_2_joint": 103,
138
+ "left_middle_1_joint": 49,
139
+ "left_middle_2_joint": 147,
140
+ "left_ring_1_joint": 51,
141
+ "left_ring_2_joint": 145,
142
+ "left_thumb_2_joint": 2,
143
+ "left_thumb_3_joint": 1,
144
+ "left_thumb_4_joint": 2,
145
+ "right_index_1_joint": 37,
146
+ "right_index_2_joint": 5,
147
+ "right_little_1_joint": 12,
148
+ "right_little_2_joint": 4,
149
+ "right_middle_1_joint": 41,
150
+ "right_ring_1_joint": 26,
151
+ "right_ring_2_joint": 6,
152
+ "right_thumb_1_joint": 3,
153
+ "right_thumb_2_joint": 91,
154
+ "right_thumb_3_joint": 2,
155
+ "right_thumb_4_joint": 12
156
+ },
157
+ "joint_limit_hard_episode_events": 204,
158
+ "finger_joint_limit_saturation_counts": {
159
+ "left_index_1_joint": 451,
160
+ "left_little_1_joint": 555,
161
+ "left_middle_1_joint": 696,
162
+ "left_ring_1_joint": 711,
163
+ "left_thumb_4_joint": 48,
164
+ "right_index_1_joint": 308,
165
+ "right_little_1_joint": 212,
166
+ "right_little_2_joint": 50,
167
+ "right_middle_1_joint": 409,
168
+ "right_ring_1_joint": 271,
169
+ "right_ring_2_joint": 1,
170
+ "right_thumb_2_joint": 3544,
171
+ "right_thumb_3_joint": 32,
172
+ "right_thumb_4_joint": 85
173
+ },
174
+ "joint_limit_max_excess_by_joint": {
175
+ "left_index_1_joint": 0.9544789791107178,
176
+ "left_index_2_joint": 0.04764508083462715,
177
+ "left_little_1_joint": 1.2584928274154663,
178
+ "left_little_2_joint": 0.054968733340501785,
179
+ "left_middle_1_joint": 0.9923037886619568,
180
+ "left_middle_2_joint": 0.058840252459049225,
181
+ "left_ring_1_joint": 1.3025926351547241,
182
+ "left_ring_2_joint": 0.10544086992740631,
183
+ "left_thumb_1_joint": 1.544841506984085e-05,
184
+ "left_thumb_2_joint": 0.03194544091820717,
185
+ "left_thumb_3_joint": 0.09500502794981003,
186
+ "left_thumb_4_joint": 0.6240297555923462,
187
+ "right_index_1_joint": 1.1282039880752563,
188
+ "right_index_2_joint": 0.03369522467255592,
189
+ "right_little_1_joint": 1.412924885749817,
190
+ "right_little_2_joint": 0.4270598888397217,
191
+ "right_middle_1_joint": 1.2678178548812866,
192
+ "right_middle_2_joint": 0.0028164153918623924,
193
+ "right_ring_1_joint": 1.374912977218628,
194
+ "right_ring_2_joint": 0.2026987075805664,
195
+ "right_thumb_1_joint": 0.11635511368513107,
196
+ "right_thumb_2_joint": 0.5441575646400452,
197
+ "right_thumb_3_joint": 0.963283121585846,
198
+ "right_thumb_4_joint": 0.8909096717834473
199
+ },
200
+ "coverage": {
201
+ "hard_limit_source": {
202
+ "available": 50,
203
+ "unavailable": 0,
204
+ "unknown": 0
205
+ },
206
+ "joint_role_manifest": {
207
+ "available": 50,
208
+ "unavailable": 0,
209
+ "unknown": 0
210
+ }
211
+ },
212
+ "hard_limit_sources": {
213
+ "runtime.joint_pos_limits": 50
214
+ },
215
+ "soft_limit_sources": {
216
+ "runtime.soft_joint_pos_limits": 50
217
+ },
218
+ "unknown_runtime_joint_names": [],
219
+ "baseline_checks": {
220
+ "post_home": {
221
+ "checked": 50,
222
+ "hard_violation_episodes": 0
223
+ },
224
+ "post_reset": {
225
+ "checked": 50,
226
+ "hard_violation_episodes": 0
227
+ }
228
+ }
229
+ }
230
+ },
231
+ "robot_motion": {
232
+ "joint_acceleration_rms": 52.21611146173608,
233
+ "joint_effort_rms": 39.28649673851511,
234
+ "joint_jerk_rms": 5342.885461600006,
235
+ "joint_velocity_rms": 1.0655054731084979
236
+ },
237
+ "grasp": {
238
+ "kinematic_grasp_stability_index": 0.9273259012535624,
239
+ "mean_kinematic_grasp_stability": 0.40215106377518006,
240
+ "per_object_best_kinematic_grasp_stability": {
241
+ "obj_011_banana_2": 0.6395973651645397,
242
+ "obj_013_apple_3": 0.8733573378015763,
243
+ "obj_013_apple_4": 0.24484866411101405
244
+ },
245
+ "grasp_gsi_diagnostics": {
246
+ "mean_anti_slip_score": 0.9403808577395145,
247
+ "mean_duration_score": 0.12878994894613335,
248
+ "mean_hold_score": 0.1272510926422046,
249
+ "mean_kinematic_stability_score": 0.08770831379074026,
250
+ "mean_legacy_gsi": 0.2942340031851498,
251
+ "mean_lift_score": 0.21294563204297604,
252
+ "mean_motion_stability_score": 0.11859687821588534
253
+ }
254
+ },
255
+ "tool": {
256
+ "tool_selection_accuracy": null,
257
+ "tool_switch_success_rate": null,
258
+ "tool_switch_total": 0,
259
+ "tool_switch_successful": 0
260
+ },
261
+ "runtime": {
262
+ "avg_steps": 1534.8,
263
+ "error_count": 0
264
+ }
265
+ }
266
+ }
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_cov/summary.json ADDED
@@ -0,0 +1,262 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 200600000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 35,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1920,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.7,
50
+ "success_rate_ci": {
51
+ "mean": 0.7,
52
+ "ci_low": 0.56,
53
+ "ci_high": 0.82
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.96
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 22.702380952380945
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.6,
64
+ "hard_violation_rate": 0.36,
65
+ "drop_rate": 0.28,
66
+ "high_speed_violation_rate": 0.2,
67
+ "safety_violation_step_rate": 0.00040536122915985615
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.78,
74
+ "at_end_success_rate": 0.7,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.7
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.88,
81
+ "current_chain_depth_progress_score": 0.9,
82
+ "latched_chain_depth_progress_score": 0.9666666666666667,
83
+ "avg_current_chain_depth": 2.7,
84
+ "avg_latched_chain_depth": 2.9
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9425765108665877,
88
+ "expert_time_s": 21.333333333333332,
89
+ "expert_time_step": 1280,
90
+ "avg_steps_to_success": 1362.142857142857,
91
+ "avg_policy_steps_to_success": 454.37142857142857,
92
+ "avg_policy_queries_to_success": 454.37142857142857
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00036613272311212814,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_011_banana_2": 4,
98
+ "drop_to_table_obj_013_apple_3": 7,
99
+ "drop_to_table_obj_013_apple_4": 6,
100
+ "high_speed_obj_013_apple_3": 8,
101
+ "high_speed_obj_013_apple_4": 3
102
+ },
103
+ "robot_constraint_violation_rate": 1.0,
104
+ "finger_joint_limit_saturation_rate": 1.0,
105
+ "joint_limit_active_step_rate": 0.1308009153318078,
106
+ "finger_joint_limit_saturation_step_rate": 0.05962732919254658,
107
+ "max_joint_limit_excess_rad": 1.5251175165176392,
108
+ "robot_constraint_diagnostics": {
109
+ "joint_limit_soft_counts": {
110
+ "left_index_1_joint": 1212,
111
+ "left_index_2_joint": 64,
112
+ "left_little_1_joint": 1536,
113
+ "left_little_2_joint": 258,
114
+ "left_middle_1_joint": 1894,
115
+ "left_middle_2_joint": 148,
116
+ "left_ring_1_joint": 1940,
117
+ "left_ring_2_joint": 301,
118
+ "left_thumb_2_joint": 16,
119
+ "left_thumb_3_joint": 18,
120
+ "left_thumb_4_joint": 55,
121
+ "right_index_1_joint": 399,
122
+ "right_index_2_joint": 14,
123
+ "right_little_1_joint": 385,
124
+ "right_little_2_joint": 65,
125
+ "right_middle_1_joint": 487,
126
+ "right_middle_2_joint": 41,
127
+ "right_ring_1_joint": 312,
128
+ "right_ring_2_joint": 47,
129
+ "right_thumb_2_joint": 7021,
130
+ "right_thumb_4_joint": 23
131
+ },
132
+ "joint_limit_hard_counts": {
133
+ "left_index_1_joint": 74,
134
+ "left_index_2_joint": 117,
135
+ "left_little_1_joint": 55,
136
+ "left_little_2_joint": 189,
137
+ "left_middle_1_joint": 62,
138
+ "left_middle_2_joint": 174,
139
+ "left_ring_1_joint": 64,
140
+ "left_ring_2_joint": 210,
141
+ "left_thumb_2_joint": 1,
142
+ "left_thumb_3_joint": 1,
143
+ "left_thumb_4_joint": 4,
144
+ "right_index_1_joint": 18,
145
+ "right_index_2_joint": 9,
146
+ "right_little_1_joint": 24,
147
+ "right_little_2_joint": 3,
148
+ "right_middle_1_joint": 35,
149
+ "right_middle_2_joint": 15,
150
+ "right_ring_1_joint": 11,
151
+ "right_ring_2_joint": 12,
152
+ "right_thumb_2_joint": 98,
153
+ "right_thumb_4_joint": 17
154
+ },
155
+ "joint_limit_hard_episode_events": 212,
156
+ "finger_joint_limit_saturation_counts": {
157
+ "left_index_1_joint": 345,
158
+ "left_little_1_joint": 889,
159
+ "left_little_2_joint": 3,
160
+ "left_middle_1_joint": 1064,
161
+ "left_ring_1_joint": 1105,
162
+ "left_thumb_3_joint": 4,
163
+ "left_thumb_4_joint": 7,
164
+ "right_index_1_joint": 140,
165
+ "right_little_1_joint": 281,
166
+ "right_little_2_joint": 21,
167
+ "right_middle_1_joint": 118,
168
+ "right_ring_1_joint": 213,
169
+ "right_thumb_2_joint": 2773
170
+ },
171
+ "joint_limit_max_excess_by_joint": {
172
+ "left_index_1_joint": 0.688579261302948,
173
+ "left_index_2_joint": 0.043488599359989166,
174
+ "left_little_1_joint": 0.9149285554885864,
175
+ "left_little_2_joint": 0.10644107311964035,
176
+ "left_middle_1_joint": 0.9101478457450867,
177
+ "left_middle_2_joint": 0.0673137977719307,
178
+ "left_ring_1_joint": 1.0870126485824585,
179
+ "left_ring_2_joint": 0.07230380177497864,
180
+ "left_thumb_2_joint": 0.08802511543035507,
181
+ "left_thumb_3_joint": 0.27836883068084717,
182
+ "left_thumb_4_joint": 0.24408695101737976,
183
+ "right_index_1_joint": 0.40721237659454346,
184
+ "right_index_2_joint": 0.040965937077999115,
185
+ "right_little_1_joint": 1.5251175165176392,
186
+ "right_little_2_joint": 0.31902050971984863,
187
+ "right_middle_1_joint": 0.6464768052101135,
188
+ "right_middle_2_joint": 0.05805111303925514,
189
+ "right_ring_1_joint": 0.9161616563796997,
190
+ "right_ring_2_joint": 0.06698006391525269,
191
+ "right_thumb_1_joint": 3.44770924129989e-05,
192
+ "right_thumb_2_joint": 0.4074215888977051,
193
+ "right_thumb_3_joint": 0.003978325519710779,
194
+ "right_thumb_4_joint": 0.10082778334617615
195
+ },
196
+ "coverage": {
197
+ "hard_limit_source": {
198
+ "available": 50,
199
+ "unavailable": 0,
200
+ "unknown": 0
201
+ },
202
+ "joint_role_manifest": {
203
+ "available": 50,
204
+ "unavailable": 0,
205
+ "unknown": 0
206
+ }
207
+ },
208
+ "hard_limit_sources": {
209
+ "runtime.joint_pos_limits": 50
210
+ },
211
+ "soft_limit_sources": {
212
+ "runtime.soft_joint_pos_limits": 50
213
+ },
214
+ "unknown_runtime_joint_names": [],
215
+ "baseline_checks": {
216
+ "post_home": {
217
+ "checked": 50,
218
+ "hard_violation_episodes": 0
219
+ },
220
+ "post_reset": {
221
+ "checked": 50,
222
+ "hard_violation_episodes": 0
223
+ }
224
+ }
225
+ }
226
+ },
227
+ "robot_motion": {
228
+ "joint_acceleration_rms": 57.10186569644383,
229
+ "joint_effort_rms": 39.54838460960955,
230
+ "joint_jerk_rms": 5824.59062138594,
231
+ "joint_velocity_rms": 1.1155436797056792
232
+ },
233
+ "grasp": {
234
+ "kinematic_grasp_stability_index": 0.87834275622815,
235
+ "mean_kinematic_grasp_stability": 0.3492165926559069,
236
+ "per_object_best_kinematic_grasp_stability": {
237
+ "obj_011_banana_2": 0.7352310792790759,
238
+ "obj_013_apple_3": 0.7839122971730303,
239
+ "obj_013_apple_4": 0.2921300953589021
240
+ },
241
+ "grasp_gsi_diagnostics": {
242
+ "mean_anti_slip_score": 0.9101997831849488,
243
+ "mean_duration_score": 0.13256012007426948,
244
+ "mean_hold_score": 0.11791556839915074,
245
+ "mean_kinematic_stability_score": 0.08020551502541955,
246
+ "mean_legacy_gsi": 0.2840924027990203,
247
+ "mean_lift_score": 0.20547626752524703,
248
+ "mean_motion_stability_score": 0.11039383990355804
249
+ }
250
+ },
251
+ "tool": {
252
+ "tool_selection_accuracy": null,
253
+ "tool_switch_success_rate": null,
254
+ "tool_switch_total": 0,
255
+ "tool_switch_successful": 0
256
+ },
257
+ "runtime": {
258
+ "avg_steps": 1529.5,
259
+ "error_count": 0
260
+ }
261
+ }
262
+ }
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/inv_only/summary.json ADDED
@@ -0,0 +1,257 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100600000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 41,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1920,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.82,
50
+ "success_rate_ci": {
51
+ "mean": 0.82,
52
+ "ci_low": 0.7,
53
+ "ci_high": 0.92
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.93
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 22.392276422764237
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.66,
64
+ "hard_violation_rate": 0.34,
65
+ "drop_rate": 0.2,
66
+ "high_speed_violation_rate": 0.24,
67
+ "safety_violation_step_rate": 0.0003592897118772887
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.82,
74
+ "at_end_success_rate": 0.82,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.82
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.91,
81
+ "current_chain_depth_progress_score": 0.94,
82
+ "latched_chain_depth_progress_score": 0.9466666666666667,
83
+ "avg_current_chain_depth": 2.82,
84
+ "avg_latched_chain_depth": 2.84
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.953735245333645,
88
+ "expert_time_s": 21.333333333333332,
89
+ "expert_time_step": 1280,
90
+ "avg_steps_to_success": 1343.5365853658536,
91
+ "avg_policy_steps_to_success": 448.1707317073171,
92
+ "avg_policy_queries_to_success": 448.1707317073171
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00037310854694949213,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_013_apple_3": 6,
98
+ "drop_to_table_obj_013_apple_4": 8,
99
+ "high_speed_obj_013_apple_3": 9,
100
+ "high_speed_obj_013_apple_4": 4
101
+ },
102
+ "robot_constraint_violation_rate": 1.0,
103
+ "finger_joint_limit_saturation_rate": 0.98,
104
+ "joint_limit_active_step_rate": 0.12919228909002972,
105
+ "finger_joint_limit_saturation_step_rate": 0.060775236647550614,
106
+ "max_joint_limit_excess_rad": 1.5756245851516724,
107
+ "robot_constraint_diagnostics": {
108
+ "joint_limit_soft_counts": {
109
+ "left_index_1_joint": 1183,
110
+ "left_index_2_joint": 46,
111
+ "left_little_1_joint": 1388,
112
+ "left_little_2_joint": 265,
113
+ "left_middle_1_joint": 1636,
114
+ "left_middle_2_joint": 148,
115
+ "left_ring_1_joint": 1684,
116
+ "left_ring_2_joint": 299,
117
+ "left_thumb_2_joint": 11,
118
+ "left_thumb_3_joint": 2,
119
+ "left_thumb_4_joint": 17,
120
+ "right_index_1_joint": 424,
121
+ "right_index_2_joint": 8,
122
+ "right_little_1_joint": 283,
123
+ "right_little_2_joint": 51,
124
+ "right_middle_1_joint": 524,
125
+ "right_middle_2_joint": 24,
126
+ "right_ring_1_joint": 265,
127
+ "right_ring_2_joint": 13,
128
+ "right_thumb_2_joint": 6716,
129
+ "right_thumb_4_joint": 33
130
+ },
131
+ "joint_limit_hard_counts": {
132
+ "left_index_1_joint": 59,
133
+ "left_index_2_joint": 85,
134
+ "left_little_1_joint": 48,
135
+ "left_little_2_joint": 179,
136
+ "left_middle_1_joint": 50,
137
+ "left_middle_2_joint": 154,
138
+ "left_ring_1_joint": 49,
139
+ "left_ring_2_joint": 197,
140
+ "left_thumb_2_joint": 1,
141
+ "left_thumb_3_joint": 1,
142
+ "left_thumb_4_joint": 2,
143
+ "right_index_1_joint": 16,
144
+ "right_index_2_joint": 5,
145
+ "right_little_1_joint": 11,
146
+ "right_little_2_joint": 4,
147
+ "right_middle_1_joint": 27,
148
+ "right_middle_2_joint": 15,
149
+ "right_ring_1_joint": 14,
150
+ "right_ring_2_joint": 3,
151
+ "right_thumb_2_joint": 99,
152
+ "right_thumb_4_joint": 12
153
+ },
154
+ "joint_limit_hard_episode_events": 191,
155
+ "finger_joint_limit_saturation_counts": {
156
+ "left_index_1_joint": 452,
157
+ "left_little_1_joint": 760,
158
+ "left_middle_1_joint": 939,
159
+ "left_ring_1_joint": 921,
160
+ "right_index_1_joint": 261,
161
+ "right_little_1_joint": 243,
162
+ "right_middle_1_joint": 267,
163
+ "right_ring_1_joint": 169,
164
+ "right_thumb_2_joint": 2660
165
+ },
166
+ "joint_limit_max_excess_by_joint": {
167
+ "left_index_1_joint": 0.71388840675354,
168
+ "left_index_2_joint": 0.02974337711930275,
169
+ "left_little_1_joint": 0.7726544737815857,
170
+ "left_little_2_joint": 0.08912867307662964,
171
+ "left_middle_1_joint": 1.0820238590240479,
172
+ "left_middle_2_joint": 0.039869461208581924,
173
+ "left_ring_1_joint": 0.9920606017112732,
174
+ "left_ring_2_joint": 0.0692342221736908,
175
+ "left_thumb_2_joint": 0.033564914017915726,
176
+ "left_thumb_3_joint": 0.027722299098968506,
177
+ "left_thumb_4_joint": 0.15004417300224304,
178
+ "right_index_1_joint": 0.5022022724151611,
179
+ "right_index_2_joint": 0.023375432938337326,
180
+ "right_little_1_joint": 1.5756245851516724,
181
+ "right_little_2_joint": 0.1778852939605713,
182
+ "right_middle_1_joint": 0.7972173094749451,
183
+ "right_middle_2_joint": 0.04769420251250267,
184
+ "right_ring_1_joint": 0.8915323615074158,
185
+ "right_ring_2_joint": 0.05484747514128685,
186
+ "right_thumb_1_joint": 0.00018919340800493956,
187
+ "right_thumb_2_joint": 0.3858295679092407,
188
+ "right_thumb_3_joint": 0.00019537062325980514,
189
+ "right_thumb_4_joint": 0.15239492058753967
190
+ },
191
+ "coverage": {
192
+ "hard_limit_source": {
193
+ "available": 50,
194
+ "unavailable": 0,
195
+ "unknown": 0
196
+ },
197
+ "joint_role_manifest": {
198
+ "available": 50,
199
+ "unavailable": 0,
200
+ "unknown": 0
201
+ }
202
+ },
203
+ "hard_limit_sources": {
204
+ "runtime.joint_pos_limits": 50
205
+ },
206
+ "soft_limit_sources": {
207
+ "runtime.soft_joint_pos_limits": 50
208
+ },
209
+ "unknown_runtime_joint_names": [],
210
+ "baseline_checks": {
211
+ "post_home": {
212
+ "checked": 50,
213
+ "hard_violation_episodes": 0
214
+ },
215
+ "post_reset": {
216
+ "checked": 50,
217
+ "hard_violation_episodes": 0
218
+ }
219
+ }
220
+ }
221
+ },
222
+ "robot_motion": {
223
+ "joint_acceleration_rms": 58.180073991574496,
224
+ "joint_effort_rms": 39.965325025301595,
225
+ "joint_jerk_rms": 5951.447830016606,
226
+ "joint_velocity_rms": 1.1348407938025868
227
+ },
228
+ "grasp": {
229
+ "kinematic_grasp_stability_index": 0.8897562553422163,
230
+ "mean_kinematic_grasp_stability": 0.3563319230283136,
231
+ "per_object_best_kinematic_grasp_stability": {
232
+ "obj_011_banana_2": 0.7278564307511154,
233
+ "obj_013_apple_3": 0.8001711648531005,
234
+ "obj_013_apple_4": 0.3270453281879376
235
+ },
236
+ "grasp_gsi_diagnostics": {
237
+ "mean_anti_slip_score": 0.9044621410295516,
238
+ "mean_duration_score": 0.1313233855312158,
239
+ "mean_hold_score": 0.1197297852056927,
240
+ "mean_kinematic_stability_score": 0.07727047794374294,
241
+ "mean_legacy_gsi": 0.2838011160129997,
242
+ "mean_lift_score": 0.20608263514622818,
243
+ "mean_motion_stability_score": 0.11245154924459355
244
+ }
245
+ },
246
+ "tool": {
247
+ "tool_selection_accuracy": null,
248
+ "tool_switch_success_rate": null,
249
+ "tool_switch_total": 0,
250
+ "tool_switch_successful": 0
251
+ },
252
+ "runtime": {
253
+ "avg_steps": 1447.3,
254
+ "error_count": 0
255
+ }
256
+ }
257
+ }
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/06_fruit_bowl_loading/none/summary.json ADDED
@@ -0,0 +1,261 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100600000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 44,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1920,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.88,
50
+ "success_rate_ci": {
51
+ "mean": 0.88,
52
+ "ci_low": 0.78,
53
+ "ci_high": 0.96
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.955
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 22.99696969696967
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.56,
64
+ "hard_violation_rate": 0.42,
65
+ "drop_rate": 0.14,
66
+ "high_speed_violation_rate": 0.38,
67
+ "safety_violation_step_rate": 0.0004983940635729316
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.88,
74
+ "at_end_success_rate": 0.88,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.88
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.92,
81
+ "current_chain_depth_progress_score": 0.9333333333333333,
82
+ "latched_chain_depth_progress_score": 0.9666666666666667,
83
+ "avg_current_chain_depth": 2.8,
84
+ "avg_latched_chain_depth": 2.9
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9316518591651985,
88
+ "expert_time_s": 21.333333333333332,
89
+ "expert_time_step": 1280,
90
+ "avg_steps_to_success": 1379.8181818181818,
91
+ "avg_policy_steps_to_success": 460.25,
92
+ "avg_policy_queries_to_success": 460.25
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00044301694539816146,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_011_banana_2": 2,
98
+ "drop_to_table_obj_013_apple_3": 3,
99
+ "drop_to_table_obj_013_apple_4": 6,
100
+ "high_speed_obj_011_banana_2": 1,
101
+ "high_speed_obj_013_apple_3": 8,
102
+ "high_speed_obj_013_apple_4": 12
103
+ },
104
+ "robot_constraint_violation_rate": 0.98,
105
+ "finger_joint_limit_saturation_rate": 0.96,
106
+ "joint_limit_active_step_rate": 0.12570605825672831,
107
+ "finger_joint_limit_saturation_step_rate": 0.05642928342009082,
108
+ "max_joint_limit_excess_rad": 0.972054123878479,
109
+ "robot_constraint_diagnostics": {
110
+ "joint_limit_soft_counts": {
111
+ "left_index_1_joint": 894,
112
+ "left_index_2_joint": 40,
113
+ "left_little_1_joint": 1498,
114
+ "left_little_2_joint": 188,
115
+ "left_middle_1_joint": 1574,
116
+ "left_middle_2_joint": 189,
117
+ "left_ring_1_joint": 1680,
118
+ "left_ring_2_joint": 332,
119
+ "left_thumb_2_joint": 162,
120
+ "left_thumb_4_joint": 11,
121
+ "right_index_1_joint": 168,
122
+ "right_index_2_joint": 3,
123
+ "right_little_1_joint": 85,
124
+ "right_middle_1_joint": 317,
125
+ "right_middle_2_joint": 9,
126
+ "right_ring_1_joint": 144,
127
+ "right_ring_2_joint": 16,
128
+ "right_thumb_2_joint": 6574,
129
+ "right_thumb_3_joint": 1,
130
+ "right_thumb_4_joint": 63
131
+ },
132
+ "joint_limit_hard_counts": {
133
+ "left_index_1_joint": 48,
134
+ "left_index_2_joint": 63,
135
+ "left_little_1_joint": 51,
136
+ "left_little_2_joint": 175,
137
+ "left_middle_1_joint": 54,
138
+ "left_middle_2_joint": 164,
139
+ "left_ring_1_joint": 51,
140
+ "left_ring_2_joint": 180,
141
+ "left_thumb_2_joint": 9,
142
+ "left_thumb_4_joint": 2,
143
+ "right_index_1_joint": 18,
144
+ "right_index_2_joint": 2,
145
+ "right_little_1_joint": 8,
146
+ "right_middle_1_joint": 32,
147
+ "right_middle_2_joint": 4,
148
+ "right_ring_1_joint": 9,
149
+ "right_ring_2_joint": 5,
150
+ "right_thumb_2_joint": 93,
151
+ "right_thumb_3_joint": 1,
152
+ "right_thumb_4_joint": 10
153
+ },
154
+ "joint_limit_hard_episode_events": 199,
155
+ "finger_joint_limit_saturation_counts": {
156
+ "left_index_1_joint": 309,
157
+ "left_little_1_joint": 833,
158
+ "left_middle_1_joint": 711,
159
+ "left_middle_2_joint": 3,
160
+ "left_ring_1_joint": 792,
161
+ "left_ring_2_joint": 43,
162
+ "left_thumb_2_joint": 24,
163
+ "right_index_1_joint": 73,
164
+ "right_little_1_joint": 55,
165
+ "right_middle_1_joint": 108,
166
+ "right_ring_1_joint": 72,
167
+ "right_thumb_2_joint": 2798
168
+ },
169
+ "joint_limit_max_excess_by_joint": {
170
+ "left_index_1_joint": 0.8251557946205139,
171
+ "left_index_2_joint": 0.039083871990442276,
172
+ "left_little_1_joint": 0.8335281014442444,
173
+ "left_little_2_joint": 0.07381156086921692,
174
+ "left_middle_1_joint": 0.9602811336517334,
175
+ "left_middle_2_joint": 0.11868715286254883,
176
+ "left_ring_1_joint": 0.972054123878479,
177
+ "left_ring_2_joint": 0.7195554375648499,
178
+ "left_thumb_1_joint": 1.4882676623528823e-05,
179
+ "left_thumb_2_joint": 0.22311218082904816,
180
+ "left_thumb_3_joint": 0.00031279659015126526,
181
+ "left_thumb_4_joint": 0.07767227292060852,
182
+ "right_index_1_joint": 0.8659150004386902,
183
+ "right_index_2_joint": 0.023793375119566917,
184
+ "right_little_1_joint": 0.48093992471694946,
185
+ "right_little_2_joint": 1.7764557924238034e-05,
186
+ "right_middle_1_joint": 0.8673695921897888,
187
+ "right_middle_2_joint": 0.035325340926647186,
188
+ "right_ring_1_joint": 0.8179689049720764,
189
+ "right_ring_2_joint": 0.08797579258680344,
190
+ "right_thumb_1_joint": 8.618831634521484e-05,
191
+ "right_thumb_2_joint": 0.31030362844467163,
192
+ "right_thumb_3_joint": 0.017132222652435303,
193
+ "right_thumb_4_joint": 0.154220849275589
194
+ },
195
+ "coverage": {
196
+ "hard_limit_source": {
197
+ "available": 50,
198
+ "unavailable": 0,
199
+ "unknown": 0
200
+ },
201
+ "joint_role_manifest": {
202
+ "available": 50,
203
+ "unavailable": 0,
204
+ "unknown": 0
205
+ }
206
+ },
207
+ "hard_limit_sources": {
208
+ "runtime.joint_pos_limits": 50
209
+ },
210
+ "soft_limit_sources": {
211
+ "runtime.soft_joint_pos_limits": 50
212
+ },
213
+ "unknown_runtime_joint_names": [],
214
+ "baseline_checks": {
215
+ "post_home": {
216
+ "checked": 50,
217
+ "hard_violation_episodes": 0
218
+ },
219
+ "post_reset": {
220
+ "checked": 50,
221
+ "hard_violation_episodes": 0
222
+ }
223
+ }
224
+ }
225
+ },
226
+ "robot_motion": {
227
+ "joint_acceleration_rms": 55.35695659123332,
228
+ "joint_effort_rms": 39.797297035345125,
229
+ "joint_jerk_rms": 5658.523985705526,
230
+ "joint_velocity_rms": 1.1120583905938077
231
+ },
232
+ "grasp": {
233
+ "kinematic_grasp_stability_index": 0.8909102055369971,
234
+ "mean_kinematic_grasp_stability": 0.35234361267073,
235
+ "per_object_best_kinematic_grasp_stability": {
236
+ "obj_011_banana_2": 0.6924046957783577,
237
+ "obj_013_apple_3": 0.8111089965240547,
238
+ "obj_013_apple_4": 0.24059681512430783
239
+ },
240
+ "grasp_gsi_diagnostics": {
241
+ "mean_anti_slip_score": 0.9196922721544439,
242
+ "mean_duration_score": 0.11593373131486409,
243
+ "mean_hold_score": 0.11276026175850268,
244
+ "mean_kinematic_stability_score": 0.07309666207752648,
245
+ "mean_legacy_gsi": 0.27952238686948144,
246
+ "mean_lift_score": 0.19883059749088797,
247
+ "mean_motion_stability_score": 0.10513085372288246
248
+ }
249
+ },
250
+ "tool": {
251
+ "tool_selection_accuracy": null,
252
+ "tool_switch_success_rate": null,
253
+ "tool_switch_total": 0,
254
+ "tool_switch_successful": 0
255
+ },
256
+ "runtime": {
257
+ "avg_steps": 1444.64,
258
+ "error_count": 0
259
+ }
260
+ }
261
+ }
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/cov_only/summary.json ADDED
@@ -0,0 +1,271 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100700000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 43,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1806,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.86,
50
+ "success_rate_ci": {
51
+ "mean": 0.86,
52
+ "ci_low": 0.76,
53
+ "ci_high": 0.94
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.98
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.939922480620076
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.86,
64
+ "hard_violation_rate": 0.0,
65
+ "drop_rate": 0.0,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.0
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.96,
74
+ "at_end_success_rate": 0.86,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.86
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.955,
81
+ "current_chain_depth_progress_score": 0.99,
82
+ "latched_chain_depth_progress_score": 0.99,
83
+ "avg_current_chain_depth": 1.98,
84
+ "avg_latched_chain_depth": 1.98
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.8392274223809373,
88
+ "expert_time_s": 20.05,
89
+ "expert_time_step": 1203,
90
+ "avg_steps_to_success": 1436.3953488372092,
91
+ "avg_policy_steps_to_success": 479.13953488372096,
92
+ "avg_policy_queries_to_success": 479.13953488372096
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0,
96
+ "violation_counts": {},
97
+ "robot_constraint_violation_rate": 1.0,
98
+ "finger_joint_limit_saturation_rate": 1.0,
99
+ "joint_limit_active_step_rate": 0.10399559181259828,
100
+ "finger_joint_limit_saturation_step_rate": 0.03068259706748021,
101
+ "max_joint_limit_excess_rad": 0.9809489250183105,
102
+ "robot_constraint_diagnostics": {
103
+ "joint_limit_soft_counts": {
104
+ "left_middle_dip_joint": 326,
105
+ "left_middle_pip_joint": 559,
106
+ "left_middle_pitch_joint": 21,
107
+ "left_pinky_dip_joint": 37,
108
+ "left_pinky_mcp_joint": 1621,
109
+ "left_pinky_pip_joint": 19,
110
+ "left_ring_dip_joint": 751,
111
+ "left_ring_pip_joint": 1220,
112
+ "left_thumb_dip_joint": 18,
113
+ "right_index_dip_joint": 29,
114
+ "right_index_pip_joint": 155,
115
+ "right_index_pitch_joint": 63,
116
+ "right_middle_dip_joint": 422,
117
+ "right_middle_pip_joint": 888,
118
+ "right_middle_pitch_joint": 87,
119
+ "right_pinky_mcp_joint": 415,
120
+ "right_pinky_pip_joint": 1063,
121
+ "right_ring_dip_joint": 892,
122
+ "right_ring_pip_joint": 2695,
123
+ "right_thumb_dip_joint": 236,
124
+ "right_thumb_mcp_joint": 2463,
125
+ "right_thumb_pip_joint": 33
126
+ },
127
+ "joint_limit_hard_counts": {
128
+ "left_middle_dip_joint": 144,
129
+ "left_middle_pip_joint": 116,
130
+ "left_middle_pitch_joint": 2,
131
+ "left_pinky_dip_joint": 2,
132
+ "left_pinky_mcp_joint": 146,
133
+ "left_pinky_pip_joint": 2,
134
+ "left_ring_dip_joint": 184,
135
+ "left_ring_mcp_joint": 1,
136
+ "left_ring_pip_joint": 143,
137
+ "left_thumb_dip_joint": 6,
138
+ "right_index_dip_joint": 16,
139
+ "right_index_pip_joint": 13,
140
+ "right_index_pitch_joint": 12,
141
+ "right_middle_dip_joint": 180,
142
+ "right_middle_pip_joint": 97,
143
+ "right_middle_pitch_joint": 30,
144
+ "right_pinky_dip_joint": 3,
145
+ "right_pinky_mcp_joint": 33,
146
+ "right_pinky_pip_joint": 40,
147
+ "right_ring_dip_joint": 90,
148
+ "right_ring_mcp_joint": 8,
149
+ "right_ring_pip_joint": 109,
150
+ "right_thumb_dip_joint": 10,
151
+ "right_thumb_mcp_joint": 90,
152
+ "right_thumb_pip_joint": 2
153
+ },
154
+ "joint_limit_hard_episode_events": 440,
155
+ "finger_joint_limit_saturation_counts": {
156
+ "left_middle_pip_joint": 34,
157
+ "left_pinky_dip_joint": 2,
158
+ "left_pinky_mcp_joint": 342,
159
+ "left_ring_dip_joint": 38,
160
+ "left_ring_pip_joint": 228,
161
+ "right_index_pip_joint": 58,
162
+ "right_index_pitch_joint": 22,
163
+ "right_middle_pip_joint": 207,
164
+ "right_middle_pitch_joint": 16,
165
+ "right_pinky_mcp_joint": 28,
166
+ "right_pinky_pip_joint": 722,
167
+ "right_ring_dip_joint": 101,
168
+ "right_ring_pip_joint": 1620,
169
+ "right_thumb_dip_joint": 199,
170
+ "right_thumb_mcp_joint": 5
171
+ },
172
+ "joint_limit_max_excess_by_joint": {
173
+ "left_index_dip_joint": 1.3141128185623074e-08,
174
+ "left_index_pip_joint": 0.000269099953584373,
175
+ "left_index_pitch_joint": 0.0036064807791262865,
176
+ "left_middle_dip_joint": 0.07069453597068787,
177
+ "left_middle_pip_joint": 0.32633844017982483,
178
+ "left_middle_pitch_joint": 0.06922262907028198,
179
+ "left_pinky_dip_joint": 0.23911011219024658,
180
+ "left_pinky_mcp_joint": 0.34184738993644714,
181
+ "left_pinky_pip_joint": 0.18609635531902313,
182
+ "left_ring_dip_joint": 0.24877914786338806,
183
+ "left_ring_mcp_joint": 0.005153956823050976,
184
+ "left_ring_pip_joint": 0.4548375606536865,
185
+ "left_thumb_dip_joint": 0.08335696160793304,
186
+ "left_thumb_yaw_joint": 4.4135878852102906e-05,
187
+ "right_index_dip_joint": 0.07233132421970367,
188
+ "right_index_pip_joint": 0.31643998622894287,
189
+ "right_index_pitch_joint": 0.6898936629295349,
190
+ "right_middle_dip_joint": 0.15184268355369568,
191
+ "right_middle_pip_joint": 0.47695982456207275,
192
+ "right_middle_pitch_joint": 0.3540981709957123,
193
+ "right_pinky_dip_joint": 0.008350329473614693,
194
+ "right_pinky_mcp_joint": 0.24723505973815918,
195
+ "right_pinky_pip_joint": 0.8552443385124207,
196
+ "right_ring_dip_joint": 0.3416437804698944,
197
+ "right_ring_mcp_joint": 0.009221380576491356,
198
+ "right_ring_pip_joint": 0.8993977308273315,
199
+ "right_thumb_dip_joint": 0.9809489250183105,
200
+ "right_thumb_mcp_joint": 0.24519293010234833,
201
+ "right_thumb_pip_joint": 0.0903400182723999,
202
+ "right_thumb_yaw_joint": 6.558656605193391e-05
203
+ },
204
+ "coverage": {
205
+ "hard_limit_source": {
206
+ "available": 50,
207
+ "unavailable": 0,
208
+ "unknown": 0
209
+ },
210
+ "joint_role_manifest": {
211
+ "available": 50,
212
+ "unavailable": 0,
213
+ "unknown": 0
214
+ }
215
+ },
216
+ "hard_limit_sources": {
217
+ "runtime.joint_pos_limits": 50
218
+ },
219
+ "soft_limit_sources": {
220
+ "runtime.soft_joint_pos_limits": 50
221
+ },
222
+ "unknown_runtime_joint_names": [],
223
+ "baseline_checks": {
224
+ "post_home": {
225
+ "checked": 50,
226
+ "hard_violation_episodes": 0
227
+ },
228
+ "post_reset": {
229
+ "checked": 50,
230
+ "hard_violation_episodes": 0
231
+ }
232
+ }
233
+ }
234
+ },
235
+ "robot_motion": {
236
+ "joint_acceleration_rms": 85.0646854431312,
237
+ "joint_effort_rms": 23.33456933501183,
238
+ "joint_jerk_rms": 8725.312609788629,
239
+ "joint_velocity_rms": 1.129253746485012
240
+ },
241
+ "grasp": {
242
+ "kinematic_grasp_stability_index": 0.7559797948886013,
243
+ "mean_kinematic_grasp_stability": 0.40882948136700303,
244
+ "per_object_best_kinematic_grasp_stability": {
245
+ "obj_014_lemon_2": 0.5076502516994731,
246
+ "obj_014_lemon_3": 0.5833973528020631,
247
+ "obj_017_orange_4": 0.3295176562922766,
248
+ "obj_017_orange_5": 0.399483290592393
249
+ },
250
+ "grasp_gsi_diagnostics": {
251
+ "mean_anti_slip_score": 0.6560201954653455,
252
+ "mean_duration_score": 0.4259089904250056,
253
+ "mean_hold_score": 0.36993109828083776,
254
+ "mean_kinematic_stability_score": 0.09167451254719615,
255
+ "mean_legacy_gsi": 0.43670519314638073,
256
+ "mean_lift_score": 0.4404033162607579,
257
+ "mean_motion_stability_score": 0.3571173733286086
258
+ }
259
+ },
260
+ "tool": {
261
+ "tool_selection_accuracy": null,
262
+ "tool_switch_success_rate": null,
263
+ "tool_switch_total": 0,
264
+ "tool_switch_successful": 0
265
+ },
266
+ "runtime": {
267
+ "avg_steps": 1488.14,
268
+ "error_count": 0
269
+ }
270
+ }
271
+ }
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_cov/summary.json ADDED
@@ -0,0 +1,284 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 200700000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 43,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1806,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.86,
50
+ "success_rate_ci": {
51
+ "mean": 0.86,
52
+ "ci_low": 0.76,
53
+ "ci_high": 0.94
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.92
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.499224806201497
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.86,
64
+ "hard_violation_rate": 0.04,
65
+ "drop_rate": 0.04,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 2.7296301351166918e-05
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.88,
74
+ "at_end_success_rate": 0.86,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.86
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.905,
81
+ "current_chain_depth_progress_score": 0.95,
82
+ "latched_chain_depth_progress_score": 0.95,
83
+ "avg_current_chain_depth": 1.9,
84
+ "avg_latched_chain_depth": 1.9
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.8540494948398192,
88
+ "expert_time_s": 20.05,
89
+ "expert_time_step": 1203,
90
+ "avg_steps_to_success": 1409.953488372093,
91
+ "avg_policy_steps_to_success": 470.3255813953488,
92
+ "avg_policy_queries_to_success": 470.3255813953488
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 2.7296301351166918e-05,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_014_lemon_3": 1,
98
+ "drop_to_table_obj_017_orange_5": 1
99
+ },
100
+ "robot_constraint_violation_rate": 1.0,
101
+ "finger_joint_limit_saturation_rate": 1.0,
102
+ "joint_limit_active_step_rate": 0.13076293162276512,
103
+ "finger_joint_limit_saturation_step_rate": 0.033042172785587555,
104
+ "max_joint_limit_excess_rad": 0.8837183117866516,
105
+ "robot_constraint_diagnostics": {
106
+ "joint_limit_soft_counts": {
107
+ "left_index_dip_joint": 20,
108
+ "left_index_pip_joint": 22,
109
+ "left_index_pitch_joint": 202,
110
+ "left_middle_dip_joint": 327,
111
+ "left_middle_pip_joint": 586,
112
+ "left_middle_pitch_joint": 214,
113
+ "left_pinky_mcp_joint": 2048,
114
+ "left_pinky_pip_joint": 64,
115
+ "left_ring_dip_joint": 832,
116
+ "left_ring_mcp_joint": 206,
117
+ "left_ring_pip_joint": 1440,
118
+ "left_thumb_dip_joint": 95,
119
+ "right_index_dip_joint": 39,
120
+ "right_index_pip_joint": 76,
121
+ "right_index_pitch_joint": 124,
122
+ "right_middle_dip_joint": 291,
123
+ "right_middle_pip_joint": 595,
124
+ "right_middle_pitch_joint": 85,
125
+ "right_pinky_dip_joint": 5,
126
+ "right_pinky_mcp_joint": 500,
127
+ "right_pinky_pip_joint": 890,
128
+ "right_ring_dip_joint": 826,
129
+ "right_ring_mcp_joint": 99,
130
+ "right_ring_pip_joint": 2621,
131
+ "right_thumb_dip_joint": 127,
132
+ "right_thumb_mcp_joint": 3608,
133
+ "right_thumb_pip_joint": 13
134
+ },
135
+ "joint_limit_hard_counts": {
136
+ "left_index_dip_joint": 11,
137
+ "left_index_pip_joint": 18,
138
+ "left_index_pitch_joint": 4,
139
+ "left_middle_dip_joint": 152,
140
+ "left_middle_pip_joint": 156,
141
+ "left_middle_pitch_joint": 11,
142
+ "left_pinky_mcp_joint": 156,
143
+ "left_pinky_pip_joint": 6,
144
+ "left_ring_dip_joint": 231,
145
+ "left_ring_mcp_joint": 13,
146
+ "left_ring_pip_joint": 177,
147
+ "left_thumb_dip_joint": 7,
148
+ "right_index_dip_joint": 15,
149
+ "right_index_pip_joint": 14,
150
+ "right_index_pitch_joint": 9,
151
+ "right_middle_dip_joint": 149,
152
+ "right_middle_pip_joint": 94,
153
+ "right_middle_pitch_joint": 22,
154
+ "right_pinky_dip_joint": 2,
155
+ "right_pinky_mcp_joint": 33,
156
+ "right_pinky_pip_joint": 40,
157
+ "right_ring_dip_joint": 76,
158
+ "right_ring_mcp_joint": 10,
159
+ "right_ring_pip_joint": 112,
160
+ "right_thumb_dip_joint": 10,
161
+ "right_thumb_mcp_joint": 90,
162
+ "right_thumb_pip_joint": 1
163
+ },
164
+ "joint_limit_hard_episode_events": 432,
165
+ "finger_joint_limit_saturation_counts": {
166
+ "left_index_pitch_joint": 135,
167
+ "left_middle_pip_joint": 72,
168
+ "left_middle_pitch_joint": 111,
169
+ "left_pinky_mcp_joint": 476,
170
+ "left_pinky_pip_joint": 1,
171
+ "left_ring_dip_joint": 2,
172
+ "left_ring_mcp_joint": 110,
173
+ "left_ring_pip_joint": 228,
174
+ "left_thumb_dip_joint": 54,
175
+ "right_index_pip_joint": 27,
176
+ "right_index_pitch_joint": 84,
177
+ "right_middle_pip_joint": 113,
178
+ "right_pinky_mcp_joint": 170,
179
+ "right_pinky_pip_joint": 492,
180
+ "right_ring_dip_joint": 145,
181
+ "right_ring_mcp_joint": 23,
182
+ "right_ring_pip_joint": 1455,
183
+ "right_thumb_dip_joint": 80,
184
+ "right_thumb_mcp_joint": 43
185
+ },
186
+ "joint_limit_max_excess_by_joint": {
187
+ "left_index_dip_joint": 0.05909399315714836,
188
+ "left_index_pip_joint": 0.07746269553899765,
189
+ "left_index_pitch_joint": 0.548991858959198,
190
+ "left_middle_dip_joint": 0.09415020048618317,
191
+ "left_middle_pip_joint": 0.43800756335258484,
192
+ "left_middle_pitch_joint": 0.614388644695282,
193
+ "left_pinky_mcp_joint": 0.6025779247283936,
194
+ "left_pinky_pip_joint": 0.17798109352588654,
195
+ "left_ring_dip_joint": 0.19265790283679962,
196
+ "left_ring_mcp_joint": 0.5234812498092651,
197
+ "left_ring_pip_joint": 0.5063444375991821,
198
+ "left_thumb_dip_joint": 0.47156935930252075,
199
+ "left_thumb_yaw_joint": 2.8825117624364793e-05,
200
+ "right_index_dip_joint": 0.06651128083467484,
201
+ "right_index_pip_joint": 0.36407333612442017,
202
+ "right_index_pitch_joint": 0.7054418325424194,
203
+ "right_middle_dip_joint": 0.11492473632097244,
204
+ "right_middle_pip_joint": 0.5587264895439148,
205
+ "right_middle_pitch_joint": 0.11083249002695084,
206
+ "right_pinky_dip_joint": 0.03582749515771866,
207
+ "right_pinky_mcp_joint": 0.7528288960456848,
208
+ "right_pinky_pip_joint": 0.8389469981193542,
209
+ "right_ring_dip_joint": 0.3736472427845001,
210
+ "right_ring_mcp_joint": 0.21354295313358307,
211
+ "right_ring_pip_joint": 0.8837183117866516,
212
+ "right_thumb_dip_joint": 0.6680051684379578,
213
+ "right_thumb_mcp_joint": 0.17633181810379028,
214
+ "right_thumb_pip_joint": 0.06911215931177139,
215
+ "right_thumb_yaw_joint": 6.988152745179832e-05
216
+ },
217
+ "coverage": {
218
+ "hard_limit_source": {
219
+ "available": 50,
220
+ "unavailable": 0,
221
+ "unknown": 0
222
+ },
223
+ "joint_role_manifest": {
224
+ "available": 50,
225
+ "unavailable": 0,
226
+ "unknown": 0
227
+ }
228
+ },
229
+ "hard_limit_sources": {
230
+ "runtime.joint_pos_limits": 50
231
+ },
232
+ "soft_limit_sources": {
233
+ "runtime.soft_joint_pos_limits": 50
234
+ },
235
+ "unknown_runtime_joint_names": [],
236
+ "baseline_checks": {
237
+ "post_home": {
238
+ "checked": 50,
239
+ "hard_violation_episodes": 0
240
+ },
241
+ "post_reset": {
242
+ "checked": 50,
243
+ "hard_violation_episodes": 0
244
+ }
245
+ }
246
+ }
247
+ },
248
+ "robot_motion": {
249
+ "joint_acceleration_rms": 85.62071788145099,
250
+ "joint_effort_rms": 23.84921944664871,
251
+ "joint_jerk_rms": 8770.851225563985,
252
+ "joint_velocity_rms": 1.1522906550163983
253
+ },
254
+ "grasp": {
255
+ "kinematic_grasp_stability_index": 0.7814069179654677,
256
+ "mean_kinematic_grasp_stability": 0.45397072841776426,
257
+ "per_object_best_kinematic_grasp_stability": {
258
+ "obj_014_lemon_2": 0.5633900533981074,
259
+ "obj_014_lemon_3": 0.5507761716396838,
260
+ "obj_017_orange_4": 0.4954747146567751,
261
+ "obj_017_orange_5": 0.36737843641062007
262
+ },
263
+ "grasp_gsi_diagnostics": {
264
+ "mean_anti_slip_score": 0.7079281332869088,
265
+ "mean_duration_score": 0.3935085356100021,
266
+ "mean_hold_score": 0.35504850588567327,
267
+ "mean_kinematic_stability_score": 0.118169020369024,
268
+ "mean_legacy_gsi": 0.43335109787172854,
269
+ "mean_lift_score": 0.42591376680475057,
270
+ "mean_motion_stability_score": 0.3458257178143905
271
+ }
272
+ },
273
+ "tool": {
274
+ "tool_selection_accuracy": null,
275
+ "tool_switch_success_rate": null,
276
+ "tool_switch_total": 0,
277
+ "tool_switch_successful": 0
278
+ },
279
+ "runtime": {
280
+ "avg_steps": 1465.4,
281
+ "error_count": 0
282
+ }
283
+ }
284
+ }
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/inv_only/summary.json ADDED
@@ -0,0 +1,293 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100700000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 44,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1806,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.88,
50
+ "success_rate_ci": {
51
+ "mean": 0.88,
52
+ "ci_low": 0.78,
53
+ "ci_high": 0.96
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.945
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.847727272727198
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.86,
64
+ "hard_violation_rate": 0.04,
65
+ "drop_rate": 0.04,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 2.7102474455917824e-05
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.9,
74
+ "at_end_success_rate": 0.88,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.88
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.94,
81
+ "current_chain_depth_progress_score": 0.97,
82
+ "latched_chain_depth_progress_score": 0.97,
83
+ "avg_current_chain_depth": 1.94,
84
+ "avg_latched_chain_depth": 1.94
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.8428982359879864,
88
+ "expert_time_s": 20.05,
89
+ "expert_time_step": 1203,
90
+ "avg_steps_to_success": 1430.8636363636363,
91
+ "avg_policy_steps_to_success": 477.29545454545456,
92
+ "avg_policy_queries_to_success": 477.29545454545456
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 2.7102474455917824e-05,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_014_lemon_3": 1,
98
+ "drop_to_table_obj_017_orange_5": 1
99
+ },
100
+ "robot_constraint_violation_rate": 1.0,
101
+ "finger_joint_limit_saturation_rate": 1.0,
102
+ "joint_limit_active_step_rate": 0.13262595875003388,
103
+ "finger_joint_limit_saturation_step_rate": 0.0374014147491666,
104
+ "max_joint_limit_excess_rad": 1.207065463066101,
105
+ "robot_constraint_diagnostics": {
106
+ "joint_limit_soft_counts": {
107
+ "left_index_dip_joint": 12,
108
+ "left_index_pip_joint": 43,
109
+ "left_index_pitch_joint": 145,
110
+ "left_middle_dip_joint": 338,
111
+ "left_middle_pip_joint": 651,
112
+ "left_middle_pitch_joint": 16,
113
+ "left_pinky_dip_joint": 12,
114
+ "left_pinky_mcp_joint": 1873,
115
+ "left_pinky_pip_joint": 122,
116
+ "left_ring_dip_joint": 842,
117
+ "left_ring_mcp_joint": 61,
118
+ "left_ring_pip_joint": 1343,
119
+ "left_thumb_dip_joint": 125,
120
+ "left_thumb_pip_joint": 22,
121
+ "right_index_dip_joint": 58,
122
+ "right_index_pip_joint": 164,
123
+ "right_index_pitch_joint": 62,
124
+ "right_middle_dip_joint": 474,
125
+ "right_middle_pip_joint": 985,
126
+ "right_middle_pitch_joint": 109,
127
+ "right_pinky_dip_joint": 3,
128
+ "right_pinky_mcp_joint": 394,
129
+ "right_pinky_pip_joint": 1193,
130
+ "right_ring_dip_joint": 970,
131
+ "right_ring_mcp_joint": 54,
132
+ "right_ring_pip_joint": 2936,
133
+ "right_thumb_dip_joint": 328,
134
+ "right_thumb_mcp_joint": 4049,
135
+ "right_thumb_pip_joint": 2,
136
+ "right_thumb_yaw_joint": 6
137
+ },
138
+ "joint_limit_hard_counts": {
139
+ "left_index_dip_joint": 3,
140
+ "left_index_pip_joint": 3,
141
+ "left_index_pitch_joint": 4,
142
+ "left_middle_dip_joint": 147,
143
+ "left_middle_pip_joint": 112,
144
+ "left_middle_pitch_joint": 1,
145
+ "left_pinky_dip_joint": 2,
146
+ "left_pinky_mcp_joint": 128,
147
+ "left_pinky_pip_joint": 8,
148
+ "left_ring_dip_joint": 217,
149
+ "left_ring_mcp_joint": 3,
150
+ "left_ring_pip_joint": 174,
151
+ "left_thumb_dip_joint": 8,
152
+ "left_thumb_pip_joint": 1,
153
+ "right_index_dip_joint": 14,
154
+ "right_index_pip_joint": 17,
155
+ "right_index_pitch_joint": 12,
156
+ "right_middle_dip_joint": 204,
157
+ "right_middle_pip_joint": 112,
158
+ "right_middle_pitch_joint": 16,
159
+ "right_pinky_dip_joint": 2,
160
+ "right_pinky_mcp_joint": 54,
161
+ "right_pinky_pip_joint": 45,
162
+ "right_ring_dip_joint": 92,
163
+ "right_ring_mcp_joint": 4,
164
+ "right_ring_pip_joint": 106,
165
+ "right_thumb_dip_joint": 10,
166
+ "right_thumb_mcp_joint": 93,
167
+ "right_thumb_pip_joint": 1,
168
+ "right_thumb_yaw_joint": 1
169
+ },
170
+ "joint_limit_hard_episode_events": 419,
171
+ "finger_joint_limit_saturation_counts": {
172
+ "left_index_pip_joint": 1,
173
+ "left_index_pitch_joint": 72,
174
+ "left_middle_pip_joint": 126,
175
+ "left_pinky_mcp_joint": 454,
176
+ "left_pinky_pip_joint": 40,
177
+ "left_ring_dip_joint": 4,
178
+ "left_ring_pip_joint": 316,
179
+ "left_thumb_dip_joint": 91,
180
+ "left_thumb_pip_joint": 3,
181
+ "right_index_pip_joint": 65,
182
+ "right_index_pitch_joint": 7,
183
+ "right_middle_pip_joint": 171,
184
+ "right_middle_pitch_joint": 26,
185
+ "right_pinky_mcp_joint": 36,
186
+ "right_pinky_pip_joint": 789,
187
+ "right_ring_dip_joint": 174,
188
+ "right_ring_mcp_joint": 44,
189
+ "right_ring_pip_joint": 1826,
190
+ "right_thumb_dip_joint": 288,
191
+ "right_thumb_mcp_joint": 70
192
+ },
193
+ "joint_limit_max_excess_by_joint": {
194
+ "left_index_dip_joint": 0.047060526907444,
195
+ "left_index_pip_joint": 0.144338458776474,
196
+ "left_index_pitch_joint": 0.7024231553077698,
197
+ "left_middle_dip_joint": 0.09270153194665909,
198
+ "left_middle_pip_joint": 0.5002774000167847,
199
+ "left_middle_pitch_joint": 0.13047239184379578,
200
+ "left_pinky_dip_joint": 0.052039504051208496,
201
+ "left_pinky_mcp_joint": 0.7251112461090088,
202
+ "left_pinky_pip_joint": 0.6358579993247986,
203
+ "left_ring_dip_joint": 0.19424408674240112,
204
+ "left_ring_mcp_joint": 0.146920308470726,
205
+ "left_ring_pip_joint": 0.4143218994140625,
206
+ "left_thumb_dip_joint": 1.207065463066101,
207
+ "left_thumb_pip_joint": 0.17888569831848145,
208
+ "left_thumb_yaw_joint": 3.5364038922125474e-05,
209
+ "right_index_dip_joint": 0.14194084703922272,
210
+ "right_index_pip_joint": 0.38331130146980286,
211
+ "right_index_pitch_joint": 0.14838723838329315,
212
+ "right_middle_dip_joint": 0.14147964119911194,
213
+ "right_middle_pip_joint": 0.6018909811973572,
214
+ "right_middle_pitch_joint": 0.27007946372032166,
215
+ "right_pinky_dip_joint": 0.015862150117754936,
216
+ "right_pinky_mcp_joint": 0.418827623128891,
217
+ "right_pinky_pip_joint": 0.7498195171356201,
218
+ "right_ring_dip_joint": 0.3484480082988739,
219
+ "right_ring_mcp_joint": 0.49124646186828613,
220
+ "right_ring_pip_joint": 0.7407041788101196,
221
+ "right_thumb_dip_joint": 1.1835674047470093,
222
+ "right_thumb_mcp_joint": 0.12698711454868317,
223
+ "right_thumb_pip_joint": 0.016096115112304688,
224
+ "right_thumb_yaw_joint": 0.02201257459819317
225
+ },
226
+ "coverage": {
227
+ "hard_limit_source": {
228
+ "available": 50,
229
+ "unavailable": 0,
230
+ "unknown": 0
231
+ },
232
+ "joint_role_manifest": {
233
+ "available": 50,
234
+ "unavailable": 0,
235
+ "unknown": 0
236
+ }
237
+ },
238
+ "hard_limit_sources": {
239
+ "runtime.joint_pos_limits": 50
240
+ },
241
+ "soft_limit_sources": {
242
+ "runtime.soft_joint_pos_limits": 50
243
+ },
244
+ "unknown_runtime_joint_names": [],
245
+ "baseline_checks": {
246
+ "post_home": {
247
+ "checked": 50,
248
+ "hard_violation_episodes": 0
249
+ },
250
+ "post_reset": {
251
+ "checked": 50,
252
+ "hard_violation_episodes": 0
253
+ }
254
+ }
255
+ }
256
+ },
257
+ "robot_motion": {
258
+ "joint_acceleration_rms": 87.25096926386014,
259
+ "joint_effort_rms": 24.249471739258222,
260
+ "joint_jerk_rms": 8960.659012657421,
261
+ "joint_velocity_rms": 1.1603185682962887
262
+ },
263
+ "grasp": {
264
+ "kinematic_grasp_stability_index": 0.7083304399695026,
265
+ "mean_kinematic_grasp_stability": 0.3892406840777212,
266
+ "per_object_best_kinematic_grasp_stability": {
267
+ "obj_014_lemon_2": 0.5225279013463037,
268
+ "obj_014_lemon_3": 0.5052918247721466,
269
+ "obj_017_orange_4": 0.33891392175278057,
270
+ "obj_017_orange_5": 0.37591922130243094
271
+ },
272
+ "grasp_gsi_diagnostics": {
273
+ "mean_anti_slip_score": 0.6356095096731629,
274
+ "mean_duration_score": 0.43040281139038106,
275
+ "mean_hold_score": 0.380413005572529,
276
+ "mean_kinematic_stability_score": 0.08746276174623271,
277
+ "mean_legacy_gsi": 0.43993698667653414,
278
+ "mean_lift_score": 0.449880907603702,
279
+ "mean_motion_stability_score": 0.3678163101567465
280
+ }
281
+ },
282
+ "tool": {
283
+ "tool_selection_accuracy": null,
284
+ "tool_switch_success_rate": null,
285
+ "tool_switch_total": 0,
286
+ "tool_switch_successful": 0
287
+ },
288
+ "runtime": {
289
+ "avg_steps": 1475.88,
290
+ "error_count": 0
291
+ }
292
+ }
293
+ }
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/07_citrus_plate_loading/none/summary.json ADDED
@@ -0,0 +1,284 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100700000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 45,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1806,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.9,
50
+ "success_rate_ci": {
51
+ "mean": 0.9,
52
+ "ci_low": 0.82,
53
+ "ci_high": 0.98
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.98
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.956296296296216
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.9,
64
+ "hard_violation_rate": 0.04,
65
+ "drop_rate": 0.04,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 2.7132624267419144e-05
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.96,
74
+ "at_end_success_rate": 0.9,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.9
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.965,
81
+ "current_chain_depth_progress_score": 0.99,
82
+ "latched_chain_depth_progress_score": 0.99,
83
+ "avg_current_chain_depth": 1.98,
84
+ "avg_latched_chain_depth": 1.98
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.8395180862761689,
88
+ "expert_time_s": 20.05,
89
+ "expert_time_step": 1203,
90
+ "avg_steps_to_success": 1437.3777777777777,
91
+ "avg_policy_steps_to_success": 479.53333333333336,
92
+ "avg_policy_queries_to_success": 479.53333333333336
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 2.7132624267419144e-05,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_017_orange_5": 2
98
+ },
99
+ "robot_constraint_violation_rate": 1.0,
100
+ "finger_joint_limit_saturation_rate": 1.0,
101
+ "joint_limit_active_step_rate": 0.1108232038202735,
102
+ "finger_joint_limit_saturation_step_rate": 0.03196223138701975,
103
+ "max_joint_limit_excess_rad": 0.8491209745407104,
104
+ "robot_constraint_diagnostics": {
105
+ "joint_limit_soft_counts": {
106
+ "left_index_dip_joint": 16,
107
+ "left_index_pip_joint": 20,
108
+ "left_index_pitch_joint": 42,
109
+ "left_middle_dip_joint": 361,
110
+ "left_middle_pip_joint": 602,
111
+ "left_middle_pitch_joint": 49,
112
+ "left_pinky_dip_joint": 36,
113
+ "left_pinky_mcp_joint": 1772,
114
+ "left_pinky_pip_joint": 30,
115
+ "left_ring_dip_joint": 765,
116
+ "left_ring_pip_joint": 1283,
117
+ "left_thumb_dip_joint": 50,
118
+ "right_index_dip_joint": 68,
119
+ "right_index_pip_joint": 200,
120
+ "right_index_pitch_joint": 4,
121
+ "right_middle_dip_joint": 395,
122
+ "right_middle_pip_joint": 873,
123
+ "right_middle_pitch_joint": 57,
124
+ "right_pinky_dip_joint": 4,
125
+ "right_pinky_mcp_joint": 411,
126
+ "right_pinky_pip_joint": 1048,
127
+ "right_ring_dip_joint": 929,
128
+ "right_ring_mcp_joint": 66,
129
+ "right_ring_pip_joint": 2713,
130
+ "right_thumb_dip_joint": 223,
131
+ "right_thumb_mcp_joint": 2804,
132
+ "right_thumb_pip_joint": 11,
133
+ "right_thumb_yaw_joint": 67
134
+ },
135
+ "joint_limit_hard_counts": {
136
+ "left_index_dip_joint": 5,
137
+ "left_index_pip_joint": 5,
138
+ "left_index_pitch_joint": 1,
139
+ "left_middle_dip_joint": 156,
140
+ "left_middle_pip_joint": 111,
141
+ "left_middle_pitch_joint": 2,
142
+ "left_pinky_dip_joint": 1,
143
+ "left_pinky_mcp_joint": 114,
144
+ "left_pinky_pip_joint": 2,
145
+ "left_ring_dip_joint": 171,
146
+ "left_ring_pip_joint": 123,
147
+ "left_thumb_dip_joint": 11,
148
+ "right_index_dip_joint": 19,
149
+ "right_index_pip_joint": 13,
150
+ "right_index_pitch_joint": 8,
151
+ "right_middle_dip_joint": 151,
152
+ "right_middle_pip_joint": 81,
153
+ "right_middle_pitch_joint": 22,
154
+ "right_pinky_dip_joint": 5,
155
+ "right_pinky_mcp_joint": 36,
156
+ "right_pinky_pip_joint": 43,
157
+ "right_ring_dip_joint": 86,
158
+ "right_ring_mcp_joint": 7,
159
+ "right_ring_pip_joint": 98,
160
+ "right_thumb_dip_joint": 11,
161
+ "right_thumb_mcp_joint": 82,
162
+ "right_thumb_pip_joint": 2,
163
+ "right_thumb_yaw_joint": 2
164
+ },
165
+ "joint_limit_hard_episode_events": 345,
166
+ "finger_joint_limit_saturation_counts": {
167
+ "left_index_pitch_joint": 23,
168
+ "left_middle_pip_joint": 66,
169
+ "left_middle_pitch_joint": 20,
170
+ "left_pinky_dip_joint": 30,
171
+ "left_pinky_mcp_joint": 453,
172
+ "left_pinky_pip_joint": 10,
173
+ "left_ring_dip_joint": 18,
174
+ "left_ring_pip_joint": 324,
175
+ "right_index_pip_joint": 137,
176
+ "right_middle_pip_joint": 309,
177
+ "right_pinky_mcp_joint": 86,
178
+ "right_pinky_pip_joint": 639,
179
+ "right_ring_dip_joint": 111,
180
+ "right_ring_mcp_joint": 19,
181
+ "right_ring_pip_joint": 1555,
182
+ "right_thumb_dip_joint": 187,
183
+ "right_thumb_mcp_joint": 14
184
+ },
185
+ "joint_limit_max_excess_by_joint": {
186
+ "left_index_dip_joint": 0.05969364941120148,
187
+ "left_index_pip_joint": 0.10146134346723557,
188
+ "left_index_pitch_joint": 0.27332088351249695,
189
+ "left_middle_dip_joint": 0.08895663917064667,
190
+ "left_middle_pip_joint": 0.32017526030540466,
191
+ "left_middle_pitch_joint": 0.216016948223114,
192
+ "left_pinky_dip_joint": 0.5906717777252197,
193
+ "left_pinky_mcp_joint": 0.4250677525997162,
194
+ "left_pinky_pip_joint": 0.2856070399284363,
195
+ "left_ring_dip_joint": 0.2037249207496643,
196
+ "left_ring_mcp_joint": 6.32929295534268e-05,
197
+ "left_ring_pip_joint": 0.47824564576148987,
198
+ "left_thumb_dip_joint": 0.10979275405406952,
199
+ "left_thumb_yaw_joint": 4.2589879740262404e-05,
200
+ "right_index_dip_joint": 0.0806288942694664,
201
+ "right_index_pip_joint": 0.38403570652008057,
202
+ "right_index_pitch_joint": 0.02323862910270691,
203
+ "right_middle_dip_joint": 0.14636723697185516,
204
+ "right_middle_pip_joint": 0.533318817615509,
205
+ "right_middle_pitch_joint": 0.0602438822388649,
206
+ "right_pinky_dip_joint": 0.03488396480679512,
207
+ "right_pinky_mcp_joint": 0.3615626096725464,
208
+ "right_pinky_pip_joint": 0.6839773654937744,
209
+ "right_ring_dip_joint": 0.3320575952529907,
210
+ "right_ring_mcp_joint": 0.22439317405223846,
211
+ "right_ring_pip_joint": 0.7183210253715515,
212
+ "right_thumb_dip_joint": 0.8491209745407104,
213
+ "right_thumb_mcp_joint": 0.2440502792596817,
214
+ "right_thumb_pip_joint": 0.06143748760223389,
215
+ "right_thumb_yaw_joint": 0.1361614316701889
216
+ },
217
+ "coverage": {
218
+ "hard_limit_source": {
219
+ "available": 50,
220
+ "unavailable": 0,
221
+ "unknown": 0
222
+ },
223
+ "joint_role_manifest": {
224
+ "available": 50,
225
+ "unavailable": 0,
226
+ "unknown": 0
227
+ }
228
+ },
229
+ "hard_limit_sources": {
230
+ "runtime.joint_pos_limits": 50
231
+ },
232
+ "soft_limit_sources": {
233
+ "runtime.soft_joint_pos_limits": 50
234
+ },
235
+ "unknown_runtime_joint_names": [],
236
+ "baseline_checks": {
237
+ "post_home": {
238
+ "checked": 50,
239
+ "hard_violation_episodes": 0
240
+ },
241
+ "post_reset": {
242
+ "checked": 50,
243
+ "hard_violation_episodes": 0
244
+ }
245
+ }
246
+ }
247
+ },
248
+ "robot_motion": {
249
+ "joint_acceleration_rms": 84.28350292588628,
250
+ "joint_effort_rms": 23.483385449163777,
251
+ "joint_jerk_rms": 8644.790004008662,
252
+ "joint_velocity_rms": 1.1278087556541898
253
+ },
254
+ "grasp": {
255
+ "kinematic_grasp_stability_index": 0.7879875122333718,
256
+ "mean_kinematic_grasp_stability": 0.47321701922321135,
257
+ "per_object_best_kinematic_grasp_stability": {
258
+ "obj_014_lemon_2": 0.6033152260159483,
259
+ "obj_014_lemon_3": 0.576670326725777,
260
+ "obj_017_orange_4": 0.4157642701629602,
261
+ "obj_017_orange_5": 0.3808947532938913
262
+ },
263
+ "grasp_gsi_diagnostics": {
264
+ "mean_anti_slip_score": 0.6872853288378528,
265
+ "mean_duration_score": 0.431157119146507,
266
+ "mean_hold_score": 0.3854486122735531,
267
+ "mean_kinematic_stability_score": 0.12692451888938236,
268
+ "mean_legacy_gsi": 0.454838655405977,
269
+ "mean_lift_score": 0.4586264595852593,
270
+ "mean_motion_stability_score": 0.3761122067152371
271
+ }
272
+ },
273
+ "tool": {
274
+ "tool_selection_accuracy": null,
275
+ "tool_switch_success_rate": null,
276
+ "tool_switch_total": 0,
277
+ "tool_switch_successful": 0
278
+ },
279
+ "runtime": {
280
+ "avg_steps": 1474.24,
281
+ "error_count": 0
282
+ }
283
+ }
284
+ }
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/cov_only/summary.json ADDED
@@ -0,0 +1,236 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100800000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 46,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3432,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.92,
50
+ "success_rate_ci": {
51
+ "mean": 0.92,
52
+ "ci_low": 0.84,
53
+ "ci_high": 0.98
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.94
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 29.41159420289816
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.72,
64
+ "hard_violation_rate": 0.22,
65
+ "drop_rate": 0.22,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.00011590660035404198
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.92,
74
+ "at_end_success_rate": 0.92,
75
+ "task_family_success_rate": {
76
+ "pour_tilt": 0.92
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.0,
81
+ "current_chain_depth_progress_score": 0.0,
82
+ "latched_chain_depth_progress_score": 0.96,
83
+ "avg_current_chain_depth": 0,
84
+ "avg_latched_chain_depth": 0.96
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.298647916184152,
88
+ "expert_time_s": 38.13333333333333,
89
+ "expert_time_step": 2288,
90
+ "avg_steps_to_success": 1764.695652173913,
91
+ "avg_policy_steps_to_success": 588.5652173913044,
92
+ "avg_policy_queries_to_success": 588.5652173913044
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00011590660035404198,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_172_soy_sauce_1": 11
98
+ },
99
+ "robot_constraint_violation_rate": 1.0,
100
+ "finger_joint_limit_saturation_rate": 0.22,
101
+ "joint_limit_active_step_rate": 0.029219000252887126,
102
+ "finger_joint_limit_saturation_step_rate": 0.0012012138582146168,
103
+ "max_joint_limit_excess_rad": 0.5132278203964233,
104
+ "robot_constraint_diagnostics": {
105
+ "joint_limit_soft_counts": {
106
+ "left_hand_Thumb_Flexion": 265,
107
+ "left_hand_j3": 66,
108
+ "right_hand_Index_Finger_Distal": 375,
109
+ "right_hand_Index_Finger_Proximal": 60,
110
+ "right_hand_Middle_Finger_Distal": 405,
111
+ "right_hand_Middle_Finger_Proximal": 471,
112
+ "right_hand_Pinky": 38,
113
+ "right_hand_Ring_Finger": 493,
114
+ "right_hand_Thumb_Flexion": 1649,
115
+ "right_hand_j12": 38,
116
+ "right_hand_j13": 26,
117
+ "right_hand_j3": 56
118
+ },
119
+ "joint_limit_hard_counts": {
120
+ "left_hand_Thumb_Flexion": 6,
121
+ "left_hand_Thumb_Opposition": 3,
122
+ "left_hand_index_spread": 1,
123
+ "left_hand_j3": 3,
124
+ "right_hand_Index_Finger_Distal": 99,
125
+ "right_hand_Index_Finger_Proximal": 23,
126
+ "right_hand_Middle_Finger_Distal": 134,
127
+ "right_hand_Middle_Finger_Proximal": 187,
128
+ "right_hand_Pinky": 15,
129
+ "right_hand_Ring_Finger": 92,
130
+ "right_hand_Thumb_Flexion": 124,
131
+ "right_hand_j12": 31,
132
+ "right_hand_j13": 13,
133
+ "right_hand_j3": 18
134
+ },
135
+ "joint_limit_hard_episode_events": 425,
136
+ "finger_joint_limit_saturation_counts": {
137
+ "left_hand_j3": 24,
138
+ "right_hand_Middle_Finger_Proximal": 28,
139
+ "right_hand_Ring_Finger": 88
140
+ },
141
+ "joint_limit_max_excess_by_joint": {
142
+ "left_hand_Finger_Spread": 0.0005879159434698522,
143
+ "left_hand_Index_Finger_Distal": 0.0002469142200425267,
144
+ "left_hand_Middle_Finger_Proximal": 0.0012230873107910156,
145
+ "left_hand_Thumb_Flexion": 0.09769544750452042,
146
+ "left_hand_Thumb_Opposition": 0.00774567574262619,
147
+ "left_hand_index_spread": 0.006606868468225002,
148
+ "left_hand_j14": 4.389960668049753e-05,
149
+ "left_hand_j3": 0.41080188751220703,
150
+ "left_hand_j4": 1.12001066199241e-08,
151
+ "left_hand_j5": 0.004091720096766949,
152
+ "left_hand_ring_spread": 0.001642384915612638,
153
+ "right_hand_Finger_Spread": 0.0010961228981614113,
154
+ "right_hand_Index_Finger_Distal": 0.15557320415973663,
155
+ "right_hand_Index_Finger_Proximal": 0.17341288924217224,
156
+ "right_hand_Middle_Finger_Distal": 0.20617221295833588,
157
+ "right_hand_Middle_Finger_Proximal": 0.4011339843273163,
158
+ "right_hand_Pinky": 0.133636012673378,
159
+ "right_hand_Ring_Finger": 0.5132278203964233,
160
+ "right_hand_Thumb_Flexion": 0.07913890480995178,
161
+ "right_hand_index_spread": 0.0011869793524965644,
162
+ "right_hand_j12": 0.11219929158687592,
163
+ "right_hand_j13": 0.1142551526427269,
164
+ "right_hand_j14": 0.0010773956310003996,
165
+ "right_hand_j15": 0.00017703449702821672,
166
+ "right_hand_j16": 7.669040496693924e-05,
167
+ "right_hand_j17": 4.359794547781348e-05,
168
+ "right_hand_j3": 0.08059211820363998,
169
+ "right_hand_ring_spread": 0.002806097036227584
170
+ },
171
+ "coverage": {
172
+ "hard_limit_source": {
173
+ "available": 50,
174
+ "unavailable": 0,
175
+ "unknown": 0
176
+ },
177
+ "joint_role_manifest": {
178
+ "available": 50,
179
+ "unavailable": 0,
180
+ "unknown": 0
181
+ }
182
+ },
183
+ "hard_limit_sources": {
184
+ "runtime.joint_pos_limits": 50
185
+ },
186
+ "soft_limit_sources": {
187
+ "runtime.soft_joint_pos_limits": 50
188
+ },
189
+ "unknown_runtime_joint_names": [],
190
+ "baseline_checks": {
191
+ "post_home": {
192
+ "checked": 50,
193
+ "hard_violation_episodes": 0
194
+ },
195
+ "post_reset": {
196
+ "checked": 50,
197
+ "hard_violation_episodes": 0
198
+ }
199
+ }
200
+ }
201
+ },
202
+ "robot_motion": {
203
+ "joint_acceleration_rms": 12.362786889150172,
204
+ "joint_effort_rms": 23.803155663464892,
205
+ "joint_jerk_rms": 1197.858248764796,
206
+ "joint_velocity_rms": 0.34491669678404147
207
+ },
208
+ "grasp": {
209
+ "kinematic_grasp_stability_index": 0.9217487830943563,
210
+ "mean_kinematic_grasp_stability": 0.41485641383604166,
211
+ "per_object_best_kinematic_grasp_stability": {
212
+ "obj_146_olive_oil_1": 0.8890348283554951,
213
+ "obj_172_soy_sauce_1": 0.8305182193069449
214
+ },
215
+ "grasp_gsi_diagnostics": {
216
+ "mean_anti_slip_score": 0.9691492413899511,
217
+ "mean_duration_score": 0.3770936366126462,
218
+ "mean_hold_score": 0.12244862961491712,
219
+ "mean_kinematic_stability_score": 0.07704816125279253,
220
+ "mean_legacy_gsi": 0.34048221964265596,
221
+ "mean_lift_score": 0.17680622248440592,
222
+ "mean_motion_stability_score": 0.1085091684024342
223
+ }
224
+ },
225
+ "tool": {
226
+ "tool_selection_accuracy": 1.0,
227
+ "tool_switch_success_rate": 0.0,
228
+ "tool_switch_total": 243,
229
+ "tool_switch_successful": 0
230
+ },
231
+ "runtime": {
232
+ "avg_steps": 1898.08,
233
+ "error_count": 0
234
+ }
235
+ }
236
+ }
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_cov/summary.json ADDED
@@ -0,0 +1,239 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 200800000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 42,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3432,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.84,
50
+ "success_rate_ci": {
51
+ "mean": 0.84,
52
+ "ci_low": 0.72,
53
+ "ci_high": 0.94
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.9
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 29.718253968253563
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.64,
64
+ "hard_violation_rate": 0.26,
65
+ "drop_rate": 0.26,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.00012702010826021534
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.84,
74
+ "at_end_success_rate": 0.84,
75
+ "task_family_success_rate": {
76
+ "pour_tilt": 0.84
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.0,
81
+ "current_chain_depth_progress_score": 0.0,
82
+ "latched_chain_depth_progress_score": 0.96,
83
+ "avg_current_chain_depth": 0,
84
+ "avg_latched_chain_depth": 0.96
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.2875608278400994,
88
+ "expert_time_s": 38.13333333333333,
89
+ "expert_time_step": 2288,
90
+ "avg_steps_to_success": 1783.095238095238,
91
+ "avg_policy_steps_to_success": 594.7380952380952,
92
+ "avg_policy_queries_to_success": 594.7380952380952
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00012702010826021534,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_146_olive_oil_1": 2,
98
+ "drop_to_table_obj_172_soy_sauce_1": 11
99
+ },
100
+ "robot_constraint_violation_rate": 1.0,
101
+ "finger_joint_limit_saturation_rate": 0.16,
102
+ "joint_limit_active_step_rate": 0.021007171750727922,
103
+ "finger_joint_limit_saturation_step_rate": 0.0005569343208332519,
104
+ "max_joint_limit_excess_rad": 0.358076274394989,
105
+ "robot_constraint_diagnostics": {
106
+ "joint_limit_soft_counts": {
107
+ "left_hand_Thumb_Flexion": 120,
108
+ "left_hand_j3": 22,
109
+ "right_hand_Index_Finger_Distal": 426,
110
+ "right_hand_Index_Finger_Proximal": 22,
111
+ "right_hand_Middle_Finger_Distal": 463,
112
+ "right_hand_Middle_Finger_Proximal": 542,
113
+ "right_hand_Pinky": 33,
114
+ "right_hand_Ring_Finger": 589,
115
+ "right_hand_Thumb_Flexion": 1179,
116
+ "right_hand_j12": 12,
117
+ "right_hand_j13": 24,
118
+ "right_hand_j3": 24
119
+ },
120
+ "joint_limit_hard_counts": {
121
+ "left_hand_Thumb_Flexion": 2,
122
+ "left_hand_Thumb_Opposition": 2,
123
+ "left_hand_index_spread": 1,
124
+ "left_hand_j3": 3,
125
+ "right_hand_Index_Finger_Distal": 106,
126
+ "right_hand_Index_Finger_Proximal": 11,
127
+ "right_hand_Middle_Finger_Distal": 137,
128
+ "right_hand_Middle_Finger_Proximal": 180,
129
+ "right_hand_Pinky": 14,
130
+ "right_hand_Ring_Finger": 94,
131
+ "right_hand_Thumb_Flexion": 131,
132
+ "right_hand_j12": 25,
133
+ "right_hand_j13": 17,
134
+ "right_hand_j3": 14
135
+ },
136
+ "joint_limit_hard_episode_events": 410,
137
+ "finger_joint_limit_saturation_counts": {
138
+ "left_hand_Thumb_Flexion": 7,
139
+ "right_hand_Middle_Finger_Proximal": 9,
140
+ "right_hand_Ring_Finger": 47,
141
+ "right_hand_Thumb_Flexion": 2
142
+ },
143
+ "joint_limit_max_excess_by_joint": {
144
+ "left_hand_Finger_Spread": 0.0005116399261169136,
145
+ "left_hand_Index_Finger_Distal": 0.00020662408496718854,
146
+ "left_hand_Index_Finger_Proximal": 0.0014279484748840332,
147
+ "left_hand_Middle_Finger_Proximal": 0.00028592348098754883,
148
+ "left_hand_Thumb_Flexion": 0.10653100907802582,
149
+ "left_hand_Thumb_Opposition": 0.006616934202611446,
150
+ "left_hand_index_spread": 0.006160148419439793,
151
+ "left_hand_j14": 2.709794534894172e-05,
152
+ "left_hand_j3": 0.019199179485440254,
153
+ "left_hand_j4": 1.1358388896098859e-08,
154
+ "left_hand_j5": 0.004905383102595806,
155
+ "left_hand_ring_spread": 0.002877138089388609,
156
+ "right_hand_Finger_Spread": 0.0010858584428206086,
157
+ "right_hand_Index_Finger_Distal": 0.1453433781862259,
158
+ "right_hand_Index_Finger_Proximal": 0.0852048397064209,
159
+ "right_hand_Middle_Finger_Distal": 0.164153054356575,
160
+ "right_hand_Middle_Finger_Proximal": 0.26833483576774597,
161
+ "right_hand_Pinky": 0.0574755035340786,
162
+ "right_hand_Ring_Finger": 0.358076274394989,
163
+ "right_hand_Thumb_Flexion": 0.13494816422462463,
164
+ "right_hand_index_spread": 0.00134497438557446,
165
+ "right_hand_j12": 0.03448190540075302,
166
+ "right_hand_j13": 0.04841304197907448,
167
+ "right_hand_j14": 0.000998900388367474,
168
+ "right_hand_j15": 0.00014427250425796956,
169
+ "right_hand_j16": 0.00012261513620615005,
170
+ "right_hand_j17": 7.388457743218169e-05,
171
+ "right_hand_j3": 0.030036840587854385,
172
+ "right_hand_ring_spread": 0.0024274573661386967
173
+ },
174
+ "coverage": {
175
+ "hard_limit_source": {
176
+ "available": 50,
177
+ "unavailable": 0,
178
+ "unknown": 0
179
+ },
180
+ "joint_role_manifest": {
181
+ "available": 50,
182
+ "unavailable": 0,
183
+ "unknown": 0
184
+ }
185
+ },
186
+ "hard_limit_sources": {
187
+ "runtime.joint_pos_limits": 50
188
+ },
189
+ "soft_limit_sources": {
190
+ "runtime.soft_joint_pos_limits": 50
191
+ },
192
+ "unknown_runtime_joint_names": [],
193
+ "baseline_checks": {
194
+ "post_home": {
195
+ "checked": 50,
196
+ "hard_violation_episodes": 0
197
+ },
198
+ "post_reset": {
199
+ "checked": 50,
200
+ "hard_violation_episodes": 0
201
+ }
202
+ }
203
+ }
204
+ },
205
+ "robot_motion": {
206
+ "joint_acceleration_rms": 11.792238288665263,
207
+ "joint_effort_rms": 23.649055229948377,
208
+ "joint_jerk_rms": 1146.102490649346,
209
+ "joint_velocity_rms": 0.3326468455177981
210
+ },
211
+ "grasp": {
212
+ "kinematic_grasp_stability_index": 0.8997033591607417,
213
+ "mean_kinematic_grasp_stability": 0.414472364768168,
214
+ "per_object_best_kinematic_grasp_stability": {
215
+ "obj_146_olive_oil_1": 0.8880000099800393,
216
+ "obj_172_soy_sauce_1": 0.8259950323226042
217
+ },
218
+ "grasp_gsi_diagnostics": {
219
+ "mean_anti_slip_score": 0.9737284359309228,
220
+ "mean_duration_score": 0.3956826331062602,
221
+ "mean_hold_score": 0.11013182946682396,
222
+ "mean_kinematic_stability_score": 0.0691246131984264,
223
+ "mean_legacy_gsi": 0.3385516151766398,
224
+ "mean_lift_score": 0.16427089528421862,
225
+ "mean_motion_stability_score": 0.09848555469815135
226
+ }
227
+ },
228
+ "tool": {
229
+ "tool_selection_accuracy": 0.98,
230
+ "tool_switch_success_rate": 0.0,
231
+ "tool_switch_total": 252,
232
+ "tool_switch_successful": 0
233
+ },
234
+ "runtime": {
235
+ "avg_steps": 2046.92,
236
+ "error_count": 0
237
+ }
238
+ }
239
+ }
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/inv_only/summary.json ADDED
@@ -0,0 +1,234 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100800000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 46,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3432,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.92,
50
+ "success_rate_ci": {
51
+ "mean": 0.92,
52
+ "ci_low": 0.84,
53
+ "ci_high": 0.98
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.97
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 29.428260869564827
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.74,
64
+ "hard_violation_rate": 0.2,
65
+ "drop_rate": 0.2,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.00010531858873091101
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.92,
74
+ "at_end_success_rate": 0.92,
75
+ "task_family_success_rate": {
76
+ "pour_tilt": 0.92
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.0,
81
+ "current_chain_depth_progress_score": 0.0,
82
+ "latched_chain_depth_progress_score": 1.0,
83
+ "avg_current_chain_depth": 0,
84
+ "avg_latched_chain_depth": 1
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.2974824741591338,
88
+ "expert_time_s": 38.13333333333333,
89
+ "expert_time_step": 2288,
90
+ "avg_steps_to_success": 1765.695652173913,
91
+ "avg_policy_steps_to_success": 588.8913043478261,
92
+ "avg_policy_queries_to_success": 588.8913043478261
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00010531858873091101,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_146_olive_oil_1": 1,
98
+ "drop_to_table_obj_172_soy_sauce_1": 9
99
+ },
100
+ "robot_constraint_violation_rate": 1.0,
101
+ "finger_joint_limit_saturation_rate": 0.16,
102
+ "joint_limit_active_step_rate": 0.021569246972090575,
103
+ "finger_joint_limit_saturation_step_rate": 0.0009478672985781991,
104
+ "max_joint_limit_excess_rad": 0.9229612946510315,
105
+ "robot_constraint_diagnostics": {
106
+ "joint_limit_soft_counts": {
107
+ "left_hand_j3": 4,
108
+ "right_hand_Index_Finger_Distal": 407,
109
+ "right_hand_Index_Finger_Proximal": 26,
110
+ "right_hand_Middle_Finger_Distal": 462,
111
+ "right_hand_Middle_Finger_Proximal": 501,
112
+ "right_hand_Pinky": 81,
113
+ "right_hand_Ring_Finger": 519,
114
+ "right_hand_Thumb_Flexion": 1229,
115
+ "right_hand_j12": 16,
116
+ "right_hand_j13": 72,
117
+ "right_hand_j3": 20
118
+ },
119
+ "joint_limit_hard_counts": {
120
+ "left_hand_j3": 1,
121
+ "right_hand_Index_Finger_Distal": 110,
122
+ "right_hand_Index_Finger_Proximal": 25,
123
+ "right_hand_Middle_Finger_Distal": 118,
124
+ "right_hand_Middle_Finger_Proximal": 193,
125
+ "right_hand_Pinky": 22,
126
+ "right_hand_Ring_Finger": 79,
127
+ "right_hand_Thumb_Flexion": 118,
128
+ "right_hand_j12": 22,
129
+ "right_hand_j13": 21,
130
+ "right_hand_j3": 10
131
+ },
132
+ "joint_limit_hard_episode_events": 398,
133
+ "finger_joint_limit_saturation_counts": {
134
+ "right_hand_Pinky": 38,
135
+ "right_hand_Ring_Finger": 45,
136
+ "right_hand_j13": 44
137
+ },
138
+ "joint_limit_max_excess_by_joint": {
139
+ "left_hand_Finger_Spread": 0.0007436490268446505,
140
+ "left_hand_Index_Finger_Distal": 0.00019865839567501098,
141
+ "left_hand_Index_Finger_Proximal": 0.0009461641311645508,
142
+ "left_hand_Middle_Finger_Proximal": 0.0005312561988830566,
143
+ "left_hand_Thumb_Flexion": 0.00026161063578911126,
144
+ "left_hand_Thumb_Opposition": 0.0037443393375724554,
145
+ "left_hand_index_spread": 0.0037924214266240597,
146
+ "left_hand_j14": 1.9059898477280512e-05,
147
+ "left_hand_j3": 0.022458622232079506,
148
+ "left_hand_j4": 1.069313082524559e-08,
149
+ "left_hand_j5": 0.0033391204196959734,
150
+ "left_hand_ring_spread": 0.0025065806694328785,
151
+ "right_hand_Finger_Spread": 0.0013409005478024483,
152
+ "right_hand_Index_Finger_Distal": 0.15362444519996643,
153
+ "right_hand_Index_Finger_Proximal": 0.07561922073364258,
154
+ "right_hand_Middle_Finger_Distal": 0.16407990455627441,
155
+ "right_hand_Middle_Finger_Proximal": 0.22367317974567413,
156
+ "right_hand_Pinky": 0.5998358130455017,
157
+ "right_hand_Ring_Finger": 0.3684403896331787,
158
+ "right_hand_Thumb_Flexion": 0.0927312970161438,
159
+ "right_hand_index_spread": 0.0018890518695116043,
160
+ "right_hand_j12": 0.03391297534108162,
161
+ "right_hand_j13": 0.9229612946510315,
162
+ "right_hand_j14": 0.0011280913604423404,
163
+ "right_hand_j15": 0.0004935747128911316,
164
+ "right_hand_j16": 0.00010464079969096929,
165
+ "right_hand_j17": 5.9021000197390094e-05,
166
+ "right_hand_j3": 0.07269296050071716,
167
+ "right_hand_ring_spread": 0.001699715619906783
168
+ },
169
+ "coverage": {
170
+ "hard_limit_source": {
171
+ "available": 50,
172
+ "unavailable": 0,
173
+ "unknown": 0
174
+ },
175
+ "joint_role_manifest": {
176
+ "available": 50,
177
+ "unavailable": 0,
178
+ "unknown": 0
179
+ }
180
+ },
181
+ "hard_limit_sources": {
182
+ "runtime.joint_pos_limits": 50
183
+ },
184
+ "soft_limit_sources": {
185
+ "runtime.soft_joint_pos_limits": 50
186
+ },
187
+ "unknown_runtime_joint_names": [],
188
+ "baseline_checks": {
189
+ "post_home": {
190
+ "checked": 50,
191
+ "hard_violation_episodes": 0
192
+ },
193
+ "post_reset": {
194
+ "checked": 50,
195
+ "hard_violation_episodes": 0
196
+ }
197
+ }
198
+ }
199
+ },
200
+ "robot_motion": {
201
+ "joint_acceleration_rms": 12.304212929422514,
202
+ "joint_effort_rms": 23.981134502390272,
203
+ "joint_jerk_rms": 1196.111260887531,
204
+ "joint_velocity_rms": 0.347243559970784
205
+ },
206
+ "grasp": {
207
+ "kinematic_grasp_stability_index": 0.9211428676998469,
208
+ "mean_kinematic_grasp_stability": 0.41568631865253086,
209
+ "per_object_best_kinematic_grasp_stability": {
210
+ "obj_146_olive_oil_1": 0.9012839915912564,
211
+ "obj_172_soy_sauce_1": 0.8522570085976181
212
+ },
213
+ "grasp_gsi_diagnostics": {
214
+ "mean_anti_slip_score": 0.9804246210342135,
215
+ "mean_duration_score": 0.38126016211267616,
216
+ "mean_hold_score": 0.11463282394309499,
217
+ "mean_kinematic_stability_score": 0.07331048243441803,
218
+ "mean_legacy_gsi": 0.3395696477690703,
219
+ "mean_lift_score": 0.17004849282723658,
220
+ "mean_motion_stability_score": 0.10212488970780825
221
+ }
222
+ },
223
+ "tool": {
224
+ "tool_selection_accuracy": 1.0,
225
+ "tool_switch_success_rate": 0.0,
226
+ "tool_switch_total": 245,
227
+ "tool_switch_successful": 0
228
+ },
229
+ "runtime": {
230
+ "avg_steps": 1899,
231
+ "error_count": 0
232
+ }
233
+ }
234
+ }
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/08_frypan_stand_pour/none/summary.json ADDED
@@ -0,0 +1,237 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100800000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 46,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3432,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.92,
50
+ "success_rate_ci": {
51
+ "mean": 0.92,
52
+ "ci_low": 0.84,
53
+ "ci_high": 0.98
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.95
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 29.522101449274967
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.76,
64
+ "hard_violation_rate": 0.2,
65
+ "drop_rate": 0.2,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.00010503208730267097
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.92,
74
+ "at_end_success_rate": 0.92,
75
+ "task_family_success_rate": {
76
+ "pour_tilt": 0.92
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.0,
81
+ "current_chain_depth_progress_score": 0.0,
82
+ "latched_chain_depth_progress_score": 0.98,
83
+ "avg_current_chain_depth": 0,
84
+ "avg_latched_chain_depth": 0.98
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 1.2947400579566428,
88
+ "expert_time_s": 38.13333333333333,
89
+ "expert_time_step": 2288,
90
+ "avg_steps_to_success": 1771.3260869565217,
91
+ "avg_policy_steps_to_success": 590.7391304347826,
92
+ "avg_policy_queries_to_success": 590.7391304347826
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00010503208730267097,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_146_olive_oil_1": 1,
98
+ "drop_to_table_obj_172_soy_sauce_1": 9
99
+ },
100
+ "robot_constraint_violation_rate": 1.0,
101
+ "finger_joint_limit_saturation_rate": 0.24,
102
+ "joint_limit_active_step_rate": 0.02026068964068523,
103
+ "finger_joint_limit_saturation_step_rate": 0.002384228381770631,
104
+ "max_joint_limit_excess_rad": 0.8928852081298828,
105
+ "robot_constraint_diagnostics": {
106
+ "joint_limit_soft_counts": {
107
+ "left_hand_Thumb_Flexion": 88,
108
+ "left_hand_j3": 67,
109
+ "right_hand_Index_Finger_Distal": 332,
110
+ "right_hand_Index_Finger_Proximal": 81,
111
+ "right_hand_Middle_Finger_Distal": 366,
112
+ "right_hand_Middle_Finger_Proximal": 397,
113
+ "right_hand_Pinky": 45,
114
+ "right_hand_Ring_Finger": 455,
115
+ "right_hand_Thumb_Flexion": 1046,
116
+ "right_hand_j12": 36,
117
+ "right_hand_j13": 27,
118
+ "right_hand_j3": 89
119
+ },
120
+ "joint_limit_hard_counts": {
121
+ "left_hand_Thumb_Flexion": 8,
122
+ "left_hand_Thumb_Opposition": 5,
123
+ "left_hand_j3": 3,
124
+ "right_hand_Index_Finger_Distal": 92,
125
+ "right_hand_Index_Finger_Proximal": 17,
126
+ "right_hand_Middle_Finger_Distal": 115,
127
+ "right_hand_Middle_Finger_Proximal": 148,
128
+ "right_hand_Pinky": 13,
129
+ "right_hand_Ring_Finger": 73,
130
+ "right_hand_Thumb_Flexion": 78,
131
+ "right_hand_j12": 24,
132
+ "right_hand_j13": 11,
133
+ "right_hand_j3": 27
134
+ },
135
+ "joint_limit_hard_episode_events": 357,
136
+ "finger_joint_limit_saturation_counts": {
137
+ "right_hand_Middle_Finger_Proximal": 48,
138
+ "right_hand_Ring_Finger": 93,
139
+ "right_hand_Thumb_Flexion": 125
140
+ },
141
+ "joint_limit_max_excess_by_joint": {
142
+ "left_hand_Finger_Spread": 0.0009490541415289044,
143
+ "left_hand_Index_Finger_Distal": 0.0022865093778818846,
144
+ "left_hand_Index_Finger_Proximal": 0.0008803009986877441,
145
+ "left_hand_Middle_Finger_Proximal": 0.0002664327621459961,
146
+ "left_hand_Thumb_Flexion": 0.07644931226968765,
147
+ "left_hand_Thumb_Opposition": 0.008481916040182114,
148
+ "left_hand_index_spread": 0.00429605832323432,
149
+ "left_hand_j14": 1.8486472981749102e-05,
150
+ "left_hand_j3": 0.08389672636985779,
151
+ "left_hand_j4": 1.21193970414879e-08,
152
+ "left_hand_j5": 0.004469777923077345,
153
+ "left_hand_ring_spread": 0.002679330063983798,
154
+ "right_hand_Finger_Spread": 0.0011678230948746204,
155
+ "right_hand_Index_Finger_Distal": 0.16022998094558716,
156
+ "right_hand_Index_Finger_Proximal": 0.17510204017162323,
157
+ "right_hand_Middle_Finger_Distal": 0.18705473840236664,
158
+ "right_hand_Middle_Finger_Proximal": 0.43274712562561035,
159
+ "right_hand_Pinky": 0.13001397252082825,
160
+ "right_hand_Ring_Finger": 0.5133461356163025,
161
+ "right_hand_Thumb_Flexion": 0.8928852081298828,
162
+ "right_hand_index_spread": 0.0013408118393272161,
163
+ "right_hand_j12": 0.07300031930208206,
164
+ "right_hand_j13": 0.10882633179426193,
165
+ "right_hand_j14": 0.0012215382885187864,
166
+ "right_hand_j15": 0.00017079741519410163,
167
+ "right_hand_j16": 9.44543571677059e-05,
168
+ "right_hand_j17": 3.754894351004623e-05,
169
+ "right_hand_j3": 0.08871258050203323,
170
+ "right_hand_ring_spread": 0.002199087757617235
171
+ },
172
+ "coverage": {
173
+ "hard_limit_source": {
174
+ "available": 50,
175
+ "unavailable": 0,
176
+ "unknown": 0
177
+ },
178
+ "joint_role_manifest": {
179
+ "available": 50,
180
+ "unavailable": 0,
181
+ "unknown": 0
182
+ }
183
+ },
184
+ "hard_limit_sources": {
185
+ "runtime.joint_pos_limits": 50
186
+ },
187
+ "soft_limit_sources": {
188
+ "runtime.soft_joint_pos_limits": 50
189
+ },
190
+ "unknown_runtime_joint_names": [],
191
+ "baseline_checks": {
192
+ "post_home": {
193
+ "checked": 50,
194
+ "hard_violation_episodes": 0
195
+ },
196
+ "post_reset": {
197
+ "checked": 50,
198
+ "hard_violation_episodes": 0
199
+ }
200
+ }
201
+ }
202
+ },
203
+ "robot_motion": {
204
+ "joint_acceleration_rms": 12.071473326800463,
205
+ "joint_effort_rms": 23.579778840025604,
206
+ "joint_jerk_rms": 1167.7899500366761,
207
+ "joint_velocity_rms": 0.34543107159964254
208
+ },
209
+ "grasp": {
210
+ "kinematic_grasp_stability_index": 0.9182448918749921,
211
+ "mean_kinematic_grasp_stability": 0.43102443200858265,
212
+ "per_object_best_kinematic_grasp_stability": {
213
+ "obj_146_olive_oil_1": 0.8967199334771675,
214
+ "obj_172_soy_sauce_1": 0.8622229686058217
215
+ },
216
+ "grasp_gsi_diagnostics": {
217
+ "mean_anti_slip_score": 0.9867382621613335,
218
+ "mean_duration_score": 0.3799952344086119,
219
+ "mean_hold_score": 0.11702260748272522,
220
+ "mean_kinematic_stability_score": 0.0746752546922635,
221
+ "mean_legacy_gsi": 0.34088103623424787,
222
+ "mean_lift_score": 0.16930378973529323,
223
+ "mean_motion_stability_score": 0.10165760404633771
224
+ }
225
+ },
226
+ "tool": {
227
+ "tool_selection_accuracy": 1.0,
228
+ "tool_switch_success_rate": 0.0,
229
+ "tool_switch_total": 251,
230
+ "tool_switch_successful": 0
231
+ },
232
+ "runtime": {
233
+ "avg_steps": 1904.18,
234
+ "error_count": 0
235
+ }
236
+ }
237
+ }
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/cov_only/summary.json ADDED
@@ -0,0 +1,229 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100900000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 39,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1926,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.78,
50
+ "success_rate_ci": {
51
+ "mean": 0.78,
52
+ "ci_low": 0.66,
53
+ "ci_high": 0.88
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.98
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.060256410256383
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.78,
64
+ "hard_violation_rate": 0.02,
65
+ "drop_rate": 0.02,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 1.3307251121135908e-05
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.78,
74
+ "at_end_success_rate": 0.78,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.78
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.88,
81
+ "current_chain_depth_progress_score": 0.97,
82
+ "latched_chain_depth_progress_score": 0.98,
83
+ "avg_current_chain_depth": 1.94,
84
+ "avg_latched_chain_depth": 1.96
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9306172059015765,
88
+ "expert_time_s": 21.383333333333333,
89
+ "expert_time_step": 1283,
90
+ "avg_steps_to_success": 1383.6153846153845,
91
+ "avg_policy_steps_to_success": 461.53846153846155,
92
+ "avg_policy_queries_to_success": 461.53846153846155
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 1.3307251121135908e-05,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_209_soap_3": 1
98
+ },
99
+ "robot_constraint_violation_rate": 1.0,
100
+ "finger_joint_limit_saturation_rate": 0.92,
101
+ "joint_limit_active_step_rate": 0.08561885371338843,
102
+ "finger_joint_limit_saturation_step_rate": 0.04298242112126898,
103
+ "max_joint_limit_excess_rad": 0.8817054033279419,
104
+ "robot_constraint_diagnostics": {
105
+ "joint_limit_soft_counts": {
106
+ "leap_l_10": 708,
107
+ "leap_l_2": 716,
108
+ "leap_l_6": 114,
109
+ "multi_leap_r_1": 833,
110
+ "multi_leap_r_10": 2286,
111
+ "multi_leap_r_12": 89,
112
+ "multi_leap_r_13": 786,
113
+ "multi_leap_r_2": 1520,
114
+ "multi_leap_r_5": 1132,
115
+ "multi_leap_r_6": 219,
116
+ "multi_leap_r_9": 1731
117
+ },
118
+ "joint_limit_hard_counts": {
119
+ "leap_l_10": 72,
120
+ "leap_l_11": 1,
121
+ "leap_l_2": 529,
122
+ "leap_l_6": 26,
123
+ "multi_leap_r_1": 25,
124
+ "multi_leap_r_10": 270,
125
+ "multi_leap_r_12": 6,
126
+ "multi_leap_r_13": 10,
127
+ "multi_leap_r_2": 151,
128
+ "multi_leap_r_5": 28,
129
+ "multi_leap_r_6": 10,
130
+ "multi_leap_r_9": 44
131
+ },
132
+ "joint_limit_hard_episode_events": 880,
133
+ "finger_joint_limit_saturation_counts": {
134
+ "leap_l_10": 254,
135
+ "leap_l_2": 37,
136
+ "multi_leap_r_1": 341,
137
+ "multi_leap_r_10": 1377,
138
+ "multi_leap_r_12": 60,
139
+ "multi_leap_r_13": 699,
140
+ "multi_leap_r_2": 671,
141
+ "multi_leap_r_5": 425,
142
+ "multi_leap_r_6": 63,
143
+ "multi_leap_r_9": 649
144
+ },
145
+ "joint_limit_max_excess_by_joint": {
146
+ "leap_l_10": 0.6566107273101807,
147
+ "leap_l_11": 0.005630463361740112,
148
+ "leap_l_13": 0.005555152893066406,
149
+ "leap_l_15": 0.00033593177795410156,
150
+ "leap_l_2": 0.35798460245132446,
151
+ "leap_l_3": 3.3468008041381836e-05,
152
+ "leap_l_6": 0.22603267431259155,
153
+ "multi_leap_r_1": 0.37076932191848755,
154
+ "multi_leap_r_10": 0.7364250421524048,
155
+ "multi_leap_r_11": 0.0037478208541870117,
156
+ "multi_leap_r_12": 0.5515403747558594,
157
+ "multi_leap_r_13": 0.8817054033279419,
158
+ "multi_leap_r_2": 0.6888589859008789,
159
+ "multi_leap_r_3": 0.002137809991836548,
160
+ "multi_leap_r_5": 0.5842604041099548,
161
+ "multi_leap_r_6": 0.3784955143928528,
162
+ "multi_leap_r_9": 0.6324809789657593
163
+ },
164
+ "coverage": {
165
+ "hard_limit_source": {
166
+ "available": 50,
167
+ "unavailable": 0,
168
+ "unknown": 0
169
+ },
170
+ "joint_role_manifest": {
171
+ "available": 50,
172
+ "unavailable": 0,
173
+ "unknown": 0
174
+ }
175
+ },
176
+ "hard_limit_sources": {
177
+ "runtime.joint_pos_limits": 50
178
+ },
179
+ "soft_limit_sources": {
180
+ "runtime.soft_joint_pos_limits": 50
181
+ },
182
+ "unknown_runtime_joint_names": [],
183
+ "baseline_checks": {
184
+ "post_home": {
185
+ "checked": 50,
186
+ "hard_violation_episodes": 0
187
+ },
188
+ "post_reset": {
189
+ "checked": 50,
190
+ "hard_violation_episodes": 0
191
+ }
192
+ }
193
+ }
194
+ },
195
+ "robot_motion": {
196
+ "joint_acceleration_rms": 19.95293092685631,
197
+ "joint_effort_rms": 65.70173692732727,
198
+ "joint_jerk_rms": 1940.6074568003714,
199
+ "joint_velocity_rms": 0.7102099215469648
200
+ },
201
+ "grasp": {
202
+ "kinematic_grasp_stability_index": 0.9600902997124623,
203
+ "mean_kinematic_grasp_stability": 0.4950071957626701,
204
+ "per_object_best_kinematic_grasp_stability": {
205
+ "obj_200_cleaner_2": 0.9006776625417269,
206
+ "obj_209_soap_3": 0.8804730382129634
207
+ },
208
+ "grasp_gsi_diagnostics": {
209
+ "mean_anti_slip_score": 0.7624127517367337,
210
+ "mean_duration_score": 0.43548394086052455,
211
+ "mean_hold_score": 0.4193152388822984,
212
+ "mean_kinematic_stability_score": 0.13946654523377386,
213
+ "mean_legacy_gsi": 0.485273646619065,
214
+ "mean_lift_score": 0.4798715615310679,
215
+ "mean_motion_stability_score": 0.3886600580024861
216
+ }
217
+ },
218
+ "tool": {
219
+ "tool_selection_accuracy": null,
220
+ "tool_switch_success_rate": null,
221
+ "tool_switch_total": 0,
222
+ "tool_switch_successful": 0
223
+ },
224
+ "runtime": {
225
+ "avg_steps": 1502.94,
226
+ "error_count": 0
227
+ }
228
+ }
229
+ }
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_cov/summary.json ADDED
@@ -0,0 +1,239 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 200900000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 37,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1926,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.74,
50
+ "success_rate_ci": {
51
+ "mean": 0.74,
52
+ "ci_low": 0.62,
53
+ "ci_high": 0.86
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.96
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.040090090090064
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.74,
64
+ "hard_violation_rate": 0.02,
65
+ "drop_rate": 0.02,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 1.312559885544778e-05
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.8,
74
+ "at_end_success_rate": 0.74,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.74
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.86,
81
+ "current_chain_depth_progress_score": 0.91,
82
+ "latched_chain_depth_progress_score": 0.96,
83
+ "avg_current_chain_depth": 1.82,
84
+ "avg_latched_chain_depth": 1.92
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9295476152795806,
88
+ "expert_time_s": 21.383333333333333,
89
+ "expert_time_step": 1283,
90
+ "avg_steps_to_success": 1382.4054054054054,
91
+ "avg_policy_steps_to_success": 461.1081081081081,
92
+ "avg_policy_queries_to_success": 461.1081081081081
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 1.312559885544778e-05,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_209_soap_3": 1
98
+ },
99
+ "robot_constraint_violation_rate": 1.0,
100
+ "finger_joint_limit_saturation_rate": 1.0,
101
+ "joint_limit_active_step_rate": 0.09319175187367924,
102
+ "finger_joint_limit_saturation_step_rate": 0.05252864661950202,
103
+ "max_joint_limit_excess_rad": 0.9820003509521484,
104
+ "robot_constraint_diagnostics": {
105
+ "joint_limit_soft_counts": {
106
+ "leap_l_10": 576,
107
+ "leap_l_2": 674,
108
+ "leap_l_6": 122,
109
+ "leap_l_9": 128,
110
+ "multi_leap_r_1": 540,
111
+ "multi_leap_r_10": 3256,
112
+ "multi_leap_r_12": 255,
113
+ "multi_leap_r_13": 1415,
114
+ "multi_leap_r_14": 7,
115
+ "multi_leap_r_15": 3,
116
+ "multi_leap_r_2": 2427,
117
+ "multi_leap_r_5": 1065,
118
+ "multi_leap_r_6": 70,
119
+ "multi_leap_r_9": 1303
120
+ },
121
+ "joint_limit_hard_counts": {
122
+ "leap_l_10": 58,
123
+ "leap_l_2": 479,
124
+ "leap_l_6": 28,
125
+ "leap_l_9": 1,
126
+ "multi_leap_r_1": 14,
127
+ "multi_leap_r_10": 277,
128
+ "multi_leap_r_12": 7,
129
+ "multi_leap_r_13": 20,
130
+ "multi_leap_r_14": 1,
131
+ "multi_leap_r_15": 1,
132
+ "multi_leap_r_2": 164,
133
+ "multi_leap_r_3": 1,
134
+ "multi_leap_r_5": 33,
135
+ "multi_leap_r_6": 5,
136
+ "multi_leap_r_9": 37
137
+ },
138
+ "joint_limit_hard_episode_events": 832,
139
+ "finger_joint_limit_saturation_counts": {
140
+ "leap_l_10": 152,
141
+ "leap_l_2": 31,
142
+ "leap_l_9": 116,
143
+ "multi_leap_r_1": 252,
144
+ "multi_leap_r_10": 1981,
145
+ "multi_leap_r_12": 178,
146
+ "multi_leap_r_13": 1262,
147
+ "multi_leap_r_2": 1554,
148
+ "multi_leap_r_5": 302,
149
+ "multi_leap_r_6": 5,
150
+ "multi_leap_r_9": 390
151
+ },
152
+ "joint_limit_max_excess_by_joint": {
153
+ "leap_l_10": 0.6981533765792847,
154
+ "leap_l_11": 0.0011772513389587402,
155
+ "leap_l_13": 0.004713296890258789,
156
+ "leap_l_15": 0.00024449825286865234,
157
+ "leap_l_2": 0.41226738691329956,
158
+ "leap_l_3": 2.1845102310180664e-05,
159
+ "leap_l_6": 0.18395817279815674,
160
+ "leap_l_9": 0.7540305256843567,
161
+ "multi_leap_r_1": 0.3734046220779419,
162
+ "multi_leap_r_10": 0.7699466943740845,
163
+ "multi_leap_r_11": 0.003787696361541748,
164
+ "multi_leap_r_12": 0.8998863697052002,
165
+ "multi_leap_r_13": 0.9525907039642334,
166
+ "multi_leap_r_14": 0.08535146713256836,
167
+ "multi_leap_r_15": 0.016042470932006836,
168
+ "multi_leap_r_2": 0.9820003509521484,
169
+ "multi_leap_r_3": 0.008730798959732056,
170
+ "multi_leap_r_5": 0.3814393877983093,
171
+ "multi_leap_r_6": 0.16107743978500366,
172
+ "multi_leap_r_9": 0.41611117124557495
173
+ },
174
+ "coverage": {
175
+ "hard_limit_source": {
176
+ "available": 50,
177
+ "unavailable": 0,
178
+ "unknown": 0
179
+ },
180
+ "joint_role_manifest": {
181
+ "available": 50,
182
+ "unavailable": 0,
183
+ "unknown": 0
184
+ }
185
+ },
186
+ "hard_limit_sources": {
187
+ "runtime.joint_pos_limits": 50
188
+ },
189
+ "soft_limit_sources": {
190
+ "runtime.soft_joint_pos_limits": 50
191
+ },
192
+ "unknown_runtime_joint_names": [],
193
+ "baseline_checks": {
194
+ "post_home": {
195
+ "checked": 50,
196
+ "hard_violation_episodes": 0
197
+ },
198
+ "post_reset": {
199
+ "checked": 50,
200
+ "hard_violation_episodes": 0
201
+ }
202
+ }
203
+ }
204
+ },
205
+ "robot_motion": {
206
+ "joint_acceleration_rms": 19.43570872399307,
207
+ "joint_effort_rms": 66.87008586862443,
208
+ "joint_jerk_rms": 1887.0248941540428,
209
+ "joint_velocity_rms": 0.7066941779355986
210
+ },
211
+ "grasp": {
212
+ "kinematic_grasp_stability_index": 0.9681196909507807,
213
+ "mean_kinematic_grasp_stability": 0.5042801243290936,
214
+ "per_object_best_kinematic_grasp_stability": {
215
+ "obj_200_cleaner_2": 0.8986257538932036,
216
+ "obj_209_soap_3": 0.9125742956937086
217
+ },
218
+ "grasp_gsi_diagnostics": {
219
+ "mean_anti_slip_score": 0.7421124558984913,
220
+ "mean_duration_score": 0.45325946061065386,
221
+ "mean_hold_score": 0.42890947094249654,
222
+ "mean_kinematic_stability_score": 0.13281229460412555,
223
+ "mean_legacy_gsi": 0.49074641808186864,
224
+ "mean_lift_score": 0.49119936911039247,
225
+ "mean_motion_stability_score": 0.39794950241639765
226
+ }
227
+ },
228
+ "tool": {
229
+ "tool_selection_accuracy": null,
230
+ "tool_switch_success_rate": null,
231
+ "tool_switch_total": 0,
232
+ "tool_switch_successful": 0
233
+ },
234
+ "runtime": {
235
+ "avg_steps": 1523.74,
236
+ "error_count": 0
237
+ }
238
+ }
239
+ }
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/inv_only/summary.json ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100900000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 36,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1926,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.72,
50
+ "success_rate_ci": {
51
+ "mean": 0.72,
52
+ "ci_low": 0.6,
53
+ "ci_high": 0.84
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.98
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.219444444444406
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.72,
64
+ "hard_violation_rate": 0.0,
65
+ "drop_rate": 0.0,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.0
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.72,
74
+ "at_end_success_rate": 0.72,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.72
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.86,
81
+ "current_chain_depth_progress_score": 0.98,
82
+ "latched_chain_depth_progress_score": 0.98,
83
+ "avg_current_chain_depth": 1.96,
84
+ "avg_latched_chain_depth": 1.96
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9236297453227216,
88
+ "expert_time_s": 21.383333333333333,
89
+ "expert_time_step": 1283,
90
+ "avg_steps_to_success": 1393.1666666666667,
91
+ "avg_policy_steps_to_success": 464.75,
92
+ "avg_policy_queries_to_success": 464.75
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0,
96
+ "violation_counts": {},
97
+ "robot_constraint_violation_rate": 1.0,
98
+ "finger_joint_limit_saturation_rate": 1.0,
99
+ "joint_limit_active_step_rate": 0.09344121994865012,
100
+ "finger_joint_limit_saturation_step_rate": 0.05188153219741176,
101
+ "max_joint_limit_excess_rad": 1.3442001342773438,
102
+ "robot_constraint_diagnostics": {
103
+ "joint_limit_soft_counts": {
104
+ "leap_l_10": 747,
105
+ "leap_l_2": 672,
106
+ "leap_l_6": 84,
107
+ "multi_leap_r_1": 659,
108
+ "multi_leap_r_10": 2342,
109
+ "multi_leap_r_12": 593,
110
+ "multi_leap_r_13": 965,
111
+ "multi_leap_r_2": 1973,
112
+ "multi_leap_r_5": 1314,
113
+ "multi_leap_r_6": 30,
114
+ "multi_leap_r_9": 1980
115
+ },
116
+ "joint_limit_hard_counts": {
117
+ "leap_l_10": 65,
118
+ "leap_l_2": 469,
119
+ "leap_l_6": 20,
120
+ "multi_leap_r_1": 20,
121
+ "multi_leap_r_10": 192,
122
+ "multi_leap_r_12": 4,
123
+ "multi_leap_r_13": 9,
124
+ "multi_leap_r_2": 148,
125
+ "multi_leap_r_5": 39,
126
+ "multi_leap_r_6": 5,
127
+ "multi_leap_r_9": 52
128
+ },
129
+ "joint_limit_hard_episode_events": 752,
130
+ "finger_joint_limit_saturation_counts": {
131
+ "leap_l_10": 324,
132
+ "leap_l_2": 45,
133
+ "multi_leap_r_1": 240,
134
+ "multi_leap_r_10": 1382,
135
+ "multi_leap_r_12": 542,
136
+ "multi_leap_r_13": 886,
137
+ "multi_leap_r_2": 1077,
138
+ "multi_leap_r_5": 463,
139
+ "multi_leap_r_9": 739
140
+ },
141
+ "joint_limit_max_excess_by_joint": {
142
+ "leap_l_10": 0.8291943073272705,
143
+ "leap_l_11": 0.001100689172744751,
144
+ "leap_l_13": 0.0028162002563476562,
145
+ "leap_l_15": 0.0012218952178955078,
146
+ "leap_l_2": 0.4221232533454895,
147
+ "leap_l_3": 5.307793617248535e-05,
148
+ "leap_l_6": 0.1923193335533142,
149
+ "multi_leap_r_1": 0.4613455533981323,
150
+ "multi_leap_r_10": 0.625508189201355,
151
+ "multi_leap_r_11": 0.0024254322052001953,
152
+ "multi_leap_r_12": 1.3442001342773438,
153
+ "multi_leap_r_13": 0.8589392900466919,
154
+ "multi_leap_r_2": 0.8465927839279175,
155
+ "multi_leap_r_3": 0.0013757944107055664,
156
+ "multi_leap_r_5": 0.4275851845741272,
157
+ "multi_leap_r_6": 0.09731262922286987,
158
+ "multi_leap_r_9": 0.48914384841918945
159
+ },
160
+ "coverage": {
161
+ "hard_limit_source": {
162
+ "available": 50,
163
+ "unavailable": 0,
164
+ "unknown": 0
165
+ },
166
+ "joint_role_manifest": {
167
+ "available": 50,
168
+ "unavailable": 0,
169
+ "unknown": 0
170
+ }
171
+ },
172
+ "hard_limit_sources": {
173
+ "runtime.joint_pos_limits": 50
174
+ },
175
+ "soft_limit_sources": {
176
+ "runtime.soft_joint_pos_limits": 50
177
+ },
178
+ "unknown_runtime_joint_names": [],
179
+ "baseline_checks": {
180
+ "post_home": {
181
+ "checked": 50,
182
+ "hard_violation_episodes": 0
183
+ },
184
+ "post_reset": {
185
+ "checked": 50,
186
+ "hard_violation_episodes": 0
187
+ }
188
+ }
189
+ }
190
+ },
191
+ "robot_motion": {
192
+ "joint_acceleration_rms": 19.40242171824975,
193
+ "joint_effort_rms": 66.50204655183835,
194
+ "joint_jerk_rms": 1890.6197933523554,
195
+ "joint_velocity_rms": 0.7014668876990304
196
+ },
197
+ "grasp": {
198
+ "kinematic_grasp_stability_index": 0.9660479851668408,
199
+ "mean_kinematic_grasp_stability": 0.5190378980985876,
200
+ "per_object_best_kinematic_grasp_stability": {
201
+ "obj_200_cleaner_2": 0.9075564489565543,
202
+ "obj_209_soap_3": 0.8998850185380606
203
+ },
204
+ "grasp_gsi_diagnostics": {
205
+ "mean_anti_slip_score": 0.7486323631585352,
206
+ "mean_duration_score": 0.4503614280360579,
207
+ "mean_hold_score": 0.43351209113219336,
208
+ "mean_kinematic_stability_score": 0.13974007464917101,
209
+ "mean_legacy_gsi": 0.4937539242551601,
210
+ "mean_lift_score": 0.491841137572851,
211
+ "mean_motion_stability_score": 0.40244879026018643
212
+ }
213
+ },
214
+ "tool": {
215
+ "tool_selection_accuracy": null,
216
+ "tool_switch_success_rate": null,
217
+ "tool_switch_total": 0,
218
+ "tool_switch_successful": 0
219
+ },
220
+ "runtime": {
221
+ "avg_steps": 1542.36,
222
+ "error_count": 0
223
+ }
224
+ }
225
+ }
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/09_cleaner_moisturizer_box_loading/none/summary.json ADDED
@@ -0,0 +1,228 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 100900000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 42,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 1926,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.84,
50
+ "success_rate_ci": {
51
+ "mean": 0.84,
52
+ "ci_low": 0.74,
53
+ "ci_high": 0.94
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.98
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 23.153571428571396
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.84,
64
+ "hard_violation_rate": 0.0,
65
+ "drop_rate": 0.0,
66
+ "high_speed_violation_rate": 0.0,
67
+ "safety_violation_step_rate": 0.0
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.86,
74
+ "at_end_success_rate": 0.84,
75
+ "task_family_success_rate": {
76
+ "rigid_pick_place": 0.84
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.92,
81
+ "current_chain_depth_progress_score": 0.97,
82
+ "latched_chain_depth_progress_score": 0.98,
83
+ "avg_current_chain_depth": 1.94,
84
+ "avg_latched_chain_depth": 1.96
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9284321598490018,
88
+ "expert_time_s": 21.383333333333333,
89
+ "expert_time_step": 1283,
90
+ "avg_steps_to_success": 1389.2142857142858,
91
+ "avg_policy_steps_to_success": 463.35714285714283,
92
+ "avg_policy_queries_to_success": 463.35714285714283
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0,
96
+ "violation_counts": {},
97
+ "robot_constraint_violation_rate": 1.0,
98
+ "finger_joint_limit_saturation_rate": 0.94,
99
+ "joint_limit_active_step_rate": 0.09307843536031456,
100
+ "finger_joint_limit_saturation_step_rate": 0.04590875194902041,
101
+ "max_joint_limit_excess_rad": 2.4507758617401123,
102
+ "robot_constraint_diagnostics": {
103
+ "joint_limit_soft_counts": {
104
+ "leap_l_10": 1336,
105
+ "leap_l_2": 662,
106
+ "leap_l_6": 146,
107
+ "multi_leap_r_1": 1084,
108
+ "multi_leap_r_10": 2293,
109
+ "multi_leap_r_12": 127,
110
+ "multi_leap_r_13": 322,
111
+ "multi_leap_r_2": 1555,
112
+ "multi_leap_r_5": 954,
113
+ "multi_leap_r_6": 388,
114
+ "multi_leap_r_9": 1789
115
+ },
116
+ "joint_limit_hard_counts": {
117
+ "leap_l_10": 83,
118
+ "leap_l_2": 493,
119
+ "leap_l_6": 28,
120
+ "multi_leap_r_1": 24,
121
+ "multi_leap_r_10": 208,
122
+ "multi_leap_r_12": 3,
123
+ "multi_leap_r_13": 6,
124
+ "multi_leap_r_2": 138,
125
+ "multi_leap_r_5": 24,
126
+ "multi_leap_r_6": 14,
127
+ "multi_leap_r_9": 41
128
+ },
129
+ "joint_limit_hard_episode_events": 806,
130
+ "finger_joint_limit_saturation_counts": {
131
+ "leap_l_10": 634,
132
+ "leap_l_2": 44,
133
+ "leap_l_6": 25,
134
+ "multi_leap_r_1": 406,
135
+ "multi_leap_r_10": 1309,
136
+ "multi_leap_r_12": 98,
137
+ "multi_leap_r_13": 286,
138
+ "multi_leap_r_2": 921,
139
+ "multi_leap_r_5": 364,
140
+ "multi_leap_r_6": 222,
141
+ "multi_leap_r_9": 659
142
+ },
143
+ "joint_limit_max_excess_by_joint": {
144
+ "leap_l_10": 0.8450669050216675,
145
+ "leap_l_11": 0.0008967220783233643,
146
+ "leap_l_13": 0.005696296691894531,
147
+ "leap_l_15": 0.0003108978271484375,
148
+ "leap_l_2": 0.364902138710022,
149
+ "leap_l_3": 8.39531421661377e-05,
150
+ "leap_l_6": 0.5563956499099731,
151
+ "multi_leap_r_1": 0.4487890601158142,
152
+ "multi_leap_r_10": 0.7905848026275635,
153
+ "multi_leap_r_11": 0.003816276788711548,
154
+ "multi_leap_r_12": 2.4507758617401123,
155
+ "multi_leap_r_13": 0.8960253000259399,
156
+ "multi_leap_r_2": 0.9739581346511841,
157
+ "multi_leap_r_3": 0.0014981627464294434,
158
+ "multi_leap_r_5": 0.5018796920776367,
159
+ "multi_leap_r_6": 0.4307611584663391,
160
+ "multi_leap_r_7": 0.0008980333805084229,
161
+ "multi_leap_r_9": 0.5537928342819214
162
+ },
163
+ "coverage": {
164
+ "hard_limit_source": {
165
+ "available": 50,
166
+ "unavailable": 0,
167
+ "unknown": 0
168
+ },
169
+ "joint_role_manifest": {
170
+ "available": 50,
171
+ "unavailable": 0,
172
+ "unknown": 0
173
+ }
174
+ },
175
+ "hard_limit_sources": {
176
+ "runtime.joint_pos_limits": 50
177
+ },
178
+ "soft_limit_sources": {
179
+ "runtime.soft_joint_pos_limits": 50
180
+ },
181
+ "unknown_runtime_joint_names": [],
182
+ "baseline_checks": {
183
+ "post_home": {
184
+ "checked": 50,
185
+ "hard_violation_episodes": 0
186
+ },
187
+ "post_reset": {
188
+ "checked": 50,
189
+ "hard_violation_episodes": 0
190
+ }
191
+ }
192
+ }
193
+ },
194
+ "robot_motion": {
195
+ "joint_acceleration_rms": 20.055854886332252,
196
+ "joint_effort_rms": 66.37089692843377,
197
+ "joint_jerk_rms": 1958.9416532249338,
198
+ "joint_velocity_rms": 0.7080286412923932
199
+ },
200
+ "grasp": {
201
+ "kinematic_grasp_stability_index": 0.9604311928794026,
202
+ "mean_kinematic_grasp_stability": 0.4854185789543128,
203
+ "per_object_best_kinematic_grasp_stability": {
204
+ "obj_200_cleaner_2": 0.902860546459245,
205
+ "obj_209_soap_3": 0.8673472347863544
206
+ },
207
+ "grasp_gsi_diagnostics": {
208
+ "mean_anti_slip_score": 0.7525602612846038,
209
+ "mean_duration_score": 0.41627282686226086,
210
+ "mean_hold_score": 0.4027618105404881,
211
+ "mean_kinematic_stability_score": 0.11251937900957219,
212
+ "mean_legacy_gsi": 0.4703562252994722,
213
+ "mean_lift_score": 0.46527997530970455,
214
+ "mean_motion_stability_score": 0.37337070415163004
215
+ }
216
+ },
217
+ "tool": {
218
+ "tool_selection_accuracy": null,
219
+ "tool_switch_success_rate": null,
220
+ "tool_switch_total": 0,
221
+ "tool_switch_successful": 0
222
+ },
223
+ "runtime": {
224
+ "avg_steps": 1475.1,
225
+ "error_count": 0
226
+ }
227
+ }
228
+ }
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/cov_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/cov_only/summary.json ADDED
@@ -0,0 +1,266 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 101200000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 26,
10
+ "error_count": 1,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3447,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.52,
50
+ "success_rate_ci": {
51
+ "mean": 0.52,
52
+ "ci_low": 0.38,
53
+ "ci_high": 0.66
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.85
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 39.4012820512811
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.3,
64
+ "hard_violation_rate": 0.54,
65
+ "drop_rate": 0.24,
66
+ "high_speed_violation_rate": 0.4,
67
+ "safety_violation_step_rate": 0.002953023882493385
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.54,
74
+ "at_end_success_rate": 0.52,
75
+ "task_family_success_rate": {
76
+ "tool_precision_place": 0.52
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.82,
81
+ "current_chain_depth_progress_score": 0.825,
82
+ "latched_chain_depth_progress_score": 0.85,
83
+ "avg_current_chain_depth": 3.3,
84
+ "avg_latched_chain_depth": 3.4
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9780704580529452,
88
+ "expert_time_s": 38.28333333333333,
89
+ "expert_time_step": 2297,
90
+ "avg_steps_to_success": 2364.076923076923,
91
+ "avg_policy_steps_to_success": 788.3076923076923,
92
+ "avg_policy_queries_to_success": 788.3076923076923
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.0003001891890005096,
96
+ "violation_counts": {
97
+ "drop_obj_048_hammer_3": 1,
98
+ "drop_to_table_obj_043_phillips_screwdriver_1": 3,
99
+ "drop_to_table_obj_048_hammer_3": 12,
100
+ "high_speed_obj_043_phillips_screwdriver_1": 12,
101
+ "high_speed_obj_044_flat_screwdriver_2": 2,
102
+ "high_speed_obj_048_hammer_3": 13
103
+ },
104
+ "robot_constraint_violation_rate": 1.0,
105
+ "finger_joint_limit_saturation_rate": 1.0,
106
+ "joint_limit_active_step_rate": 0.19297976166374622,
107
+ "finger_joint_limit_saturation_step_rate": 0.14459345308322222,
108
+ "max_joint_limit_excess_rad": 1.3115074634552002,
109
+ "robot_constraint_diagnostics": {
110
+ "joint_limit_soft_counts": {
111
+ "left_index_1_joint": 3729,
112
+ "left_index_2_joint": 15,
113
+ "left_little_1_joint": 6392,
114
+ "left_little_2_joint": 80,
115
+ "left_middle_1_joint": 7909,
116
+ "left_middle_2_joint": 77,
117
+ "left_ring_1_joint": 8226,
118
+ "left_ring_2_joint": 95,
119
+ "left_thumb_2_joint": 7547,
120
+ "left_thumb_3_joint": 3001,
121
+ "left_thumb_4_joint": 3277,
122
+ "right_index_1_joint": 6942,
123
+ "right_little_1_joint": 1684,
124
+ "right_middle_1_joint": 4598,
125
+ "right_middle_2_joint": 8,
126
+ "right_ring_1_joint": 2188,
127
+ "right_ring_2_joint": 14,
128
+ "right_thumb_2_joint": 210,
129
+ "right_thumb_3_joint": 258,
130
+ "right_thumb_4_joint": 480
131
+ },
132
+ "joint_limit_hard_counts": {
133
+ "left_index_1_joint": 129,
134
+ "left_index_2_joint": 23,
135
+ "left_little_1_joint": 159,
136
+ "left_little_2_joint": 67,
137
+ "left_middle_1_joint": 187,
138
+ "left_middle_2_joint": 44,
139
+ "left_ring_1_joint": 190,
140
+ "left_ring_2_joint": 49,
141
+ "left_thumb_2_joint": 152,
142
+ "left_thumb_3_joint": 53,
143
+ "left_thumb_4_joint": 131,
144
+ "right_index_1_joint": 260,
145
+ "right_little_1_joint": 45,
146
+ "right_little_2_joint": 8,
147
+ "right_middle_1_joint": 121,
148
+ "right_middle_2_joint": 11,
149
+ "right_ring_1_joint": 66,
150
+ "right_ring_2_joint": 20,
151
+ "right_thumb_2_joint": 18,
152
+ "right_thumb_3_joint": 17,
153
+ "right_thumb_4_joint": 40
154
+ },
155
+ "joint_limit_hard_episode_events": 616,
156
+ "finger_joint_limit_saturation_counts": {
157
+ "left_index_1_joint": 1851,
158
+ "left_little_1_joint": 4363,
159
+ "left_little_2_joint": 3,
160
+ "left_middle_1_joint": 5427,
161
+ "left_middle_2_joint": 34,
162
+ "left_ring_1_joint": 5870,
163
+ "left_ring_2_joint": 47,
164
+ "left_thumb_2_joint": 5269,
165
+ "left_thumb_3_joint": 2400,
166
+ "left_thumb_4_joint": 2287,
167
+ "right_index_1_joint": 4524,
168
+ "right_little_1_joint": 939,
169
+ "right_middle_1_joint": 2713,
170
+ "right_ring_1_joint": 1109,
171
+ "right_thumb_2_joint": 10,
172
+ "right_thumb_3_joint": 96,
173
+ "right_thumb_4_joint": 147
174
+ },
175
+ "joint_limit_max_excess_by_joint": {
176
+ "left_index_1_joint": 0.8876303434371948,
177
+ "left_index_2_joint": 0.03147381916642189,
178
+ "left_little_1_joint": 1.194810390472412,
179
+ "left_little_2_joint": 0.13955436646938324,
180
+ "left_middle_1_joint": 0.9944871068000793,
181
+ "left_middle_2_joint": 0.31187304854393005,
182
+ "left_ring_1_joint": 1.1062034368515015,
183
+ "left_ring_2_joint": 0.4842528998851776,
184
+ "left_thumb_1_joint": 1.583285120432265e-05,
185
+ "left_thumb_2_joint": 0.5784367918968201,
186
+ "left_thumb_3_joint": 1.3091992139816284,
187
+ "left_thumb_4_joint": 1.3115074634552002,
188
+ "right_index_1_joint": 0.6335136294364929,
189
+ "right_index_2_joint": 0.001141366665251553,
190
+ "right_little_1_joint": 0.7750421166419983,
191
+ "right_little_2_joint": 0.008474268950521946,
192
+ "right_middle_1_joint": 0.958519697189331,
193
+ "right_middle_2_joint": 0.024957530200481415,
194
+ "right_ring_1_joint": 0.5307520627975464,
195
+ "right_ring_2_joint": 0.024919383227825165,
196
+ "right_thumb_2_joint": 0.14785043895244598,
197
+ "right_thumb_3_joint": 0.36655503511428833,
198
+ "right_thumb_4_joint": 1.0262298583984375
199
+ },
200
+ "coverage": {
201
+ "hard_limit_source": {
202
+ "available": 50,
203
+ "unavailable": 0,
204
+ "unknown": 0
205
+ },
206
+ "joint_role_manifest": {
207
+ "available": 50,
208
+ "unavailable": 0,
209
+ "unknown": 0
210
+ }
211
+ },
212
+ "hard_limit_sources": {
213
+ "runtime.joint_pos_limits": 50
214
+ },
215
+ "soft_limit_sources": {
216
+ "runtime.soft_joint_pos_limits": 50
217
+ },
218
+ "unknown_runtime_joint_names": [],
219
+ "baseline_checks": {
220
+ "post_home": {
221
+ "checked": 50,
222
+ "hard_violation_episodes": 0
223
+ },
224
+ "post_reset": {
225
+ "checked": 50,
226
+ "hard_violation_episodes": 0
227
+ }
228
+ }
229
+ }
230
+ },
231
+ "robot_motion": {
232
+ "joint_acceleration_rms": 59.51643336935102,
233
+ "joint_effort_rms": 52.6126499045288,
234
+ "joint_jerk_rms": 6101.994166461753,
235
+ "joint_velocity_rms": 1.2448206343830217
236
+ },
237
+ "grasp": {
238
+ "kinematic_grasp_stability_index": 0.9743028273492969,
239
+ "mean_kinematic_grasp_stability": 0.49150601917342024,
240
+ "per_object_best_kinematic_grasp_stability": {
241
+ "obj_043_phillips_screwdriver_1": 0.837567723669333,
242
+ "obj_044_flat_screwdriver_2": 0.8037437949736153,
243
+ "obj_048_hammer_3": 0.6730401214149229
244
+ },
245
+ "grasp_gsi_diagnostics": {
246
+ "mean_anti_slip_score": 0.7158532251473051,
247
+ "mean_duration_score": 0.5073132194801812,
248
+ "mean_hold_score": 0.4378100649588933,
249
+ "mean_kinematic_stability_score": 0.18837491378166277,
250
+ "mean_legacy_gsi": 0.5106962946100225,
251
+ "mean_lift_score": 0.5661183718981532,
252
+ "mean_motion_stability_score": 0.4249677638882656
253
+ }
254
+ },
255
+ "tool": {
256
+ "tool_selection_accuracy": 0.0,
257
+ "tool_switch_success_rate": 0.0,
258
+ "tool_switch_total": 149,
259
+ "tool_switch_successful": 0
260
+ },
261
+ "runtime": {
262
+ "avg_steps": 2864.86,
263
+ "error_count": 1
264
+ }
265
+ }
266
+ }
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_cov/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_cov/summary.json ADDED
@@ -0,0 +1,265 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 201200000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 30,
10
+ "error_count": 0,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3447,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.6,
50
+ "success_rate_ci": {
51
+ "mean": 0.6,
52
+ "ci_low": 0.46,
53
+ "ci_high": 0.74
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.83
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 39.73555555555458
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.32,
64
+ "hard_violation_rate": 0.54,
65
+ "drop_rate": 0.24,
66
+ "high_speed_violation_rate": 0.38,
67
+ "safety_violation_step_rate": 0.0005125868549948741
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.62,
74
+ "at_end_success_rate": 0.6,
75
+ "task_family_success_rate": {
76
+ "tool_precision_place": 0.6
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.81,
81
+ "current_chain_depth_progress_score": 0.82,
82
+ "latched_chain_depth_progress_score": 0.83,
83
+ "avg_current_chain_depth": 3.28,
84
+ "avg_latched_chain_depth": 3.32
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9682165172579809,
88
+ "expert_time_s": 38.28333333333333,
89
+ "expert_time_step": 2297,
90
+ "avg_steps_to_success": 2384.133333333333,
91
+ "avg_policy_steps_to_success": 795.0333333333333,
92
+ "avg_policy_queries_to_success": 795.0333333333333
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00032748604624672513,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_043_phillips_screwdriver_1": 3,
98
+ "drop_to_table_obj_044_flat_screwdriver_2": 1,
99
+ "drop_to_table_obj_048_hammer_3": 10,
100
+ "high_speed_distractor_00_clutter_033_spatula": 6,
101
+ "high_speed_obj_043_phillips_screwdriver_1": 12,
102
+ "high_speed_obj_044_flat_screwdriver_2": 2,
103
+ "high_speed_obj_048_hammer_3": 11,
104
+ "high_speed_obj_154_wooden_box_1": 1
105
+ },
106
+ "robot_constraint_violation_rate": 1.0,
107
+ "finger_joint_limit_saturation_rate": 1.0,
108
+ "joint_limit_active_step_rate": 0.19082469529559176,
109
+ "finger_joint_limit_saturation_step_rate": 0.14777451873789726,
110
+ "max_joint_limit_excess_rad": 1.721112608909607,
111
+ "robot_constraint_diagnostics": {
112
+ "joint_limit_soft_counts": {
113
+ "left_index_1_joint": 3134,
114
+ "left_index_2_joint": 18,
115
+ "left_little_1_joint": 5821,
116
+ "left_little_2_joint": 57,
117
+ "left_middle_1_joint": 8034,
118
+ "left_middle_2_joint": 11,
119
+ "left_ring_1_joint": 8087,
120
+ "left_ring_2_joint": 12,
121
+ "left_thumb_1_joint": 76,
122
+ "left_thumb_2_joint": 8206,
123
+ "left_thumb_3_joint": 3531,
124
+ "left_thumb_4_joint": 3944,
125
+ "right_index_1_joint": 6840,
126
+ "right_little_1_joint": 1552,
127
+ "right_middle_1_joint": 3365,
128
+ "right_ring_1_joint": 1997,
129
+ "right_thumb_2_joint": 881,
130
+ "right_thumb_3_joint": 269,
131
+ "right_thumb_4_joint": 134
132
+ },
133
+ "joint_limit_hard_counts": {
134
+ "left_index_1_joint": 108,
135
+ "left_index_2_joint": 1,
136
+ "left_little_1_joint": 150,
137
+ "left_little_2_joint": 75,
138
+ "left_middle_1_joint": 158,
139
+ "left_middle_2_joint": 31,
140
+ "left_ring_1_joint": 172,
141
+ "left_ring_2_joint": 32,
142
+ "left_thumb_1_joint": 1,
143
+ "left_thumb_2_joint": 151,
144
+ "left_thumb_3_joint": 61,
145
+ "left_thumb_4_joint": 160,
146
+ "right_index_1_joint": 290,
147
+ "right_little_1_joint": 49,
148
+ "right_middle_1_joint": 56,
149
+ "right_ring_1_joint": 53,
150
+ "right_ring_2_joint": 1,
151
+ "right_thumb_2_joint": 17,
152
+ "right_thumb_3_joint": 19,
153
+ "right_thumb_4_joint": 44
154
+ },
155
+ "joint_limit_hard_episode_events": 526,
156
+ "finger_joint_limit_saturation_counts": {
157
+ "left_index_1_joint": 1474,
158
+ "left_index_2_joint": 16,
159
+ "left_little_1_joint": 3930,
160
+ "left_middle_1_joint": 5587,
161
+ "left_ring_1_joint": 5743,
162
+ "left_thumb_1_joint": 73,
163
+ "left_thumb_2_joint": 6125,
164
+ "left_thumb_3_joint": 2892,
165
+ "left_thumb_4_joint": 2716,
166
+ "right_index_1_joint": 4318,
167
+ "right_little_1_joint": 687,
168
+ "right_middle_1_joint": 2400,
169
+ "right_ring_1_joint": 986,
170
+ "right_thumb_2_joint": 285,
171
+ "right_thumb_3_joint": 65,
172
+ "right_thumb_4_joint": 18
173
+ },
174
+ "joint_limit_max_excess_by_joint": {
175
+ "left_index_1_joint": 1.721112608909607,
176
+ "left_index_2_joint": 0.9208273887634277,
177
+ "left_little_1_joint": 0.9165555834770203,
178
+ "left_little_2_joint": 0.04622817039489746,
179
+ "left_middle_1_joint": 0.9195486307144165,
180
+ "left_middle_2_joint": 0.030572056770324707,
181
+ "left_ring_1_joint": 0.9073184728622437,
182
+ "left_ring_2_joint": 0.02050631307065487,
183
+ "left_thumb_1_joint": 1.4623029232025146,
184
+ "left_thumb_2_joint": 0.6359100937843323,
185
+ "left_thumb_3_joint": 1.2478957176208496,
186
+ "left_thumb_4_joint": 1.229111671447754,
187
+ "right_index_1_joint": 0.5458549857139587,
188
+ "right_index_2_joint": 2.89820445686928e-06,
189
+ "right_little_1_joint": 0.5645976066589355,
190
+ "right_little_2_joint": 3.7828931453987025e-06,
191
+ "right_middle_1_joint": 0.709052562713623,
192
+ "right_middle_2_joint": 0.000769146834500134,
193
+ "right_ring_1_joint": 0.6099717617034912,
194
+ "right_ring_2_joint": 0.006652869284152985,
195
+ "right_thumb_2_joint": 0.29265502095222473,
196
+ "right_thumb_3_joint": 0.39177048206329346,
197
+ "right_thumb_4_joint": 0.23184248805046082
198
+ },
199
+ "coverage": {
200
+ "hard_limit_source": {
201
+ "available": 50,
202
+ "unavailable": 0,
203
+ "unknown": 0
204
+ },
205
+ "joint_role_manifest": {
206
+ "available": 50,
207
+ "unavailable": 0,
208
+ "unknown": 0
209
+ }
210
+ },
211
+ "hard_limit_sources": {
212
+ "runtime.joint_pos_limits": 50
213
+ },
214
+ "soft_limit_sources": {
215
+ "runtime.soft_joint_pos_limits": 50
216
+ },
217
+ "unknown_runtime_joint_names": [],
218
+ "baseline_checks": {
219
+ "post_home": {
220
+ "checked": 50,
221
+ "hard_violation_episodes": 0
222
+ },
223
+ "post_reset": {
224
+ "checked": 50,
225
+ "hard_violation_episodes": 0
226
+ }
227
+ }
228
+ }
229
+ },
230
+ "robot_motion": {
231
+ "joint_acceleration_rms": 60.54083268321258,
232
+ "joint_effort_rms": 52.524398880409116,
233
+ "joint_jerk_rms": 6193.9154842156095,
234
+ "joint_velocity_rms": 1.2608158422892228
235
+ },
236
+ "grasp": {
237
+ "kinematic_grasp_stability_index": 0.9920513404452811,
238
+ "mean_kinematic_grasp_stability": 0.49255223784125557,
239
+ "per_object_best_kinematic_grasp_stability": {
240
+ "obj_043_phillips_screwdriver_1": 0.8601013449686868,
241
+ "obj_044_flat_screwdriver_2": 0.8560972009461396,
242
+ "obj_048_hammer_3": 0.6766763386988689
243
+ },
244
+ "grasp_gsi_diagnostics": {
245
+ "mean_anti_slip_score": 0.7336397597182205,
246
+ "mean_duration_score": 0.5031793066139518,
247
+ "mean_hold_score": 0.43041748134352503,
248
+ "mean_kinematic_stability_score": 0.19259010619149425,
249
+ "mean_legacy_gsi": 0.5083334823361455,
250
+ "mean_lift_score": 0.5635565989627579,
251
+ "mean_motion_stability_score": 0.41729682212718094
252
+ }
253
+ },
254
+ "tool": {
255
+ "tool_selection_accuracy": 0.0,
256
+ "tool_switch_success_rate": 0.0,
257
+ "tool_switch_total": 160,
258
+ "tool_switch_successful": 0
259
+ },
260
+ "runtime": {
261
+ "avg_steps": 2809.28,
262
+ "error_count": 0
263
+ }
264
+ }
265
+ }
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_only/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/inv_only/summary.json ADDED
@@ -0,0 +1,259 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metric_schema_version": "core-v1.8-policy-step-query-split",
3
+ "metadata": {
4
+ "policy_name": "GigaWorld-Policy-0.6",
5
+ "base_seed": 101200000,
6
+ "seed_policy": "episode_seed = base_seed + episode_index - 1",
7
+ "randomization_recorded": true,
8
+ "n_episodes": 50,
9
+ "successes": 40,
10
+ "error_count": 1,
11
+ "evaluation_protocol": "reach_and_stop",
12
+ "early_stop": true,
13
+ "success_definition": "terminal.raw_condition holds for dwell_time_s",
14
+ "measurement_window": "episode_start_to_stable_success_or_failure_budget",
15
+ "max_episode_steps": 3447,
16
+ "dwell_time_s": 0.5,
17
+ "field_semantics": {
18
+ "success_rate": "primary protocol success rate: stable_success for reach_and_stop; terminal at_end_success_budget for fixed_horizon",
19
+ "avg_time_to_success_s": "main efficiency metric: mean time_to_stable_success_s over reached-success episodes",
20
+ "task_efficiency": "benchmark-specific expert-normalized speed ratio: expert_time_s / time_to_stable_success_s over reached-success episodes",
21
+ "avg_steps_to_success": "diagnostic mean steps_to_stable_success over reached-success episodes, or None if no episode succeeds",
22
+ "avg_policy_steps_to_success": "diagnostic mean policy-rate control/action steps through stable success over reached-success episodes",
23
+ "avg_policy_queries_to_success": "diagnostic mean actual model inference queries through stable success over reached-success episodes; None when the first-success query count was not observed",
24
+ "safe_success_rate": "episode success rate with no task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
25
+ "hard_violation_rate": "episode rate with any task-level hard safety violation (drop or tracked-object high speed; joint limits are diagnostic by default)",
26
+ "drop_rate": "episode rate with drop_off_table or drop_to_table_after_lift violation",
27
+ "high_speed_violation_rate": "episode rate with any tracked object exceeding the configured speed threshold",
28
+ "excessive_impact_rate": "deprecated contact-force impact metric; no longer populated by core safety",
29
+ "robot_constraint_violation_rate": "diagnostic episode rate with any asset hard-joint-limit observation; excluded from safe_success_rate unless include_in_core_safety is explicitly enabled",
30
+ "finger_joint_limit_saturation_rate": "episode rate with dexterous finger joints saturated beyond finger joint-limit thresholds; diagnostic by default and not included in safe_success_rate",
31
+ "joint_limit_active_step_rate": "diagnostic mean fraction of steps with any joint exceeding the soft diagnostic margin",
32
+ "finger_joint_limit_saturation_step_rate": "diagnostic mean fraction of steps with finger joint-limit saturation",
33
+ "max_joint_limit_excess_rad": "diagnostic maximum operating/soft-limit excess observed in radians; hard-limit source and role validation are in robot_constraint_diagnostics",
34
+ "safety_violation_rate_time": "compatibility diagnostic; with fixed physics_dt this is equivalent to safety_violation_step_rate and is omitted from grouped summary output",
35
+ "safety_violation_step_rate": "core safety metric: fraction of executed steps with task-level safety violations; joint-limit observations are excluded by default",
36
+ "safety_violation_rate_event": "deprecated compatibility alias of safety_violation_events_per_step; omitted from grouped summary output",
37
+ "safety_violation_events_per_step": "diagnostic event density; may exceed 1 when multiple violations happen in one step",
38
+ "episode_violation_rate": "compatibility alias of hard_violation_rate; omitted from grouped core output",
39
+ "latched_stage_completion_rate": "main progress metric; ever-reached stage completion rate at episode end",
40
+ "current_stage_completion_rate": "diagnostic terminal-state stage completion rate at episode end",
41
+ "current_normalized_progress_score": "chain-depth diagnostic; normalized progress score at episode end",
42
+ "current_chain_depth_progress_score": "alias of current_normalized_progress_score for ordered dependency-depth diagnostics",
43
+ "latched_chain_depth_progress_score": "alias of latched_normalized_progress_score for ever-reached dependency-depth diagnostics",
44
+ "robustness": "success rates grouped by benchmark protocol category: cov, inv, and inv+cov"
45
+ }
46
+ },
47
+ "core_metrics": {
48
+ "completion": {
49
+ "success_rate": 0.8,
50
+ "success_rate_ci": {
51
+ "mean": 0.8,
52
+ "ci_low": 0.68,
53
+ "ci_high": 0.9
54
+ }
55
+ },
56
+ "progress": {
57
+ "latched_stage_completion_rate": 0.905
58
+ },
59
+ "efficiency": {
60
+ "avg_time_to_success_s": 39.21208333333239
61
+ },
62
+ "safety": {
63
+ "safe_success_rate": 0.52,
64
+ "hard_violation_rate": 0.46,
65
+ "drop_rate": 0.18,
66
+ "high_speed_violation_rate": 0.34,
67
+ "safety_violation_step_rate": 0.00029729306837740575
68
+ },
69
+ "robustness": {}
70
+ },
71
+ "diagnostic_metrics": {
72
+ "completion": {
73
+ "ever_instant_success_rate": 0.82,
74
+ "at_end_success_rate": 0.8,
75
+ "task_family_success_rate": {
76
+ "tool_precision_place": 0.8
77
+ }
78
+ },
79
+ "progress": {
80
+ "current_stage_completion_rate": 0.9,
81
+ "current_chain_depth_progress_score": 0.9,
82
+ "latched_chain_depth_progress_score": 0.905,
83
+ "avg_current_chain_depth": 3.6,
84
+ "avg_latched_chain_depth": 3.62
85
+ },
86
+ "efficiency": {
87
+ "task_efficiency": 0.9828144306663895,
88
+ "expert_time_s": 38.28333333333333,
89
+ "expert_time_step": 2297,
90
+ "avg_steps_to_success": 2352.725,
91
+ "avg_policy_steps_to_success": 784.6,
92
+ "avg_policy_queries_to_success": 784.6
93
+ },
94
+ "safety": {
95
+ "safety_violation_events_per_step": 0.00025035205758097326,
96
+ "violation_counts": {
97
+ "drop_to_table_obj_043_phillips_screwdriver_1": 2,
98
+ "drop_to_table_obj_048_hammer_3": 7,
99
+ "high_speed_obj_043_phillips_screwdriver_1": 13,
100
+ "high_speed_obj_048_hammer_3": 10
101
+ },
102
+ "robot_constraint_violation_rate": 1.0,
103
+ "finger_joint_limit_saturation_rate": 1.0,
104
+ "joint_limit_active_step_rate": 0.18580816773587858,
105
+ "finger_joint_limit_saturation_step_rate": 0.14272414332655298,
106
+ "max_joint_limit_excess_rad": 1.4388682842254639,
107
+ "robot_constraint_diagnostics": {
108
+ "joint_limit_soft_counts": {
109
+ "left_index_1_joint": 2938,
110
+ "left_index_2_joint": 14,
111
+ "left_little_1_joint": 6010,
112
+ "left_little_2_joint": 91,
113
+ "left_middle_1_joint": 6416,
114
+ "left_middle_2_joint": 5,
115
+ "left_ring_1_joint": 6430,
116
+ "left_ring_2_joint": 15,
117
+ "left_thumb_2_joint": 6294,
118
+ "left_thumb_3_joint": 3963,
119
+ "left_thumb_4_joint": 3864,
120
+ "right_index_1_joint": 5897,
121
+ "right_little_1_joint": 1743,
122
+ "right_middle_1_joint": 3861,
123
+ "right_ring_1_joint": 2299,
124
+ "right_thumb_2_joint": 577,
125
+ "right_thumb_3_joint": 501,
126
+ "right_thumb_4_joint": 182
127
+ },
128
+ "joint_limit_hard_counts": {
129
+ "left_index_1_joint": 112,
130
+ "left_index_2_joint": 5,
131
+ "left_little_1_joint": 132,
132
+ "left_little_2_joint": 59,
133
+ "left_middle_1_joint": 117,
134
+ "left_middle_2_joint": 17,
135
+ "left_ring_1_joint": 190,
136
+ "left_ring_2_joint": 41,
137
+ "left_thumb_2_joint": 154,
138
+ "left_thumb_3_joint": 61,
139
+ "left_thumb_4_joint": 144,
140
+ "right_index_1_joint": 369,
141
+ "right_little_1_joint": 50,
142
+ "right_middle_1_joint": 68,
143
+ "right_ring_1_joint": 64,
144
+ "right_ring_2_joint": 3,
145
+ "right_thumb_2_joint": 30,
146
+ "right_thumb_3_joint": 21,
147
+ "right_thumb_4_joint": 38
148
+ },
149
+ "joint_limit_hard_episode_events": 649,
150
+ "finger_joint_limit_saturation_counts": {
151
+ "left_index_1_joint": 1281,
152
+ "left_index_2_joint": 7,
153
+ "left_little_1_joint": 4134,
154
+ "left_little_2_joint": 17,
155
+ "left_middle_1_joint": 4804,
156
+ "left_ring_1_joint": 4706,
157
+ "left_thumb_2_joint": 4407,
158
+ "left_thumb_3_joint": 3248,
159
+ "left_thumb_4_joint": 2849,
160
+ "right_index_1_joint": 3631,
161
+ "right_little_1_joint": 903,
162
+ "right_middle_1_joint": 2562,
163
+ "right_ring_1_joint": 1193,
164
+ "right_thumb_2_joint": 6,
165
+ "right_thumb_3_joint": 191,
166
+ "right_thumb_4_joint": 32
167
+ },
168
+ "joint_limit_max_excess_by_joint": {
169
+ "left_index_1_joint": 1.0550408363342285,
170
+ "left_index_2_joint": 0.6662468910217285,
171
+ "left_little_1_joint": 0.9941279888153076,
172
+ "left_little_2_joint": 0.7187364101409912,
173
+ "left_middle_1_joint": 1.2752089500427246,
174
+ "left_middle_2_joint": 0.032596368342638016,
175
+ "left_ring_1_joint": 1.103806495666504,
176
+ "left_ring_2_joint": 0.044780611991882324,
177
+ "left_thumb_1_joint": 1.4731243027199525e-05,
178
+ "left_thumb_2_joint": 1.3467258214950562,
179
+ "left_thumb_3_joint": 1.2803540229797363,
180
+ "left_thumb_4_joint": 1.1851575374603271,
181
+ "right_index_1_joint": 0.5484644174575806,
182
+ "right_index_2_joint": 5.040064024797175e-06,
183
+ "right_little_1_joint": 0.5311143398284912,
184
+ "right_little_2_joint": 2.770118953776546e-06,
185
+ "right_middle_1_joint": 1.0109795331954956,
186
+ "right_middle_2_joint": 6.324823971226579e-06,
187
+ "right_ring_1_joint": 1.4388682842254639,
188
+ "right_ring_2_joint": 0.009177175350487232,
189
+ "right_thumb_2_joint": 0.10043612867593765,
190
+ "right_thumb_3_joint": 0.49819523096084595,
191
+ "right_thumb_4_joint": 0.41772422194480896
192
+ },
193
+ "coverage": {
194
+ "hard_limit_source": {
195
+ "available": 50,
196
+ "unavailable": 0,
197
+ "unknown": 0
198
+ },
199
+ "joint_role_manifest": {
200
+ "available": 50,
201
+ "unavailable": 0,
202
+ "unknown": 0
203
+ }
204
+ },
205
+ "hard_limit_sources": {
206
+ "runtime.joint_pos_limits": 50
207
+ },
208
+ "soft_limit_sources": {
209
+ "runtime.soft_joint_pos_limits": 50
210
+ },
211
+ "unknown_runtime_joint_names": [],
212
+ "baseline_checks": {
213
+ "post_home": {
214
+ "checked": 50,
215
+ "hard_violation_episodes": 0
216
+ },
217
+ "post_reset": {
218
+ "checked": 50,
219
+ "hard_violation_episodes": 0
220
+ }
221
+ }
222
+ }
223
+ },
224
+ "robot_motion": {
225
+ "joint_acceleration_rms": 63.28501680127741,
226
+ "joint_effort_rms": 54.73705314660231,
227
+ "joint_jerk_rms": 6478.355938268087,
228
+ "joint_velocity_rms": 1.3110709520651764
229
+ },
230
+ "grasp": {
231
+ "kinematic_grasp_stability_index": 0.9755920811644978,
232
+ "mean_kinematic_grasp_stability": 0.4812405531973917,
233
+ "per_object_best_kinematic_grasp_stability": {
234
+ "obj_043_phillips_screwdriver_1": 0.8794016420754402,
235
+ "obj_044_flat_screwdriver_2": 0.8174005520209241,
236
+ "obj_048_hammer_3": 0.653717688236517
237
+ },
238
+ "grasp_gsi_diagnostics": {
239
+ "mean_anti_slip_score": 0.7388472423328131,
240
+ "mean_duration_score": 0.4716969184891232,
241
+ "mean_hold_score": 0.39987580635649583,
242
+ "mean_kinematic_stability_score": 0.16912442548005774,
243
+ "mean_legacy_gsi": 0.4874706012204468,
244
+ "mean_lift_score": 0.5400715013920048,
245
+ "mean_motion_stability_score": 0.3900136932136301
246
+ }
247
+ },
248
+ "tool": {
249
+ "tool_selection_accuracy": 0.0,
250
+ "tool_switch_success_rate": 0.0,
251
+ "tool_switch_total": 154,
252
+ "tool_switch_successful": 0
253
+ },
254
+ "runtime": {
255
+ "avg_steps": 2556.4,
256
+ "error_count": 1
257
+ }
258
+ }
259
+ }
gigaworld-policy-06/per_task_eval/12_screwdriver_box_and_hammer/none/per_episode.jsonl ADDED
The diff for this file is too large to render. See raw diff