Hugging Face's logo

Datasets:
MLL-Lab
/
LAGEN-datasets

Tasks:
Robotics
Modalities:
Image
Tabular
Text
Size:
10M - 100M
Tags:
LeRobot
image
video
License:
Dataset card Data Studio Files
xet
 Community
1
Caesarrr commited on
Commit
cf2b236
·
verified ·
1 Parent(s): 979f221

Complete fine-tuning-scope asset ownership transfer before removing old repositories

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_4500.json +63 -0
  2. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_4750.json +63 -0
  3. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_500.json +63 -0
  4. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_5000.json +63 -0
  5. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_750.json +63 -0
  6. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1000.json +63 -0
  7. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1250.json +63 -0
  8. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1500.json +63 -0
  9. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1750.json +63 -0
  10. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2000.json +63 -0
  11. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2250.json +63 -0
  12. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_250.json +63 -0
  13. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2500.json +63 -0
  14. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2750.json +63 -0
  15. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3000.json +63 -0
  16. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3250.json +63 -0
  17. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3500.json +63 -0
  18. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3750.json +63 -0
  19. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4000.json +63 -0
  20. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4250.json +63 -0
  21. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4500.json +63 -0
  22. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4750.json +63 -0
  23. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_500.json +63 -0
  24. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_5000.json +63 -0
  25. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_750.json +63 -0
  26. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1000.json +63 -0
  27. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1250.json +63 -0
  28. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1500.json +63 -0
  29. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1750.json +63 -0
  30. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2000.json +63 -0
  31. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2250.json +63 -0
  32. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_250.json +63 -0
  33. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2500.json +63 -0
  34. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2750.json +63 -0
  35. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_3000.json +63 -0
  36. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_500.json +63 -0
  37. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_750.json +63 -0
  38. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1000.json +63 -0
  39. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1250.json +63 -0
  40. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1500.json +63 -0
  41. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1750.json +63 -0
  42. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2000.json +63 -0
  43. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2250.json +63 -0
  44. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_250.json +63 -0
  45. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2500.json +63 -0
  46. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2750.json +63 -0
  47. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3000.json +63 -0
  48. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3250.json +63 -0
  49. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3500.json +63 -0
  50. experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3750.json +63 -0
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_4500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14356,
26
+ "1": 3644
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_4750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14336,
26
+ "1": 3664
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 146.1799999999996,
7
+ "mean_length": 1197.4,
8
+ "std_reward": 110.17210899315654,
9
+ "std_length": 889.6151077853838,
10
+ "episode_rewards": [
11
+ 232.3999999999929,
12
+ 62.8000000000005,
13
+ 54.90000000000042,
14
+ 59.900000000000475,
15
+ 320.9000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 1894,
19
+ 522,
20
+ 461,
21
+ 502,
22
+ 2608
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 4746,
26
+ "1": 1241
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 146.1799999999996,
55
+ "mean_length": 1197.4,
56
+ "std_reward": 110.17210899315654,
57
+ "std_length": 889.6151077853838,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 146.1799999999996,
60
+ "macro_mean_length": 1197.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_5000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14371,
26
+ "1": 3629
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 5000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning/eval/mid_train/step_750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 3.8999999999999986,
7
+ "mean_length": 50.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 3.8999999999999986,
12
+ 3.8999999999999986,
13
+ 3.8999999999999986,
14
+ 3.8999999999999986,
15
+ 3.8999999999999986
16
+ ],
17
+ "episode_lengths": [
18
+ 50,
19
+ 50,
20
+ 50,
21
+ 50,
22
+ 50
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 210,
26
+ "1": 40
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 3.8999999999999986,
55
+ "mean_length": 50.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 3.8999999999999986,
60
+ "macro_mean_length": 50.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 79.39999999999868,
7
+ "mean_length": 659.2,
8
+ "std_reward": 62.80047770518764,
9
+ "std_length": 506.935656666603,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 31.900000000000162,
13
+ 171.8999999999956,
14
+ 50.40000000000037,
15
+ 134.39999999999728
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 276,
20
+ 1406,
21
+ 425,
22
+ 1103
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2647,
26
+ "1": 649
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 79.39999999999868,
55
+ "mean_length": 659.2,
56
+ "std_reward": 62.80047770518764,
57
+ "std_length": 506.935656666603,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 79.39999999999868,
60
+ "macro_mean_length": 659.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 335.94000000001523,
7
+ "mean_length": 2724.8,
8
+ "std_reward": 154.00615052654007,
9
+ "std_length": 1240.415317544894,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 295.49999999999903,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 50.40000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 2399,
20
+ 3600,
21
+ 3600,
22
+ 425
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 10894,
26
+ "1": 2730
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 335.94000000001523,
55
+ "mean_length": 2724.8,
56
+ "std_reward": 154.00615052654007,
57
+ "std_length": 1240.415317544894,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 335.94000000001523,
60
+ "macro_mean_length": 2724.8,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 292.5400000000077,
7
+ "mean_length": 2375.8,
8
+ "std_reward": 145.61193082987145,
9
+ "std_length": 1173.9889948376858,
10
+ "episode_rewards": [
11
+ 52.70000000000039,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 269.8999999999946,
15
+ 250.89999999999208
16
+ ],
17
+ "episode_lengths": [
18
+ 439,
19
+ 3600,
20
+ 3600,
21
+ 2197,
22
+ 2043
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9500,
26
+ "1": 2379
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 292.5400000000077,
55
+ "mean_length": 2375.8,
56
+ "std_reward": 145.61193082987145,
57
+ "std_length": 1173.9889948376858,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 292.5400000000077,
60
+ "macro_mean_length": 2375.8,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_1750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 390.9600000000196,
7
+ "mean_length": 3168.4,
8
+ "std_reward": 107.2800000000121,
9
+ "std_length": 863.2,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 176.3999999999954,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 1442,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 12633,
26
+ "1": 3209
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 390.9600000000196,
55
+ "mean_length": 3168.4,
56
+ "std_reward": 107.2800000000121,
57
+ "std_length": 863.2,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 390.9600000000196,
60
+ "macro_mean_length": 3168.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 282.3200000000154,
7
+ "mean_length": 2293.0,
8
+ "std_reward": 199.98049304870963,
9
+ "std_length": 1610.6818431956076,
10
+ "episode_rewards": [
11
+ 73.89999999999999,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 3.8999999999999986,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 615,
19
+ 3600,
20
+ 3600,
21
+ 50,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9138,
26
+ "1": 2327
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 282.3200000000154,
55
+ "mean_length": 2293.0,
56
+ "std_reward": 199.98049304870963,
57
+ "std_length": 1610.6818431956076,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 282.3200000000154,
60
+ "macro_mean_length": 2293.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14346,
26
+ "1": 3654
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 65.11999999999905,
7
+ "mean_length": 543.4,
8
+ "std_reward": 57.274371231815906,
9
+ "std_length": 461.5935874771225,
10
+ "episode_rewards": [
11
+ 3.8999999999999986,
12
+ 132.49999999999739,
13
+ 54.90000000000042,
14
+ 3.8999999999999986,
15
+ 130.39999999999745
16
+ ],
17
+ "episode_lengths": [
18
+ 50,
19
+ 1084,
20
+ 461,
21
+ 50,
22
+ 1072
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2142,
26
+ "1": 575
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 65.11999999999905,
55
+ "mean_length": 543.4,
56
+ "std_reward": 57.274371231815906,
57
+ "std_length": 461.5935874771225,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 65.11999999999905,
60
+ "macro_mean_length": 543.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 370.4600000000205,
7
+ "mean_length": 3003.0,
8
+ "std_reward": 148.28000000001026,
9
+ "std_length": 1194.0,
10
+ "episode_rewards": [
11
+ 73.89999999999999,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 615,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 11891,
26
+ "1": 3124
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 370.4600000000205,
55
+ "mean_length": 3003.0,
56
+ "std_reward": 148.28000000001026,
57
+ "std_length": 1194.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 370.4600000000205,
60
+ "macro_mean_length": 3003.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_2750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 218.3800000000104,
7
+ "mean_length": 1778.2,
8
+ "std_reward": 184.83810646077453,
9
+ "std_length": 1488.5559982748382,
10
+ "episode_rewards": [
11
+ 73.89999999999999,
12
+ 444.60000000002566,
13
+ 54.90000000000042,
14
+ 444.60000000002566,
15
+ 73.89999999999999
16
+ ],
17
+ "episode_lengths": [
18
+ 615,
19
+ 3600,
20
+ 461,
21
+ 3600,
22
+ 615
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 7081,
26
+ "1": 1810
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 218.3800000000104,
55
+ "mean_length": 1778.2,
56
+ "std_reward": 184.83810646077453,
57
+ "std_length": 1488.5559982748382,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 218.3800000000104,
60
+ "macro_mean_length": 1778.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 317.72000000001435,
7
+ "mean_length": 2578.6,
8
+ "std_reward": 162.01944821534275,
9
+ "std_length": 1304.6114517357264,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 54.90000000000042,
14
+ 199.89999999999435,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 461,
21
+ 1632,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 10219,
26
+ "1": 2674
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 317.72000000001435,
55
+ "mean_length": 2578.6,
56
+ "std_reward": 162.01944821534275,
57
+ "std_length": 1304.6114517357264,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 317.72000000001435,
60
+ "macro_mean_length": 2578.6,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 365.76000000002057,
7
+ "mean_length": 2965.0,
8
+ "std_reward": 157.6800000000101,
9
+ "std_length": 1270.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 50.40000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 425
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 11775,
26
+ "1": 3050
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 365.76000000002057,
55
+ "mean_length": 2965.0,
56
+ "std_reward": 157.6800000000101,
57
+ "std_length": 1270.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 365.76000000002057,
60
+ "macro_mean_length": 2965.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 408.66000000001924,
7
+ "mean_length": 3311.2,
8
+ "std_reward": 71.8800000000128,
9
+ "std_length": 577.6,
10
+ "episode_rewards": [
11
+ 264.89999999999367,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 2156,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 13206,
26
+ "1": 3350
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 408.66000000001924,
55
+ "mean_length": 3311.2,
56
+ "std_reward": 71.8800000000128,
57
+ "std_length": 577.6,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 408.66000000001924,
60
+ "macro_mean_length": 3311.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_3750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 442.26000000002523,
7
+ "mean_length": 3582.4,
8
+ "std_reward": 4.680000000000791,
9
+ "std_length": 35.2,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 432.9000000000237
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3512
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14270,
26
+ "1": 3642
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 442.26000000002523,
55
+ "mean_length": 3582.4,
56
+ "std_reward": 4.680000000000791,
57
+ "std_length": 35.2,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 442.26000000002523,
60
+ "macro_mean_length": 3582.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 442.26000000002523,
7
+ "mean_length": 3582.4,
8
+ "std_reward": 4.680000000000791,
9
+ "std_length": 35.2,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 432.9000000000237
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3512
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14256,
26
+ "1": 3656
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 442.26000000002523,
55
+ "mean_length": 3582.4,
56
+ "std_reward": 4.680000000000791,
57
+ "std_length": 35.2,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 442.26000000002523,
60
+ "macro_mean_length": 3582.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14338,
26
+ "1": 3662
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 365.76000000002057,
7
+ "mean_length": 2965.0,
8
+ "std_reward": 157.6800000000101,
9
+ "std_length": 1270.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 50.40000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 425
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 11866,
26
+ "1": 2959
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 365.76000000002057,
55
+ "mean_length": 2965.0,
56
+ "std_reward": 157.6800000000101,
57
+ "std_length": 1270.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 365.76000000002057,
60
+ "macro_mean_length": 2965.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_4750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14315,
26
+ "1": 3685
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 4750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 176.28000000000358,
7
+ "mean_length": 1440.4,
8
+ "std_reward": 163.1896246702057,
9
+ "std_length": 1314.9364395285422,
10
+ "episode_rewards": [
11
+ 270.7999999999948,
12
+ 18.200000000000014,
13
+ 17.90000000000001,
14
+ 129.89999999999748,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 2206,
19
+ 166,
20
+ 163,
21
+ 1067,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 5754,
26
+ "1": 1448
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 176.28000000000358,
55
+ "mean_length": 1440.4,
56
+ "std_reward": 163.1896246702057,
57
+ "std_length": 1314.9364395285422,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 176.28000000000358,
60
+ "macro_mean_length": 1440.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_5000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 442.26000000002523,
7
+ "mean_length": 3582.4,
8
+ "std_reward": 4.680000000000791,
9
+ "std_length": 35.2,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 432.9000000000237
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3512
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14239,
26
+ "1": 3673
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 5000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 442.26000000002523,
55
+ "mean_length": 3582.4,
56
+ "std_reward": 4.680000000000791,
57
+ "std_length": 35.2,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 442.26000000002523,
60
+ "macro_mean_length": 3582.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected/eval/mid_train/step_750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 87.69999999999818,
7
+ "mean_length": 726.0,
8
+ "std_reward": 100.46123630535064,
9
+ "std_length": 810.7465695271242,
10
+ "episode_rewards": [
11
+ 264.89999999999367,
12
+ 22.40000000000006,
13
+ 12.899999999999974,
14
+ 3.8999999999999986,
15
+ 134.39999999999728
16
+ ],
17
+ "episode_lengths": [
18
+ 2156,
19
+ 199,
20
+ 122,
21
+ 50,
22
+ 1103
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2897,
26
+ "1": 733
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 87.69999999999818,
55
+ "mean_length": 726.0,
56
+ "std_reward": 100.46123630535064,
57
+ "std_length": 810.7465695271242,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 87.69999999999818,
60
+ "macro_mean_length": 726.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 4.799999999999995,
7
+ "mean_length": 57.2,
8
+ "std_reward": 1.7999999999999952,
9
+ "std_length": 14.4,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 3.8999999999999986,
13
+ 3.8999999999999986,
14
+ 3.8999999999999986,
15
+ 3.8999999999999986
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 50,
20
+ 50,
21
+ 50,
22
+ 50
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 237,
26
+ "1": 49
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 4.799999999999995,
55
+ "mean_length": 57.2,
56
+ "std_reward": 1.7999999999999952,
57
+ "std_length": 14.4,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 4.799999999999995,
60
+ "macro_mean_length": 57.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 4.799999999999995,
7
+ "mean_length": 57.2,
8
+ "std_reward": 1.7999999999999952,
9
+ "std_length": 14.4,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 3.8999999999999986,
13
+ 3.8999999999999986,
14
+ 3.8999999999999986,
15
+ 3.8999999999999986
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 50,
20
+ 50,
21
+ 50,
22
+ 50
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 235,
26
+ "1": 51
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 4.799999999999995,
55
+ "mean_length": 57.2,
56
+ "std_reward": 1.7999999999999952,
57
+ "std_length": 14.4,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 4.799999999999995,
60
+ "macro_mean_length": 57.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 153.74000000000365,
7
+ "mean_length": 1258.2,
8
+ "std_reward": 160.43479173796118,
9
+ "std_length": 1292.4083565189449,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 444.60000000002566,
13
+ 171.8999999999956,
14
+ 3.8999999999999986,
15
+ 139.89999999999702
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 3600,
20
+ 1406,
21
+ 50,
22
+ 1149
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 4978,
26
+ "1": 1313
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 153.74000000000365,
55
+ "mean_length": 1258.2,
56
+ "std_reward": 160.43479173796118,
57
+ "std_length": 1292.4083565189449,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 153.74000000000365,
60
+ "macro_mean_length": 1258.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_1750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 217.38000000000935,
7
+ "mean_length": 1770.0,
8
+ "std_reward": 195.1522113633477,
9
+ "std_length": 1572.0778606672125,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 176.3999999999954,
13
+ 12.899999999999974,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 1442,
20
+ 122,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 7008,
26
+ "1": 1842
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 217.38000000000935,
55
+ "mean_length": 1770.0,
56
+ "std_reward": 195.1522113633477,
57
+ "std_length": 1572.0778606672125,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 217.38000000000935,
60
+ "macro_mean_length": 1770.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 116.27999999999801,
7
+ "mean_length": 956.0,
8
+ "std_reward": 102.48608490912143,
9
+ "std_length": 827.2419235024298,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 176.3999999999954,
13
+ 12.899999999999974,
14
+ 278.89999999999617,
15
+ 104.79999999999863
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 1442,
20
+ 122,
21
+ 2269,
22
+ 861
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 3842,
26
+ "1": 938
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 116.27999999999801,
55
+ "mean_length": 956.0,
56
+ "std_reward": 102.48608490912143,
57
+ "std_length": 827.2419235024298,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 116.27999999999801,
60
+ "macro_mean_length": 956.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 243.26000000001076,
7
+ "mean_length": 1978.4,
8
+ "std_reward": 199.3218462688027,
9
+ "std_length": 1605.7832481378052,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 3.8999999999999986,
15
+ 314.8000000000025
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 3600,
20
+ 3600,
21
+ 50,
22
+ 2556
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 7898,
26
+ "1": 1994
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 243.26000000001076,
55
+ "mean_length": 1978.4,
56
+ "std_reward": 199.3218462688027,
57
+ "std_length": 1605.7832481378052,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 243.26000000001076,
60
+ "macro_mean_length": 1978.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 37.10000000000025,
7
+ "mean_length": 319.0,
8
+ "std_reward": 23.528960878033065,
9
+ "std_length": 189.75668631170814,
10
+ "episode_rewards": [
11
+ 60.600000000000485,
12
+ 13.69999999999997,
13
+ 4.599999999999996,
14
+ 59.900000000000475,
15
+ 46.70000000000033
16
+ ],
17
+ "episode_lengths": [
18
+ 509,
19
+ 130,
20
+ 57,
21
+ 502,
22
+ 397
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 1243,
26
+ "1": 352
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 37.10000000000025,
55
+ "mean_length": 319.0,
56
+ "std_reward": 23.528960878033065,
57
+ "std_length": 189.75668631170814,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 37.10000000000025,
60
+ "macro_mean_length": 319.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 283.2200000000154,
7
+ "mean_length": 2300.2,
8
+ "std_reward": 198.73168242634333,
9
+ "std_length": 1600.6886517995933,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 73.89999999999999
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 615
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9213,
26
+ "1": 2288
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 283.2200000000154,
55
+ "mean_length": 2300.2,
56
+ "std_reward": 198.73168242634333,
57
+ "std_length": 1600.6886517995933,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 283.2200000000154,
60
+ "macro_mean_length": 2300.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_2750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 112.99999999999913,
7
+ "mean_length": 930.4,
8
+ "std_reward": 113.37080752998062,
9
+ "std_length": 914.9803495157696,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 176.3999999999954,
13
+ 73.89999999999999,
14
+ 3.8999999999999986,
15
+ 302.4000000000004
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 1442,
20
+ 615,
21
+ 50,
22
+ 2459
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 3696,
26
+ "1": 956
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 112.99999999999913,
55
+ "mean_length": 930.4,
56
+ "std_reward": 113.37080752998062,
57
+ "std_length": 914.9803495157696,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 112.99999999999913,
60
+ "macro_mean_length": 930.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_3000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 303.72000000001447,
7
+ "mean_length": 2465.6,
8
+ "std_reward": 180.53576266215117,
9
+ "std_length": 1454.0181016754914,
10
+ "episode_rewards": [
11
+ 8.399999999999986,
12
+ 176.3999999999954,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 86,
19
+ 1442,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9797,
26
+ "1": 2531
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 303.72000000001447,
55
+ "mean_length": 2465.6,
56
+ "std_reward": 180.53576266215117,
57
+ "std_length": 1454.0181016754914,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 303.72000000001447,
60
+ "macro_mean_length": 2465.6,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 73.41999999999845,
7
+ "mean_length": 612.0,
8
+ "std_reward": 90.13413116017405,
9
+ "std_length": 727.8139872247579,
10
+ "episode_rewards": [
11
+ 247.19999999999223,
12
+ 22.40000000000006,
13
+ 18.800000000000022,
14
+ 3.8999999999999986,
15
+ 74.79999999999994
16
+ ],
17
+ "episode_lengths": [
18
+ 2015,
19
+ 199,
20
+ 172,
21
+ 50,
22
+ 624
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2359,
26
+ "1": 701
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 73.41999999999845,
55
+ "mean_length": 612.0,
56
+ "std_reward": 90.13413116017405,
57
+ "std_length": 727.8139872247579,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 73.41999999999845,
60
+ "macro_mean_length": 612.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_3k_steps/eval/mid_train/step_750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 72.7599999999988,
7
+ "mean_length": 607.2,
8
+ "std_reward": 66.34210126307238,
9
+ "std_length": 535.8131763964003,
10
+ "episode_rewards": [
11
+ 186.79999999999492,
12
+ 102.79999999999869,
13
+ 18.900000000000023,
14
+ 3.8999999999999986,
15
+ 51.40000000000038
16
+ ],
17
+ "episode_lengths": [
18
+ 1528,
19
+ 850,
20
+ 173,
21
+ 50,
22
+ 435
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2406,
26
+ "1": 630
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 72.7599999999988,
55
+ "mean_length": 607.2,
56
+ "std_reward": 66.34210126307238,
57
+ "std_length": 535.8131763964003,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 72.7599999999988,
60
+ "macro_mean_length": 607.2,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 33.600000000000165,
7
+ "mean_length": 289.4,
8
+ "std_reward": 20.155892438689175,
9
+ "std_length": 162.70291945751927,
10
+ "episode_rewards": [
11
+ 50.40000000000037,
12
+ 22.40000000000006,
13
+ 12.899999999999974,
14
+ 64.40000000000042,
15
+ 17.90000000000001
16
+ ],
17
+ "episode_lengths": [
18
+ 425,
19
+ 199,
20
+ 122,
21
+ 538,
22
+ 163
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 1165,
26
+ "1": 282
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 33.600000000000165,
55
+ "mean_length": 289.4,
56
+ "std_reward": 20.155892438689175,
57
+ "std_length": 162.70291945751927,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 33.600000000000165,
60
+ "macro_mean_length": 289.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 221.74000000000711,
7
+ "mean_length": 1806.8,
8
+ "std_reward": 183.6014226524479,
9
+ "std_length": 1480.1807186962003,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 353.9000000000096,
13
+ 3.8999999999999986,
14
+ 3.8999999999999986,
15
+ 302.4000000000004
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 2875,
20
+ 50,
21
+ 50,
22
+ 2459
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 7195,
26
+ "1": 1839
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 221.74000000000711,
55
+ "mean_length": 1806.8,
56
+ "std_reward": 183.6014226524479,
57
+ "std_length": 1480.1807186962003,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 221.74000000000711,
60
+ "macro_mean_length": 1806.8,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 71.09999999999886,
7
+ "mean_length": 592.4,
8
+ "std_reward": 71.60558637424673,
9
+ "std_length": 577.9593757350078,
10
+ "episode_rewards": [
11
+ 73.89999999999999,
12
+ 199.89999999999435,
13
+ 3.8999999999999986,
14
+ 3.8999999999999986,
15
+ 73.89999999999999
16
+ ],
17
+ "episode_lengths": [
18
+ 615,
19
+ 1632,
20
+ 50,
21
+ 50,
22
+ 615
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 2330,
26
+ "1": 632
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 71.09999999999886,
55
+ "mean_length": 592.4,
56
+ "std_reward": 71.60558637424673,
57
+ "std_length": 577.9593757350078,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 71.09999999999886,
60
+ "macro_mean_length": 592.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_1750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 374.1000000000155,
7
+ "mean_length": 3034.4,
8
+ "std_reward": 96.30771516344092,
9
+ "std_length": 773.4529332803645,
10
+ "episode_rewards": [
11
+ 200.8999999999943,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 335.8000000000064
16
+ ],
17
+ "episode_lengths": [
18
+ 1642,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 2730
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 11876,
26
+ "1": 3296
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 1750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 374.1000000000155,
55
+ "mean_length": 3034.4,
56
+ "std_reward": 96.30771516344092,
57
+ "std_length": 773.4529332803645,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 374.1000000000155,
60
+ "macro_mean_length": 3034.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 174.0400000000039,
7
+ "mean_length": 1421.6,
8
+ "std_reward": 141.63890143602075,
9
+ "std_length": 1140.6557061620304,
10
+ "episode_rewards": [
11
+ 73.89999999999999,
12
+ 176.3999999999954,
13
+ 54.90000000000042,
14
+ 120.39999999999792,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 615,
19
+ 1442,
20
+ 461,
21
+ 990,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 5607,
26
+ "1": 1501
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 174.0400000000039,
55
+ "mean_length": 1421.6,
56
+ "std_reward": 141.63890143602075,
57
+ "std_length": 1140.6557061620304,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 174.0400000000039,
60
+ "macro_mean_length": 1421.6,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 208.66000000000992,
7
+ "mean_length": 1699.0,
8
+ "std_reward": 196.09581943530796,
9
+ "std_length": 1580.14505663246,
10
+ "episode_rewards": [
11
+ 115.89999999999812,
12
+ 444.60000000002566,
13
+ 3.8999999999999986,
14
+ 444.60000000002566,
15
+ 34.30000000000018
16
+ ],
17
+ "episode_lengths": [
18
+ 954,
19
+ 3600,
20
+ 50,
21
+ 3600,
22
+ 291
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 6763,
26
+ "1": 1732
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 208.66000000000992,
55
+ "mean_length": 1699.0,
56
+ "std_reward": 196.09581943530796,
57
+ "std_length": 1580.14505663246,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 208.66000000000992,
60
+ "macro_mean_length": 1699.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 6.119999999999993,
7
+ "mean_length": 66.8,
8
+ "std_reward": 1.8059900331950844,
9
+ "std_length": 14.4,
10
+ "episode_rewards": [
11
+ 3.8999999999999986,
12
+ 6.5999999999999925,
13
+ 8.999999999999984,
14
+ 4.4999999999999964,
15
+ 6.5999999999999925
16
+ ],
17
+ "episode_lengths": [
18
+ 50,
19
+ 68,
20
+ 92,
21
+ 56,
22
+ 68
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 269,
26
+ "1": 65
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 6.119999999999993,
55
+ "mean_length": 66.8,
56
+ "std_reward": 1.8059900331950844,
57
+ "std_length": 14.4,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 6.119999999999993,
60
+ "macro_mean_length": 66.8,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 419.8600000000212,
7
+ "mean_length": 3401.6,
8
+ "std_reward": 49.48000000000879,
9
+ "std_length": 396.8,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 320.9000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 2608
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 13543,
26
+ "1": 3465
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 419.8600000000212,
55
+ "mean_length": 3401.6,
56
+ "std_reward": 49.48000000000879,
57
+ "std_length": 396.8,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 419.8600000000212,
60
+ "macro_mean_length": 3401.6,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_2750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 290.720000000015,
7
+ "mean_length": 2360.8,
8
+ "std_reward": 191.76282642891158,
9
+ "std_length": 1544.3919709711004,
10
+ "episode_rewards": [
11
+ 115.89999999999812,
12
+ 444.60000000002566,
13
+ 3.8999999999999986,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 954,
19
+ 3600,
20
+ 50,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9347,
26
+ "1": 2457
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 2750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 290.720000000015,
55
+ "mean_length": 2360.8,
56
+ "std_reward": 191.76282642891158,
57
+ "std_length": 1544.3919709711004,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 290.720000000015,
60
+ "macro_mean_length": 2360.8,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3000.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 282.3200000000154,
7
+ "mean_length": 2293.0,
8
+ "std_reward": 199.98049304870963,
9
+ "std_length": 1610.6818431956076,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 3.8999999999999986,
14
+ 444.60000000002566,
15
+ 73.89999999999999
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 50,
21
+ 3600,
22
+ 615
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 9093,
26
+ "1": 2372
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3000,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 282.3200000000154,
55
+ "mean_length": 2293.0,
56
+ "std_reward": 199.98049304870963,
57
+ "std_length": 1610.6818431956076,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 282.3200000000154,
60
+ "macro_mean_length": 2293.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3250.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 442.26000000002523,
7
+ "mean_length": 3582.4,
8
+ "std_reward": 4.680000000000791,
9
+ "std_length": 35.2,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 432.9000000000237
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3512
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14322,
26
+ "1": 3590
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3250,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 442.26000000002523,
55
+ "mean_length": 3582.4,
56
+ "std_reward": 4.680000000000791,
57
+ "std_length": 35.2,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 442.26000000002523,
60
+ "macro_mean_length": 3582.4,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3500.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 365.76000000002057,
7
+ "mean_length": 2965.0,
8
+ "std_reward": 157.6800000000101,
9
+ "std_length": 1270.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 50.40000000000037
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 425
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 11739,
26
+ "1": 3086
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3500,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 365.76000000002057,
55
+ "mean_length": 2965.0,
56
+ "std_reward": 157.6800000000101,
57
+ "std_length": 1270.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 365.76000000002057,
60
+ "macro_mean_length": 2965.0,
61
+ "distributed_eval": false
62
+ }
63
+ }
experiments/vla-finetuning-scope/provenance/vla-finetuning-scope-data/source_evaluations/flappy_fix_latency_2_200ep_full_tuning_corrected_bs32_4k_steps/eval/mid_train/step_3750.json ADDED
@@ -0,0 +1,63 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "per_latency": {
3
+ "flappy/latency_2": {
4
+ "latency": 2,
5
+ "num_episodes": 5,
6
+ "mean_reward": 444.60000000002566,
7
+ "mean_length": 3600.0,
8
+ "std_reward": 0.0,
9
+ "std_length": 0.0,
10
+ "episode_rewards": [
11
+ 444.60000000002566,
12
+ 444.60000000002566,
13
+ 444.60000000002566,
14
+ 444.60000000002566,
15
+ 444.60000000002566
16
+ ],
17
+ "episode_lengths": [
18
+ 3600,
19
+ 3600,
20
+ 3600,
21
+ 3600,
22
+ 3600
23
+ ],
24
+ "decoded_action_hist": {
25
+ "0": 14322,
26
+ "1": 3678
27
+ },
28
+ "fixed_episode_seeds": true,
29
+ "eval_seed": 42,
30
+ "episode_seeds": [
31
+ 42,
32
+ 43,
33
+ 44,
34
+ 45,
35
+ 46
36
+ ],
37
+ "episode_indices": [
38
+ 0,
39
+ 1,
40
+ 2,
41
+ 3,
42
+ 4
43
+ ]
44
+ }
45
+ },
46
+ "aggregate": {
47
+ "stage": "mid_train",
48
+ "step": 3750,
49
+ "task": "flappy",
50
+ "model_alias": "openvla",
51
+ "fixed_episode_seeds": true,
52
+ "eval_seed": 42,
53
+ "total_episodes": 5,
54
+ "mean_reward": 444.60000000002566,
55
+ "mean_length": 3600.0,
56
+ "std_reward": 0.0,
57
+ "std_length": 0.0,
58
+ "task_count": 1,
59
+ "macro_mean_reward": 444.60000000002566,
60
+ "macro_mean_length": 3600.0,
61
+ "distributed_eval": false
62
+ }
63
+ }