anhtld commited on
Commit
3d17280
·
verified ·
1 Parent(s): 1457a8a

Auto-sync: 2026-06-28 01:15:29

Browse files
results/README.md CHANGED
@@ -24,6 +24,8 @@ Use these summaries for the current paper tables:
24
  - `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_knn4_v2_summary.md`
25
  - `h16_policy_ckpt_nonexpert_policy_bc5_summary.md`
26
  - `h16_policy_ckpt_nonexpert_policy_bc5_bestpt_field_sweep_summary.md`
 
 
27
 
28
  Known stale or incomplete files:
29
 
@@ -55,21 +57,19 @@ Completed deployment-clean diagnostics:
55
  - Job `14858327` exported the train-split field-selected no-expert target map.
56
  It selected `near_miss` for 1,820 of 2,298 train groups and non-near-miss
57
  no-expert candidates for the rest.
 
 
58
 
59
  Active field-teacher distillation jobs:
60
 
61
- Last checked: `2026-06-28 05:03 UTC`.
62
 
63
- - Job `14858328` is running 3 seeds of `field_selected_noexpert_bc5` with the
64
- train-split target map; jobs `14858329`/`14858330` and
65
- `14858331`/`14858332` evaluate direct and field-guided rollouts afterward.
66
- - Job `14858333` rebuilds `paper_table_status.*` after the train-split
67
- field-teacher summaries.
68
  - Job `14858449` completed the all-split target map export for aligned validation
69
  checkpoint selection and seed-invariant student train coverage.
70
- - Job `14858450` trains 3 seeds of `field_selected_noexpert_bc5_allmap`; jobs
71
- `14858451`/`14858452` and `14858453`/`14858454` evaluate direct and
72
- field-guided rollouts afterward.
 
73
  - Job `14858455` rebuilds `paper_table_status.*` after the allmap summaries.
74
 
75
  Paper table automation:
 
24
  - `h16_policy_ckpt_near_miss_policy_bc5_bestpt_retrieval_residual_knn4_v2_summary.md`
25
  - `h16_policy_ckpt_nonexpert_policy_bc5_summary.md`
26
  - `h16_policy_ckpt_nonexpert_policy_bc5_bestpt_field_sweep_summary.md`
27
+ - `h16_policy_ckpt_field_selected_noexpert_bc5_summary.md`
28
+ - `h16_policy_ckpt_field_selected_noexpert_bc5_bestpt_field_sweep_summary.md`
29
 
30
  Known stale or incomplete files:
31
 
 
57
  - Job `14858327` exported the train-split field-selected no-expert target map.
58
  It selected `near_miss` for 1,820 of 2,298 train groups and non-near-miss
59
  no-expert candidates for the rest.
60
+ - Jobs `14858328`-`14858333` completed train-split field-teacher distillation:
61
+ direct student rollout is 26.84%, and field-guided rollout is 27.65%.
62
 
63
  Active field-teacher distillation jobs:
64
 
65
+ Last checked: `2026-06-28 05:20 UTC`.
66
 
 
 
 
 
 
67
  - Job `14858449` completed the all-split target map export for aligned validation
68
  checkpoint selection and seed-invariant student train coverage.
69
+ - Job `14858450` completed 3 seeds of `field_selected_noexpert_bc5_allmap` with
70
+ 100% train/val target-map coverage.
71
+ - Jobs `14858451`/`14858452` and `14858453`/`14858454` are pending priority and
72
+ will evaluate direct and field-guided allmap rollouts.
73
  - Job `14858455` rebuilds `paper_table_status.*` after the allmap summaries.
74
 
75
  Paper table automation:
results/dovla_cil_run_report_2026-06-27.md CHANGED
@@ -23,6 +23,8 @@ The deployed h=16 behavior-cloning policy remains weak, but the interventional f
23
  | Trust-region field optimization | 25.39% | -4.35 pp | Negative diagnostic; field is not a generic action optimizer |
24
  | Best non-expert proposal policy | 27.88% | -1.86 pp | Broad BC targets do not solve proposal generation |
25
  | Best non-expert proposal + field | 26.49% | -3.25 pp | Broad proposal-field bridge also fails |
 
 
26
  | Train-state residual retrieval | 32.12% | +2.38 pp | Positive clean bridge, below current clean best |
27
  | KNN train-state residual retrieval | 29.91% | +0.17 pp | More retrieved tangent neighborhoods dilute the signal |
28
 
@@ -75,16 +77,16 @@ Completed since the first report:
75
  - Field-teacher target exports now exist:
76
  - train split: 2,298 groups, 1,820 `near_miss`, 478 non-near-miss no-expert targets;
77
  - all split: 2,873 groups, 2,273 `near_miss`, 600 non-near-miss no-expert targets.
 
 
78
 
79
  Active field-teacher jobs:
80
 
81
- - `14858328` is running train-split `field_selected_noexpert_bc5` training;
82
- `14858329`/`14858330` and `14858331`/`14858332` evaluate direct and
83
- field-guided rollouts.
84
- - `14858450` is pending all-split `field_selected_noexpert_bc5_allmap`
85
- training; `14858451`/`14858452` and `14858453`/`14858454` evaluate direct and
86
- field-guided rollouts.
87
- - `14858333` and `14858455` rebuild `paper_table_status.*` after the summaries.
88
 
89
  Use `results/paper_table_status.md` as the canonical current table. The job
90
  table below is historical and includes statuses that were true at earlier
 
23
  | Trust-region field optimization | 25.39% | -4.35 pp | Negative diagnostic; field is not a generic action optimizer |
24
  | Best non-expert proposal policy | 27.88% | -1.86 pp | Broad BC targets do not solve proposal generation |
25
  | Best non-expert proposal + field | 26.49% | -3.25 pp | Broad proposal-field bridge also fails |
26
+ | Field-selected no-expert policy, seed-0 train map | 26.84% | -2.90 pp | Train-split field-teacher student fails |
27
+ | Field-selected no-expert policy + field, seed-0 train map | 27.65% | -2.09 pp | Field scoring around that student remains below baseline |
28
  | Train-state residual retrieval | 32.12% | +2.38 pp | Positive clean bridge, below current clean best |
29
  | KNN train-state residual retrieval | 29.91% | +0.17 pp | More retrieved tangent neighborhoods dilute the signal |
30
 
 
77
  - Field-teacher target exports now exist:
78
  - train split: 2,298 groups, 1,820 `near_miss`, 478 non-near-miss no-expert targets;
79
  - all split: 2,873 groups, 2,273 `near_miss`, 600 non-near-miss no-expert targets.
80
+ - Train-split field-teacher distillation completed as a negative diagnostic:
81
+ direct student 26.84%, field-guided best 27.65%.
82
 
83
  Active field-teacher jobs:
84
 
85
+ - `14858450` completed all-split `field_selected_noexpert_bc5_allmap` training
86
+ with 100% train/val target coverage for all three seeds.
87
+ - `14858451`/`14858452` and `14858453`/`14858454` are pending priority to
88
+ evaluate direct and field-guided allmap rollouts.
89
+ - `14858455` rebuilds `paper_table_status.*` after the allmap summaries.
 
 
90
 
91
  Use `results/paper_table_status.md` as the canonical current table. The job
92
  table below is historical and includes statuses that were true at earlier
results/h16_policy_ckpt_field_selected_noexpert_bc5_bestpt_field_sweep_summary.json ADDED
@@ -0,0 +1,985 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_root": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep",
3
+ "baseline_h4_policy_success": 0.2967,
4
+ "baseline_h16_policy_success": 0.29739130434782607,
5
+ "num_result_files": 12,
6
+ "configs": [
7
+ {
8
+ "config": "k8_sigma0.10",
9
+ "completed_seeds": [
10
+ 0,
11
+ 1,
12
+ 2
13
+ ],
14
+ "num_completed": 3,
15
+ "mean_success": 0.27652173913043476,
16
+ "std_success": 0.04217167183645961,
17
+ "mean_progress": 0.5031051267804444,
18
+ "mean_action_mse_to_best": 0.4708352829712997,
19
+ "absolute_gain_vs_h4_baseline": -0.020178260869565257,
20
+ "absolute_gain_vs_h16_policy": -0.020869565217391306
21
+ },
22
+ {
23
+ "config": "k64_sigma0.50",
24
+ "completed_seeds": [
25
+ 0,
26
+ 1,
27
+ 2
28
+ ],
29
+ "num_completed": 3,
30
+ "mean_success": 0.2707246376811594,
31
+ "std_success": 0.040313828755278176,
32
+ "mean_progress": 0.49595623123586885,
33
+ "mean_action_mse_to_best": 0.4889948419428876,
34
+ "absolute_gain_vs_h4_baseline": -0.025975362318840622,
35
+ "absolute_gain_vs_h16_policy": -0.026666666666666672
36
+ },
37
+ {
38
+ "config": "k16_sigma0.20",
39
+ "completed_seeds": [
40
+ 0,
41
+ 1,
42
+ 2
43
+ ],
44
+ "num_completed": 3,
45
+ "mean_success": 0.2684057971014493,
46
+ "std_success": 0.04516161554449235,
47
+ "mean_progress": 0.49610337290181145,
48
+ "mean_action_mse_to_best": 0.4744613969620263,
49
+ "absolute_gain_vs_h4_baseline": -0.028294202898550724,
50
+ "absolute_gain_vs_h16_policy": -0.028985507246376774
51
+ },
52
+ {
53
+ "config": "k32_sigma0.35",
54
+ "completed_seeds": [
55
+ 0,
56
+ 1,
57
+ 2
58
+ ],
59
+ "num_completed": 3,
60
+ "mean_success": 0.2672463768115942,
61
+ "std_success": 0.044214063688506305,
62
+ "mean_progress": 0.493803747793176,
63
+ "mean_action_mse_to_best": 0.48158605349317624,
64
+ "absolute_gain_vs_h4_baseline": -0.02945362318840583,
65
+ "absolute_gain_vs_h16_policy": -0.03014492753623188
66
+ }
67
+ ],
68
+ "best": {
69
+ "config": "k8_sigma0.10",
70
+ "completed_seeds": [
71
+ 0,
72
+ 1,
73
+ 2
74
+ ],
75
+ "num_completed": 3,
76
+ "mean_success": 0.27652173913043476,
77
+ "std_success": 0.04217167183645961,
78
+ "mean_progress": 0.5031051267804444,
79
+ "mean_action_mse_to_best": 0.4708352829712997,
80
+ "absolute_gain_vs_h4_baseline": -0.020178260869565257,
81
+ "absolute_gain_vs_h16_policy": -0.020869565217391306
82
+ },
83
+ "rows": [
84
+ {
85
+ "config": "k16_sigma0.20",
86
+ "seed": 0,
87
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k16_sigma0.20/seed_0/online_rollout.json",
88
+ "selection_mode": "field",
89
+ "num_candidates": 16,
90
+ "candidate_sigma": 0.2,
91
+ "field_optim_steps": 0,
92
+ "field_optim_step_size": 0.0,
93
+ "field_optim_trust_radius": 0.0,
94
+ "field_optim_l2_penalty": 0.0,
95
+ "retrieval_neighbors": 0,
96
+ "num_groups": 575,
97
+ "policy_rollout_success_rate": 0.21913043478260869,
98
+ "policy_rollout_progress": 0.46793411430032195,
99
+ "oracle_success_rate": 0.8573913043478261,
100
+ "action_mse_to_best": 0.468760979136695,
101
+ "per_task": {
102
+ "LiftPegUpright-v1": {
103
+ "action_mse_to_best": 0.35266343150863944,
104
+ "expert_success_rate": 0.8865979381443299,
105
+ "num_groups": 97,
106
+ "oracle_success_rate": 0.9278350515463918,
107
+ "policy_expert_regret": 1.20295861394135,
108
+ "policy_oracle_regret": 1.2557163656372385,
109
+ "policy_rollout_progress": 0.5160976043067027,
110
+ "policy_rollout_success_rate": 0.14432989690721648,
111
+ "restore_max_error": 4.76837158203125e-07
112
+ },
113
+ "PickCube-v1": {
114
+ "action_mse_to_best": 0.4146603767831738,
115
+ "expert_success_rate": 0.9375,
116
+ "num_groups": 208,
117
+ "oracle_success_rate": 0.9471153846153846,
118
+ "policy_expert_regret": 1.2972651475664372,
119
+ "policy_oracle_regret": 1.3106880149929425,
120
+ "policy_rollout_progress": 0.47517616109241945,
121
+ "policy_rollout_success_rate": 0.13942307692307693,
122
+ "restore_max_error": 4.76837158203125e-07
123
+ },
124
+ "PullCube-v1": {
125
+ "action_mse_to_best": 0.7021898142509646,
126
+ "expert_success_rate": 0.19480519480519481,
127
+ "num_groups": 77,
128
+ "oracle_success_rate": 0.36363636363636365,
129
+ "policy_expert_regret": 0.2625723276604814,
130
+ "policy_oracle_regret": 0.5177143988359793,
131
+ "policy_rollout_progress": 0.267815385828726,
132
+ "policy_rollout_success_rate": 0.12987012987012986,
133
+ "restore_max_error": 4.76837158203125e-07
134
+ },
135
+ "PushCube-v1": {
136
+ "action_mse_to_best": 0.3936375088672689,
137
+ "expert_success_rate": 0.8617021276595744,
138
+ "num_groups": 94,
139
+ "oracle_success_rate": 0.9893617021276596,
140
+ "policy_expert_regret": 0.702704998882527,
141
+ "policy_oracle_regret": 0.8001218893109484,
142
+ "policy_rollout_progress": 0.6057824138948258,
143
+ "policy_rollout_success_rate": 0.574468085106383,
144
+ "restore_max_error": 4.76837158203125e-07
145
+ },
146
+ "StackCube-v1": {
147
+ "action_mse_to_best": 0.5859526287425648,
148
+ "expert_success_rate": 0.6767676767676768,
149
+ "num_groups": 99,
150
+ "oracle_success_rate": 0.8585858585858586,
151
+ "policy_expert_regret": 1.0460213774984533,
152
+ "policy_oracle_regret": 1.1990576034242457,
153
+ "policy_rollout_progress": 0.4302896462907695,
154
+ "policy_rollout_success_rate": 0.1919191919191919,
155
+ "restore_max_error": 3.948807716369629e-07
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "config": "k16_sigma0.20",
161
+ "seed": 1,
162
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k16_sigma0.20/seed_1/online_rollout.json",
163
+ "selection_mode": "field",
164
+ "num_candidates": 16,
165
+ "candidate_sigma": 0.2,
166
+ "field_optim_steps": 0,
167
+ "field_optim_step_size": 0.0,
168
+ "field_optim_trust_radius": 0.0,
169
+ "field_optim_l2_penalty": 0.0,
170
+ "retrieval_neighbors": 0,
171
+ "num_groups": 575,
172
+ "policy_rollout_success_rate": 0.2782608695652174,
173
+ "policy_rollout_progress": 0.5095858010730666,
174
+ "oracle_success_rate": 0.8695652173913043,
175
+ "action_mse_to_best": 0.4680624765392555,
176
+ "per_task": {
177
+ "LiftPegUpright-v1": {
178
+ "action_mse_to_best": 0.35002948106802273,
179
+ "expert_success_rate": 0.8584070796460177,
180
+ "num_groups": 113,
181
+ "oracle_success_rate": 0.9380530973451328,
182
+ "policy_expert_regret": 1.1798317065808626,
183
+ "policy_oracle_regret": 1.2723509379192792,
184
+ "policy_rollout_progress": 0.5208574027086781,
185
+ "policy_rollout_success_rate": 0.1504424778761062,
186
+ "restore_max_error": 4.76837158203125e-07
187
+ },
188
+ "PickCube-v1": {
189
+ "action_mse_to_best": 0.39405514559020166,
190
+ "expert_success_rate": 0.9402173913043478,
191
+ "num_groups": 184,
192
+ "oracle_success_rate": 0.9456521739130435,
193
+ "policy_expert_regret": 1.1305499031998054,
194
+ "policy_oracle_regret": 1.1380868081671551,
195
+ "policy_rollout_progress": 0.5538577482246024,
196
+ "policy_rollout_success_rate": 0.2391304347826087,
197
+ "restore_max_error": 4.76837158203125e-07
198
+ },
199
+ "PullCube-v1": {
200
+ "action_mse_to_best": 0.6909535866426794,
201
+ "expert_success_rate": 0.25,
202
+ "num_groups": 76,
203
+ "oracle_success_rate": 0.40789473684210525,
204
+ "policy_expert_regret": 0.39813052471972216,
205
+ "policy_oracle_regret": 0.6531739636005736,
206
+ "policy_rollout_progress": 0.2255952119900796,
207
+ "policy_rollout_success_rate": 0.07894736842105263,
208
+ "restore_max_error": 4.76837158203125e-07
209
+ },
210
+ "PushCube-v1": {
211
+ "action_mse_to_best": 0.41637755876725857,
212
+ "expert_success_rate": 0.8198198198198198,
213
+ "num_groups": 111,
214
+ "oracle_success_rate": 1.0,
215
+ "policy_expert_regret": 0.39772362980219694,
216
+ "policy_oracle_regret": 0.4817321421326818,
217
+ "policy_rollout_progress": 0.7705201101195704,
218
+ "policy_rollout_success_rate": 0.7477477477477478,
219
+ "restore_max_error": 4.76837158203125e-07
220
+ },
221
+ "StackCube-v1": {
222
+ "action_mse_to_best": 0.6411655412393285,
223
+ "expert_success_rate": 0.6923076923076923,
224
+ "num_groups": 91,
225
+ "oracle_success_rate": 0.8571428571428571,
226
+ "policy_expert_regret": 1.2249158738733648,
227
+ "policy_oracle_regret": 1.3844720531295944,
228
+ "policy_rollout_progress": 0.3249687373965651,
229
+ "policy_rollout_success_rate": 0.10989010989010989,
230
+ "restore_max_error": 3.948807716369629e-07
231
+ }
232
+ }
233
+ },
234
+ {
235
+ "config": "k16_sigma0.20",
236
+ "seed": 2,
237
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k16_sigma0.20/seed_2/online_rollout.json",
238
+ "selection_mode": "field",
239
+ "num_candidates": 16,
240
+ "candidate_sigma": 0.2,
241
+ "field_optim_steps": 0,
242
+ "field_optim_step_size": 0.0,
243
+ "field_optim_trust_radius": 0.0,
244
+ "field_optim_l2_penalty": 0.0,
245
+ "retrieval_neighbors": 0,
246
+ "num_groups": 575,
247
+ "policy_rollout_success_rate": 0.30782608695652175,
248
+ "policy_rollout_progress": 0.5107902033320458,
249
+ "oracle_success_rate": 0.8765217391304347,
250
+ "action_mse_to_best": 0.48656073521012844,
251
+ "per_task": {
252
+ "LiftPegUpright-v1": {
253
+ "action_mse_to_best": 0.37539764823547256,
254
+ "expert_success_rate": 0.8229166666666666,
255
+ "num_groups": 96,
256
+ "oracle_success_rate": 0.9270833333333334,
257
+ "policy_expert_regret": 1.0008620529746015,
258
+ "policy_oracle_regret": 1.1211368835841615,
259
+ "policy_rollout_progress": 0.5729093362266818,
260
+ "policy_rollout_success_rate": 0.23958333333333334,
261
+ "restore_max_error": 3.5762786865234375e-07
262
+ },
263
+ "PickCube-v1": {
264
+ "action_mse_to_best": 0.3807297901372717,
265
+ "expert_success_rate": 0.9444444444444444,
266
+ "num_groups": 198,
267
+ "oracle_success_rate": 0.9595959595959596,
268
+ "policy_expert_regret": 1.0298031912157022,
269
+ "policy_oracle_regret": 1.0368866417444107,
270
+ "policy_rollout_progress": 0.5855083816581309,
271
+ "policy_rollout_success_rate": 0.3181818181818182,
272
+ "restore_max_error": 4.76837158203125e-07
273
+ },
274
+ "PullCube-v1": {
275
+ "action_mse_to_best": 0.7341989019678699,
276
+ "expert_success_rate": 0.24444444444444444,
277
+ "num_groups": 90,
278
+ "oracle_success_rate": 0.4666666666666667,
279
+ "policy_expert_regret": 0.36530390631540083,
280
+ "policy_oracle_regret": 0.6634106457784785,
281
+ "policy_rollout_progress": 0.2576814704747651,
282
+ "policy_rollout_success_rate": 0.13333333333333333,
283
+ "restore_max_error": 4.0978193283081055e-07
284
+ },
285
+ "PushCube-v1": {
286
+ "action_mse_to_best": 0.45260379271637097,
287
+ "expert_success_rate": 0.8514851485148515,
288
+ "num_groups": 101,
289
+ "oracle_success_rate": 1.0,
290
+ "policy_expert_regret": 0.47767280514287475,
291
+ "policy_oracle_regret": 0.6053462448981729,
292
+ "policy_rollout_progress": 0.711485438270144,
293
+ "policy_rollout_success_rate": 0.6831683168316832,
294
+ "restore_max_error": 4.76837158203125e-07
295
+ },
296
+ "StackCube-v1": {
297
+ "action_mse_to_best": 0.6284318425175216,
298
+ "expert_success_rate": 0.7666666666666667,
299
+ "num_groups": 90,
300
+ "oracle_success_rate": 0.9111111111111111,
301
+ "policy_expert_regret": 1.31410144418478,
302
+ "policy_oracle_regret": 1.4628699377179146,
303
+ "policy_rollout_progress": 0.3080338829093509,
304
+ "policy_rollout_success_rate": 0.1111111111111111,
305
+ "restore_max_error": 4.76837158203125e-07
306
+ }
307
+ }
308
+ },
309
+ {
310
+ "config": "k32_sigma0.35",
311
+ "seed": 0,
312
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k32_sigma0.35/seed_0/online_rollout.json",
313
+ "selection_mode": "field",
314
+ "num_candidates": 32,
315
+ "candidate_sigma": 0.35,
316
+ "field_optim_steps": 0,
317
+ "field_optim_step_size": 0.0,
318
+ "field_optim_trust_radius": 0.0,
319
+ "field_optim_l2_penalty": 0.0,
320
+ "retrieval_neighbors": 0,
321
+ "num_groups": 575,
322
+ "policy_rollout_success_rate": 0.21913043478260869,
323
+ "policy_rollout_progress": 0.46738898912987986,
324
+ "oracle_success_rate": 0.8573913043478261,
325
+ "action_mse_to_best": 0.47375724539484665,
326
+ "per_task": {
327
+ "LiftPegUpright-v1": {
328
+ "action_mse_to_best": 0.3524797545726766,
329
+ "expert_success_rate": 0.8865979381443299,
330
+ "num_groups": 97,
331
+ "oracle_success_rate": 0.9278350515463918,
332
+ "policy_expert_regret": 1.1849500404805253,
333
+ "policy_oracle_regret": 1.2377077921764137,
334
+ "policy_rollout_progress": 0.523796899417012,
335
+ "policy_rollout_success_rate": 0.15463917525773196,
336
+ "restore_max_error": 4.76837158203125e-07
337
+ },
338
+ "PickCube-v1": {
339
+ "action_mse_to_best": 0.41800202110495704,
340
+ "expert_success_rate": 0.9375,
341
+ "num_groups": 208,
342
+ "oracle_success_rate": 0.9471153846153846,
343
+ "policy_expert_regret": 1.293418727734556,
344
+ "policy_oracle_regret": 1.3068415951610615,
345
+ "policy_rollout_progress": 0.4790225809243006,
346
+ "policy_rollout_success_rate": 0.13942307692307693,
347
+ "restore_max_error": 4.76837158203125e-07
348
+ },
349
+ "PullCube-v1": {
350
+ "action_mse_to_best": 0.7232648653595092,
351
+ "expert_success_rate": 0.19480519480519481,
352
+ "num_groups": 77,
353
+ "oracle_success_rate": 0.36363636363636365,
354
+ "policy_expert_regret": 0.2673212721985684,
355
+ "policy_oracle_regret": 0.5014150884479699,
356
+ "policy_rollout_progress": 0.27085492715014764,
357
+ "policy_rollout_success_rate": 0.14285714285714285,
358
+ "restore_max_error": 4.76837158203125e-07
359
+ },
360
+ "PushCube-v1": {
361
+ "action_mse_to_best": 0.3973129745176498,
362
+ "expert_success_rate": 0.8617021276595744,
363
+ "num_groups": 94,
364
+ "oracle_success_rate": 0.9893617021276596,
365
+ "policy_expert_regret": 0.7247732332412232,
366
+ "policy_oracle_regret": 0.8431307670283825,
367
+ "policy_rollout_progress": 0.5840501319220726,
368
+ "policy_rollout_success_rate": 0.5531914893617021,
369
+ "restore_max_error": 4.76837158203125e-07
370
+ },
371
+ "StackCube-v1": {
372
+ "action_mse_to_best": 0.5882489422355035,
373
+ "expert_success_rate": 0.6767676767676768,
374
+ "num_groups": 99,
375
+ "oracle_success_rate": 0.8585858585858586,
376
+ "policy_expert_regret": 1.046542024973667,
377
+ "policy_oracle_regret": 1.1995782508994595,
378
+ "policy_rollout_progress": 0.42976899881555575,
379
+ "policy_rollout_success_rate": 0.1919191919191919,
380
+ "restore_max_error": 3.948807716369629e-07
381
+ }
382
+ }
383
+ },
384
+ {
385
+ "config": "k32_sigma0.35",
386
+ "seed": 1,
387
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k32_sigma0.35/seed_1/online_rollout.json",
388
+ "selection_mode": "field",
389
+ "num_candidates": 32,
390
+ "candidate_sigma": 0.35,
391
+ "field_optim_steps": 0,
392
+ "field_optim_step_size": 0.0,
393
+ "field_optim_trust_radius": 0.0,
394
+ "field_optim_l2_penalty": 0.0,
395
+ "retrieval_neighbors": 0,
396
+ "num_groups": 575,
397
+ "policy_rollout_success_rate": 0.27652173913043476,
398
+ "policy_rollout_progress": 0.5041067413548412,
399
+ "oracle_success_rate": 0.8695652173913043,
400
+ "action_mse_to_best": 0.4759621399697726,
401
+ "per_task": {
402
+ "LiftPegUpright-v1": {
403
+ "action_mse_to_best": 0.349115602874024,
404
+ "expert_success_rate": 0.8584070796460177,
405
+ "num_groups": 113,
406
+ "oracle_success_rate": 0.9380530973451328,
407
+ "policy_expert_regret": 1.1531897760861742,
408
+ "policy_oracle_regret": 1.245709007424591,
409
+ "policy_rollout_progress": 0.5298002181591186,
410
+ "policy_rollout_success_rate": 0.168141592920354,
411
+ "restore_max_error": 4.76837158203125e-07
412
+ },
413
+ "PickCube-v1": {
414
+ "action_mse_to_best": 0.39974564475857693,
415
+ "expert_success_rate": 0.9402173913043478,
416
+ "num_groups": 184,
417
+ "oracle_success_rate": 0.9456521739130435,
418
+ "policy_expert_regret": 1.1354111315753392,
419
+ "policy_oracle_regret": 1.1429480365426887,
420
+ "policy_rollout_progress": 0.5489965198490687,
421
+ "policy_rollout_success_rate": 0.2391304347826087,
422
+ "restore_max_error": 4.76837158203125e-07
423
+ },
424
+ "PullCube-v1": {
425
+ "action_mse_to_best": 0.7238199750672242,
426
+ "expert_success_rate": 0.25,
427
+ "num_groups": 76,
428
+ "oracle_success_rate": 0.40789473684210525,
429
+ "policy_expert_regret": 0.451350043429748,
430
+ "policy_oracle_regret": 0.6905959643129456,
431
+ "policy_rollout_progress": 0.19615636427739733,
432
+ "policy_rollout_success_rate": 0.06578947368421052,
433
+ "restore_max_error": 4.76837158203125e-07
434
+ },
435
+ "PushCube-v1": {
436
+ "action_mse_to_best": 0.4286191267473204,
437
+ "expert_success_rate": 0.8198198198198198,
438
+ "num_groups": 111,
439
+ "oracle_success_rate": 1.0,
440
+ "policy_expert_regret": 0.4146888485631427,
441
+ "policy_oracle_regret": 0.5320024189648328,
442
+ "policy_rollout_progress": 0.7472768603144465,
443
+ "policy_rollout_success_rate": 0.7207207207207207,
444
+ "restore_max_error": 4.76837158203125e-07
445
+ },
446
+ "StackCube-v1": {
447
+ "action_mse_to_best": 0.6383288741562065,
448
+ "expert_success_rate": 0.6923076923076923,
449
+ "num_groups": 91,
450
+ "oracle_success_rate": 0.8571428571428571,
451
+ "policy_expert_regret": 1.1968848662717002,
452
+ "policy_oracle_regret": 1.3564410455279299,
453
+ "policy_rollout_progress": 0.3420107340092187,
454
+ "policy_rollout_success_rate": 0.12087912087912088,
455
+ "restore_max_error": 3.948807716369629e-07
456
+ }
457
+ }
458
+ },
459
+ {
460
+ "config": "k32_sigma0.35",
461
+ "seed": 2,
462
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k32_sigma0.35/seed_2/online_rollout.json",
463
+ "selection_mode": "field",
464
+ "num_candidates": 32,
465
+ "candidate_sigma": 0.35,
466
+ "field_optim_steps": 0,
467
+ "field_optim_step_size": 0.0,
468
+ "field_optim_trust_radius": 0.0,
469
+ "field_optim_l2_penalty": 0.0,
470
+ "retrieval_neighbors": 0,
471
+ "num_groups": 575,
472
+ "policy_rollout_success_rate": 0.3060869565217391,
473
+ "policy_rollout_progress": 0.5099155128948069,
474
+ "oracle_success_rate": 0.8765217391304347,
475
+ "action_mse_to_best": 0.49503877511490946,
476
+ "per_task": {
477
+ "LiftPegUpright-v1": {
478
+ "action_mse_to_best": 0.3736456293069447,
479
+ "expert_success_rate": 0.8229166666666666,
480
+ "num_groups": 96,
481
+ "oracle_success_rate": 0.9270833333333334,
482
+ "policy_expert_regret": 1.0010371517079573,
483
+ "policy_oracle_regret": 1.1213119823175173,
484
+ "policy_rollout_progress": 0.5755758554053804,
485
+ "policy_rollout_success_rate": 0.25,
486
+ "restore_max_error": 3.5762786865234375e-07
487
+ },
488
+ "PickCube-v1": {
489
+ "action_mse_to_best": 0.3811475902494758,
490
+ "expert_success_rate": 0.9444444444444444,
491
+ "num_groups": 198,
492
+ "oracle_success_rate": 0.9595959595959596,
493
+ "policy_expert_regret": 1.0301577876764108,
494
+ "policy_oracle_regret": 1.037241238205119,
495
+ "policy_rollout_progress": 0.5851537851974224,
496
+ "policy_rollout_success_rate": 0.3181818181818182,
497
+ "restore_max_error": 4.76837158203125e-07
498
+ },
499
+ "PullCube-v1": {
500
+ "action_mse_to_best": 0.7827393248677254,
501
+ "expert_success_rate": 0.24444444444444444,
502
+ "num_groups": 90,
503
+ "oracle_success_rate": 0.4666666666666667,
504
+ "policy_expert_regret": 0.39821494068536495,
505
+ "policy_oracle_regret": 0.6810178060084582,
506
+ "policy_rollout_progress": 0.24872837766177125,
507
+ "policy_rollout_success_rate": 0.12222222222222222,
508
+ "restore_max_error": 4.0978193283081055e-07
509
+ },
510
+ "PushCube-v1": {
511
+ "action_mse_to_best": 0.4594265000507383,
512
+ "expert_success_rate": 0.8514851485148515,
513
+ "num_groups": 101,
514
+ "oracle_success_rate": 1.0,
515
+ "policy_expert_regret": 0.4958930028842227,
516
+ "policy_oracle_regret": 0.6248585299985243,
517
+ "policy_rollout_progress": 0.7018741432688024,
518
+ "policy_rollout_success_rate": 0.6732673267326733,
519
+ "restore_max_error": 4.76837158203125e-07
520
+ },
521
+ "StackCube-v1": {
522
+ "action_mse_to_best": 0.6273497407221131,
523
+ "expert_success_rate": 0.7666666666666667,
524
+ "num_groups": 90,
525
+ "oracle_success_rate": 0.9111111111111111,
526
+ "policy_expert_regret": 1.302014817463027,
527
+ "policy_oracle_regret": 1.4507833109961616,
528
+ "policy_rollout_progress": 0.32012050963110394,
529
+ "policy_rollout_success_rate": 0.1111111111111111,
530
+ "restore_max_error": 4.76837158203125e-07
531
+ }
532
+ }
533
+ },
534
+ {
535
+ "config": "k64_sigma0.50",
536
+ "seed": 0,
537
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k64_sigma0.50/seed_0/online_rollout.json",
538
+ "selection_mode": "field",
539
+ "num_candidates": 64,
540
+ "candidate_sigma": 0.5,
541
+ "field_optim_steps": 0,
542
+ "field_optim_step_size": 0.0,
543
+ "field_optim_trust_radius": 0.0,
544
+ "field_optim_l2_penalty": 0.0,
545
+ "retrieval_neighbors": 0,
546
+ "num_groups": 575,
547
+ "policy_rollout_success_rate": 0.22782608695652173,
548
+ "policy_rollout_progress": 0.474431411483649,
549
+ "oracle_success_rate": 0.8573913043478261,
550
+ "action_mse_to_best": 0.47588635947069396,
551
+ "per_task": {
552
+ "LiftPegUpright-v1": {
553
+ "action_mse_to_best": 0.35277089376732246,
554
+ "expert_success_rate": 0.8865979381443299,
555
+ "num_groups": 97,
556
+ "oracle_success_rate": 0.9278350515463918,
557
+ "policy_expert_regret": 1.184891695214301,
558
+ "policy_oracle_regret": 1.2376494469101895,
559
+ "policy_rollout_progress": 0.5238552446832362,
560
+ "policy_rollout_success_rate": 0.15463917525773196,
561
+ "restore_max_error": 4.76837158203125e-07
562
+ },
563
+ "PickCube-v1": {
564
+ "action_mse_to_best": 0.41841672013442105,
565
+ "expert_success_rate": 0.9375,
566
+ "num_groups": 208,
567
+ "oracle_success_rate": 0.9471153846153846,
568
+ "policy_expert_regret": 1.2863650867590108,
569
+ "policy_oracle_regret": 1.2997879541855162,
570
+ "policy_rollout_progress": 0.48126852959215355,
571
+ "policy_rollout_success_rate": 0.14423076923076922,
572
+ "restore_max_error": 4.76837158203125e-07
573
+ },
574
+ "PullCube-v1": {
575
+ "action_mse_to_best": 0.7349351081126309,
576
+ "expert_success_rate": 0.19480519480519481,
577
+ "num_groups": 77,
578
+ "oracle_success_rate": 0.36363636363636365,
579
+ "policy_expert_regret": 0.2896711546761383,
580
+ "policy_oracle_regret": 0.501337112019794,
581
+ "policy_rollout_progress": 0.28172497145752967,
582
+ "policy_rollout_success_rate": 0.15584415584415584,
583
+ "restore_max_error": 4.76837158203125e-07
584
+ },
585
+ "PushCube-v1": {
586
+ "action_mse_to_best": 0.3999176624607533,
587
+ "expert_success_rate": 0.8617021276595744,
588
+ "num_groups": 94,
589
+ "oracle_success_rate": 0.9893617021276596,
590
+ "policy_expert_regret": 0.6628930673320242,
591
+ "policy_oracle_regret": 0.7818994298577309,
592
+ "policy_rollout_progress": 0.613366575475703,
593
+ "policy_rollout_success_rate": 0.5851063829787234,
594
+ "restore_max_error": 4.76837158203125e-07
595
+ },
596
+ "StackCube-v1": {
597
+ "action_mse_to_best": 0.5879084708714726,
598
+ "expert_success_rate": 0.6767676767676768,
599
+ "num_groups": 99,
600
+ "oracle_success_rate": 0.8585858585858586,
601
+ "policy_expert_regret": 1.0467052882668948,
602
+ "policy_oracle_regret": 1.199741514192687,
603
+ "policy_rollout_progress": 0.429605735522328,
604
+ "policy_rollout_success_rate": 0.1919191919191919,
605
+ "restore_max_error": 3.948807716369629e-07
606
+ }
607
+ }
608
+ },
609
+ {
610
+ "config": "k64_sigma0.50",
611
+ "seed": 1,
612
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k64_sigma0.50/seed_1/online_rollout.json",
613
+ "selection_mode": "field",
614
+ "num_candidates": 64,
615
+ "candidate_sigma": 0.5,
616
+ "field_optim_steps": 0,
617
+ "field_optim_step_size": 0.0,
618
+ "field_optim_trust_radius": 0.0,
619
+ "field_optim_l2_penalty": 0.0,
620
+ "retrieval_neighbors": 0,
621
+ "num_groups": 575,
622
+ "policy_rollout_success_rate": 0.27652173913043476,
623
+ "policy_rollout_progress": 0.5038111128841631,
624
+ "oracle_success_rate": 0.8695652173913043,
625
+ "action_mse_to_best": 0.485377110492681,
626
+ "per_task": {
627
+ "LiftPegUpright-v1": {
628
+ "action_mse_to_best": 0.3496806182470008,
629
+ "expert_success_rate": 0.8584070796460177,
630
+ "num_groups": 113,
631
+ "oracle_success_rate": 0.9380530973451328,
632
+ "policy_expert_regret": 1.169163638643459,
633
+ "policy_oracle_regret": 1.2616828699818754,
634
+ "policy_rollout_progress": 0.5226759131239579,
635
+ "policy_rollout_success_rate": 0.1592920353982301,
636
+ "restore_max_error": 4.76837158203125e-07
637
+ },
638
+ "PickCube-v1": {
639
+ "action_mse_to_best": 0.4060886063329551,
640
+ "expert_success_rate": 0.9402173913043478,
641
+ "num_groups": 184,
642
+ "oracle_success_rate": 0.9456521739130435,
643
+ "policy_expert_regret": 1.13451692196743,
644
+ "policy_oracle_regret": 1.1420538269347795,
645
+ "policy_rollout_progress": 0.549890729456978,
646
+ "policy_rollout_success_rate": 0.2391304347826087,
647
+ "restore_max_error": 4.76837158203125e-07
648
+ },
649
+ "PullCube-v1": {
650
+ "action_mse_to_best": 0.7909409432690345,
651
+ "expert_success_rate": 0.25,
652
+ "num_groups": 76,
653
+ "oracle_success_rate": 0.40789473684210525,
654
+ "policy_expert_regret": 0.43219950842555926,
655
+ "policy_oracle_regret": 0.6995100968010775,
656
+ "policy_rollout_progress": 0.18716378344975296,
657
+ "policy_rollout_success_rate": 0.06578947368421052,
658
+ "restore_max_error": 4.76837158203125e-07
659
+ },
660
+ "PushCube-v1": {
661
+ "action_mse_to_best": 0.41940284178063675,
662
+ "expert_success_rate": 0.8198198198198198,
663
+ "num_groups": 111,
664
+ "oracle_success_rate": 1.0,
665
+ "policy_expert_regret": 0.3959935593443948,
666
+ "policy_oracle_regret": 0.5129548193098189,
667
+ "policy_rollout_progress": 0.7573154509604514,
668
+ "policy_rollout_success_rate": 0.7297297297297297,
669
+ "restore_max_error": 4.76837158203125e-07
670
+ },
671
+ "StackCube-v1": {
672
+ "action_mse_to_best": 0.6394769008793346,
673
+ "expert_success_rate": 0.6923076923076923,
674
+ "num_groups": 91,
675
+ "oracle_success_rate": 0.8571428571428571,
676
+ "policy_expert_regret": 1.196448843066509,
677
+ "policy_oracle_regret": 1.3560050223227387,
678
+ "policy_rollout_progress": 0.34244675721441,
679
+ "policy_rollout_success_rate": 0.12087912087912088,
680
+ "restore_max_error": 3.948807716369629e-07
681
+ }
682
+ }
683
+ },
684
+ {
685
+ "config": "k64_sigma0.50",
686
+ "seed": 2,
687
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k64_sigma0.50/seed_2/online_rollout.json",
688
+ "selection_mode": "field",
689
+ "num_candidates": 64,
690
+ "candidate_sigma": 0.5,
691
+ "field_optim_steps": 0,
692
+ "field_optim_step_size": 0.0,
693
+ "field_optim_trust_radius": 0.0,
694
+ "field_optim_l2_penalty": 0.0,
695
+ "retrieval_neighbors": 0,
696
+ "num_groups": 575,
697
+ "policy_rollout_success_rate": 0.30782608695652175,
698
+ "policy_rollout_progress": 0.5096261693397944,
699
+ "oracle_success_rate": 0.8765217391304347,
700
+ "action_mse_to_best": 0.5057210558652878,
701
+ "per_task": {
702
+ "LiftPegUpright-v1": {
703
+ "action_mse_to_best": 0.3736456293069447,
704
+ "expert_success_rate": 0.8229166666666666,
705
+ "num_groups": 96,
706
+ "oracle_success_rate": 0.9270833333333334,
707
+ "policy_expert_regret": 1.0010371517079573,
708
+ "policy_oracle_regret": 1.1213119823175173,
709
+ "policy_rollout_progress": 0.5755758554053804,
710
+ "policy_rollout_success_rate": 0.25,
711
+ "restore_max_error": 3.5762786865234375e-07
712
+ },
713
+ "PickCube-v1": {
714
+ "action_mse_to_best": 0.3808895478946994,
715
+ "expert_success_rate": 0.9444444444444444,
716
+ "num_groups": 198,
717
+ "oracle_success_rate": 0.9595959595959596,
718
+ "policy_expert_regret": 1.0302143845443767,
719
+ "policy_oracle_regret": 1.037297835073085,
720
+ "policy_rollout_progress": 0.5850971883294566,
721
+ "policy_rollout_success_rate": 0.3181818181818182,
722
+ "restore_max_error": 4.76837158203125e-07
723
+ },
724
+ "PullCube-v1": {
725
+ "action_mse_to_best": 0.8449775661859248,
726
+ "expert_success_rate": 0.24444444444444444,
727
+ "num_groups": 90,
728
+ "oracle_success_rate": 0.4666666666666667,
729
+ "policy_expert_regret": 0.42840350720410547,
730
+ "policy_oracle_regret": 0.6953901627307965,
731
+ "policy_rollout_progress": 0.23573277904134657,
732
+ "policy_rollout_success_rate": 0.12222222222222222,
733
+ "restore_max_error": 4.0978193283081055e-07
734
+ },
735
+ "PushCube-v1": {
736
+ "action_mse_to_best": 0.4644910559648334,
737
+ "expert_success_rate": 0.8514851485148515,
738
+ "num_groups": 101,
739
+ "oracle_success_rate": 1.0,
740
+ "policy_expert_regret": 0.4763863344888876,
741
+ "policy_oracle_regret": 0.6050164121212346,
742
+ "policy_rollout_progress": 0.7118152710470823,
743
+ "policy_rollout_success_rate": 0.6831683168316832,
744
+ "restore_max_error": 4.76837158203125e-07
745
+ },
746
+ "StackCube-v1": {
747
+ "action_mse_to_best": 0.628243540186021,
748
+ "expert_success_rate": 0.7666666666666667,
749
+ "num_groups": 90,
750
+ "oracle_success_rate": 0.9111111111111111,
751
+ "policy_expert_regret": 1.3018994440635046,
752
+ "policy_oracle_regret": 1.450667937596639,
753
+ "policy_rollout_progress": 0.3202358830306265,
754
+ "policy_rollout_success_rate": 0.1111111111111111,
755
+ "restore_max_error": 4.76837158203125e-07
756
+ }
757
+ }
758
+ },
759
+ {
760
+ "config": "k8_sigma0.10",
761
+ "seed": 0,
762
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k8_sigma0.10/seed_0/online_rollout.json",
763
+ "selection_mode": "field",
764
+ "num_candidates": 8,
765
+ "candidate_sigma": 0.1,
766
+ "field_optim_steps": 0,
767
+ "field_optim_step_size": 0.0,
768
+ "field_optim_trust_radius": 0.0,
769
+ "field_optim_l2_penalty": 0.0,
770
+ "retrieval_neighbors": 0,
771
+ "num_groups": 575,
772
+ "policy_rollout_success_rate": 0.23130434782608697,
773
+ "policy_rollout_progress": 0.4804371877111818,
774
+ "oracle_success_rate": 0.8573913043478261,
775
+ "action_mse_to_best": 0.4665292473299348,
776
+ "per_task": {
777
+ "LiftPegUpright-v1": {
778
+ "action_mse_to_best": 0.3532839670930941,
779
+ "expert_success_rate": 0.8865979381443299,
780
+ "num_groups": 97,
781
+ "oracle_success_rate": 0.9278350515463918,
782
+ "policy_expert_regret": 1.1681908187792474,
783
+ "policy_oracle_regret": 1.220948570475136,
784
+ "policy_rollout_progress": 0.5405561211182899,
785
+ "policy_rollout_success_rate": 0.15463917525773196,
786
+ "restore_max_error": 4.76837158203125e-07
787
+ },
788
+ "PickCube-v1": {
789
+ "action_mse_to_best": 0.4131943359791946,
790
+ "expert_success_rate": 0.9375,
791
+ "num_groups": 208,
792
+ "oracle_success_rate": 0.9471153846153846,
793
+ "policy_expert_regret": 1.279397780675656,
794
+ "policy_oracle_regret": 1.2928206481021614,
795
+ "policy_rollout_progress": 0.48342814336781603,
796
+ "policy_rollout_success_rate": 0.14903846153846154,
797
+ "restore_max_error": 4.76837158203125e-07
798
+ },
799
+ "PullCube-v1": {
800
+ "action_mse_to_best": 0.6910990537567572,
801
+ "expert_success_rate": 0.19480519480519481,
802
+ "num_groups": 77,
803
+ "oracle_success_rate": 0.36363636363636365,
804
+ "policy_expert_regret": 0.31112918189980765,
805
+ "policy_oracle_regret": 0.5419044978432841,
806
+ "policy_rollout_progress": 0.28447308410684785,
807
+ "policy_rollout_success_rate": 0.15584415584415584,
808
+ "restore_max_error": 4.76837158203125e-07
809
+ },
810
+ "PushCube-v1": {
811
+ "action_mse_to_best": 0.39067850464043463,
812
+ "expert_success_rate": 0.8617021276595744,
813
+ "num_groups": 94,
814
+ "oracle_success_rate": 0.9893617021276596,
815
+ "policy_expert_regret": 0.6222622066419176,
816
+ "policy_oracle_regret": 0.7400320969680523,
817
+ "policy_rollout_progress": 0.6339573126207007,
818
+ "policy_rollout_success_rate": 0.6063829787234043,
819
+ "restore_max_error": 4.76837158203125e-07
820
+ },
821
+ "StackCube-v1": {
822
+ "action_mse_to_best": 0.5868984237125124,
823
+ "expert_success_rate": 0.6767676767676768,
824
+ "num_groups": 99,
825
+ "oracle_success_rate": 0.8585858585858586,
826
+ "policy_expert_regret": 1.0527102315064631,
827
+ "policy_oracle_regret": 1.2175495678728276,
828
+ "policy_rollout_progress": 0.4218986919431975,
829
+ "policy_rollout_success_rate": 0.18181818181818182,
830
+ "restore_max_error": 3.948807716369629e-07
831
+ }
832
+ }
833
+ },
834
+ {
835
+ "config": "k8_sigma0.10",
836
+ "seed": 1,
837
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k8_sigma0.10/seed_1/online_rollout.json",
838
+ "selection_mode": "field",
839
+ "num_candidates": 8,
840
+ "candidate_sigma": 0.1,
841
+ "field_optim_steps": 0,
842
+ "field_optim_step_size": 0.0,
843
+ "field_optim_trust_radius": 0.0,
844
+ "field_optim_l2_penalty": 0.0,
845
+ "retrieval_neighbors": 0,
846
+ "num_groups": 575,
847
+ "policy_rollout_success_rate": 0.28347826086956524,
848
+ "policy_rollout_progress": 0.5112803031795699,
849
+ "oracle_success_rate": 0.8695652173913043,
850
+ "action_mse_to_best": 0.46244008072125525,
851
+ "per_task": {
852
+ "LiftPegUpright-v1": {
853
+ "action_mse_to_best": 0.3500265953691462,
854
+ "expert_success_rate": 0.8584070796460177,
855
+ "num_groups": 113,
856
+ "oracle_success_rate": 0.9380530973451328,
857
+ "policy_expert_regret": 1.1728479576058093,
858
+ "policy_oracle_regret": 1.2653671889442257,
859
+ "policy_rollout_progress": 0.5189915941616076,
860
+ "policy_rollout_success_rate": 0.1592920353982301,
861
+ "restore_max_error": 4.76837158203125e-07
862
+ },
863
+ "PickCube-v1": {
864
+ "action_mse_to_best": 0.39200827625134715,
865
+ "expert_success_rate": 0.9402173913043478,
866
+ "num_groups": 184,
867
+ "oracle_success_rate": 0.9456521739130435,
868
+ "policy_expert_regret": 1.136972747793745,
869
+ "policy_oracle_regret": 1.1445096527610945,
870
+ "policy_rollout_progress": 0.5528696862393585,
871
+ "policy_rollout_success_rate": 0.23369565217391305,
872
+ "restore_max_error": 4.76837158203125e-07
873
+ },
874
+ "PullCube-v1": {
875
+ "action_mse_to_best": 0.6667322731135708,
876
+ "expert_success_rate": 0.25,
877
+ "num_groups": 76,
878
+ "oracle_success_rate": 0.40789473684210525,
879
+ "policy_expert_regret": 0.4254634598956296,
880
+ "policy_oracle_regret": 0.7053027871604028,
881
+ "policy_rollout_progress": 0.21283553334835328,
882
+ "policy_rollout_success_rate": 0.06578947368421052,
883
+ "restore_max_error": 4.76837158203125e-07
884
+ },
885
+ "PushCube-v1": {
886
+ "action_mse_to_best": 0.4085549002026652,
887
+ "expert_success_rate": 0.8198198198198198,
888
+ "num_groups": 111,
889
+ "oracle_success_rate": 1.0,
890
+ "policy_expert_regret": 0.3684367251557273,
891
+ "policy_oracle_regret": 0.4687258407876298,
892
+ "policy_rollout_progress": 0.7745174024556134,
893
+ "policy_rollout_success_rate": 0.7567567567567568,
894
+ "restore_max_error": 4.76837158203125e-07
895
+ },
896
+ "StackCube-v1": {
897
+ "action_mse_to_best": 0.639552435479485,
898
+ "expert_success_rate": 0.6923076923076923,
899
+ "num_groups": 91,
900
+ "oracle_success_rate": 0.8571428571428571,
901
+ "policy_expert_regret": 1.1711464825567308,
902
+ "policy_oracle_regret": 1.3307026618129605,
903
+ "policy_rollout_progress": 0.3457710957461661,
904
+ "policy_rollout_success_rate": 0.14285714285714285,
905
+ "restore_max_error": 3.948807716369629e-07
906
+ }
907
+ }
908
+ },
909
+ {
910
+ "config": "k8_sigma0.10",
911
+ "seed": 2,
912
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep/k8_sigma0.10/seed_2/online_rollout.json",
913
+ "selection_mode": "field",
914
+ "num_candidates": 8,
915
+ "candidate_sigma": 0.1,
916
+ "field_optim_steps": 0,
917
+ "field_optim_step_size": 0.0,
918
+ "field_optim_trust_radius": 0.0,
919
+ "field_optim_l2_penalty": 0.0,
920
+ "retrieval_neighbors": 0,
921
+ "num_groups": 575,
922
+ "policy_rollout_success_rate": 0.31478260869565217,
923
+ "policy_rollout_progress": 0.5175978894505814,
924
+ "oracle_success_rate": 0.8765217391304347,
925
+ "action_mse_to_best": 0.48353652086270893,
926
+ "per_task": {
927
+ "LiftPegUpright-v1": {
928
+ "action_mse_to_best": 0.3778388290957082,
929
+ "expert_success_rate": 0.8229166666666666,
930
+ "num_groups": 96,
931
+ "oracle_success_rate": 0.9270833333333334,
932
+ "policy_expert_regret": 1.0527732948151727,
933
+ "policy_oracle_regret": 1.1730423748182754,
934
+ "policy_rollout_progress": 0.5550954629046222,
935
+ "policy_rollout_success_rate": 0.21875,
936
+ "restore_max_error": 3.5762786865234375e-07
937
+ },
938
+ "PickCube-v1": {
939
+ "action_mse_to_best": 0.3805342139017702,
940
+ "expert_success_rate": 0.9444444444444444,
941
+ "num_groups": 198,
942
+ "oracle_success_rate": 0.9595959595959596,
943
+ "policy_expert_regret": 1.0301365685357584,
944
+ "policy_oracle_regret": 1.0372200190644667,
945
+ "policy_rollout_progress": 0.5851750043380749,
946
+ "policy_rollout_success_rate": 0.3181818181818182,
947
+ "restore_max_error": 4.76837158203125e-07
948
+ },
949
+ "PullCube-v1": {
950
+ "action_mse_to_best": 0.7137743220043679,
951
+ "expert_success_rate": 0.24444444444444444,
952
+ "num_groups": 90,
953
+ "oracle_success_rate": 0.4666666666666667,
954
+ "policy_expert_regret": 0.42930986115501985,
955
+ "policy_oracle_regret": 0.705745512040125,
956
+ "policy_rollout_progress": 0.2533828056106965,
957
+ "policy_rollout_success_rate": 0.13333333333333333,
958
+ "restore_max_error": 4.0978193283081055e-07
959
+ },
960
+ "PushCube-v1": {
961
+ "action_mse_to_best": 0.45052769396564746,
962
+ "expert_success_rate": 0.8514851485148515,
963
+ "num_groups": 101,
964
+ "oracle_success_rate": 1.0,
965
+ "policy_expert_regret": 0.43760172889964416,
966
+ "policy_oracle_regret": 0.49276997871918254,
967
+ "policy_rollout_progress": 0.7646557638550749,
968
+ "policy_rollout_success_rate": 0.7425742574257426,
969
+ "restore_max_error": 4.76837158203125e-07
970
+ },
971
+ "StackCube-v1": {
972
+ "action_mse_to_best": 0.6296912386599515,
973
+ "expert_success_rate": 0.7666666666666667,
974
+ "num_groups": 90,
975
+ "oracle_success_rate": 0.9111111111111111,
976
+ "policy_expert_regret": 1.3062432550721699,
977
+ "policy_oracle_regret": 1.4550117486053042,
978
+ "policy_rollout_progress": 0.31589207202196123,
979
+ "policy_rollout_success_rate": 0.1111111111111111,
980
+ "restore_max_error": 4.76837158203125e-07
981
+ }
982
+ }
983
+ }
984
+ ]
985
+ }
results/h16_policy_ckpt_field_selected_noexpert_bc5_bestpt_field_sweep_summary.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # h=16 Field-Guided Rollout Sweep
2
+
3
+ Result root: `/scratch/knguy52/dovla/experiments/dovla_h16_field_selected_noexpert_bc5_bestpt_field_sweep`
4
+ Completed result files: 12
5
+ Baseline h=4 policy success: 29.67%
6
+ Baseline h=16 policy success: 29.74%
7
+
8
+ | config | seeds | mean success | gain vs h=16 | progress | action MSE |
9
+ |---|---:|---:|---:|---:|---:|
10
+ | k8_sigma0.10 | 3 | 27.65% | -2.09% | 50.31% | 0.471 |
11
+ | k64_sigma0.50 | 3 | 27.07% | -2.67% | 49.60% | 0.489 |
12
+ | k16_sigma0.20 | 3 | 26.84% | -2.90% | 49.61% | 0.474 |
13
+ | k32_sigma0.35 | 3 | 26.72% | -3.01% | 49.38% | 0.482 |
14
+
15
+ Best config:
16
+ - k8_sigma0.10
17
+ - mean success: 27.65%
18
+ - gain vs h=16 policy: -2.09%
results/h16_policy_ckpt_field_selected_noexpert_bc5_summary.json ADDED
@@ -0,0 +1,277 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_root": "/scratch/knguy52/dovla/experiments/dovla_h16_policy_ckpt_runs",
3
+ "objective": "field_selected_noexpert_bc5",
4
+ "out_name": "policy_rollout.json",
5
+ "num_completed": 3,
6
+ "baseline_h4_policy_success": 0.2967,
7
+ "baseline_h16_rank_checkpoint_success": 0.29739130434782607,
8
+ "mean_success": 0.2684057971014493,
9
+ "std_success": 0.060479078751281036,
10
+ "mean_progress": 0.49737596602975453,
11
+ "mean_action_mse_to_best": 0.4531931612484049,
12
+ "gain_vs_h4": -0.028294202898550724,
13
+ "gain_vs_h16_rank_checkpoint": -0.028985507246376774,
14
+ "rows": [
15
+ {
16
+ "seed": 0,
17
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_policy_ckpt_runs/field_selected_noexpert_bc5/seed_0/policy_rollout.json",
18
+ "num_groups": 575,
19
+ "selection_mode": "policy",
20
+ "num_candidates": 1,
21
+ "candidate_sigma": 0.0,
22
+ "field_optim_steps": 0,
23
+ "field_optim_step_size": 0.0,
24
+ "field_optim_trust_radius": 0.0,
25
+ "field_optim_l2_penalty": 0.0,
26
+ "retrieval_neighbors": 0,
27
+ "policy_rollout_success_rate": 0.2,
28
+ "policy_rollout_progress": 0.4590369786613661,
29
+ "oracle_success_rate": 0.8573913043478261,
30
+ "action_mse_to_best": 0.43677694921098326,
31
+ "best_policy_val": {
32
+ "bc_loss": 0.20038707554340363,
33
+ "field_effect_loss": 0.010150647670444515,
34
+ "field_potential_loss": 0.24883300893836552,
35
+ "field_preference_loss": 0.5245596203539107,
36
+ "lattice_edges": 3833.3333333333335,
37
+ "progress_mae": 0.1846791985962126,
38
+ "rank_acc": 0.8240882919894325,
39
+ "rank_loss": 0.5245526432991028,
40
+ "regret_mae": 0.41100967427094776,
41
+ "success_accuracy": 0.8816119233767191,
42
+ "total_loss": 1.8707644608285692
43
+ },
44
+ "per_task": {
45
+ "LiftPegUpright-v1": {
46
+ "action_mse_to_best": 0.3570530945051116,
47
+ "expert_success_rate": 0.8865979381443299,
48
+ "num_groups": 97,
49
+ "oracle_success_rate": 0.9278350515463918,
50
+ "policy_expert_regret": 1.3238540102833325,
51
+ "policy_oracle_regret": 1.3634367666600906,
52
+ "policy_rollout_progress": 0.45710550156450763,
53
+ "policy_rollout_success_rate": 0.08247422680412371,
54
+ "restore_max_error": 4.76837158203125e-07
55
+ },
56
+ "PickCube-v1": {
57
+ "action_mse_to_best": 0.3344459112267941,
58
+ "expert_success_rate": 0.9375,
59
+ "num_groups": 208,
60
+ "oracle_success_rate": 0.9471153846153846,
61
+ "policy_expert_regret": 1.2608316384721547,
62
+ "policy_oracle_regret": 1.2608528731999775,
63
+ "policy_rollout_progress": 0.5009997645930315,
64
+ "policy_rollout_success_rate": 0.16346153846153846,
65
+ "restore_max_error": 4.76837158203125e-07
66
+ },
67
+ "PullCube-v1": {
68
+ "action_mse_to_best": 0.680206579047364,
69
+ "expert_success_rate": 0.19480519480519481,
70
+ "num_groups": 77,
71
+ "oracle_success_rate": 0.36363636363636365,
72
+ "policy_expert_regret": 0.3535010172900828,
73
+ "policy_oracle_regret": 0.5732873161262894,
74
+ "policy_rollout_progress": 0.2555536103567907,
75
+ "policy_rollout_success_rate": 0.09090909090909091,
76
+ "restore_max_error": 4.76837158203125e-07
77
+ },
78
+ "PushCube-v1": {
79
+ "action_mse_to_best": 0.3976371214784523,
80
+ "expert_success_rate": 0.8617021276595744,
81
+ "num_groups": 94,
82
+ "oracle_success_rate": 0.9893617021276596,
83
+ "policy_expert_regret": 0.6190679878630536,
84
+ "policy_oracle_regret": 0.7177435708172778,
85
+ "policy_rollout_progress": 0.6545968547146371,
86
+ "policy_rollout_success_rate": 0.6276595744680851,
87
+ "restore_max_error": 4.76837158203125e-07
88
+ },
89
+ "StackCube-v1": {
90
+ "action_mse_to_best": 0.5777176776618669,
91
+ "expert_success_rate": 0.6767676767676768,
92
+ "num_groups": 99,
93
+ "oracle_success_rate": 0.8585858585858586,
94
+ "policy_expert_regret": 1.2022333902241005,
95
+ "policy_oracle_regret": 1.4056525183747512,
96
+ "policy_rollout_progress": 0.34534692628817126,
97
+ "policy_rollout_success_rate": 0.0707070707070707,
98
+ "restore_max_error": 3.948807716369629e-07
99
+ }
100
+ }
101
+ },
102
+ {
103
+ "seed": 1,
104
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_policy_ckpt_runs/field_selected_noexpert_bc5/seed_1/policy_rollout.json",
105
+ "num_groups": 575,
106
+ "selection_mode": "policy",
107
+ "num_candidates": 1,
108
+ "candidate_sigma": 0.0,
109
+ "field_optim_steps": 0,
110
+ "field_optim_step_size": 0.0,
111
+ "field_optim_trust_radius": 0.0,
112
+ "field_optim_l2_penalty": 0.0,
113
+ "retrieval_neighbors": 0,
114
+ "policy_rollout_success_rate": 0.31478260869565217,
115
+ "policy_rollout_progress": 0.5297623421640499,
116
+ "oracle_success_rate": 0.8695652173913043,
117
+ "action_mse_to_best": 0.4462703319169257,
118
+ "best_policy_val": {
119
+ "bc_loss": 0.1842877868976858,
120
+ "field_effect_loss": 0.0097798108537164,
121
+ "field_potential_loss": 0.2688092821174198,
122
+ "field_preference_loss": 0.5151831226216422,
123
+ "lattice_edges": 3833.3333333333335,
124
+ "progress_mae": 0.187625582019488,
125
+ "rank_acc": 0.8292942278914981,
126
+ "rank_loss": 0.5151734915044572,
127
+ "regret_mae": 0.364786328540908,
128
+ "success_accuracy": 0.8715837796529134,
129
+ "total_loss": 1.8006369206640456
130
+ },
131
+ "per_task": {
132
+ "LiftPegUpright-v1": {
133
+ "action_mse_to_best": 0.34565994232085295,
134
+ "expert_success_rate": 0.8584070796460177,
135
+ "num_groups": 113,
136
+ "oracle_success_rate": 0.9380530973451328,
137
+ "policy_expert_regret": 1.0717079918996422,
138
+ "policy_oracle_regret": 1.1398367760455714,
139
+ "policy_rollout_progress": 0.5687134840847117,
140
+ "policy_rollout_success_rate": 0.21238938053097345,
141
+ "restore_max_error": 4.76837158203125e-07
142
+ },
143
+ "PickCube-v1": {
144
+ "action_mse_to_best": 0.38182726992375177,
145
+ "expert_success_rate": 0.9402173913043478,
146
+ "num_groups": 184,
147
+ "oracle_success_rate": 0.9456521739130435,
148
+ "policy_expert_regret": 1.1353175368402963,
149
+ "policy_oracle_regret": 1.1428379987002066,
150
+ "policy_rollout_progress": 0.5414779179365091,
151
+ "policy_rollout_success_rate": 0.2391304347826087,
152
+ "restore_max_error": 4.76837158203125e-07
153
+ },
154
+ "PullCube-v1": {
155
+ "action_mse_to_best": 0.6262720288696552,
156
+ "expert_success_rate": 0.25,
157
+ "num_groups": 76,
158
+ "oracle_success_rate": 0.40789473684210525,
159
+ "policy_expert_regret": 0.4040453817793413,
160
+ "policy_oracle_regret": 0.6600335615343953,
161
+ "policy_rollout_progress": 0.21925117011721196,
162
+ "policy_rollout_success_rate": 0.07894736842105263,
163
+ "restore_max_error": 4.76837158203125e-07
164
+ },
165
+ "PushCube-v1": {
166
+ "action_mse_to_best": 0.3934799292133198,
167
+ "expert_success_rate": 0.8198198198198198,
168
+ "num_groups": 111,
169
+ "oracle_success_rate": 1.0,
170
+ "policy_expert_regret": 0.3318567147126069,
171
+ "policy_oracle_regret": 0.39898568257555234,
172
+ "policy_rollout_progress": 0.8082215246316549,
173
+ "policy_rollout_success_rate": 0.7927927927927928,
174
+ "restore_max_error": 4.76837158203125e-07
175
+ },
176
+ "StackCube-v1": {
177
+ "action_mse_to_best": 0.6155681688706953,
178
+ "expert_success_rate": 0.6923076923076923,
179
+ "num_groups": 91,
180
+ "oracle_success_rate": 0.8571428571428571,
181
+ "policy_expert_regret": 1.086109248818932,
182
+ "policy_oracle_regret": 1.2456936374470429,
183
+ "policy_rollout_progress": 0.3773748129606247,
184
+ "policy_rollout_success_rate": 0.2087912087912088,
185
+ "restore_max_error": 3.948807716369629e-07
186
+ }
187
+ }
188
+ },
189
+ {
190
+ "seed": 2,
191
+ "path": "/scratch/knguy52/dovla/experiments/dovla_h16_policy_ckpt_runs/field_selected_noexpert_bc5/seed_2/policy_rollout.json",
192
+ "num_groups": 575,
193
+ "selection_mode": "policy",
194
+ "num_candidates": 1,
195
+ "candidate_sigma": 0.0,
196
+ "field_optim_steps": 0,
197
+ "field_optim_step_size": 0.0,
198
+ "field_optim_trust_radius": 0.0,
199
+ "field_optim_l2_penalty": 0.0,
200
+ "retrieval_neighbors": 0,
201
+ "policy_rollout_success_rate": 0.29043478260869565,
202
+ "policy_rollout_progress": 0.5033285772638476,
203
+ "oracle_success_rate": 0.8765217391304347,
204
+ "action_mse_to_best": 0.4765322026173058,
205
+ "best_policy_val": {
206
+ "bc_loss": 0.1667875602013535,
207
+ "field_effect_loss": 0.009291487564850185,
208
+ "field_potential_loss": 0.2782222537530793,
209
+ "field_preference_loss": 0.5111869358354144,
210
+ "lattice_edges": 3833.3333333333335,
211
+ "progress_mae": 0.19110752642154694,
212
+ "rank_acc": 0.8316871755652957,
213
+ "rank_loss": 0.511167930232154,
214
+ "regret_mae": 0.3905289289024141,
215
+ "success_accuracy": 0.8701276878515879,
216
+ "total_loss": 1.715771006213294
217
+ },
218
+ "per_task": {
219
+ "LiftPegUpright-v1": {
220
+ "action_mse_to_best": 0.36919080032870016,
221
+ "expert_success_rate": 0.8229166666666666,
222
+ "num_groups": 96,
223
+ "oracle_success_rate": 0.9270833333333334,
224
+ "policy_expert_regret": 1.057313250998656,
225
+ "policy_oracle_regret": 1.1238391535977523,
226
+ "policy_rollout_progress": 0.567264037206769,
227
+ "policy_rollout_success_rate": 0.22916666666666666,
228
+ "restore_max_error": 3.5762786865234375e-07
229
+ },
230
+ "PickCube-v1": {
231
+ "action_mse_to_best": 0.3894048094410788,
232
+ "expert_success_rate": 0.9444444444444444,
233
+ "num_groups": 198,
234
+ "oracle_success_rate": 0.9595959595959596,
235
+ "policy_expert_regret": 1.0942561205928072,
236
+ "policy_oracle_regret": 1.108403751062173,
237
+ "policy_rollout_progress": 0.5686413978198261,
238
+ "policy_rollout_success_rate": 0.2777777777777778,
239
+ "restore_max_error": 4.76837158203125e-07
240
+ },
241
+ "PullCube-v1": {
242
+ "action_mse_to_best": 0.7008399473244531,
243
+ "expert_success_rate": 0.24444444444444444,
244
+ "num_groups": 90,
245
+ "oracle_success_rate": 0.4666666666666667,
246
+ "policy_expert_regret": 0.4289126294147637,
247
+ "policy_oracle_regret": 0.7427257450711396,
248
+ "policy_rollout_progress": 0.2430547617582811,
249
+ "policy_rollout_success_rate": 0.1111111111111111,
250
+ "restore_max_error": 4.0978193283081055e-07
251
+ },
252
+ "PushCube-v1": {
253
+ "action_mse_to_best": 0.4208698135496366,
254
+ "expert_success_rate": 0.8514851485148515,
255
+ "num_groups": 101,
256
+ "oracle_success_rate": 1.0,
257
+ "policy_expert_regret": 0.4975291885951958,
258
+ "policy_oracle_regret": 0.6072372276299071,
259
+ "policy_rollout_progress": 0.7095944555384097,
260
+ "policy_rollout_success_rate": 0.6831683168316832,
261
+ "restore_max_error": 4.76837158203125e-07
262
+ },
263
+ "StackCube-v1": {
264
+ "action_mse_to_best": 0.620867788626088,
265
+ "expert_success_rate": 0.7666666666666667,
266
+ "num_groups": 90,
267
+ "oracle_success_rate": 0.9111111111111111,
268
+ "policy_expert_regret": 1.2906039165125953,
269
+ "policy_oracle_regret": 1.4395524981949064,
270
+ "policy_rollout_progress": 0.32024021132124797,
271
+ "policy_rollout_success_rate": 0.12222222222222222,
272
+ "restore_max_error": 4.76837158203125e-07
273
+ }
274
+ }
275
+ }
276
+ ]
277
+ }
results/h16_policy_ckpt_field_selected_noexpert_bc5_summary.md ADDED
@@ -0,0 +1,19 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # h=16 Best-Policy Checkpoint Rollout
2
+
3
+ Run root: `/scratch/knguy52/dovla/experiments/dovla_h16_policy_ckpt_runs`
4
+ Objective: `field_selected_noexpert_bc5`
5
+ Result file: `policy_rollout.json`
6
+ Completed seeds: 3
7
+ Baseline h=4 policy success: 29.67%
8
+ Baseline h=16 rank-checkpoint success: 29.74%
9
+
10
+ Mean success: 26.84% +/- 6.05%
11
+ Gain vs h=16 rank checkpoint: -2.90%
12
+ Mean progress: 49.74%
13
+ Mean action MSE to best: 0.453
14
+
15
+ | seed | mode | k | retrieval K | sigma | opt steps | trust | success | progress | oracle | action MSE |
16
+ |---:|---|---:|---:|---:|---:|---:|---:|---:|---:|---:|
17
+ | 0 | policy | 1 | 0 | 0.00 | 0 | 0.00 | 20.00% | 45.90% | 85.74% | 0.437 |
18
+ | 1 | policy | 1 | 0 | 0.00 | 0 | 0.00 | 31.48% | 52.98% | 86.96% | 0.446 |
19
+ | 2 | policy | 1 | 0 | 0.00 | 0 | 0.00 | 29.04% | 50.33% | 87.65% | 0.477 |