Linslab commited on
Commit
fbae558
·
verified ·
1 Parent(s): 5ef5618

Rewrite the root README for the merged repository (small_flow_matching_policy, pi05, astra)

Browse files

Describes the three top-level folders and the path mapping from the former ManiLadder_Models and ManiLadder_PI05_Models layouts. Adds *.mp4 to .gitattributes for the evaluation videos under small_flow_matching_policy/.

Files changed (2) hide show
  1. .gitattributes +1 -107
  2. README.md +76 -510
.gitattributes CHANGED
@@ -33,110 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200000_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
37
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200001_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
38
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200002_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
39
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200003_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
40
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200004_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
41
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200005_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
42
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200006_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
43
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200007_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
44
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200008_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
45
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200009_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
46
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200010_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
47
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200011_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
48
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200012_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
49
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200013_succ1.mp4 filter=lfs diff=lfs merge=lfs -text
50
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200014_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
51
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200015_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
52
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200016_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
53
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200017_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
54
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200018_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
55
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200019_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
56
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200020_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
57
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200021_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
58
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200022_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
59
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200023_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
60
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200024_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
61
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200025_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
62
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200026_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
63
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200027_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
64
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200028_succ1.mp4 filter=lfs diff=lfs merge=lfs -text
65
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200029_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
66
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200030_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
67
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200031_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
68
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200032_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
69
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200033_succ1.mp4 filter=lfs diff=lfs merge=lfs -text
70
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200034_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
71
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200035_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
72
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200036_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
73
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200037_succ1.mp4 filter=lfs diff=lfs merge=lfs -text
74
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200038_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
75
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200039_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
76
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200040_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
77
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200041_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
78
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200042_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
79
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200043_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
80
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200044_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
81
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200045_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
82
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200046_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
83
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200047_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
84
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200048_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
85
- small_flow_matching_policy/BagDropIn-v1/evaluation_videos/seed_200049_succ0.mp4 filter=lfs diff=lfs merge=lfs -text
86
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_000.mp4 filter=lfs diff=lfs merge=lfs -text
87
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_001.mp4 filter=lfs diff=lfs merge=lfs -text
88
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_002.mp4 filter=lfs diff=lfs merge=lfs -text
89
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_003.mp4 filter=lfs diff=lfs merge=lfs -text
90
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_004.mp4 filter=lfs diff=lfs merge=lfs -text
91
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_005.mp4 filter=lfs diff=lfs merge=lfs -text
92
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_006.mp4 filter=lfs diff=lfs merge=lfs -text
93
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_007.mp4 filter=lfs diff=lfs merge=lfs -text
94
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_008.mp4 filter=lfs diff=lfs merge=lfs -text
95
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_009.mp4 filter=lfs diff=lfs merge=lfs -text
96
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_010.mp4 filter=lfs diff=lfs merge=lfs -text
97
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_011.mp4 filter=lfs diff=lfs merge=lfs -text
98
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_012.mp4 filter=lfs diff=lfs merge=lfs -text
99
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_013.mp4 filter=lfs diff=lfs merge=lfs -text
100
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_014.mp4 filter=lfs diff=lfs merge=lfs -text
101
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_015.mp4 filter=lfs diff=lfs merge=lfs -text
102
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_016.mp4 filter=lfs diff=lfs merge=lfs -text
103
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_017.mp4 filter=lfs diff=lfs merge=lfs -text
104
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_018.mp4 filter=lfs diff=lfs merge=lfs -text
105
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_019.mp4 filter=lfs diff=lfs merge=lfs -text
106
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_020.mp4 filter=lfs diff=lfs merge=lfs -text
107
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_021.mp4 filter=lfs diff=lfs merge=lfs -text
108
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_022.mp4 filter=lfs diff=lfs merge=lfs -text
109
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_023.mp4 filter=lfs diff=lfs merge=lfs -text
110
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_024.mp4 filter=lfs diff=lfs merge=lfs -text
111
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_025.mp4 filter=lfs diff=lfs merge=lfs -text
112
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_026.mp4 filter=lfs diff=lfs merge=lfs -text
113
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_027.mp4 filter=lfs diff=lfs merge=lfs -text
114
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_028.mp4 filter=lfs diff=lfs merge=lfs -text
115
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_029.mp4 filter=lfs diff=lfs merge=lfs -text
116
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_030.mp4 filter=lfs diff=lfs merge=lfs -text
117
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_031.mp4 filter=lfs diff=lfs merge=lfs -text
118
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_032.mp4 filter=lfs diff=lfs merge=lfs -text
119
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_033.mp4 filter=lfs diff=lfs merge=lfs -text
120
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_034.mp4 filter=lfs diff=lfs merge=lfs -text
121
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_035.mp4 filter=lfs diff=lfs merge=lfs -text
122
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_036.mp4 filter=lfs diff=lfs merge=lfs -text
123
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_037.mp4 filter=lfs diff=lfs merge=lfs -text
124
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_038.mp4 filter=lfs diff=lfs merge=lfs -text
125
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_039.mp4 filter=lfs diff=lfs merge=lfs -text
126
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_040.mp4 filter=lfs diff=lfs merge=lfs -text
127
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_041.mp4 filter=lfs diff=lfs merge=lfs -text
128
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_042.mp4 filter=lfs diff=lfs merge=lfs -text
129
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_043.mp4 filter=lfs diff=lfs merge=lfs -text
130
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_044.mp4 filter=lfs diff=lfs merge=lfs -text
131
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_045.mp4 filter=lfs diff=lfs merge=lfs -text
132
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_046.mp4 filter=lfs diff=lfs merge=lfs -text
133
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_047.mp4 filter=lfs diff=lfs merge=lfs -text
134
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_048.mp4 filter=lfs diff=lfs merge=lfs -text
135
- small_flow_matching_policy/BagHanging-v1/evaluation50/episode_049.mp4 filter=lfs diff=lfs merge=lfs -text
136
- small_flow_matching_policy/BagHanging-v1/visualizations/ckpt20k_eval50_front_grid.mp4 filter=lfs diff=lfs merge=lfs -text
137
- small_flow_matching_policy/BagHanging-v1/visualizations/expert20_front_grid.mp4 filter=lfs diff=lfs merge=lfs -text
138
- small_flow_matching_policy/SwingRing-v1/evaluation/videos_standard_front/SwingRing-v1_ep000_20261028-114857_succ0_ret66.54.mp4 filter=lfs diff=lfs merge=lfs -text
139
- small_flow_matching_policy/SwingRing-v1/evaluation/videos_standard_front/SwingRing-v1_ep001_20261028-114954_succ0_ret78.49.mp4 filter=lfs diff=lfs merge=lfs -text
140
- small_flow_matching_policy/SwingRing-v1/evaluation/videos_standard_front/SwingRing-v1_ep002_20261028-115049_succ0_ret73.53.mp4 filter=lfs diff=lfs merge=lfs -text
141
- small_flow_matching_policy/SwingRing-v1/evaluation/videos_standard_front/SwingRing-v1_ep003_20261028-115144_succ0_ret65.88.mp4 filter=lfs diff=lfs merge=lfs -text
142
- small_flow_matching_policy/SwingRing-v1/evaluation/videos_standard_front/SwingRing-v1_ep004_20261028-115238_succ0_ret73.45.mp4 filter=lfs diff=lfs merge=lfs -text
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ *.mp4 filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
README.md CHANGED
@@ -2,516 +2,82 @@
2
  license: mit
3
  ---
4
 
5
- # ManiLadder PI0.5 LoRA models
6
 
7
- ## single_task
 
8
 
9
- One PI0.5 LoRA checkpoint per ManiLadder task (110 tasks), each finetuned on 20 demos.
10
- Checkpoints are selected by the highest **50-episode** success rate; ties use the earliest step.
11
-
12
- - The 60 previously published tasks use fixed evaluation seeds 1000–1049.
13
- - The 22 completed retraining runs published through 2026-10-30 use seeds 800000–800049. All reached 40,000 training steps. 15 have all eight 5k–40k evaluations; 7 have seven evaluations because one intermediate checkpoint was not retained. Their selected weights are the best among the available evaluated checkpoints.
14
-
15
- - The 3 fresh-data runs in `pi05_newdata_3tasks_20261028` use pinned dataset revision `d120457e637b01679fbb40fd4de1d2394f0d55c9`, 20 demos per task, 40,000 training steps, and all eight 5k–40k eval50 checkpoints with seeds 800000–800049. **All three scored 0/50 at every checkpoint**; step 5,000 is the earliest tied checkpoint. See [their manifest](single_task/pi05_newdata_3tasks_20261028_manifest.json).
16
- - BagHanging-v1 and BagDropIn-v1 (campaign `pi05_bagtasks_20261106`, added 2026-11-08) were trained on 20 demos regenerated with the updated task code (now in `Linslab/ManiLadder_Dataset` `raw/<task>`), 40,000 training steps, and all eight 5k–40k eval50 checkpoints with seeds 1000–1049. See [their manifest](single_task/pi05_bagtasks_20261106_manifest.json).
17
- - The 23 tasks from the HF-data retraining campaign `pi05_hf21_20261025` (added 2026-10-29; raw demos freshly downloaded from Linslab/ManiLadder_Dataset) use seeds 1000–1049. All reached 40,000 training steps and have all eight 5k–40k evaluations. See [its manifest](single_task/pi05_hf21_20261025_manifest.json).
18
-
19
- See [retraining provenance and missing steps](single_task/pi05_retrain_eval50_20261022_manifest.json). These scores use different seed sets across the two groups and are not a controlled comparison.
20
-
21
- Each `single_task/<task>/` folder contains:
22
-
23
- - `lora_step_XXXXXXX.npz` — LoRA + trainable weights
24
- - `norm_stats.json` — normalization statistics used in training
25
- - `eval_result.json` — 50-episode evaluation of this checkpoint
26
-
27
- `single_task/results.csv` lists the selected checkpoint per task; `single_task/all_checkpoints_eval.csv` has the
28
- 50-episode result of every evaluated checkpoint (870 total).
29
-
30
- Evaluation: `exec_steps=16`, `sample_steps=10`. Tasks whose checkpoints all scored 0/50 are still included;
31
- for them the step choice is arbitrary.
32
-
33
- | task | level | step | success (50 ep) |
34
- |---|---|---|---|
35
- | DrapeCloth-v1 | 1 | 10000 | 1.00 |
36
- | DualArmButtonPress-v1 | 1 | 5000 | 0.98 |
37
- | DualArmCoverCloth-v1 | 1 | 15000 | 0.80 |
38
- | DualArmDoubleDoorBox-v1 | 1 | 20000 | 0.82 |
39
- | DualArmFoldHandKerchief-v1 | 1 | 5000 | 0.74 |
40
- | DualArmHammerHandover-v1 | 1 | 25000 | 0.74 |
41
- | DualArmLiftSoftDex-v1 | 1 | 25000 | 1.00 |
42
- | DualArmOpenPen-v1 | 1 | 35000 | 0.36 |
43
- | DualArmOpenScissor-v1 | 1 | 15000 | 0.86 |
44
- | DualArmPouringWater-v1 | 1 | 35000 | 0.68 |
45
- | DualArmPressButton-v1 | 1 | 15000 | 0.80 |
46
- | DualArmPushBox-v1 | 1 | 30000 | 0.78 |
47
- | DualArmRopeCylinderWarpping-v1 | 1 | 10000 | 0.62 |
48
- | DualArmRopeStretch-v1 | 1 | 10000 | 0.92 |
49
- | DualArmSwingCup-v1 | 1 | 20000 | 0.36 |
50
- | GraspCylinder-v1 | 1 | 15000 | 0.94 |
51
- | LiftBucket-v1 | 1 | 40000 | 0.96 |
52
- | OpenDoor-v1 | 1 | 5000 | 0.30 |
53
- | PickPlacePen-v1 | 1 | 20000 | 0.36 |
54
- | PushSoccer-v1 | 1 | 40000 | 0.80 |
55
- | ReorientCube-v1 | 1 | 5000 | 1.00 |
56
- | RopePalmCageHoldDex-v1 | 1 | 5000 | 0.94 |
57
- | ShapeTToI-v1 | 1 | 40000 | 0.50 |
58
- | SqueezeClay-v1 | 1 | 15000 | 0.84 |
59
- | StackBlock-v1 | 1 | 15000 | 0.46 |
60
- | StraightenRope-v1 | 1 | 35000 | 0.84 |
61
- | TurnOnFaucet-v1 | 1 | 10000 | 0.92 |
62
- | UnfoldShirt-v1 | 1 | 5000 | 0.78 |
63
- | ClothWipe-v1 | 2 | 25000 | 0.52 |
64
- | DexOpenDoor-v1 | 2 | 5000 | 0.56 |
65
- | DualArmBookPickPlace-v1 | 2 | 15000 | 0.56 |
66
- | DualArmCleanupPotEnv-v1 | 2 | 35000 | 0.38 |
67
- | DualArmClothAlign-v1 | 2 | 40000 | 0.38 |
68
- | DualArmJugglingBall-v1 | 2 | 40000 | 0.06 |
69
- | DualArmLiftBall-v1 | 2 | 35000 | 0.54 |
70
- | DualArmOpenBottle-v1 | 2 | 30000 | 0.06 |
71
- | DualArmOpenLaptop-v1 | 2 | 25000 | 0.42 |
72
- | DualArmPegInHole-v1 | 2 | 5000 | 0.18 |
73
- | DualArmPlaceRopeOverRod-v1 | 2 | 40000 | 0.46 |
74
- | DualArmPouringTwoCups-v1 | 2 | 5000 | 0.00 |
75
- | DualArmPushObject-v1 | 2 | 25000 | 0.06 |
76
- | DualArmStraightenRope-v1 | 2 | 35000 | 0.48 |
77
- | DualArmStuffCloth-v1 | 2 | 15000 | 0.40 |
78
- | DualArmThrowBlock-v1 | 2 | 5000 | 0.00 |
79
- | FoldShirt-v1 | 2 | 40000 | 0.72 |
80
- | GraspTwoBlocks-v1 | 2 | 5000 | 0.72 |
81
- | HammerNailOut-v1 | 2 | 40000 | 0.44 |
82
- | HangMug-v1 | 2 | 10000 | 0.70 |
83
- | KitAssembly-v1 | 2 | 15000 | 0.06 |
84
- | OpenBoxRetrieval-v1 | 2 | 25000 | 0.38 |
85
- | PivotBook-v1 | 2 | 15000 | 0.38 |
86
- | ReorientPen-v1 | 2 | 25000 | 0.38 |
87
- | RollingPin-v1 | 2 | 25000 | 0.96 |
88
- | RopeBridgeDex-v1 | 2 | 30000 | 0.78 |
89
- | ScoopWater-v1 | 2 | 15000 | 0.44 |
90
- | StraightenTwoRopesSequential-v1 | 2 | 15000 | 0.36 |
91
- | ClutterGrasp-v1 | 3 | 15000 | 0.16 |
92
- | DualArmAssembleBracket-v1 | 3 | 5000 | 0.00 |
93
- | DualArmFoldCarton-v1 | 3 | 25000 | 0.14 |
94
- | DualArmLayerClayInDish-v1 | 3 | 5000 | 0.00 |
95
- | DualArmMakeLine-v1 | 3 | 5000 | 0.00 |
96
- | DualArmOpenFridge-v1 | 3 | 5000 | 0.00 |
97
- | DualArmOrganizeCabinet-v1 | 3 | 5000 | 0.00 |
98
- | DualArmPingpongBalance-v1 | 3 | 15000 | 0.20 |
99
- | DualArmPlugCharger-v1 | 3 | 5000 | 0.00 |
100
- | DualArmSetupTable-v1 | 3 | 25000 | 0.04 |
101
- | DualArmShapeDoll-v1 | 3 | 30000 | 0.04 |
102
- | DualArmThreadCloth-v1 | 3 | 5000 | 0.12 |
103
- | DualArmThreadRopeThroughTube-v1 | 3 | 35000 | 0.16 |
104
- | DualArmThreading-v1 | 3 | 20000 | 0.26 |
105
- | DualArmWeaveCloth-v1 | 3 | 20000 | 0.32 |
106
- | GroceryPrepare-v1 | 3 | 5000 | 0.00 |
107
- | HammerNail-v1 | 3 | 35000 | 0.12 |
108
- | HitBox-v1 | 3 | 25000 | 0.06 |
109
- | LiquidMeasuring-v1 | 3 | 35000 | 0.32 |
110
- | MobilePickPlace-v1 | 3 | 15000 | 0.08 |
111
- | PressFitSoftPeg-v1 | 3 | 20000 | 0.08 |
112
- | PushTargetPose-v1 | 3 | 30000 | 0.02 |
113
- | SeparateTwoRopesEnv-v1 | 3 | 35000 | 0.26 |
114
- | Stack6Blocks-v1 | 3 | 5000 | 0.00 |
115
- | TweezersPickPlace-v1 | 3 | 5000 | 0.00 |
116
- | BagDropIn-v1 | 3 | 15000 | 0.70 |
117
- | BagHanging-v1 | 3 | 10000 | 0.88 |
118
- | BuildTent-v1 | 4 | 10000 | 0.06 |
119
- | ClaySealStamp-v1 | 4 | 5000 | 0.00 |
120
- | CoverPattern-v1 | 4 | 30000 | 0.02 |
121
- | DualArmBallClassify-v1 | 4 | 10000 | 0.00 |
122
- | DualArmBilliardBridge-v1 | 4 | 5000 | 0.00 |
123
- | DualArmClothCylinderWipe-v1 | 4 | 5000 | 0.02 |
124
- | DualArmClothHanger-v1 | 4 | 5000 | 0.00 |
125
- | DualArmFlipTable-v1 | 4 | 5000 | 0.00 |
126
- | DualArmJumpRope-v1 | 4 | 20000 | 0.02 |
127
- | DualArmKeyboardType-v1 | 4 | 5000 | 0.00 |
128
- | DualArmLiquidMixing-v1 | 4 | 5000 | 0.00 |
129
- | DualArmMahjongSetup-v1 | 4 | 5000 | 0.02 |
130
- | DualArmSandShovel-v1 | 4 | 15000 | 0.02 |
131
- | DualArmSendPingpong-v1 | 4 | 40000 | 0.08 |
132
- | DualArmShootBasketball-v1 | 4 | 5000 | 0.00 |
133
- | DualArmWallRings-v1 | 4 | 5000 | 0.00 |
134
- | FitnessBall-v1 | 4 | 5000 | 0.00 |
135
- | FlipPan-v1 | 4 | 40000 | 0.04 |
136
- | HitBilliard-v1 | 4 | 5000 | 0.00 |
137
- | PingPongBallTransport-v1 | 4 | 40000 | 0.16 |
138
- | RingTossing-v1 | 4 | 5000 | 0.00 |
139
- | RodThroughHoles-v1 | 4 | 5000 | 0.02 |
140
- | SpinPenHora-v1 | 4 | 5000 | 0.00 |
141
- | StampRoll-v1 | 4 | 5000 | 0.00 |
142
- | SwingRing-v1 | 4 | 15000 | 0.04 |
143
- | TennisRally-v1 | 4 | 15000 | 0.06 |
144
- | WrapRope-v1 | 4 | 5000 | 0.02 |
145
-
146
- <!-- multi_task:begin -->
147
-
148
- ## multi_task (level-wise)
149
-
150
- One PI0.5 LoRA policy per ManiLadder level, finetuned jointly on **every task of that level** (20 demos per task, 109 tasks in total) and evaluated on each of its tasks for **50 episodes** (seeds 1000–1049). Only the final checkpoint (step 80,000) was evaluated; no checkpoint selection.
151
-
152
- Recipe: batch 64, lr 5e-5 with 1k warmup and cosine decay over 80k steps, 8×A100-80GB data parallel, no online eval. `action_dim` follows the widest embodiment in the level; narrower embodiments are zero-padded and the normalization statistics ignore padded dimensions (see `metadata.json`). Evaluation: `exec_steps=16`, `sample_steps=10`.
153
-
154
- Not included: BagHanging-v1 and BagDropIn-v1 (level 3) and SwingRing-v1 (level 4) on request; OpenCoffer-v1, ClayMolding-v1, PickSoft-v1, DualArmBuildCanel-v1, DualArmRopeScissorCut-v1 and DualArmSqueezeClayDex-v1 have no demos. MazeBlockPush-v1 and DualArmOpenScissor-v1 were trained on local demos newer than the HF copies (details in `metadata.json`).
155
-
156
- Each `multi_task/level_<N>/` folder contains `lora_step_0080000.npz`, `norm_stats.json`, `metadata.json`, `eval_result.json` (per-task 50-episode results) and `eval_reports/<task>.json` (per-episode rows). `multi_task/results.csv` lists every (level, task) row.
157
-
158
- | level | tasks | evaluated | action_dim | mean success (50 ep/task) |
159
- |---|---|---|---|---|
160
- | level_1 | 28 | 28 | 44 | 0.579 |
161
- | level_2 | 28 | 28 | 44 | 0.241 |
162
- | level_3 | 26 | 26 | 50 | 0.005 |
163
- | level_4 | 27 | 27 | 50 | 0.001 |
164
-
165
- | task | level | embodiment | success (50 ep) |
166
- |---|---|---|---|
167
- | DualArmLiftSoftDex-v1 | 1 | deformable/dual_arm_dex | 1.00 |
168
- | DualArmOpenScissor-v1 | 1 | rigid_body/dual_arm_dex | 0.98 |
169
- | DrapeCloth-v1 | 1 | deformable/single_arm_dex | 0.96 |
170
- | RopePalmCageHoldDex-v1 | 1 | deformable/single_arm_dex | 0.92 |
171
- | DualArmButtonPress-v1 | 1 | rigid_body/dual_arm_gripper | 0.90 |
172
- | GraspCylinder-v1 | 1 | rigid_body/single_arm_dex | 0.86 |
173
- | ReorientCube-v1 | 1 | rigid_body/single_arm_dex | 0.84 |
174
- | TurnOnFaucet-v1 | 1 | rigid_body/single_arm_dex | 0.84 |
175
- | DualArmPouringWater-v1 | 1 | deformable/dual_arm_gripper | 0.72 |
176
- | DualArmDoubleDoorBox-v1 | 1 | rigid_body/dual_arm_gripper | 0.70 |
177
- | DualArmRopeCylinderWarpping-v1 | 1 | deformable/dual_arm_gripper | 0.66 |
178
- | StraightenRope-v1 | 1 | deformable/single_arm_gripper | 0.66 |
179
- | DualArmFoldHandKerchief-v1 | 1 | deformable/dual_arm_gripper | 0.60 |
180
- | DualArmPushBox-v1 | 1 | rigid_body/dual_arm_gripper | 0.60 |
181
- | PushSoccer-v1 | 1 | rigid_body/single_arm_gripper | 0.60 |
182
- | LiftBucket-v1 | 1 | rigid_body/single_arm_dex | 0.54 |
183
- | StackBlock-v1 | 1 | rigid_body/single_arm_gripper | 0.54 |
184
- | DualArmRopeStretch-v1 | 1 | deformable/dual_arm_dex | 0.48 |
185
- | DualArmSwingCup-v1 | 1 | rigid_body/dual_arm_dex | 0.42 |
186
- | SqueezeClay-v1 | 1 | deformable/single_arm_dex | 0.40 |
187
- | DualArmPressButton-v1 | 1 | rigid_body/dual_arm_dex | 0.38 |
188
- | PickPlacePen-v1 | 1 | rigid_body/single_arm_gripper | 0.34 |
189
- | DualArmOpenPen-v1 | 1 | rigid_body/dual_arm_dex | 0.30 |
190
- | UnfoldShirt-v1 | 1 | deformable/single_arm_gripper | 0.28 |
191
- | DualArmCoverCloth-v1 | 1 | deformable/dual_arm_dex | 0.24 |
192
- | DualArmHammerHandover-v1 | 1 | rigid_body/dual_arm_gripper | 0.24 |
193
- | OpenDoor-v1 | 1 | rigid_body/single_arm_gripper | 0.12 |
194
- | ShapeTToI-v1 | 1 | deformable/single_arm_gripper | 0.10 |
195
- | FoldShirt-v1 | 2 | deformable/single_arm_gripper | 0.62 |
196
- | GraspTwoBlocks-v1 | 2 | rigid_body/single_arm_dex | 0.56 |
197
- | DualArmClothAlign-v1 | 2 | deformable/dual_arm_gripper | 0.50 |
198
- | DualArmPlaceRopeOverRod-v1 | 2 | deformable/dual_arm_dex | 0.48 |
199
- | DualArmStuffCloth-v1 | 2 | deformable/dual_arm_dex | 0.44 |
200
- | HangMug-v1 | 2 | rigid_body/single_arm_gripper | 0.42 |
201
- | RollingPin-v1 | 2 | deformable/single_arm_dex | 0.38 |
202
- | DualArmLiftBall-v1 | 2 | rigid_body/dual_arm_gripper | 0.36 |
203
- | DualArmOpenLaptop-v1 | 2 | rigid_body/dual_arm_dex | 0.34 |
204
- | HammerNailOut-v1 | 2 | rigid_body/single_arm_dex | 0.32 |
205
- | RopeBridgeDex-v1 | 2 | deformable/single_arm_dex | 0.32 |
206
- | ReorientPen-v1 | 2 | rigid_body/single_arm_dex | 0.30 |
207
- | DexOpenDoor-v1 | 2 | rigid_body/single_arm_dex | 0.28 |
208
- | ScoopWater-v1 | 2 | deformable/single_arm_gripper | 0.26 |
209
- | ClothWipe-v1 | 2 | deformable/single_arm_dex | 0.24 |
210
- | DualArmStraightenRope-v1 | 2 | deformable/dual_arm_gripper | 0.24 |
211
- | DualArmThrowBlock-v1 | 2 | rigid_body/dual_arm_dex | 0.20 |
212
- | DualArmPegInHole-v1 | 2 | rigid_body/dual_arm_gripper | 0.18 |
213
- | PivotBook-v1 | 2 | rigid_body/single_arm_gripper | 0.16 |
214
- | StraightenTwoRopesSequential-v1 | 2 | deformable/single_arm_gripper | 0.12 |
215
- | OpenBoxRetrieval-v1 | 2 | rigid_body/single_arm_gripper | 0.04 |
216
- | DualArmBookPickPlace-v1 | 2 | rigid_body/dual_arm_gripper | 0.00 |
217
- | DualArmCleanupPotEnv-v1 | 2 | rigid_body/dual_arm_gripper | 0.00 |
218
- | DualArmJugglingBall-v1 | 2 | rigid_body/dual_arm_dex | 0.00 |
219
- | DualArmOpenBottle-v1 | 2 | rigid_body/dual_arm_dex | 0.00 |
220
- | DualArmPouringTwoCups-v1 | 2 | deformable/dual_arm_dex | 0.00 |
221
- | DualArmPushObject-v1 | 2 | deformable/dual_arm_gripper | 0.00 |
222
- | KitAssembly-v1 | 2 | rigid_body/single_arm_gripper | 0.00 |
223
- | SeparateTwoRopesEnv-v1 | 3 | deformable/single_arm_gripper | 0.06 |
224
- | ClutterGrasp-v1 | 3 | rigid_body/single_arm_dex | 0.04 |
225
- | DualArmThreadCloth-v1 | 3 | deformable/dual_arm_dex | 0.04 |
226
- | DualArmAssembleBracket-v1 | 3 | rigid_body/dual_arm_gripper | 0.00 |
227
- | DualArmFoldCarton-v1 | 3 | rigid_body/dual_arm_gripper | 0.00 |
228
- | DualArmLayerClayInDish-v1 | 3 | deformable/dual_arm_dex | 0.00 |
229
- | DualArmMakeLine-v1 | 3 | rigid_body/dual_arm_dex | 0.00 |
230
- | DualArmOpenFridge-v1 | 3 | rigid_body/dual_arm_gripper | 0.00 |
231
- | DualArmOrganizeCabinet-v1 | 3 | rigid_body/dual_arm_gripper | 0.00 |
232
- | DualArmPingpongBalance-v1 | 3 | rigid_body/dual_arm_dex | 0.00 |
233
- | DualArmPlugCharger-v1 | 3 | rigid_body/dual_arm_dex | 0.00 |
234
- | DualArmSetupTable-v1 | 3 | rigid_body/dual_arm_dex | 0.00 |
235
- | DualArmShapeDoll-v1 | 3 | deformable/dual_arm_gripper | 0.00 |
236
- | DualArmThreadRopeThroughTube-v1 | 3 | deformable/dual_arm_gripper | 0.00 |
237
- | DualArmThreading-v1 | 3 | deformable/dual_arm_dex | 0.00 |
238
- | DualArmWeaveCloth-v1 | 3 | deformable/dual_arm_gripper | 0.00 |
239
- | GroceryPrepare-v1 | 3 | rigid_body/single_arm_gripper | 0.00 |
240
- | HammerNail-v1 | 3 | rigid_body/single_arm_dex | 0.00 |
241
- | HitBox-v1 | 3 | deformable/single_arm_dex | 0.00 |
242
- | LiquidMeasuring-v1 | 3 | deformable/single_arm_gripper | 0.00 |
243
- | MazeBlockPush-v1 | 3 | rigid_body/single_arm_gripper | 0.00 |
244
- | MobilePickPlace-v1 | 3 | rigid_body/single_arm_dex | 0.00 |
245
- | PressFitSoftPeg-v1 | 3 | deformable/single_arm_dex | 0.00 |
246
- | PushTargetPose-v1 | 3 | rigid_body/single_arm_gripper | 0.00 |
247
- | Stack6Blocks-v1 | 3 | rigid_body/single_arm_gripper | 0.00 |
248
- | TweezersPickPlace-v1 | 3 | rigid_body/single_arm_dex | 0.00 |
249
- | DualArmClothCylinderWipe-v1 | 4 | deformable/dual_arm_dex | 0.02 |
250
- | BuildTent-v1 | 4 | deformable/single_arm_gripper | 0.00 |
251
- | ClaySealStamp-v1 | 4 | deformable/single_arm_dex | 0.00 |
252
- | CoverPattern-v1 | 4 | deformable/single_arm_dex | 0.00 |
253
- | DualArmBallClassify-v1 | 4 | rigid_body/dual_arm_gripper | 0.00 |
254
- | DualArmBilliardBridge-v1 | 4 | rigid_body/dual_arm_gripper | 0.00 |
255
- | DualArmClothHanger-v1 | 4 | deformable/dual_arm_gripper | 0.00 |
256
- | DualArmFlipTable-v1 | 4 | rigid_body/dual_arm_dex | 0.00 |
257
- | DualArmJumpRope-v1 | 4 | deformable/dual_arm_dex | 0.00 |
258
- | DualArmKeyboardType-v1 | 4 | rigid_body/dual_arm_dex | 0.00 |
259
- | DualArmLiquidMixing-v1 | 4 | deformable/dual_arm_dex | 0.00 |
260
- | DualArmMahjongSetup-v1 | 4 | rigid_body/dual_arm_gripper | 0.00 |
261
- | DualArmRubikCube-v1 | 4 | rigid_body/dual_arm_gripper | 0.00 |
262
- | DualArmSandShovel-v1 | 4 | deformable/dual_arm_gripper | 0.00 |
263
- | DualArmSendPingpong-v1 | 4 | rigid_body/dual_arm_dex | 0.00 |
264
- | DualArmShootBasketball-v1 | 4 | rigid_body/dual_arm_dex | 0.00 |
265
- | DualArmWallRings-v1 | 4 | deformable/dual_arm_gripper | 0.00 |
266
- | FitnessBall-v1 | 4 | rigid_body/single_arm_dex | 0.00 |
267
- | FlipPan-v1 | 4 | rigid_body/single_arm_gripper | 0.00 |
268
- | HitBilliard-v1 | 4 | rigid_body/single_arm_gripper | 0.00 |
269
- | PingPongBallTransport-v1 | 4 | rigid_body/single_arm_gripper | 0.00 |
270
- | RingTossing-v1 | 4 | rigid_body/single_arm_gripper | 0.00 |
271
- | RodThroughHoles-v1 | 4 | rigid_body/single_arm_dex | 0.00 |
272
- | SpinPenHora-v1 | 4 | rigid_body/single_arm_dex | 0.00 |
273
- | StampRoll-v1 | 4 | deformable/single_arm_gripper | 0.00 |
274
- | TennisRally-v1 | 4 | rigid_body/single_arm_dex | 0.00 |
275
- | WrapRope-v1 | 4 | deformable/single_arm_gripper | 0.00 |
276
-
277
- <!-- multi_task:end -->
278
-
279
-
280
- ## Previous single-task checkpoints classified as reusable
281
-
282
- The [`previous_single_task/`](previous_single_task) collection preserves one best checkpoint from each of the 62 earlier single-task runs classified as reusable in the 2026-10-22 audit. It is separate from newer retraining runs in `single_task/`. See the [CSV](previous_single_task/manifest.csv) or [JSON manifest](previous_single_task/manifest.json), sorted by level, suite, and task.
283
-
284
- For 60 tasks, selection uses the available 50-episode reevaluations (seeds 1000–1049). The original **DualArmPouringWater-v1** peak is **19/20 (95%) at 30k**; the original **MazeBlockPush-v1** peak is **13/20 (65%) at 30k**. These two are selected from historical 20-episode online evaluations and must not be presented as eval50 scores. Each folder contains the selected weights, normalization statistics, evaluation summary/history, and audit provenance.
285
-
286
- Non-finite reward means in the historical reports are represented as null/blank and marked `return_mean_finite=false`; checkpoint selection uses the reported success counts.
287
-
288
- The reusable classification was based on timestamp and source-change review; the original training runs did not save complete source snapshots.
289
- ## multi_task (level-2 subsets)
290
-
291
- Six multi-task PI0.5 LoRA policies, each finetuned for 50,000 steps (batch 64, lr 5e-5, cosine decay) on a subset of the 28 level-2 tasks (20 demos per task); every level-2 task belongs to exactly three subsets. Only the final checkpoint was evaluated (50 episodes per task, seeds 1000–1049); the 10k/20k/30k/40k snapshots are included unevaluated. Campaign `pi05_lv2sub_20261029` (2026-10-29/30). See [the manifest](multi_task/pi05_lv2sub_20261029_manifest.json) and `multi_task/results.csv`.
292
-
293
- Each `multi_task/<suite>/` folder contains:
294
-
295
- - `lora_step_0010000.npz` … `lora_step_0050000.npz` — LoRA + trainable weights every 10k steps (`0050000` is the evaluated one)
296
- - `norm_stats.json` — normalization statistics used in training (action/state dim 44, zero-padded for narrower embodiments)
297
- - `metadata.json` — task list, training config, checkpoint hashes, env source hashes
298
- - `eval_result.json` — per-task 50-episode success of the final checkpoint; `eval_reports/<task>.json` has per-episode results
299
-
300
- | suite | tasks | mean success (50 ep/task) |
301
  |---|---|---|
302
- | level2_rigid | 16 | 0.173 |
303
- | level2_deformable | 12 | 0.167 |
304
- | level2_single_arm | 14 | 0.241 |
305
- | level2_dual_arm | 14 | 0.120 |
306
- | level2_gripper | 14 | 0.166 |
307
- | level2_dexterous | 14 | 0.210 |
308
-
309
- | task | level2_rigid | level2_deformable | level2_single_arm | level2_dual_arm | level2_gripper | level2_dexterous |
310
- |---|---|---|---|---|---|---|
311
- | ClothWipe-v1 | | 0.24 | 0.18 | | | 0.10 |
312
- | DexOpenDoor-v1 | 0.38 | | 0.18 | | | 0.26 |
313
- | DualArmBookPickPlace-v1 | 0.00 | | | 0.00 | 0.16 | |
314
- | DualArmCleanupPotEnv-v1 | 0.00 | | | 0.00 | 0.00 | |
315
- | DualArmClothAlign-v1 | | 0.30 | | 0.26 | 0.34 | |
316
- | DualArmJugglingBall-v1 | 0.04 | | | 0.12 | | 0.00 |
317
- | DualArmLiftBall-v1 | 0.20 | | | 0.28 | 0.16 | |
318
- | DualArmOpenBottle-v1 | 0.00 | | | 0.00 | | 0.00 |
319
- | DualArmOpenLaptop-v1 | 0.04 | | | 0.14 | | 0.08 |
320
- | DualArmPegInHole-v1 | 0.10 | | | 0.08 | 0.16 | |
321
- | DualArmPlaceRopeOverRod-v1 | | 0.12 | | 0.16 | | 0.58 |
322
- | DualArmPouringTwoCups-v1 | | 0.00 | | 0.00 | | 0.00 |
323
- | DualArmPushObject-v1 | | 0.00 | | 0.00 | 0.04 | |
324
- | DualArmStraightenRope-v1 | | 0.10 | | 0.20 | 0.10 | |
325
- | DualArmStuffCloth-v1 | | 0.28 | | 0.12 | | 0.24 |
326
- | DualArmThrowBlock-v1 | 0.22 | | | 0.32 | | 0.20 |
327
- | FoldShirt-v1 | | 0.22 | 0.50 | | 0.62 | |
328
- | GraspTwoBlocks-v1 | 0.56 | | 0.70 | | | 0.64 |
329
- | HammerNailOut-v1 | 0.08 | | 0.34 | | | 0.22 |
330
- | HangMug-v1 | 0.62 | | 0.46 | | 0.50 | |
331
- | KitAssembly-v1 | 0.06 | | 0.06 | | 0.00 | |
332
- | OpenBoxRetrieval-v1 | 0.02 | | 0.04 | | 0.04 | |
333
- | PivotBook-v1 | 0.04 | | 0.18 | | 0.16 | |
334
- | ReorientPen-v1 | 0.40 | | 0.16 | | | 0.26 |
335
- | RollingPin-v1 | | 0.18 | 0.10 | | | 0.20 |
336
- | RopeBridgeDex-v1 | | 0.22 | 0.40 | | | 0.16 |
337
- | ScoopWater-v1 | | 0.26 | 0.00 | | 0.00 | |
338
- | StraightenTwoRopesSequential-v1 | | 0.08 | 0.08 | | 0.04 | |
339
- ## multi_task (level-2 suite transfer)
340
-
341
- Six suite-transfer runs: each starts from **all parameters of the final checkpoint** of one level-2 subset policy above (`pi05_lv2sub_20261029`) and is finetuned for another 50,000 steps (fresh schedule: batch 64, lr 5e-5, 1k warmup, cosine decay) on the tasks of a *different* grouping of the same 28 level-2 tasks, then evaluated on those tasks (final checkpoint only, 50 episodes per task, seeds 1000–1049). The baseline is the from-scratch policy of the target grouping from the section above (same tasks, same seeds). Campaign `pi05_lv2xfer_20261101` (2026-11-01). See [the manifest](multi_task/pi05_lv2xfer_20261101_manifest.json) and `multi_task/results.csv` (column `campaign`).
342
-
343
- Each `multi_task/<target>_from_<source>/` folder has the same files as the subset suites (`lora_step_0010000.npz` … `lora_step_0050000.npz`, `norm_stats.json`, `metadata.json` with `init_from`, `eval_result.json`, `eval_reports/<task>.json`).
344
-
345
- | suite | source policy | target tasks | transfer mean (50 ep/task) | from-scratch baseline |
346
- |---|---|---|---|---|
347
- | level2_rigid_from_deformable | level2_deformable | 16 | 0.199 | 0.173 |
348
- | level2_deformable_from_rigid | level2_rigid | 12 | 0.258 | 0.167 |
349
- | level2_dual_arm_from_single_arm | level2_single_arm | 14 | 0.139 | 0.120 |
350
- | level2_single_arm_from_dual_arm | level2_dual_arm | 14 | 0.240 | 0.241 |
351
- | level2_dexterous_from_gripper | level2_gripper | 14 | 0.280 | 0.210 |
352
- | level2_gripper_from_dexterous | level2_dexterous | 14 | 0.227 | 0.166 |
353
-
354
- Per task: transfer success rate (from-scratch baseline in parentheses).
355
-
356
- | task | level2_rigid_from_deformable | level2_deformable_from_rigid | level2_dual_arm_from_single_arm | level2_single_arm_from_dual_arm | level2_dexterous_from_gripper | level2_gripper_from_dexterous |
357
- |---|---|---|---|---|---|---|
358
- | ClothWipe-v1 | | 0.12 (0.24) | | 0.34 (0.18) | 0.10 (0.10) | |
359
- | DexOpenDoor-v1 | 0.18 (0.38) | | | 0.10 (0.18) | 0.20 (0.26) | |
360
- | DualArmBookPickPlace-v1 | 0.26 (0.00) | | 0.00 (0.00) | | | 0.42 (0.16) |
361
- | DualArmCleanupPotEnv-v1 | 0.00 (0.00) | | 0.00 (0.00) | | | 0.02 (0.00) |
362
- | DualArmClothAlign-v1 | | 0.42 (0.30) | 0.20 (0.26) | | | 0.34 (0.34) |
363
- | DualArmJugglingBall-v1 | 0.04 (0.04) | | 0.14 (0.12) | | 0.12 (0.00) | |
364
- | DualArmLiftBall-v1 | 0.24 (0.20) | | 0.40 (0.28) | | | 0.22 (0.16) |
365
- | DualArmOpenBottle-v1 | 0.00 (0.00) | | 0.00 (0.00) | | 0.00 (0.00) | |
366
- | DualArmOpenLaptop-v1 | 0.18 (0.04) | | 0.40 (0.14) | | 0.30 (0.08) | |
367
- | DualArmPegInHole-v1 | 0.20 (0.10) | | 0.14 (0.08) | | | 0.22 (0.16) |
368
- | DualArmPlaceRopeOverRod-v1 | | 0.44 (0.12) | 0.16 (0.16) | | 0.54 (0.58) | |
369
- | DualArmPouringTwoCups-v1 | | 0.00 (0.00) | 0.00 (0.00) | | 0.00 (0.00) | |
370
- | DualArmPushObject-v1 | | 0.00 (0.00) | 0.00 (0.00) | | | 0.04 (0.04) |
371
- | DualArmStraightenRope-v1 | | 0.22 (0.10) | 0.08 (0.20) | | | 0.04 (0.10) |
372
- | DualArmStuffCloth-v1 | | 0.10 (0.28) | 0.38 (0.12) | | 0.48 (0.24) | |
373
- | DualArmThrowBlock-v1 | 0.16 (0.22) | | 0.04 (0.32) | | 0.26 (0.20) | |
374
- | FoldShirt-v1 | | 0.38 (0.22) | | 0.38 (0.50) | | 0.48 (0.62) |
375
- | GraspTwoBlocks-v1 | 0.54 (0.56) | | | 0.46 (0.70) | 0.52 (0.64) | |
376
- | HammerNailOut-v1 | 0.32 (0.08) | | | 0.30 (0.34) | 0.42 (0.22) | |
377
- | HangMug-v1 | 0.52 (0.62) | | | 0.36 (0.46) | | 0.40 (0.50) |
378
- | KitAssembly-v1 | 0.00 (0.06) | | | 0.10 (0.06) | | 0.00 (0.00) |
379
- | OpenBoxRetrieval-v1 | 0.04 (0.02) | | | 0.04 (0.04) | | 0.28 (0.04) |
380
- | PivotBook-v1 | 0.12 (0.04) | | | 0.12 (0.18) | | 0.22 (0.16) |
381
- | ReorientPen-v1 | 0.38 (0.40) | | | 0.26 (0.16) | 0.20 (0.26) | |
382
- | RollingPin-v1 | | 0.62 (0.18) | | 0.06 (0.10) | 0.28 (0.20) | |
383
- | RopeBridgeDex-v1 | | 0.34 (0.22) | | 0.48 (0.40) | 0.50 (0.16) | |
384
- | ScoopWater-v1 | | 0.24 (0.26) | | 0.18 (0.00) | | 0.30 (0.00) |
385
- | StraightenTwoRopesSequential-v1 | | 0.22 (0.08) | | 0.18 (0.08) | | 0.20 (0.04) |
386
-
387
- <!-- multi_task_level_transfer:begin -->
388
-
389
- ## multi_task (level-to-level transfer)
390
-
391
- Transfer experiments between the level-wise multi-task policies above: each policy in `multi_task/level_<B>/` (trained 80k steps from pi05_base) was used as the **initialisation** for a further **50,000 LoRA steps on every task of another level** (campaign `pi05_transfer_20261101`), then evaluated on the target level's tasks for **50 episodes** each (seeds 1000–1049, final checkpoint only). The baseline column is the target level's own policy from `multi_task/level_<B>/` on the same tasks.
392
-
393
- Recipe: identical to the level-wise runs except for the initialisation and length: batch 64, lr 5e-5 with 1k warmup and cosine decay to 5e-6 over 50k steps, fresh AdamW, 8×A100-80GB data parallel, target-level dataset and normalization statistics. Where source and target `action_dim` differ (44 ↔ 50) only the three action projection tensors are partially copied (overlapping slice kept, new slots freshly initialised); see `metadata.json`.
394
-
395
- Each `multi_task/level<B>_from_level<A>/` folder contains `lora_step_0050000.npz`, `norm_stats.json`, `metadata.json`, `eval_result.json` (per-task results with baselines) and `eval_reports/<task>.json` (per-episode rows). Rows are also in `multi_task/results.csv` (campaign `pi05_transfer_20261101`).
396
-
397
- | folder | init from | trained + evaluated on | tasks | action_dim | mean success (50 ep/task) | baseline (level_<B> from pi05_base) |
398
- |---|---|---|---|---|---|---|
399
- | level2_from_level1 | level_1 @80k | level_2 (28 tasks) | 28/28 | 44 | 0.218 | 0.241 |
400
- | level1_from_level2 | level_2 @80k | level_1 (28 tasks) | 28/28 | 44 | 0.620 | 0.579 |
401
- | level3_from_level2 | level_2 @80k | level_3 (26 tasks) | 26/26 | 50 | 0.002 | 0.005 |
402
- | level2_from_level3 | level_3 @80k | level_2 (28 tasks) | 28/28 | 44 | 0.197 | 0.241 |
403
-
404
- | task | folder | embodiment | transfer success (50 ep) | baseline success (50 ep) |
405
- |---|---|---|---|---|
406
- | DualArmPlaceRopeOverRod-v1 | level2_from_level1 | deformable/dual_arm_dex | 0.74 | 0.48 |
407
- | GraspTwoBlocks-v1 | level2_from_level1 | rigid_body/single_arm_dex | 0.64 | 0.56 |
408
- | FoldShirt-v1 | level2_from_level1 | deformable/single_arm_gripper | 0.50 | 0.62 |
409
- | HangMug-v1 | level2_from_level1 | rigid_body/single_arm_gripper | 0.42 | 0.42 |
410
- | RollingPin-v1 | level2_from_level1 | deformable/single_arm_dex | 0.38 | 0.38 |
411
- | HammerNailOut-v1 | level2_from_level1 | rigid_body/single_arm_dex | 0.32 | 0.32 |
412
- | RopeBridgeDex-v1 | level2_from_level1 | deformable/single_arm_dex | 0.32 | 0.32 |
413
- | DualArmOpenLaptop-v1 | level2_from_level1 | rigid_body/dual_arm_dex | 0.30 | 0.34 |
414
- | ReorientPen-v1 | level2_from_level1 | rigid_body/single_arm_dex | 0.30 | 0.30 |
415
- | DualArmPegInHole-v1 | level2_from_level1 | rigid_body/dual_arm_gripper | 0.28 | 0.18 |
416
- | ClothWipe-v1 | level2_from_level1 | deformable/single_arm_dex | 0.24 | 0.24 |
417
- | DualArmThrowBlock-v1 | level2_from_level1 | rigid_body/dual_arm_dex | 0.24 | 0.20 |
418
- | DualArmClothAlign-v1 | level2_from_level1 | deformable/dual_arm_gripper | 0.22 | 0.50 |
419
- | DualArmJugglingBall-v1 | level2_from_level1 | rigid_body/dual_arm_dex | 0.20 | 0.00 |
420
- | DexOpenDoor-v1 | level2_from_level1 | rigid_body/single_arm_dex | 0.18 | 0.28 |
421
- | DualArmLiftBall-v1 | level2_from_level1 | rigid_body/dual_arm_gripper | 0.18 | 0.36 |
422
- | DualArmStraightenRope-v1 | level2_from_level1 | deformable/dual_arm_gripper | 0.14 | 0.24 |
423
- | PivotBook-v1 | level2_from_level1 | rigid_body/single_arm_gripper | 0.14 | 0.16 |
424
- | ScoopWater-v1 | level2_from_level1 | deformable/single_arm_gripper | 0.14 | 0.26 |
425
- | DualArmStuffCloth-v1 | level2_from_level1 | deformable/dual_arm_dex | 0.10 | 0.44 |
426
- | OpenBoxRetrieval-v1 | level2_from_level1 | rigid_body/single_arm_gripper | 0.06 | 0.04 |
427
- | StraightenTwoRopesSequential-v1 | level2_from_level1 | deformable/single_arm_gripper | 0.06 | 0.12 |
428
- | DualArmBookPickPlace-v1 | level2_from_level1 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
429
- | DualArmCleanupPotEnv-v1 | level2_from_level1 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
430
- | DualArmOpenBottle-v1 | level2_from_level1 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
431
- | DualArmPouringTwoCups-v1 | level2_from_level1 | deformable/dual_arm_dex | 0.00 | 0.00 |
432
- | DualArmPushObject-v1 | level2_from_level1 | deformable/dual_arm_gripper | 0.00 | 0.00 |
433
- | KitAssembly-v1 | level2_from_level1 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
434
- | ReorientCube-v1 | level1_from_level2 | rigid_body/single_arm_dex | 1.00 | 0.84 |
435
- | DrapeCloth-v1 | level1_from_level2 | deformable/single_arm_dex | 0.98 | 0.96 |
436
- | DualArmLiftSoftDex-v1 | level1_from_level2 | deformable/dual_arm_dex | 0.98 | 1.00 |
437
- | TurnOnFaucet-v1 | level1_from_level2 | rigid_body/single_arm_dex | 0.98 | 0.84 |
438
- | DualArmButtonPress-v1 | level1_from_level2 | rigid_body/dual_arm_gripper | 0.94 | 0.90 |
439
- | GraspCylinder-v1 | level1_from_level2 | rigid_body/single_arm_dex | 0.94 | 0.86 |
440
- | RopePalmCageHoldDex-v1 | level1_from_level2 | deformable/single_arm_dex | 0.94 | 0.92 |
441
- | DualArmOpenScissor-v1 | level1_from_level2 | rigid_body/dual_arm_dex | 0.86 | 0.98 |
442
- | DualArmRopeStretch-v1 | level1_from_level2 | deformable/dual_arm_dex | 0.82 | 0.48 |
443
- | LiftBucket-v1 | level1_from_level2 | rigid_body/single_arm_dex | 0.82 | 0.54 |
444
- | DualArmRopeCylinderWarpping-v1 | level1_from_level2 | deformable/dual_arm_gripper | 0.70 | 0.66 |
445
- | PushSoccer-v1 | level1_from_level2 | rigid_body/single_arm_gripper | 0.70 | 0.60 |
446
- | DualArmFoldHandKerchief-v1 | level1_from_level2 | deformable/dual_arm_gripper | 0.66 | 0.60 |
447
- | StackBlock-v1 | level1_from_level2 | rigid_body/single_arm_gripper | 0.66 | 0.54 |
448
- | DualArmPouringWater-v1 | level1_from_level2 | deformable/dual_arm_gripper | 0.64 | 0.72 |
449
- | StraightenRope-v1 | level1_from_level2 | deformable/single_arm_gripper | 0.62 | 0.66 |
450
- | SqueezeClay-v1 | level1_from_level2 | deformable/single_arm_dex | 0.60 | 0.40 |
451
- | DualArmDoubleDoorBox-v1 | level1_from_level2 | rigid_body/dual_arm_gripper | 0.56 | 0.70 |
452
- | PickPlacePen-v1 | level1_from_level2 | rigid_body/single_arm_gripper | 0.50 | 0.34 |
453
- | DualArmPushBox-v1 | level1_from_level2 | rigid_body/dual_arm_gripper | 0.42 | 0.60 |
454
- | DualArmSwingCup-v1 | level1_from_level2 | rigid_body/dual_arm_dex | 0.42 | 0.42 |
455
- | DualArmCoverCloth-v1 | level1_from_level2 | deformable/dual_arm_dex | 0.36 | 0.24 |
456
- | ShapeTToI-v1 | level1_from_level2 | deformable/single_arm_gripper | 0.32 | 0.10 |
457
- | DualArmOpenPen-v1 | level1_from_level2 | rigid_body/dual_arm_dex | 0.28 | 0.30 |
458
- | UnfoldShirt-v1 | level1_from_level2 | deformable/single_arm_gripper | 0.26 | 0.28 |
459
- | DualArmPressButton-v1 | level1_from_level2 | rigid_body/dual_arm_dex | 0.18 | 0.38 |
460
- | DualArmHammerHandover-v1 | level1_from_level2 | rigid_body/dual_arm_gripper | 0.12 | 0.24 |
461
- | OpenDoor-v1 | level1_from_level2 | rigid_body/single_arm_gripper | 0.10 | 0.12 |
462
- | DualArmThreading-v1 | level3_from_level2 | deformable/dual_arm_dex | 0.04 | 0.00 |
463
- | DualArmThreadCloth-v1 | level3_from_level2 | deformable/dual_arm_dex | 0.02 | 0.04 |
464
- | ClutterGrasp-v1 | level3_from_level2 | rigid_body/single_arm_dex | 0.00 | 0.04 |
465
- | DualArmAssembleBracket-v1 | level3_from_level2 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
466
- | DualArmFoldCarton-v1 | level3_from_level2 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
467
- | DualArmLayerClayInDish-v1 | level3_from_level2 | deformable/dual_arm_dex | 0.00 | 0.00 |
468
- | DualArmMakeLine-v1 | level3_from_level2 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
469
- | DualArmOpenFridge-v1 | level3_from_level2 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
470
- | DualArmOrganizeCabinet-v1 | level3_from_level2 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
471
- | DualArmPingpongBalance-v1 | level3_from_level2 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
472
- | DualArmPlugCharger-v1 | level3_from_level2 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
473
- | DualArmSetupTable-v1 | level3_from_level2 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
474
- | DualArmShapeDoll-v1 | level3_from_level2 | deformable/dual_arm_gripper | 0.00 | 0.00 |
475
- | DualArmThreadRopeThroughTube-v1 | level3_from_level2 | deformable/dual_arm_gripper | 0.00 | 0.00 |
476
- | DualArmWeaveCloth-v1 | level3_from_level2 | deformable/dual_arm_gripper | 0.00 | 0.00 |
477
- | GroceryPrepare-v1 | level3_from_level2 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
478
- | HammerNail-v1 | level3_from_level2 | rigid_body/single_arm_dex | 0.00 | 0.00 |
479
- | HitBox-v1 | level3_from_level2 | deformable/single_arm_dex | 0.00 | 0.00 |
480
- | LiquidMeasuring-v1 | level3_from_level2 | deformable/single_arm_gripper | 0.00 | 0.00 |
481
- | MazeBlockPush-v1 | level3_from_level2 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
482
- | MobilePickPlace-v1 | level3_from_level2 | rigid_body/single_arm_dex | 0.00 | 0.00 |
483
- | PressFitSoftPeg-v1 | level3_from_level2 | deformable/single_arm_dex | 0.00 | 0.00 |
484
- | PushTargetPose-v1 | level3_from_level2 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
485
- | SeparateTwoRopesEnv-v1 | level3_from_level2 | deformable/single_arm_gripper | 0.00 | 0.06 |
486
- | Stack6Blocks-v1 | level3_from_level2 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
487
- | TweezersPickPlace-v1 | level3_from_level2 | rigid_body/single_arm_dex | 0.00 | 0.00 |
488
- | RopeBridgeDex-v1 | level2_from_level3 | deformable/single_arm_dex | 0.62 | 0.32 |
489
- | HangMug-v1 | level2_from_level3 | rigid_body/single_arm_gripper | 0.60 | 0.42 |
490
- | ClothWipe-v1 | level2_from_level3 | deformable/single_arm_dex | 0.58 | 0.24 |
491
- | FoldShirt-v1 | level2_from_level3 | deformable/single_arm_gripper | 0.46 | 0.62 |
492
- | DexOpenDoor-v1 | level2_from_level3 | rigid_body/single_arm_dex | 0.40 | 0.28 |
493
- | GraspTwoBlocks-v1 | level2_from_level3 | rigid_body/single_arm_dex | 0.38 | 0.56 |
494
- | ReorientPen-v1 | level2_from_level3 | rigid_body/single_arm_dex | 0.36 | 0.30 |
495
- | HammerNailOut-v1 | level2_from_level3 | rigid_body/single_arm_dex | 0.30 | 0.32 |
496
- | DualArmOpenLaptop-v1 | level2_from_level3 | rigid_body/dual_arm_dex | 0.26 | 0.34 |
497
- | DualArmPlaceRopeOverRod-v1 | level2_from_level3 | deformable/dual_arm_dex | 0.24 | 0.48 |
498
- | PivotBook-v1 | level2_from_level3 | rigid_body/single_arm_gripper | 0.18 | 0.16 |
499
- | DualArmJugglingBall-v1 | level2_from_level3 | rigid_body/dual_arm_dex | 0.14 | 0.00 |
500
- | DualArmLiftBall-v1 | level2_from_level3 | rigid_body/dual_arm_gripper | 0.14 | 0.36 |
501
- | DualArmStraightenRope-v1 | level2_from_level3 | deformable/dual_arm_gripper | 0.14 | 0.24 |
502
- | DualArmThrowBlock-v1 | level2_from_level3 | rigid_body/dual_arm_dex | 0.14 | 0.20 |
503
- | DualArmClothAlign-v1 | level2_from_level3 | deformable/dual_arm_gripper | 0.12 | 0.50 |
504
- | DualArmStuffCloth-v1 | level2_from_level3 | deformable/dual_arm_dex | 0.12 | 0.44 |
505
- | DualArmPegInHole-v1 | level2_from_level3 | rigid_body/dual_arm_gripper | 0.10 | 0.18 |
506
- | StraightenTwoRopesSequential-v1 | level2_from_level3 | deformable/single_arm_gripper | 0.08 | 0.12 |
507
- | RollingPin-v1 | level2_from_level3 | deformable/single_arm_dex | 0.06 | 0.38 |
508
- | ScoopWater-v1 | level2_from_level3 | deformable/single_arm_gripper | 0.06 | 0.26 |
509
- | OpenBoxRetrieval-v1 | level2_from_level3 | rigid_body/single_arm_gripper | 0.04 | 0.04 |
510
- | DualArmBookPickPlace-v1 | level2_from_level3 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
511
- | DualArmCleanupPotEnv-v1 | level2_from_level3 | rigid_body/dual_arm_gripper | 0.00 | 0.00 |
512
- | DualArmOpenBottle-v1 | level2_from_level3 | rigid_body/dual_arm_dex | 0.00 | 0.00 |
513
- | DualArmPouringTwoCups-v1 | level2_from_level3 | deformable/dual_arm_dex | 0.00 | 0.00 |
514
- | DualArmPushObject-v1 | level2_from_level3 | deformable/dual_arm_gripper | 0.00 | 0.00 |
515
- | KitAssembly-v1 | level2_from_level3 | rigid_body/single_arm_gripper | 0.00 | 0.00 |
516
-
517
- <!-- multi_task_level_transfer:end -->
 
2
  license: mit
3
  ---
4
 
5
+ # ManiLadder Models
6
 
7
+ Baseline policies for the 112 ManiLadder manipulation tasks. The repository has three top-level folders, one
8
+ per method:
9
 
10
+ | Folder | Method | What is inside |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
  |---|---|---|
12
+ | [`small_flow_matching_policy/`](small_flow_matching_policy) | Small single-task policy: frozen RADIO backbone + transformer Flow Matching head | one folder per task with the selected checkpoint and its normalization statistics (111 tasks; PressFitSoftPeg-v1 has none) |
13
+ | [`pi05/`](pi05) | PI0.5 finetuned with LoRA | sub-folders by training scope: single task, level, suite, all tasks, plus transfer and data-scaling runs |
14
+ | [`astra/`](astra) | Astra: frozen script policies written by a GPT-6 Astra coding agent (Agent-D20-E20) | one folder per task with the frozen controller and its reference data (112 tasks), plus shared protocol files |
15
+
16
+ Each folder has its own README with the file layout, training and evaluation details and result tables:
17
+ [small_flow_matching_policy/README.md](small_flow_matching_policy/README.md),
18
+ [pi05/README.md](pi05/README.md), [astra/README.md](astra/README.md).
19
+
20
+ ## Layout
21
+
22
+ ```text
23
+ small_flow_matching_policy/<task>/ ckpt_step_<N>.pt, norm_stats.json, optional release notes
24
+ pi05/
25
+ single_task/<task>/ best checkpoint per task (111 tasks)
26
+ previous_single_task/<task>/ earlier single-task runs kept for reference (62 tasks)
27
+ level/level_<1..4>/ one policy per difficulty level
28
+ level_transfer/level<B>_from_level<A>/
29
+ suite/level2_<suite>/ one policy per level-2 suite (6 suites)
30
+ suite_transfer/level2_<x>_from_<y>/
31
+ suite_data_scaling/level2_rigid_n<N>/ level-2 rigid suite at 20 / 50 / 100 / 200 demos per task
32
+ all_tasks/all112/ one policy for all 112 tasks
33
+ astra/
34
+ <task>/ frozen script policy of one task
35
+ shared/, manifest.json, results_snapshot_20261105.csv
36
+ ```
37
+
38
+ ## Download part of the repository
39
+
40
+ ```python
41
+ from huggingface_hub import snapshot_download
42
+
43
+ # one PI0.5 single-task checkpoint
44
+ snapshot_download(
45
+ repo_id="Linslab/ManiLadder_Models",
46
+ allow_patterns="pi05/single_task/StackBlock-v1/*",
47
+ local_dir="./models",
48
+ )
49
+ ```
50
+
51
+ Replace the pattern with, for example, `small_flow_matching_policy/StackBlock-v1/*`, `pi05/level/level_1/*`
52
+ or `astra/StackBlock-v1/*`. The whole repository is about 744 GB (PI0.5 431 GB, Astra 269 GB, small
53
+ flow-matching 44 GB), most of it PI0.5 LoRA weights and the demonstrations inside the Astra bundles.
54
+
55
+ ## Repository merge (2026-11-16)
56
+
57
+ This repository now combines the former `Linslab/ManiLadder_Models` (small flow-matching checkpoints) and
58
+ `Linslab/ManiLadder_PI05_Models` (PI0.5 LoRA checkpoints and frozen agent policies). The old
59
+ `ManiLadder_PI05_Models` address redirects here. File contents were not changed; paths moved as follows:
60
+
61
+ | Previous location | Current path |
62
+ |---|---|
63
+ | `ManiLadder_Models`: `<task>/...` | `small_flow_matching_policy/<task>/...` |
64
+ | `ManiLadder_PI05_Models`: `single_task/...` | `pi05/single_task/...` |
65
+ | `ManiLadder_PI05_Models`: `previous_single_task/...` | `pi05/previous_single_task/...` |
66
+ | `ManiLadder_PI05_Models`: `multi_task/level_<N>/...` | `pi05/level/level_<N>/...` |
67
+ | `ManiLadder_PI05_Models`: `multi_task/level<B>_from_level<A>/...` | `pi05/level_transfer/level<B>_from_level<A>/...` |
68
+ | `ManiLadder_PI05_Models`: `multi_task/level2_<suite>/...` | `pi05/suite/level2_<suite>/...` |
69
+ | `ManiLadder_PI05_Models`: `multi_task/level2_<x>_from_<y>/...` | `pi05/suite_transfer/level2_<x>_from_<y>/...` |
70
+ | `ManiLadder_PI05_Models`: `multi_task/results.csv` | `pi05/multi_task_results.csv` |
71
+ | `ManiLadder_PI05_Models`: `agent_d20_e20/frozen_20261102/tasks/<task>/frozen/...` | `astra/<task>/...` |
72
+
73
+ Added in the merge: the PI0.5 single-task run for DualArmRubikCube-v1, the all-task policy
74
+ (`pi05/all_tasks/all112/`), the level-2 rigid data-scaling policies (`pi05/suite_data_scaling/`), and twelve
75
+ further Astra policies. Index files (`results.csv`, manifests) were updated to the new paths. Notes and
76
+ per-run metadata written before the merge may still quote the previous paths, and dataset paths of the form
77
+ `raw/<task>` or `preprocessed/<task>`, which are now `<task>/` in the datasets linked below.
78
+
79
+ ## Related
80
+
81
+ - Raw demonstrations: [Linslab/ManiLadder_Dataset](https://huggingface.co/datasets/Linslab/ManiLadder_Dataset)
82
+ - Preprocessed demonstrations: [ChongkaiGao/ManiLadder_preprocessed_dataset](https://huggingface.co/datasets/ChongkaiGao/ManiLadder_preprocessed_dataset)
83
+ - Code: <https://github.com/HeegerGao/maniladder>