JIAlonglong commited on
Commit
062cf4c
·
verified ·
1 Parent(s): ebe262d

Publish GR1 six-task pilot protocol and verified-format results (part 2)

Browse files
gr1-six-task-mini-v1.zip CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:3bcec0db5154c274b03ff60c3c0babd9a712d2a1ca42bd4df447617e5a03a312
3
- size 61446068
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7d2fb9a78d5ac657aec04e29b940d498ebe7ca045714253d95a8ef833894b737
3
+ size 61476317
gr1-six-task-mini-v1.zip.sha256 CHANGED
@@ -1 +1 @@
1
- 3bcec0db5154c274b03ff60c3c0babd9a712d2a1ca42bd4df447617e5a03a312 gr1-six-task-mini-v1.zip
 
1
+ 7d2fb9a78d5ac657aec04e29b940d498ebe7ca045714253d95a8ef833894b737 gr1-six-task-mini-v1.zip
human-transfer/protocol.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "track": "robot-human-transfer",
3
+ "version": "human-transfer-v1",
4
+ "robot_training_split_sha256": "8b1ea19b1f2ab44d9a44d99a46cc31396f0ec71f6b88002162512e7b4a102794",
5
+ "robot_evaluation_protocol_sha256": "29f2dd387cc5b9df43711ef65837868f3dc837d65594242bef02c70cf55334fa",
6
+ "human_subset_sha256": "68a6a019ee5b9b41a608a0e550aaf6fce3b7b8ea0d04e95e2322a9ffa8630468",
7
+ "human_clips": 480,
8
+ "robot_train_demos": 480,
9
+ "human_modalities": [
10
+ "video_only",
11
+ "video_and_hand_pose"
12
+ ],
13
+ "comparisons": [
14
+ "equal_gpu_hours",
15
+ "equal_robot_exposure"
16
+ ],
17
+ "main_comparison": "equal_gpu_hours",
18
+ "recommended_budget_gpu_hours_per_arm": 192,
19
+ "optional_smoke_gpu_hours_per_arm": 16,
20
+ "two_arm_budget_with_smoke": 416,
21
+ "robot_only_baseline_required": true,
22
+ "no_external_robot_task_weights": true,
23
+ "status": "submission_track_ready; training_adapter_and_mixed_loader_not_validated; no human training launched"
24
+ }
human-transfer/subset.json ADDED
The diff for this file is too large to render. See raw diff