cy0307 commited on
Commit
f129a25
·
verified ·
1 Parent(s): efd1936

Add files using upload-large-folder tool

Browse files
Files changed (23) hide show
  1. TASK_METHOD_20_GAP_AUDIT.md +16 -20
  2. TASK_METHOD_20_RESULT_MATRIX.md +10 -10
  3. assets/charts/episode128_task_model_radar.svg +14 -4
  4. assets/charts/unified_task_model_radar.svg +13 -3
  5. data/episode128_task_model_radar.json +190 -191
  6. data/mirror_parity.json +0 -0
  7. data/public_surface_qa.json +4 -4
  8. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json +59 -0
  9. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/per_class_metrics.csv +0 -0
  10. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/predictions.csv +0 -0
  11. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/confusion_matrix.csv +0 -0
  12. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/per_class_metrics.csv +1212 -0
  13. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/predictions.csv +0 -0
  14. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/confusion_matrix.csv +0 -0
  15. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json +59 -0
  16. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/per_class_metrics.csv +892 -0
  17. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/predictions.csv +0 -0
  18. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json +1 -0
  19. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json +50 -0
  20. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/predictions.csv +0 -0
  21. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json +48 -0
  22. results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/predictions.csv +0 -0
  23. scripts/omni/run_128_task_baselines.py +413 -14
TASK_METHOD_20_GAP_AUDIT.md CHANGED
@@ -1,6 +1,6 @@
1
  # Task Method 20-Result Gap Audit
2
 
3
- Generated: `2026-06-18T11:15:34+00:00`
4
 
5
  This audit is the explicit gap ledger for the 9-method x 20-task result matrix.
6
  It keeps missing cells visible while preserving the rule that a numeric score
@@ -9,8 +9,8 @@ requires a real task target and source artifact.
9
  ## Score Summary
10
 
11
  - Method-task records: `180`
12
- - Numeric scored records: `123`
13
- - Scoreless records: `57`
14
  - Proxy-scored records: `4`
15
  - Source matrix: [`docs/data/task_method_20_result_matrix.json`](docs/data/task_method_20_result_matrix.json)
16
 
@@ -20,8 +20,8 @@ requires a real task target and source artifact.
20
  | --- | --- | --- | --- | --- | --- |
21
  | Minimal | minimal | 20/20 | 0 | 0 | scored: 20 |
22
  | Neural MLP | neural_mlp | 20/20 | 0 | 0 | scored: 20 |
23
- | 128ep Metadata Simple | metadata128_simple | 8/20 | 12 | 0 | not_supported_by_metadata_only_package: 8, scored: 8, unsupported_without_required_target: 4 |
24
- | 128ep Metadata NN | metadata128_neural_mlp | 8/20 | 12 | 0 | not_supported_by_metadata_only_package: 12, scored: 8 |
25
  | 128ep Raw Simple | raw128_simple | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
26
  | 128ep Raw NN | raw128_neural_mlp | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
27
  | Qwen3-Omni v6 LoRA | qwen3_omni_v6_lora | 15/20 | 5 | 0 | not_evaluated_in_verified_package: 5, scored: 15 |
@@ -33,14 +33,17 @@ requires a real task target and source artifact.
33
  | Status | Count | Next step |
34
  | --- | --- | --- |
35
  | not_evaluated_in_verified_package | 33 | Generate verified model outputs for this task contract and score them against the held-out labels. |
36
- | not_supported_by_metadata_only_package | 20 | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
37
- | unsupported_without_required_target | 4 | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
38
 
39
  ## Scoreless Records
40
 
41
  | Task | Task label | Method | Status | Required evidence |
42
  | --- | --- | --- | --- | --- |
 
 
43
  | 02 | Procedure Step Recognition | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 
44
  | 05 | Hand Trajectory Forecasting | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
45
  | 05 | Hand Trajectory Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
46
  | 05 | Hand Trajectory Forecasting | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
@@ -63,38 +66,31 @@ requires a real task target and source artifact.
63
  | 12 | Multimodal Synchronization Detection | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
64
  | 12 | Multimodal Synchronization Detection | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
65
  | 12 | Multimodal Synchronization Detection | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
66
- | 13 | Long-Horizon Next-Action Forecasting | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
67
- | 13 | Long-Horizon Next-Action Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
68
  | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
69
  | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
70
- | 14 | Long-Horizon Next-Subtask Forecasting | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
71
- | 14 | Long-Horizon Next-Subtask Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
72
  | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
73
  | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
74
- | 15 | Interaction Text Prediction | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
75
  | 15 | Interaction Text Prediction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
76
  | 15 | Interaction Text Prediction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
77
  | 15 | Interaction Text Prediction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
78
  | 15 | Interaction Text Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
79
- | 16 | Action-Object Relation Prediction | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
80
- | 16 | Action-Object Relation Prediction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
81
  | 16 | Action-Object Relation Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
82
- | 17 | Future Object-Set Forecasting | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
83
- | 17 | Future Object-Set Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
84
  | 17 | Future Object-Set Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
85
  | 17 | Future Object-Set Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
86
- | 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
87
  | 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
88
  | 18 | IMU-to-Hand Pose Reconstruction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
89
  | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
90
  | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
91
- | 19 | Camera-View Synchronization Retrieval | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
92
  | 19 | Camera-View Synchronization Retrieval | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
93
  | 19 | Camera-View Synchronization Retrieval | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
94
  | 19 | Camera-View Synchronization Retrieval | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
95
  | 19 | Camera-View Synchronization Retrieval | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
96
- | 20 | Time-to-Next-Transition Regression | 128ep Metadata Simple | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
97
- | 20 | Time-to-Next-Transition Regression | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
98
  | 20 | Time-to-Next-Transition Regression | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
99
  | 20 | Time-to-Next-Transition Regression | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
100
 
 
1
  # Task Method 20-Result Gap Audit
2
 
3
+ Generated: `2026-06-18T12:07:14+00:00`
4
 
5
  This audit is the explicit gap ledger for the 9-method x 20-task result matrix.
6
  It keeps missing cells visible while preserving the rule that a numeric score
 
9
  ## Score Summary
10
 
11
  - Method-task records: `180`
12
+ - Numeric scored records: `127`
13
+ - Scoreless records: `53`
14
  - Proxy-scored records: `4`
15
  - Source matrix: [`docs/data/task_method_20_result_matrix.json`](docs/data/task_method_20_result_matrix.json)
16
 
 
20
  | --- | --- | --- | --- | --- | --- |
21
  | Minimal | minimal | 20/20 | 0 | 0 | scored: 20 |
22
  | Neural MLP | neural_mlp | 20/20 | 0 | 0 | scored: 20 |
23
+ | 128ep Metadata Simple | metadata128_simple | 13/20 | 7 | 0 | scored: 13, unsupported_without_required_target: 7 |
24
+ | 128ep Metadata NN | metadata128_neural_mlp | 7/20 | 13 | 0 | not_supported_by_metadata_only_package: 7, scored: 7, unsupported_without_required_target: 6 |
25
  | 128ep Raw Simple | raw128_simple | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
26
  | 128ep Raw NN | raw128_neural_mlp | 20/20 | 0 | 2 | proxy_scored: 2, scored: 18 |
27
  | Qwen3-Omni v6 LoRA | qwen3_omni_v6_lora | 15/20 | 5 | 0 | not_evaluated_in_verified_package: 5, scored: 15 |
 
33
  | Status | Count | Next step |
34
  | --- | --- | --- |
35
  | not_evaluated_in_verified_package | 33 | Generate verified model outputs for this task contract and score them against the held-out labels. |
36
+ | not_supported_by_metadata_only_package | 7 | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
37
+ | unsupported_without_required_target | 13 | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
38
 
39
  ## Scoreless Records
40
 
41
  | Task | Task label | Method | Status | Required evidence |
42
  | --- | --- | --- | --- | --- |
43
+ | 01 | Action Recognition | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
44
+ | 02 | Procedure Step Recognition | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
45
  | 02 | Procedure Step Recognition | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
46
+ | 04 | Next-Action Prediction | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
47
  | 05 | Hand Trajectory Forecasting | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
48
  | 05 | Hand Trajectory Forecasting | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
49
  | 05 | Hand Trajectory Forecasting | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 
66
  | 12 | Multimodal Synchronization Detection | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
67
  | 12 | Multimodal Synchronization Detection | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
68
  | 12 | Multimodal Synchronization Detection | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
69
+ | 13 | Long-Horizon Next-Action Forecasting | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 
70
  | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
71
  | 13 | Long-Horizon Next-Action Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
72
+ | 14 | Long-Horizon Next-Subtask Forecasting | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 
73
  | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
74
  | 14 | Long-Horizon Next-Subtask Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
75
+ | 15 | Interaction Text Prediction | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
76
  | 15 | Interaction Text Prediction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
77
  | 15 | Interaction Text Prediction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
78
  | 15 | Interaction Text Prediction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
79
  | 15 | Interaction Text Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
80
+ | 16 | Action-Object Relation Prediction | 128ep Metadata NN | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
 
81
  | 16 | Action-Object Relation Prediction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 
 
82
  | 17 | Future Object-Set Forecasting | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
83
  | 17 | Future Object-Set Forecasting | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
84
+ | 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
85
  | 18 | IMU-to-Hand Pose Reconstruction | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
86
  | 18 | IMU-to-Hand Pose Reconstruction | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
87
  | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
88
  | 18 | IMU-to-Hand Pose Reconstruction | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
89
+ | 19 | Camera-View Synchronization Retrieval | 128ep Metadata Simple | unsupported | Export the missing target field for this 128-episode method, then rerun the same train/validation/test split. |
90
  | 19 | Camera-View Synchronization Retrieval | 128ep Metadata NN | not supported | Run the task with raw sensor-feature blocks or add a task-specific metadata target builder before assigning a numeric score. |
91
  | 19 | Camera-View Synchronization Retrieval | Qwen3-Omni v6 LoRA | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
92
  | 19 | Camera-View Synchronization Retrieval | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
93
  | 19 | Camera-View Synchronization Retrieval | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
 
 
94
  | 20 | Time-to-Next-Transition Regression | Cosmos3-Super Reasoner | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
95
  | 20 | Time-to-Next-Transition Regression | Cosmos3-Nano Future Window | not evaluated | Generate verified model outputs for this task contract and score them against the held-out labels. |
96
 
TASK_METHOD_20_RESULT_MATRIX.md CHANGED
@@ -8,8 +8,8 @@ Legend: `score` = numeric task score, `proxy` = documented raw128 compact proxy
8
  | --- | ---: | ---: | ---: | ---: | --- |
9
  | Minimal | 20 | 20 | 0 | 0 | scored 20 |
10
  | Neural MLP | 20 | 20 | 0 | 0 | scored 20 |
11
- | 128ep Metadata Simple | 20 | 8 | 0 | 12 | not supported 8, scored 8, unsupported 4 |
12
- | 128ep Metadata NN | 20 | 8 | 0 | 12 | not supported 12, scored 8 |
13
  | 128ep Raw Simple | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
14
  | 128ep Raw NN | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
15
  | Qwen3-Omni v6 LoRA | 20 | 15 | 0 | 5 | not evaluated 5, scored 15 |
@@ -30,13 +30,13 @@ Legend: `score` = numeric task score, `proxy` = documented raw128 compact proxy
30
  | 10 | Cross-Modal Reconstruction | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
31
  | 11 | Temporal Order Verification | score | score | score | score | score | score | score | not evaluated | not evaluated |
32
  | 12 | Multimodal Synchronization Detection | score | score | unsupported | not supported | score | score | score | not evaluated | not evaluated |
33
- | 13 | Long-Horizon Next-Action Forecasting | score | score | not supported | not supported | score | score | score | not evaluated | not evaluated |
34
- | 14 | Long-Horizon Next-Subtask Forecasting | score | score | not supported | not supported | score | score | score | not evaluated | not evaluated |
35
- | 15 | Interaction Text Prediction | score | score | not supported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
36
- | 16 | Action-Object Relation Prediction | score | score | not supported | not supported | score | score | score | score | not evaluated |
37
- | 17 | Future Object-Set Forecasting | score | score | not supported | not supported | score | score | score | not evaluated | not evaluated |
38
- | 18 | IMU-to-Hand Pose Reconstruction | score | score | not supported | not supported | score | score | not evaluated | not evaluated | not evaluated |
39
- | 19 | Camera-View Synchronization Retrieval | score | score | not supported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
40
- | 20 | Time-to-Next-Transition Regression | score | score | not supported | not supported | score | score | score | not evaluated | not evaluated |
41
 
42
  Sources and raw values are in `docs/data/task_method_20_result_matrix.json` and `docs/data/unified_task_model_radar.json`.
 
8
  | --- | ---: | ---: | ---: | ---: | --- |
9
  | Minimal | 20 | 20 | 0 | 0 | scored 20 |
10
  | Neural MLP | 20 | 20 | 0 | 0 | scored 20 |
11
+ | 128ep Metadata Simple | 20 | 13 | 0 | 7 | scored 13, unsupported 7 |
12
+ | 128ep Metadata NN | 20 | 13 | 0 | 7 | not supported 7, scored 13 |
13
  | 128ep Raw Simple | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
14
  | 128ep Raw NN | 20 | 20 | 2 | 0 | proxy scored 2, scored 18 |
15
  | Qwen3-Omni v6 LoRA | 20 | 15 | 0 | 5 | not evaluated 5, scored 15 |
 
30
  | 10 | Cross-Modal Reconstruction | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
31
  | 11 | Temporal Order Verification | score | score | score | score | score | score | score | not evaluated | not evaluated |
32
  | 12 | Multimodal Synchronization Detection | score | score | unsupported | not supported | score | score | score | not evaluated | not evaluated |
33
+ | 13 | Long-Horizon Next-Action Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
34
+ | 14 | Long-Horizon Next-Subtask Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
35
+ | 15 | Interaction Text Prediction | score | score | unsupported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
36
+ | 16 | Action-Object Relation Prediction | score | score | score | score | score | score | score | score | not evaluated |
37
+ | 17 | Future Object-Set Forecasting | score | score | score | score | score | score | score | not evaluated | not evaluated |
38
+ | 18 | IMU-to-Hand Pose Reconstruction | score | score | unsupported | not supported | score | score | not evaluated | not evaluated | not evaluated |
39
+ | 19 | Camera-View Synchronization Retrieval | score | score | unsupported | not supported | proxy | proxy | not evaluated | not evaluated | not evaluated |
40
+ | 20 | Time-to-Next-Transition Regression | score | score | score | score | score | score | score | not evaluated | not evaluated |
41
 
42
  Sources and raw values are in `docs/data/task_method_20_result_matrix.json` and `docs/data/unified_task_model_radar.json`.
assets/charts/episode128_task_model_radar.svg CHANGED
assets/charts/unified_task_model_radar.svg CHANGED
data/episode128_task_model_radar.json CHANGED
@@ -1,12 +1,12 @@
1
  {
2
  "title": "128-Episode 20-Task Radar",
3
  "status": "pass",
4
- "generated_at_utc": "2026-06-18T11:15:02+00:00",
5
  "description": "Selected 128-episode metadata/raw baselines plus verified Qwen3/Cosmos branches. Every method has 20 records; numeric scores appear only where the public artifact produced that task target.",
6
  "task_count": 20,
7
  "method_count": 7,
8
  "method_task_record_count": 140,
9
- "scored_method_task_count": 83,
10
  "normalization_policy": {
11
  "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
12
  "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
@@ -30,18 +30,17 @@
30
  "method_detail": "128-episode JSONL metadata/text simple baselines.",
31
  "plotted_as": "colored point overlay",
32
  "result_record_count": 20,
33
- "scored_task_count": 8,
34
- "covered_task_count": 8,
35
  "proxy_scored_task_count": 0,
36
- "scoreless_task_count": 12,
37
- "unsupported_task_count": 12,
38
  "not_evaluated_task_count": 0,
39
  "status_counts": {
40
- "not_supported_by_metadata_only_package": 8,
41
- "scored": 8,
42
- "unsupported_without_required_target": 4
43
  },
44
- "coverage_fraction": 0.4,
45
  "result_record_fraction": 1.0
46
  },
47
  {
@@ -55,17 +54,17 @@
55
  "method_detail": "128-episode JSONL metadata/text MLP baselines.",
56
  "plotted_as": "colored point overlay",
57
  "result_record_count": 20,
58
- "scored_task_count": 8,
59
- "covered_task_count": 8,
60
  "proxy_scored_task_count": 0,
61
- "scoreless_task_count": 12,
62
- "unsupported_task_count": 12,
63
  "not_evaluated_task_count": 0,
64
  "status_counts": {
65
- "not_supported_by_metadata_only_package": 12,
66
- "scored": 8
67
  },
68
- "coverage_fraction": 0.4,
69
  "result_record_fraction": 1.0
70
  },
71
  {
@@ -1295,26 +1294,26 @@
1295
  "raw128_proxy_axis": false,
1296
  "values": {
1297
  "metadata128_simple": {
1298
- "raw": null,
1299
  "metric_key": "macro_f1",
1300
- "source": null,
1301
  "scope": "multi_episode_128_metadata_baseline",
1302
- "status": "not_supported_by_metadata_only_package",
1303
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1304
- "normalized_score": null,
1305
- "raw_text": "n/a",
1306
- "status_label": "not supported"
1307
  },
1308
  "metadata128_neural_mlp": {
1309
- "raw": null,
1310
  "metric_key": "macro_f1",
1311
- "source": null,
1312
  "scope": "multi_episode_128_metadata_baseline",
1313
- "status": "not_supported_by_metadata_only_package",
1314
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1315
- "normalized_score": null,
1316
- "raw_text": "n/a",
1317
- "status_label": "not supported"
1318
  },
1319
  "raw128_simple": {
1320
  "raw": 0.0024280172369056294,
@@ -1386,26 +1385,26 @@
1386
  "raw128_proxy_axis": false,
1387
  "values": {
1388
  "metadata128_simple": {
1389
- "raw": null,
1390
  "metric_key": "macro_f1",
1391
- "source": null,
1392
  "scope": "multi_episode_128_metadata_baseline",
1393
- "status": "not_supported_by_metadata_only_package",
1394
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1395
- "normalized_score": null,
1396
- "raw_text": "n/a",
1397
- "status_label": "not supported"
1398
  },
1399
  "metadata128_neural_mlp": {
1400
- "raw": null,
1401
  "metric_key": "macro_f1",
1402
- "source": null,
1403
  "scope": "multi_episode_128_metadata_baseline",
1404
- "status": "not_supported_by_metadata_only_package",
1405
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1406
- "normalized_score": null,
1407
- "raw_text": "n/a",
1408
- "status_label": "not supported"
1409
  },
1410
  "raw128_simple": {
1411
  "raw": 0.0,
@@ -1479,13 +1478,13 @@
1479
  "metadata128_simple": {
1480
  "raw": null,
1481
  "metric_key": "macro_f1",
1482
- "source": null,
1483
  "scope": "multi_episode_128_metadata_baseline",
1484
- "status": "not_supported_by_metadata_only_package",
1485
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1486
  "normalized_score": null,
1487
  "raw_text": "n/a",
1488
- "status_label": "not supported"
1489
  },
1490
  "metadata128_neural_mlp": {
1491
  "raw": null,
@@ -1568,26 +1567,26 @@
1568
  "raw128_proxy_axis": false,
1569
  "values": {
1570
  "metadata128_simple": {
1571
- "raw": null,
1572
  "metric_key": "macro_f1",
1573
- "source": null,
1574
  "scope": "multi_episode_128_metadata_baseline",
1575
- "status": "not_supported_by_metadata_only_package",
1576
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1577
- "normalized_score": null,
1578
- "raw_text": "n/a",
1579
- "status_label": "not supported"
1580
  },
1581
  "metadata128_neural_mlp": {
1582
- "raw": null,
1583
  "metric_key": "macro_f1",
1584
- "source": null,
1585
  "scope": "multi_episode_128_metadata_baseline",
1586
- "status": "not_supported_by_metadata_only_package",
1587
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1588
- "normalized_score": null,
1589
- "raw_text": "n/a",
1590
- "status_label": "not supported"
1591
  },
1592
  "raw128_simple": {
1593
  "raw": 0.0,
@@ -1659,26 +1658,26 @@
1659
  "raw128_proxy_axis": false,
1660
  "values": {
1661
  "metadata128_simple": {
1662
- "raw": null,
1663
  "metric_key": "micro_f1",
1664
- "source": null,
1665
  "scope": "multi_episode_128_metadata_baseline",
1666
- "status": "not_supported_by_metadata_only_package",
1667
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1668
- "normalized_score": null,
1669
- "raw_text": "n/a",
1670
- "status_label": "not supported"
1671
  },
1672
  "metadata128_neural_mlp": {
1673
- "raw": null,
1674
  "metric_key": "micro_f1",
1675
- "source": null,
1676
  "scope": "multi_episode_128_metadata_baseline",
1677
- "status": "not_supported_by_metadata_only_package",
1678
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1679
- "normalized_score": null,
1680
- "raw_text": "n/a",
1681
- "status_label": "not supported"
1682
  },
1683
  "raw128_simple": {
1684
  "raw": 0.06469493412657774,
@@ -1752,13 +1751,13 @@
1752
  "metadata128_simple": {
1753
  "raw": null,
1754
  "metric_key": "mae",
1755
- "source": null,
1756
  "scope": "multi_episode_128_metadata_baseline",
1757
- "status": "not_supported_by_metadata_only_package",
1758
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1759
  "normalized_score": null,
1760
  "raw_text": "n/a",
1761
- "status_label": "not supported"
1762
  },
1763
  "metadata128_neural_mlp": {
1764
  "raw": null,
@@ -1843,13 +1842,13 @@
1843
  "metadata128_simple": {
1844
  "raw": null,
1845
  "metric_key": "mrr",
1846
- "source": null,
1847
  "scope": "multi_episode_128_metadata_baseline",
1848
- "status": "not_supported_by_metadata_only_package",
1849
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1850
  "normalized_score": null,
1851
  "raw_text": "n/a",
1852
- "status_label": "not supported"
1853
  },
1854
  "metadata128_neural_mlp": {
1855
  "raw": null,
@@ -1932,26 +1931,26 @@
1932
  "raw128_proxy_axis": false,
1933
  "values": {
1934
  "metadata128_simple": {
1935
- "raw": null,
1936
  "metric_key": "mae",
1937
- "source": null,
1938
  "scope": "multi_episode_128_metadata_baseline",
1939
- "status": "not_supported_by_metadata_only_package",
1940
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1941
- "normalized_score": null,
1942
- "raw_text": "n/a",
1943
- "status_label": "not supported"
1944
  },
1945
  "metadata128_neural_mlp": {
1946
- "raw": null,
1947
  "metric_key": "mae",
1948
- "source": null,
1949
  "scope": "multi_episode_128_metadata_baseline",
1950
- "status": "not_supported_by_metadata_only_package",
1951
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required",
1952
- "normalized_score": null,
1953
- "raw_text": "n/a",
1954
- "status_label": "not supported"
1955
  },
1956
  "raw128_simple": {
1957
  "raw": 52.32759475708008,
@@ -3530,17 +3529,17 @@
3530
  "task_label": "Long-Horizon Next-Action Forecasting",
3531
  "series_id": "metadata128_simple",
3532
  "method": "128ep Metadata Simple",
3533
- "status": "not_supported_by_metadata_only_package",
3534
- "status_label": "not supported",
3535
- "scored": false,
3536
  "proxy_scored": false,
3537
- "raw": null,
3538
- "raw_text": "n/a",
3539
- "normalized_score": null,
3540
  "metric_key": "macro_f1",
3541
- "source": null,
3542
  "scope": "multi_episode_128_metadata_baseline",
3543
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3544
  },
3545
  {
3546
  "task_number": 13,
@@ -3548,17 +3547,17 @@
3548
  "task_label": "Long-Horizon Next-Action Forecasting",
3549
  "series_id": "metadata128_neural_mlp",
3550
  "method": "128ep Metadata NN",
3551
- "status": "not_supported_by_metadata_only_package",
3552
- "status_label": "not supported",
3553
- "scored": false,
3554
  "proxy_scored": false,
3555
- "raw": null,
3556
- "raw_text": "n/a",
3557
- "normalized_score": null,
3558
  "metric_key": "macro_f1",
3559
- "source": null,
3560
  "scope": "multi_episode_128_metadata_baseline",
3561
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3562
  },
3563
  {
3564
  "task_number": 13,
@@ -3656,17 +3655,17 @@
3656
  "task_label": "Long-Horizon Next-Subtask Forecasting",
3657
  "series_id": "metadata128_simple",
3658
  "method": "128ep Metadata Simple",
3659
- "status": "not_supported_by_metadata_only_package",
3660
- "status_label": "not supported",
3661
- "scored": false,
3662
  "proxy_scored": false,
3663
- "raw": null,
3664
- "raw_text": "n/a",
3665
- "normalized_score": null,
3666
  "metric_key": "macro_f1",
3667
- "source": null,
3668
  "scope": "multi_episode_128_metadata_baseline",
3669
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3670
  },
3671
  {
3672
  "task_number": 14,
@@ -3674,17 +3673,17 @@
3674
  "task_label": "Long-Horizon Next-Subtask Forecasting",
3675
  "series_id": "metadata128_neural_mlp",
3676
  "method": "128ep Metadata NN",
3677
- "status": "not_supported_by_metadata_only_package",
3678
- "status_label": "not supported",
3679
- "scored": false,
3680
  "proxy_scored": false,
3681
- "raw": null,
3682
- "raw_text": "n/a",
3683
- "normalized_score": null,
3684
  "metric_key": "macro_f1",
3685
- "source": null,
3686
  "scope": "multi_episode_128_metadata_baseline",
3687
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3688
  },
3689
  {
3690
  "task_number": 14,
@@ -3782,17 +3781,17 @@
3782
  "task_label": "Interaction Text Prediction",
3783
  "series_id": "metadata128_simple",
3784
  "method": "128ep Metadata Simple",
3785
- "status": "not_supported_by_metadata_only_package",
3786
- "status_label": "not supported",
3787
  "scored": false,
3788
  "proxy_scored": false,
3789
  "raw": null,
3790
  "raw_text": "n/a",
3791
  "normalized_score": null,
3792
  "metric_key": "macro_f1",
3793
- "source": null,
3794
  "scope": "multi_episode_128_metadata_baseline",
3795
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3796
  },
3797
  {
3798
  "task_number": 15,
@@ -3908,17 +3907,17 @@
3908
  "task_label": "Action-Object Relation Prediction",
3909
  "series_id": "metadata128_simple",
3910
  "method": "128ep Metadata Simple",
3911
- "status": "not_supported_by_metadata_only_package",
3912
- "status_label": "not supported",
3913
- "scored": false,
3914
  "proxy_scored": false,
3915
- "raw": null,
3916
- "raw_text": "n/a",
3917
- "normalized_score": null,
3918
  "metric_key": "macro_f1",
3919
- "source": null,
3920
  "scope": "multi_episode_128_metadata_baseline",
3921
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3922
  },
3923
  {
3924
  "task_number": 16,
@@ -3926,17 +3925,17 @@
3926
  "task_label": "Action-Object Relation Prediction",
3927
  "series_id": "metadata128_neural_mlp",
3928
  "method": "128ep Metadata NN",
3929
- "status": "not_supported_by_metadata_only_package",
3930
- "status_label": "not supported",
3931
- "scored": false,
3932
  "proxy_scored": false,
3933
- "raw": null,
3934
- "raw_text": "n/a",
3935
- "normalized_score": null,
3936
  "metric_key": "macro_f1",
3937
- "source": null,
3938
  "scope": "multi_episode_128_metadata_baseline",
3939
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
3940
  },
3941
  {
3942
  "task_number": 16,
@@ -4034,17 +4033,17 @@
4034
  "task_label": "Future Object-Set Forecasting",
4035
  "series_id": "metadata128_simple",
4036
  "method": "128ep Metadata Simple",
4037
- "status": "not_supported_by_metadata_only_package",
4038
- "status_label": "not supported",
4039
- "scored": false,
4040
  "proxy_scored": false,
4041
- "raw": null,
4042
- "raw_text": "n/a",
4043
- "normalized_score": null,
4044
  "metric_key": "micro_f1",
4045
- "source": null,
4046
  "scope": "multi_episode_128_metadata_baseline",
4047
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4048
  },
4049
  {
4050
  "task_number": 17,
@@ -4052,17 +4051,17 @@
4052
  "task_label": "Future Object-Set Forecasting",
4053
  "series_id": "metadata128_neural_mlp",
4054
  "method": "128ep Metadata NN",
4055
- "status": "not_supported_by_metadata_only_package",
4056
- "status_label": "not supported",
4057
- "scored": false,
4058
  "proxy_scored": false,
4059
- "raw": null,
4060
- "raw_text": "n/a",
4061
- "normalized_score": null,
4062
  "metric_key": "micro_f1",
4063
- "source": null,
4064
  "scope": "multi_episode_128_metadata_baseline",
4065
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4066
  },
4067
  {
4068
  "task_number": 17,
@@ -4160,17 +4159,17 @@
4160
  "task_label": "IMU-to-Hand Pose Reconstruction",
4161
  "series_id": "metadata128_simple",
4162
  "method": "128ep Metadata Simple",
4163
- "status": "not_supported_by_metadata_only_package",
4164
- "status_label": "not supported",
4165
  "scored": false,
4166
  "proxy_scored": false,
4167
  "raw": null,
4168
  "raw_text": "n/a",
4169
  "normalized_score": null,
4170
  "metric_key": "mae",
4171
- "source": null,
4172
  "scope": "multi_episode_128_metadata_baseline",
4173
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4174
  },
4175
  {
4176
  "task_number": 18,
@@ -4286,17 +4285,17 @@
4286
  "task_label": "Camera-View Synchronization Retrieval",
4287
  "series_id": "metadata128_simple",
4288
  "method": "128ep Metadata Simple",
4289
- "status": "not_supported_by_metadata_only_package",
4290
- "status_label": "not supported",
4291
  "scored": false,
4292
  "proxy_scored": false,
4293
  "raw": null,
4294
  "raw_text": "n/a",
4295
  "normalized_score": null,
4296
  "metric_key": "mrr",
4297
- "source": null,
4298
  "scope": "multi_episode_128_metadata_baseline",
4299
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4300
  },
4301
  {
4302
  "task_number": 19,
@@ -4412,17 +4411,17 @@
4412
  "task_label": "Time-to-Next-Transition Regression",
4413
  "series_id": "metadata128_simple",
4414
  "method": "128ep Metadata Simple",
4415
- "status": "not_supported_by_metadata_only_package",
4416
- "status_label": "not supported",
4417
- "scored": false,
4418
  "proxy_scored": false,
4419
- "raw": null,
4420
- "raw_text": "n/a",
4421
- "normalized_score": null,
4422
  "metric_key": "mae",
4423
- "source": null,
4424
  "scope": "multi_episode_128_metadata_baseline",
4425
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4426
  },
4427
  {
4428
  "task_number": 20,
@@ -4430,17 +4429,17 @@
4430
  "task_label": "Time-to-Next-Transition Regression",
4431
  "series_id": "metadata128_neural_mlp",
4432
  "method": "128ep Metadata NN",
4433
- "status": "not_supported_by_metadata_only_package",
4434
- "status_label": "not supported",
4435
- "scored": false,
4436
  "proxy_scored": false,
4437
- "raw": null,
4438
- "raw_text": "n/a",
4439
- "normalized_score": null,
4440
  "metric_key": "mae",
4441
- "source": null,
4442
  "scope": "multi_episode_128_metadata_baseline",
4443
- "reason": "the 128-episode metadata/text rerun did not produce this task target; raw sensor blocks or a task-specific metadata target builder are required"
4444
  },
4445
  {
4446
  "task_number": 20,
 
1
  {
2
  "title": "128-Episode 20-Task Radar",
3
  "status": "pass",
4
+ "generated_at_utc": "2026-06-18T12:07:15+00:00",
5
  "description": "Selected 128-episode metadata/raw baselines plus verified Qwen3/Cosmos branches. Every method has 20 records; numeric scores appear only where the public artifact produced that task target.",
6
  "task_count": 20,
7
  "method_count": 7,
8
  "method_task_record_count": 140,
9
+ "scored_method_task_count": 93,
10
  "normalization_policy": {
11
  "higher_is_better": "bounded metrics are plotted directly on 0-1 axes after clipping to [0, 1]",
12
  "lower_is_better": "lower-error metrics are converted to best_observed_value / raw_value within the same task",
 
30
  "method_detail": "128-episode JSONL metadata/text simple baselines.",
31
  "plotted_as": "colored point overlay",
32
  "result_record_count": 20,
33
+ "scored_task_count": 13,
34
+ "covered_task_count": 13,
35
  "proxy_scored_task_count": 0,
36
+ "scoreless_task_count": 7,
37
+ "unsupported_task_count": 7,
38
  "not_evaluated_task_count": 0,
39
  "status_counts": {
40
+ "scored": 13,
41
+ "unsupported_without_required_target": 7
 
42
  },
43
+ "coverage_fraction": 0.65,
44
  "result_record_fraction": 1.0
45
  },
46
  {
 
54
  "method_detail": "128-episode JSONL metadata/text MLP baselines.",
55
  "plotted_as": "colored point overlay",
56
  "result_record_count": 20,
57
+ "scored_task_count": 13,
58
+ "covered_task_count": 13,
59
  "proxy_scored_task_count": 0,
60
+ "scoreless_task_count": 7,
61
+ "unsupported_task_count": 7,
62
  "not_evaluated_task_count": 0,
63
  "status_counts": {
64
+ "not_supported_by_metadata_only_package": 7,
65
+ "scored": 13
66
  },
67
+ "coverage_fraction": 0.65,
68
  "result_record_fraction": 1.0
69
  },
70
  {
 
1294
  "raw128_proxy_axis": false,
1295
  "values": {
1296
  "metadata128_simple": {
1297
+ "raw": 0.004579592783699693,
1298
  "metric_key": "macro_f1",
1299
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
1300
  "scope": "multi_episode_128_metadata_baseline",
1301
+ "status": "scored",
1302
+ "reason": null,
1303
+ "normalized_score": 0.004579592783699693,
1304
+ "raw_text": "0.0046",
1305
+ "status_label": "scored"
1306
  },
1307
  "metadata128_neural_mlp": {
1308
+ "raw": 0.0029821307969142615,
1309
  "metric_key": "macro_f1",
1310
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
1311
  "scope": "multi_episode_128_metadata_baseline",
1312
+ "status": "scored",
1313
+ "reason": null,
1314
+ "normalized_score": 0.0029821307969142615,
1315
+ "raw_text": "0.0030",
1316
+ "status_label": "scored"
1317
  },
1318
  "raw128_simple": {
1319
  "raw": 0.0024280172369056294,
 
1385
  "raw128_proxy_axis": false,
1386
  "values": {
1387
  "metadata128_simple": {
1388
+ "raw": 0.0001206030150753769,
1389
  "metric_key": "macro_f1",
1390
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
1391
  "scope": "multi_episode_128_metadata_baseline",
1392
+ "status": "scored",
1393
+ "reason": null,
1394
+ "normalized_score": 0.0001206030150753769,
1395
+ "raw_text": "0.0001",
1396
+ "status_label": "scored"
1397
  },
1398
  "metadata128_neural_mlp": {
1399
+ "raw": 2.086049543676662e-05,
1400
  "metric_key": "macro_f1",
1401
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
1402
  "scope": "multi_episode_128_metadata_baseline",
1403
+ "status": "scored",
1404
+ "reason": null,
1405
+ "normalized_score": 2.086049543676662e-05,
1406
+ "raw_text": "0.0000",
1407
+ "status_label": "scored"
1408
  },
1409
  "raw128_simple": {
1410
  "raw": 0.0,
 
1478
  "metadata128_simple": {
1479
  "raw": null,
1480
  "metric_key": "macro_f1",
1481
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
1482
  "scope": "multi_episode_128_metadata_baseline",
1483
+ "status": "unsupported_without_required_target",
1484
+ "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
1485
  "normalized_score": null,
1486
  "raw_text": "n/a",
1487
+ "status_label": "unsupported"
1488
  },
1489
  "metadata128_neural_mlp": {
1490
  "raw": null,
 
1567
  "raw128_proxy_axis": false,
1568
  "values": {
1569
  "metadata128_simple": {
1570
+ "raw": 0.0,
1571
  "metric_key": "macro_f1",
1572
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
1573
  "scope": "multi_episode_128_metadata_baseline",
1574
+ "status": "scored",
1575
+ "reason": null,
1576
+ "normalized_score": 0.0,
1577
+ "raw_text": "0.0000",
1578
+ "status_label": "scored"
1579
  },
1580
  "metadata128_neural_mlp": {
1581
+ "raw": 0.0,
1582
  "metric_key": "macro_f1",
1583
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
1584
  "scope": "multi_episode_128_metadata_baseline",
1585
+ "status": "scored",
1586
+ "reason": null,
1587
+ "normalized_score": 0.0,
1588
+ "raw_text": "0.0000",
1589
+ "status_label": "scored"
1590
  },
1591
  "raw128_simple": {
1592
  "raw": 0.0,
 
1658
  "raw128_proxy_axis": false,
1659
  "values": {
1660
  "metadata128_simple": {
1661
+ "raw": 0.17656983343047333,
1662
  "metric_key": "micro_f1",
1663
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
1664
  "scope": "multi_episode_128_metadata_baseline",
1665
+ "status": "scored",
1666
+ "reason": null,
1667
+ "normalized_score": 0.17656983343047333,
1668
+ "raw_text": "0.1766",
1669
+ "status_label": "scored"
1670
  },
1671
  "metadata128_neural_mlp": {
1672
+ "raw": 0.17418550827844048,
1673
  "metric_key": "micro_f1",
1674
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
1675
  "scope": "multi_episode_128_metadata_baseline",
1676
+ "status": "scored",
1677
+ "reason": null,
1678
+ "normalized_score": 0.17418550827844048,
1679
+ "raw_text": "0.1742",
1680
+ "status_label": "scored"
1681
  },
1682
  "raw128_simple": {
1683
  "raw": 0.06469493412657774,
 
1751
  "metadata128_simple": {
1752
  "raw": null,
1753
  "metric_key": "mae",
1754
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
1755
  "scope": "multi_episode_128_metadata_baseline",
1756
+ "status": "unsupported_without_required_target",
1757
+ "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
1758
  "normalized_score": null,
1759
  "raw_text": "n/a",
1760
+ "status_label": "unsupported"
1761
  },
1762
  "metadata128_neural_mlp": {
1763
  "raw": null,
 
1842
  "metadata128_simple": {
1843
  "raw": null,
1844
  "metric_key": "mrr",
1845
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
1846
  "scope": "multi_episode_128_metadata_baseline",
1847
+ "status": "unsupported_without_required_target",
1848
+ "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
1849
  "normalized_score": null,
1850
  "raw_text": "n/a",
1851
+ "status_label": "unsupported"
1852
  },
1853
  "metadata128_neural_mlp": {
1854
  "raw": null,
 
1931
  "raw128_proxy_axis": false,
1932
  "values": {
1933
  "metadata128_simple": {
1934
+ "raw": 624.8108520507812,
1935
  "metric_key": "mae",
1936
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
1937
  "scope": "multi_episode_128_metadata_baseline",
1938
+ "status": "scored",
1939
+ "reason": null,
1940
+ "normalized_score": 0.016864874132806403,
1941
+ "raw_text": "624.81",
1942
+ "status_label": "scored"
1943
  },
1944
  "metadata128_neural_mlp": {
1945
+ "raw": 41.4664421081543,
1946
  "metric_key": "mae",
1947
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
1948
  "scope": "multi_episode_128_metadata_baseline",
1949
+ "status": "scored",
1950
+ "reason": null,
1951
+ "normalized_score": 0.25411768748242325,
1952
+ "raw_text": "41.47",
1953
+ "status_label": "scored"
1954
  },
1955
  "raw128_simple": {
1956
  "raw": 52.32759475708008,
 
3529
  "task_label": "Long-Horizon Next-Action Forecasting",
3530
  "series_id": "metadata128_simple",
3531
  "method": "128ep Metadata Simple",
3532
+ "status": "scored",
3533
+ "status_label": "scored",
3534
+ "scored": true,
3535
  "proxy_scored": false,
3536
+ "raw": 0.004579592783699693,
3537
+ "raw_text": "0.0046",
3538
+ "normalized_score": 0.004579592783699693,
3539
  "metric_key": "macro_f1",
3540
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/long_horizon_next_action/metrics.json",
3541
  "scope": "multi_episode_128_metadata_baseline",
3542
+ "reason": null
3543
  },
3544
  {
3545
  "task_number": 13,
 
3547
  "task_label": "Long-Horizon Next-Action Forecasting",
3548
  "series_id": "metadata128_neural_mlp",
3549
  "method": "128ep Metadata NN",
3550
+ "status": "scored",
3551
+ "status_label": "scored",
3552
+ "scored": true,
3553
  "proxy_scored": false,
3554
+ "raw": 0.0029821307969142615,
3555
+ "raw_text": "0.0030",
3556
+ "normalized_score": 0.0029821307969142615,
3557
  "metric_key": "macro_f1",
3558
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/metrics.json",
3559
  "scope": "multi_episode_128_metadata_baseline",
3560
+ "reason": null
3561
  },
3562
  {
3563
  "task_number": 13,
 
3655
  "task_label": "Long-Horizon Next-Subtask Forecasting",
3656
  "series_id": "metadata128_simple",
3657
  "method": "128ep Metadata Simple",
3658
+ "status": "scored",
3659
+ "status_label": "scored",
3660
+ "scored": true,
3661
  "proxy_scored": false,
3662
+ "raw": 0.0001206030150753769,
3663
+ "raw_text": "0.0001",
3664
+ "normalized_score": 0.0001206030150753769,
3665
  "metric_key": "macro_f1",
3666
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/next_subtask_forecast/metrics.json",
3667
  "scope": "multi_episode_128_metadata_baseline",
3668
+ "reason": null
3669
  },
3670
  {
3671
  "task_number": 14,
 
3673
  "task_label": "Long-Horizon Next-Subtask Forecasting",
3674
  "series_id": "metadata128_neural_mlp",
3675
  "method": "128ep Metadata NN",
3676
+ "status": "scored",
3677
+ "status_label": "scored",
3678
+ "scored": true,
3679
  "proxy_scored": false,
3680
+ "raw": 2.086049543676662e-05,
3681
+ "raw_text": "0.0000",
3682
+ "normalized_score": 2.086049543676662e-05,
3683
  "metric_key": "macro_f1",
3684
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json",
3685
  "scope": "multi_episode_128_metadata_baseline",
3686
+ "reason": null
3687
  },
3688
  {
3689
  "task_number": 14,
 
3781
  "task_label": "Interaction Text Prediction",
3782
  "series_id": "metadata128_simple",
3783
  "method": "128ep Metadata Simple",
3784
+ "status": "unsupported_without_required_target",
3785
+ "status_label": "unsupported",
3786
  "scored": false,
3787
  "proxy_scored": false,
3788
  "raw": null,
3789
  "raw_text": "n/a",
3790
  "normalized_score": null,
3791
  "metric_key": "macro_f1",
3792
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/interaction_text_prediction/metrics.json",
3793
  "scope": "multi_episode_128_metadata_baseline",
3794
+ "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata"
3795
  },
3796
  {
3797
  "task_number": 15,
 
3907
  "task_label": "Action-Object Relation Prediction",
3908
  "series_id": "metadata128_simple",
3909
  "method": "128ep Metadata Simple",
3910
+ "status": "scored",
3911
+ "status_label": "scored",
3912
+ "scored": true,
3913
  "proxy_scored": false,
3914
+ "raw": 0.0,
3915
+ "raw_text": "0.0000",
3916
+ "normalized_score": 0.0,
3917
  "metric_key": "macro_f1",
3918
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/action_object_relation/metrics.json",
3919
  "scope": "multi_episode_128_metadata_baseline",
3920
+ "reason": null
3921
  },
3922
  {
3923
  "task_number": 16,
 
3925
  "task_label": "Action-Object Relation Prediction",
3926
  "series_id": "metadata128_neural_mlp",
3927
  "method": "128ep Metadata NN",
3928
+ "status": "scored",
3929
+ "status_label": "scored",
3930
+ "scored": true,
3931
  "proxy_scored": false,
3932
+ "raw": 0.0,
3933
+ "raw_text": "0.0000",
3934
+ "normalized_score": 0.0,
3935
  "metric_key": "macro_f1",
3936
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json",
3937
  "scope": "multi_episode_128_metadata_baseline",
3938
+ "reason": null
3939
  },
3940
  {
3941
  "task_number": 16,
 
4033
  "task_label": "Future Object-Set Forecasting",
4034
  "series_id": "metadata128_simple",
4035
  "method": "128ep Metadata Simple",
4036
+ "status": "scored",
4037
+ "status_label": "scored",
4038
+ "scored": true,
4039
  "proxy_scored": false,
4040
+ "raw": 0.17656983343047333,
4041
+ "raw_text": "0.1766",
4042
+ "normalized_score": 0.17656983343047333,
4043
  "metric_key": "micro_f1",
4044
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/object_set_forecast/metrics.json",
4045
  "scope": "multi_episode_128_metadata_baseline",
4046
+ "reason": null
4047
  },
4048
  {
4049
  "task_number": 17,
 
4051
  "task_label": "Future Object-Set Forecasting",
4052
  "series_id": "metadata128_neural_mlp",
4053
  "method": "128ep Metadata NN",
4054
+ "status": "scored",
4055
+ "status_label": "scored",
4056
+ "scored": true,
4057
  "proxy_scored": false,
4058
+ "raw": 0.17418550827844048,
4059
+ "raw_text": "0.1742",
4060
+ "normalized_score": 0.17418550827844048,
4061
  "metric_key": "micro_f1",
4062
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json",
4063
  "scope": "multi_episode_128_metadata_baseline",
4064
+ "reason": null
4065
  },
4066
  {
4067
  "task_number": 17,
 
4159
  "task_label": "IMU-to-Hand Pose Reconstruction",
4160
  "series_id": "metadata128_simple",
4161
  "method": "128ep Metadata Simple",
4162
+ "status": "unsupported_without_required_target",
4163
+ "status_label": "unsupported",
4164
  "scored": false,
4165
  "proxy_scored": false,
4166
  "raw": null,
4167
  "raw_text": "n/a",
4168
  "normalized_score": null,
4169
  "metric_key": "mae",
4170
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/imu_to_hand_pose/metrics.json",
4171
  "scope": "multi_episode_128_metadata_baseline",
4172
+ "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package"
4173
  },
4174
  {
4175
  "task_number": 18,
 
4285
  "task_label": "Camera-View Synchronization Retrieval",
4286
  "series_id": "metadata128_simple",
4287
  "method": "128ep Metadata Simple",
4288
+ "status": "unsupported_without_required_target",
4289
+ "status_label": "unsupported",
4290
  "scored": false,
4291
  "proxy_scored": false,
4292
  "raw": null,
4293
  "raw_text": "n/a",
4294
  "normalized_score": null,
4295
  "metric_key": "mrr",
4296
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/camera_view_sync_retrieval/metrics.json",
4297
  "scope": "multi_episode_128_metadata_baseline",
4298
+ "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package"
4299
  },
4300
  {
4301
  "task_number": 19,
 
4411
  "task_label": "Time-to-Next-Transition Regression",
4412
  "series_id": "metadata128_simple",
4413
  "method": "128ep Metadata Simple",
4414
+ "status": "scored",
4415
+ "status_label": "scored",
4416
+ "scored": true,
4417
  "proxy_scored": false,
4418
+ "raw": 624.8108520507812,
4419
+ "raw_text": "624.81",
4420
+ "normalized_score": 0.016864874132806403,
4421
  "metric_key": "mae",
4422
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/time_to_transition/metrics.json",
4423
  "scope": "multi_episode_128_metadata_baseline",
4424
+ "reason": null
4425
  },
4426
  {
4427
  "task_number": 20,
 
4429
  "task_label": "Time-to-Next-Transition Regression",
4430
  "series_id": "metadata128_neural_mlp",
4431
  "method": "128ep Metadata NN",
4432
+ "status": "scored",
4433
+ "status_label": "scored",
4434
+ "scored": true,
4435
  "proxy_scored": false,
4436
+ "raw": 41.4664421081543,
4437
+ "raw_text": "41.47",
4438
+ "normalized_score": 0.25411768748242325,
4439
  "metric_key": "mae",
4440
+ "source": "results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json",
4441
  "scope": "multi_episode_128_metadata_baseline",
4442
+ "reason": null
4443
  },
4444
  {
4445
  "task_number": 20,
data/mirror_parity.json CHANGED
The diff for this file is too large to render. See raw diff
 
data/public_surface_qa.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "title": "Ropedia Xperience-10M Public Project Surface",
3
  "status": "pass",
4
- "generated_at_utc": "2026-06-18T11:41:42+00:00",
5
  "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
6
  "checks": [
7
  {
@@ -18,7 +18,7 @@
18
  "website_integrity": {
19
  "exists": true,
20
  "status": "pass",
21
- "generated_at_utc": "2026-06-18T11:18:05+00:00"
22
  },
23
  "rendered_site_check": {
24
  "exists": true,
@@ -43,12 +43,12 @@
43
  "publication_package": {
44
  "exists": true,
45
  "status": "pass",
46
- "generated_at_utc": "2026-06-18T11:18:57+00:00"
47
  },
48
  "mirror_parity": {
49
  "exists": true,
50
  "status": "pass",
51
- "generated_at_utc": "2026-06-18T11:21:54+00:00"
52
  }
53
  },
54
  "failures": {}
 
1
  {
2
  "title": "Ropedia Xperience-10M Public Project Surface",
3
  "status": "pass",
4
+ "generated_at_utc": "2026-06-18T12:09:24+00:00",
5
  "scope": "Repo README, GitHub Pages HTML, Hugging Face Space card, artifact dataset card, and model card.",
6
  "checks": [
7
  {
 
18
  "website_integrity": {
19
  "exists": true,
20
  "status": "pass",
21
+ "generated_at_utc": "2026-06-18T11:41:43+00:00"
22
  },
23
  "rendered_site_check": {
24
  "exists": true,
 
43
  "publication_package": {
44
  "exists": true,
45
  "status": "pass",
46
+ "generated_at_utc": "2026-06-18T11:42:48+00:00"
47
  },
48
  "mirror_parity": {
49
  "exists": true,
50
  "status": "pass",
51
+ "generated_at_utc": "2026-06-18T11:43:59+00:00"
52
  }
53
  },
54
  "failures": {}
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/metrics.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "status": "pass",
3
+ "task": "action_object_relation",
4
+ "task_display_name": "Action Object Relation",
5
+ "model_family": "neural_mlp_metadata",
6
+ "source": "128_episode_qwen_jsonl_metadata",
7
+ "input_features": "frame/context metadata plus hashed prompt/options/main_task text; answer_json fields are excluded from inputs",
8
+ "split_policy": "train on train split, report held-out test split",
9
+ "num_train_windows": 25488,
10
+ "num_val_windows": 4569,
11
+ "num_test_windows": 4014,
12
+ "num_classes": 4149,
13
+ "num_train_classes": 3058,
14
+ "history": [
15
+ {
16
+ "epoch": 1,
17
+ "loss": 7.944652971050983,
18
+ "train_accuracy": 0.012986503452605148
19
+ },
20
+ {
21
+ "epoch": 7,
22
+ "loss": 0.9540771609645778,
23
+ "train_accuracy": 0.698642498430634
24
+ },
25
+ {
26
+ "epoch": 14,
27
+ "loss": 0.3690779313529279,
28
+ "train_accuracy": 0.8495370370370371
29
+ },
30
+ {
31
+ "epoch": 21,
32
+ "loss": 0.26653273193176186,
33
+ "train_accuracy": 0.888653483992467
34
+ },
35
+ {
36
+ "epoch": 28,
37
+ "loss": 0.20744486586222916,
38
+ "train_accuracy": 0.9080743879472694
39
+ },
40
+ {
41
+ "epoch": 35,
42
+ "loss": 0.18528367040651966,
43
+ "train_accuracy": 0.9144303201506592
44
+ }
45
+ ],
46
+ "device": "cuda",
47
+ "splits": {
48
+ "test": {
49
+ "accuracy": 0.0,
50
+ "balanced_accuracy": 0.0,
51
+ "macro_f1": 0.0,
52
+ "weighted_f1": 0.0,
53
+ "num_eval_windows": 4014,
54
+ "num_classes": 4149
55
+ }
56
+ },
57
+ "primary_metric": "macro_f1",
58
+ "primary_score": 0.0
59
+ }
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/per_class_metrics.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/action_object_relation/predictions.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/confusion_matrix.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/per_class_metrics.csv ADDED
@@ -0,0 +1,1212 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ class_id,class_name,support,predicted,precision,recall,f1
2
+ 0,Place jar on shelf,0,0,0.0,0.0,0.0
3
+ 1,Reach for item in box,0,0,0.0,0.0,0.0
4
+ 2,Pick up product from box,0,3,0.0,0.0,0.0
5
+ 3,Place product on shelf,0,288,0.0,0.0,0.0
6
+ 4,Pick up product from bin,0,0,0.0,0.0,0.0
7
+ 5,Reach for next product,0,5,0.0,0.0,0.0
8
+ 6,Place canned product on shelf,0,0,0.0,0.0,0.0
9
+ 7,Pick up canned product,0,15,0.0,0.0,0.0
10
+ 8,Arrange canned products on shelf,0,0,0.0,0.0,0.0
11
+ 9,Move bin to shelf area,0,0,0.0,0.0,0.0
12
+ 10,Place can on shelf,20,2,0.0,0.0,0.0
13
+ 11,Retract hand,0,0,0.0,0.0,0.0
14
+ 12,Hold and wipe product,0,0,0.0,0.0,0.0
15
+ 13,Wipe down shelf,0,0,0.0,0.0,0.0
16
+ 14,Pick up product,0,6,0.0,0.0,0.0
17
+ 15,Wipe product,0,1,0.0,0.0,0.0
18
+ 16,Wipe food product,0,1,0.0,0.0,0.0
19
+ 17,Reach into box,9,31,0.0,0.0,0.0
20
+ 18,Wipe jar,0,0,0.0,0.0,0.0
21
+ 19,Wipe shelf,0,0,0.0,0.0,0.0
22
+ 20,Hold pickle jar,0,0,0.0,0.0,0.0
23
+ 21,Release pickle jar,0,1,0.0,0.0,0.0
24
+ 22,Hold cleaning cloth,0,0,0.0,0.0,0.0
25
+ 23,Wipe the shelf,0,9,0.0,0.0,0.0
26
+ 24,Wipe the product jar,0,0,0.0,0.0,0.0
27
+ 25,Move to next section,0,0,0.0,0.0,0.0
28
+ 26,Place product in box,0,0,0.0,0.0,0.0
29
+ 27,Hold product,0,6,0.0,0.0,0.0
30
+ 28,Hold item and adjust posture,0,3,0.0,0.0,0.0
31
+ 29,Grasp product from box,0,0,0.0,0.0,0.0
32
+ 30,Prepare to place product,0,0,0.0,0.0,0.0
33
+ 31,Move product to shelf,0,0,0.0,0.0,0.0
34
+ 32,Move product to box,0,8,0.0,0.0,0.0
35
+ 33,Grasp next item,0,0,0.0,0.0,0.0
36
+ 34,Align button in row,0,0,0.0,0.0,0.0
37
+ 35,Pick up button,9,0,0.0,0.0,0.0
38
+ 36,Place button in row,0,0,0.0,0.0,0.0
39
+ 37,Align button,0,0,0.0,0.0,0.0
40
+ 38,Arrange button cluster,0,0,0.0,0.0,0.0
41
+ 39,Align button row,0,0,0.0,0.0,0.0
42
+ 40,Arrange buttons on table,0,0,0.0,0.0,0.0
43
+ 41,Adjust red button in row,0,0,0.0,0.0,0.0
44
+ 42,Select button from pile,0,0,0.0,0.0,0.0
45
+ 43,Adjust red button,0,0,0.0,0.0,0.0
46
+ 44,Pull back hand,0,0,0.0,0.0,0.0
47
+ 45,Survey the table,0,0,0.0,0.0,0.0
48
+ 46,Arrange black buttons,0,0,0.0,0.0,0.0
49
+ 47,Pick up black button,0,0,0.0,0.0,0.0
50
+ 48,Move black button,0,0,0.0,0.0,0.0
51
+ 49,Pick up red button,0,0,0.0,0.0,0.0
52
+ 50,Reach for multicolored buttons,0,0,0.0,0.0,0.0
53
+ 51,Arrange buttons in row,0,0,0.0,0.0,0.0
54
+ 52,Touch device,0,0,0.0,0.0,0.0
55
+ 53,Align red button in row,0,0,0.0,0.0,0.0
56
+ 54,Reach and sort buttons,0,0,0.0,0.0,0.0
57
+ 55,Reach for button,0,0,0.0,0.0,0.0
58
+ 56,Place button,31,0,0.0,0.0,0.0
59
+ 57,Cut cardboard piece,40,69,0.028985507246376812,0.05,0.03669724770642201
60
+ 58,Manipulate cardboard piece,0,0,0.0,0.0,0.0
61
+ 59,Position scissors to cut cardboard,0,0,0.0,0.0,0.0
62
+ 60,Cut cardboard,174,390,0.046153846153846156,0.10344827586206896,0.06382978723404256
63
+ 61,Release scissors,4,0,0.0,0.0,0.0
64
+ 62,Position scissors,0,0,0.0,0.0,0.0
65
+ 63,Pick up scissors,0,12,0.0,0.0,0.0
66
+ 64,Move away from workstation,0,10,0.0,0.0,0.0
67
+ 65,Walk through corridor,0,0,0.0,0.0,0.0
68
+ 66,Mark cardboard with ruler and pen,0,0,0.0,0.0,0.0
69
+ 67,Mark cardboard with pen and ruler,0,0,0.0,0.0,0.0
70
+ 68,Cut cardboard with scissors,0,2,0.0,0.0,0.0
71
+ 69,Place down scissors,0,4,0.0,0.0,0.0
72
+ 70,Position ruler on cardboard,0,0,0.0,0.0,0.0
73
+ 71,Hold ruler steady,0,0,0.0,0.0,0.0
74
+ 72,Mark line on cardboard,0,0,0.0,0.0,0.0
75
+ 73,Move marker and ruler,0,0,0.0,0.0,0.0
76
+ 74,Pick up smartphone,6,0,0.0,0.0,0.0
77
+ 75,Mark cardboard,0,13,0.0,0.0,0.0
78
+ 76,Hold cardboard pieces,0,8,0.0,0.0,0.0
79
+ 77,Place cardboard,0,0,0.0,0.0,0.0
80
+ 78,Align ruler on cardboard,0,0,0.0,0.0,0.0
81
+ 79,Mark cardboard with pen,0,6,0.0,0.0,0.0
82
+ 80,Plug cable into portable charger,0,0,0.0,0.0,0.0
83
+ 81,Pick up portable charger,0,0,0.0,0.0,0.0
84
+ 82,Walk towards room,0,0,0.0,0.0,0.0
85
+ 83,Hold portable charger,0,0,0.0,0.0,0.0
86
+ 84,Place charger on table,0,0,0.0,0.0,0.0
87
+ 85,Hold charger and cable,0,0,0.0,0.0,0.0
88
+ 86,Manipulate power cable plug,0,0,0.0,0.0,0.0
89
+ 87,Insert plug into power adapter,0,0,0.0,0.0,0.0
90
+ 88,Hold power adapter,0,0,0.0,0.0,0.0
91
+ 89,Align charging cable,0,0,0.0,0.0,0.0
92
+ 90,Insert charging cable,0,0,0.0,0.0,0.0
93
+ 91,Observe desktop layout,0,0,0.0,0.0,0.0
94
+ 92,Retract camera/reposition view,0,0,0.0,0.0,0.0
95
+ 93,Manipulate paper strip,152,4,0.25,0.006578947368421052,0.012820512820512818
96
+ 94,Fold paper strip,0,0,0.0,0.0,0.0
97
+ 95,Pick up yellow paper strip,0,0,0.0,0.0,0.0
98
+ 96,Pick up phone,0,0,0.0,0.0,0.0
99
+ 97,Use phone,31,4,0.0,0.0,0.0
100
+ 98,Place phone on desk,0,31,0.0,0.0,0.0
101
+ 99,Interact with phone,0,0,0.0,0.0,0.0
102
+ 100,Place phone down,11,1,0.0,0.0,0.0
103
+ 101,Adjust paper strip,0,0,0.0,0.0,0.0
104
+ 102,Release folded paper,0,0,0.0,0.0,0.0
105
+ 103,Fold paper strip into knot,0,0,0.0,0.0,0.0
106
+ 104,Fold paper strip into lucky star,0,0,0.0,0.0,0.0
107
+ 105,Inflate paper star,0,0,0.0,0.0,0.0
108
+ 106,Fold purple paper strip,0,0,0.0,0.0,0.0
109
+ 107,Fold purple paper,0,2,0.0,0.0,0.0
110
+ 108,Hold and crease purple paper,0,0,0.0,0.0,0.0
111
+ 109,Release paper,0,0,0.0,0.0,0.0
112
+ 110,Reach for phone,0,0,0.0,0.0,0.0
113
+ 111,Retrieve paper strips,0,0,0.0,0.0,0.0
114
+ 112,Fold and organize paper strips,0,0,0.0,0.0,0.0
115
+ 113,Hold charger,0,0,0.0,0.0,0.0
116
+ 114,Separate cardboard piece,0,0,0.0,0.0,0.0
117
+ 115,Cut cardboard with utility knife,0,25,0.0,0.0,0.0
118
+ 116,Mark cardboard piece,85,25,0.08,0.023529411764705882,0.03636363636363636
119
+ 117,Retract hand from bag,0,0,0.0,0.0,0.0
120
+ 118,Open small case,0,12,0.0,0.0,0.0
121
+ 119,Measure cardboard with ruler,0,0,0.0,0.0,0.0
122
+ 120,Move smartphone,28,0,0.0,0.0,0.0
123
+ 121,Hold ruler on cardboard,0,0,0.0,0.0,0.0
124
+ 122,Remove ruler,0,0,0.0,0.0,0.0
125
+ 123,Walk towards table,0,12,0.0,0.0,0.0
126
+ 124,Approach desk,0,0,0.0,0.0,0.0
127
+ 125,Position hands for work,0,2,0.0,0.0,0.0
128
+ 126,Manipulate quilling strip,0,0,0.0,0.0,0.0
129
+ 127,Release quilling strip,0,6,0.0,0.0,0.0
130
+ 128,Release paper strip,35,16,0.0,0.0,0.0
131
+ 129,Begin rolling quilling strip,0,0,0.0,0.0,0.0
132
+ 130,Type on smartphone,0,0,0.0,0.0,0.0
133
+ 131,Use phone to check stock,0,0,0.0,0.0,0.0
134
+ 132,Grasp package,0,0,0.0,0.0,0.0
135
+ 133,Place item on shelf,23,111,0.0,0.0,0.0
136
+ 134,Observe shelf status,0,0,0.0,0.0,0.0
137
+ 135,Walk towards aisle,0,0,0.0,0.0,0.0
138
+ 136,Move towards shelf,0,57,0.0,0.0,0.0
139
+ 137,Adjust item on shelf,26,2,0.0,0.0,0.0
140
+ 138,Reach for another item,18,0,0.0,0.0,0.0
141
+ 139,Remove item from shelf,0,0,0.0,0.0,0.0
142
+ 140,Discard item into bin,0,0,0.0,0.0,0.0
143
+ 141,Sweep debris,0,0,0.0,0.0,0.0
144
+ 142,Touch shelf edge,0,0,0.0,0.0,0.0
145
+ 143,Reach for product,0,31,0.0,0.0,0.0
146
+ 144,Release label,0,0,0.0,0.0,0.0
147
+ 145,Remove shelf label,0,20,0.0,0.0,0.0
148
+ 146,Move along shelf,0,0,0.0,0.0,0.0
149
+ 147,Carry stool to next shelf,0,0,0.0,0.0,0.0
150
+ 148,Place stool on floor,0,0,0.0,0.0,0.0
151
+ 149,Observe shelf,0,0,0.0,0.0,0.0
152
+ 150,Walk towards next aisle,0,0,0.0,0.0,0.0
153
+ 151,Reach for product labels,0,1,0.0,0.0,0.0
154
+ 152,Hold product labels,0,0,0.0,0.0,0.0
155
+ 153,Examine labels,0,0,0.0,0.0,0.0
156
+ 154,Place sauce in container,0,0,0.0,0.0,0.0
157
+ 155,Pick up supplement bottle,0,1,0.0,0.0,0.0
158
+ 156,Hold supplement bottle,0,0,0.0,0.0,0.0
159
+ 157,Open supplement bottle,0,0,0.0,0.0,0.0
160
+ 158,Walk through store,0,0,0.0,0.0,0.0
161
+ 159,Reach for item on shelf,0,0,0.0,0.0,0.0
162
+ 160,Examine item,0,0,0.0,0.0,0.0
163
+ 161,Place item in container,0,0,0.0,0.0,0.0
164
+ 162,Pick up another item,0,0,0.0,0.0,0.0
165
+ 163,Pick up oil bottle,0,0,0.0,0.0,0.0
166
+ 164,Place oil in container,0,0,0.0,0.0,0.0
167
+ 165,Inspect supplement bottle,0,0,0.0,0.0,0.0
168
+ 166,Place supplement bottle in container,0,0,0.0,0.0,0.0
169
+ 167,Pick up spice jar,0,0,0.0,0.0,0.0
170
+ 168,Place spice jar in container,0,0,0.0,0.0,0.0
171
+ 169,Sort beads,16,16,0.0,0.0,0.0
172
+ 170,Adjust hand position,0,0,0.0,0.0,0.0
173
+ 171,Arrange star-shaped beads,0,0,0.0,0.0,0.0
174
+ 172,Move pen,0,0,0.0,0.0,0.0
175
+ 173,Sort beads by color,0,135,0.0,0.0,0.0
176
+ 174,Move towards table,0,0,0.0,0.0,0.0
177
+ 175,Observe room,0,0,0.0,0.0,0.0
178
+ 176,Check watch,0,0,0.0,0.0,0.0
179
+ 177,Prepare to sort beads,0,1,0.0,0.0,0.0
180
+ 178,Align ruler,0,0,0.0,0.0,0.0
181
+ 179,Adjust grip,0,0,0.0,0.0,0.0
182
+ 180,Move ruler,0,0,0.0,0.0,0.0
183
+ 181,Adjust ruler position,0,2,0.0,0.0,0.0
184
+ 182,Mark cardboard with marker,0,0,0.0,0.0,0.0
185
+ 183,Draw lines on cardboard,0,0,0.0,0.0,0.0
186
+ 184,Drawing lines on cardboard,0,0,0.0,0.0,0.0
187
+ 185,Reposition marker,0,0,0.0,0.0,0.0
188
+ 186,Mark lines with marker,0,0,0.0,0.0,0.0
189
+ 187,Position the ruler,0,0,0.0,0.0,0.0
190
+ 188,Stack cardboard pieces,0,0,0.0,0.0,0.0
191
+ 189,Walking in the workspace,0,0,0.0,0.0,0.0
192
+ 190,Insert charging cable into power bank,0,0,0.0,0.0,0.0
193
+ 191,Manipulate and inspect colorful pieces,0,0,0.0,0.0,0.0
194
+ 192,Manipulate colorful pieces,0,0,0.0,0.0,0.0
195
+ 193,Sort colorful pieces,0,0,0.0,0.0,0.0
196
+ 194,Hold power bank and cable,0,0,0.0,0.0,0.0
197
+ 195,Touch pieces in box,0,0,0.0,0.0,0.0
198
+ 196,Hold small white box,0,0,0.0,0.0,0.0
199
+ 197,Place white box on table,0,0,0.0,0.0,0.0
200
+ 198,Adjust smartphone and sort pieces,0,0,0.0,0.0,0.0
201
+ 199,Sort small colorful pieces,0,0,0.0,0.0,0.0
202
+ 200,Sorting colorful paper pieces,0,0,0.0,0.0,0.0
203
+ 201,Use phone to check instructions,0,3,0.0,0.0,0.0
204
+ 202,Trace pattern on cardboard,0,0,0.0,0.0,0.0
205
+ 203,Remove cardboard pattern,0,1,0.0,0.0,0.0
206
+ 204,Remove cardboard pattern piece,0,3,0.0,0.0,0.0
207
+ 205,Cut out cardboard pattern,0,1,0.0,0.0,0.0
208
+ 206,Cut cardboard pattern,0,16,0.0,0.0,0.0
209
+ 207,Adjust cardboard position,0,0,0.0,0.0,0.0
210
+ 208,Interact with smartphone screen,0,0,0.0,0.0,0.0
211
+ 209,Pick up metal ruler,0,0,0.0,0.0,0.0
212
+ 210,Pick up pen,8,0,0.0,0.0,0.0
213
+ 211,Move pen aside,0,0,0.0,0.0,0.0
214
+ 212,Reposition and cut,0,0,0.0,0.0,0.0
215
+ 213,Hold quilling paper,0,0,0.0,0.0,0.0
216
+ 214,Roll quilling paper,0,0,0.0,0.0,0.0
217
+ 215,Release paper coil,0,0,0.0,0.0,0.0
218
+ 216,Pick up paper strip,0,0,0.0,0.0,0.0
219
+ 217,Manipulate quilled paper strip,0,2,0.0,0.0,0.0
220
+ 218,Release and prepare new strip,0,57,0.0,0.0,0.0
221
+ 219,Manipulate small paper segment,0,2,0.0,0.0,0.0
222
+ 220,Place down paper segment,0,0,0.0,0.0,0.0
223
+ 221,Reach for paper strips,0,11,0.0,0.0,0.0
224
+ 222,Browse and interact with phone interface,0,0,0.0,0.0,0.0
225
+ 223,Interacting with phone screen,0,0,0.0,0.0,0.0
226
+ 224,Pick up light blue strip,0,10,0.0,0.0,0.0
227
+ 225,Inspect strip,0,14,0.0,0.0,0.0
228
+ 226,Manipulate light blue strip,0,0,0.0,0.0,0.0
229
+ 227,Cut cardboard tube,0,0,0.0,0.0,0.0
230
+ 228,Stacking cardboard pieces,0,0,0.0,0.0,0.0
231
+ 229,Moving hand,0,0,0.0,0.0,0.0
232
+ 230,Position cardboard for cutting,0,0,0.0,0.0,0.0
233
+ 231,Place cardboard piece,0,226,0.0,0.0,0.0
234
+ 232,Pick up cardboard piece,0,0,0.0,0.0,0.0
235
+ 233,Cut cardboard piece with scissors,0,0,0.0,0.0,0.0
236
+ 234,Release cardboard piece,0,0,0.0,0.0,0.0
237
+ 235,Walk across office,0,0,0.0,0.0,0.0
238
+ 236,Cut cardboard into triangles,0,0,0.0,0.0,0.0
239
+ 237,Cut cardboard shape,0,101,0.0,0.0,0.0
240
+ 238,Pick up cardboard cutout,0,0,0.0,0.0,0.0
241
+ 239,Walk with cardboard cutout,0,0,0.0,0.0,0.0
242
+ 240,Approach workstation,0,2,0.0,0.0,0.0
243
+ 241,Organize tools and materials,0,0,0.0,0.0,0.0
244
+ 242,Cut cardboard triangle,0,0,0.0,0.0,0.0
245
+ 243,Holding marker,0,0,0.0,0.0,0.0
246
+ 244,Lift utility knife,0,0,0.0,0.0,0.0
247
+ 245,Inspect cardboard piece,0,22,0.0,0.0,0.0
248
+ 246,Position cardboard piece,0,0,0.0,0.0,0.0
249
+ 247,Align scissors,0,0,0.0,0.0,0.0
250
+ 248,Cut cardboard strip,0,61,0.0,0.0,0.0
251
+ 249,Position cardboard strip,0,0,0.0,0.0,0.0
252
+ 250,Inspect cardboard strip,0,0,0.0,0.0,0.0
253
+ 251,Align cardboard piece,0,0,0.0,0.0,0.0
254
+ 252,Complete the cut,0,0,0.0,0.0,0.0
255
+ 253,Put down utility knife,0,1,0.0,0.0,0.0
256
+ 254,Fold cardboard,0,0,0.0,0.0,0.0
257
+ 255,Pick up utility knife,18,0,0.0,0.0,0.0
258
+ 256,Hold utility knife,0,0,0.0,0.0,0.0
259
+ 257,Pick up cardboard strip,0,10,0.0,0.0,0.0
260
+ 258,Place cardboard strip,0,0,0.0,0.0,0.0
261
+ 259,Place cans into box,0,0,0.0,0.0,0.0
262
+ 260,Arrange cans in box,0,0,0.0,0.0,0.0
263
+ 261,Reach for can,0,1,0.0,0.0,0.0
264
+ 262,Arrange cans on shelf,0,0,0.0,0.0,0.0
265
+ 263,Reach for additional items,0,0,0.0,0.0,0.0
266
+ 264,Reach for container,0,6,0.0,0.0,0.0
267
+ 265,Adjust position,0,0,0.0,0.0,0.0
268
+ 266,Prepare to pick up item,0,0,0.0,0.0,0.0
269
+ 267,Place container in bin,0,0,0.0,0.0,0.0
270
+ 268,Adjust cans in bin,0,0,0.0,0.0,0.0
271
+ 269,Hold and inspect can,0,0,0.0,0.0,0.0
272
+ 270,Adjust perspective,0,0,0.0,0.0,0.0
273
+ 271,Inspect shelf and organize stock,0,0,0.0,0.0,0.0
274
+ 272,Placing stock on shelf,0,0,0.0,0.0,0.0
275
+ 273,Hold small product bag,0,10,0.0,0.0,0.0
276
+ 274,Position shelving divider,0,0,0.0,0.0,0.0
277
+ 275,Move away from shelf,0,0,0.0,0.0,0.0
278
+ 276,Pick up container,0,55,0.0,0.0,0.0
279
+ 277,Pick up cleaning cloth,0,1,0.0,0.0,0.0
280
+ 278,Pick up product box,0,0,0.0,0.0,0.0
281
+ 279,Place box on shelf,0,0,0.0,0.0,0.0
282
+ 280,Reach for next item,9,5,0.0,0.0,0.0
283
+ 281,Place plush toy on shelf,0,0,0.0,0.0,0.0
284
+ 282,Adjust placement on shelf,0,0,0.0,0.0,0.0
285
+ 283,Move plush toy,0,0,0.0,0.0,0.0
286
+ 284,Reach for product on shelf,0,0,0.0,0.0,0.0
287
+ 285,Hold cardboard,0,0,0.0,0.0,0.0
288
+ 286,Arrange cardboard,0,0,0.0,0.0,0.0
289
+ 287,Walk with marker,0,0,0.0,0.0,0.0
290
+ 288,Pick up small object,0,0,0.0,0.0,0.0
291
+ 289,Walk across room,0,0,0.0,0.0,0.0
292
+ 290,Place cardboard square on stack,0,0,0.0,0.0,0.0
293
+ 291,Arrange cardboard squares,0,1,0.0,0.0,0.0
294
+ 292,Stacking cardboard squares,0,0,0.0,0.0,0.0
295
+ 293,Stacking cardboard square,0,0,0.0,0.0,0.0
296
+ 294,Stack cardboard square,0,0,0.0,0.0,0.0
297
+ 295,Stack cardboard squares,0,0,0.0,0.0,0.0
298
+ 296,Sorting paper stars,0,0,0.0,0.0,0.0
299
+ 297,Place star,0,0,0.0,0.0,0.0
300
+ 298,Sort paper star,0,1,0.0,0.0,0.0
301
+ 299,Sort paper stars,0,1,0.0,0.0,0.0
302
+ 300,Place paper star,0,0,0.0,0.0,0.0
303
+ 301,Walk away,0,0,0.0,0.0,0.0
304
+ 302,Open door,0,0,0.0,0.0,0.0
305
+ 303,Walk through doorway,0,0,0.0,0.0,0.0
306
+ 304,Pick up object,0,0,0.0,0.0,0.0
307
+ 305,Place item on table,0,0,0.0,0.0,0.0
308
+ 306,Move phone,24,0,0.0,0.0,0.0
309
+ 307,Sort and place paper star,0,0,0.0,0.0,0.0
310
+ 308,Hold cardboard strip,0,2,0.0,0.0,0.0
311
+ 309,Align cardboard strip,0,6,0.0,0.0,0.0
312
+ 310,Hold cardboard with ruler,0,8,0.0,0.0,0.0
313
+ 311,Move utility knife along ruler,0,0,0.0,0.0,0.0
314
+ 312,Slide utility knife along ruler,0,11,0.0,0.0,0.0
315
+ 313,Guide utility knife along ruler,0,0,0.0,0.0,0.0
316
+ 314,Draw line on cardboard,0,128,0.0,0.0,0.0
317
+ 315,Marking lines on cardboard,0,2,0.0,0.0,0.0
318
+ 316,Hold craft tool,0,7,0.0,0.0,0.0
319
+ 317,Approach table,0,2,0.0,0.0,0.0
320
+ 318,Place tool on table,0,16,0.0,0.0,0.0
321
+ 319,Move hand toward craft materials,0,3,0.0,0.0,0.0
322
+ 320,Manipulate paper strips,0,31,0.0,0.0,0.0
323
+ 321,Pick up blue paper strip,0,10,0.0,0.0,0.0
324
+ 322,Hold and bend paper strip,0,0,0.0,0.0,0.0
325
+ 323,Hold small object,0,1,0.0,0.0,0.0
326
+ 324,Move hand away from workspace,0,0,0.0,0.0,0.0
327
+ 325,Observe workspace,13,78,0.0,0.0,0.0
328
+ 326,Place puzzle piece,21,9,0.0,0.0,0.0
329
+ 327,Release puzzle piece,4,0,0.0,0.0,0.0
330
+ 328,Scan for next piece,0,48,0.0,0.0,0.0
331
+ 329,Positioning puzzle piece,0,0,0.0,0.0,0.0
332
+ 330,Manipulate puzzle pieces,35,82,0.024390243902439025,0.05714285714285714,0.03418803418803419
333
+ 331,Adjust puzzle piece,11,51,0.19607843137254902,0.9090909090909091,0.3225806451612903
334
+ 332,Adjusting puzzle piece,0,1,0.0,0.0,0.0
335
+ 333,Adjusting a puzzle piece,0,0,0.0,0.0,0.0
336
+ 334,Draw line along ruler,0,0,0.0,0.0,0.0
337
+ 335,Reposition ruler,0,12,0.0,0.0,0.0
338
+ 336,Hold ruler and pen steady,0,0,0.0,0.0,0.0
339
+ 337,Mark lines on cardboard,0,0,0.0,0.0,0.0
340
+ 338,Place marker down,0,0,0.0,0.0,0.0
341
+ 339,Walk across the room,0,0,0.0,0.0,0.0
342
+ 340,Approach packing area,0,0,0.0,0.0,0.0
343
+ 341,Pack beads into box,0,0,0.0,0.0,0.0
344
+ 342,Pick up beads,0,0,0.0,0.0,0.0
345
+ 343,Deposit beads into box,0,0,0.0,0.0,0.0
346
+ 344,Pick up cardboard tray,0,0,0.0,0.0,0.0
347
+ 345,Move tray towards packing area,0,0,0.0,0.0,0.0
348
+ 346,Position cardboard tray,0,0,0.0,0.0,0.0
349
+ 347,Cut light green fabric,0,0,0.0,0.0,0.0
350
+ 348,Continue cutting fabric,0,0,0.0,0.0,0.0
351
+ 349,Cut fabric with scissors,0,1,0.0,0.0,0.0
352
+ 350,Adjusting fabric for cutting,0,0,0.0,0.0,0.0
353
+ 351,Cutting fabric,0,0,0.0,0.0,0.0
354
+ 352,Mark fabric with pen,0,0,0.0,0.0,0.0
355
+ 353,Mark fabric,0,0,0.0,0.0,0.0
356
+ 354,Mark fabric with pen and ruler,0,0,0.0,0.0,0.0
357
+ 355,Carry cardboard piece,0,7,0.0,0.0,0.0
358
+ 356,Pick up electronic accessory from box,0,0,0.0,0.0,0.0
359
+ 357,Place accessory on shelf,0,0,0.0,0.0,0.0
360
+ 358,Pick up accessory,0,0,0.0,0.0,0.0
361
+ 359,Reach towards shelf,0,0,0.0,0.0,0.0
362
+ 360,Place accessory into box,0,0,0.0,0.0,0.0
363
+ 361,Pick up new electronic product,0,0,0.0,0.0,0.0
364
+ 362,Pick up electronic product,0,0,0.0,0.0,0.0
365
+ 363,Move hand back to box,0,0,0.0,0.0,0.0
366
+ 364,Move product towards shelf,0,0,0.0,0.0,0.0
367
+ 365,Walk with shopping bag,0,0,0.0,0.0,0.0
368
+ 366,Pick up item from box,0,0,0.0,0.0,0.0
369
+ 367,Move towards box,0,0,0.0,0.0,0.0
370
+ 368,Hold items,0,0,0.0,0.0,0.0
371
+ 369,Place items on shelf,0,2,0.0,0.0,0.0
372
+ 370,Move to box,0,1,0.0,0.0,0.0
373
+ 371,Move box to next position,0,0,0.0,0.0,0.0
374
+ 372,Hold snack package,0,3,0.0,0.0,0.0
375
+ 373,Place snack package on shelf,0,44,0.0,0.0,0.0
376
+ 374,Place snack package in box,0,0,0.0,0.0,0.0
377
+ 375,Hold snack packages,0,0,0.0,0.0,0.0
378
+ 376,Pick up snack packages,0,0,0.0,0.0,0.0
379
+ 377,Walk towards shelves,9,4,0.0,0.0,0.0
380
+ 378,Pick up snack package,0,0,0.0,0.0,0.0
381
+ 379,Organize snacks in box,0,0,0.0,0.0,0.0
382
+ 380,Adjust snack package,0,0,0.0,0.0,0.0
383
+ 381,Open cardboard box,0,0,0.0,0.0,0.0
384
+ 382,Remove cardboard flap,0,5,0.0,0.0,0.0
385
+ 383,Align plastic containers,0,0,0.0,0.0,0.0
386
+ 384,Reach for items,0,0,0.0,0.0,0.0
387
+ 385,Adjust containers on shelf,0,0,0.0,0.0,0.0
388
+ 386,Adjust container position,0,0,0.0,0.0,0.0
389
+ 387,Withdraw hand,0,0,0.0,0.0,0.0
390
+ 388,Place container on shelf,0,35,0.0,0.0,0.0
391
+ 389,Place item in shopping bag,0,95,0.0,0.0,0.0
392
+ 390,Grasp item,0,0,0.0,0.0,0.0
393
+ 391,Move item to bag,0,0,0.0,0.0,0.0
394
+ 392,Pick up plush toy,0,0,0.0,0.0,0.0
395
+ 393,Place plush toy into bag,0,0,0.0,0.0,0.0
396
+ 394,Grasp shopping bag,0,0,0.0,0.0,0.0
397
+ 395,Prepare to place item in bag,0,0,0.0,0.0,0.0
398
+ 396,Organize bag contents,0,0,0.0,0.0,0.0
399
+ 397,Grasp and retrieve item,0,0,0.0,0.0,0.0
400
+ 398,Place item into shopping bag,0,0,0.0,0.0,0.0
401
+ 399,Sort star-shaped beads,16,0,0.0,0.0,0.0
402
+ 400,Sort beads on the table,0,0,0.0,0.0,0.0
403
+ 401,Sort beads on table,0,0,0.0,0.0,0.0
404
+ 402,Hold instructional sign,0,0,0.0,0.0,0.0
405
+ 403,Pick up star-shaped bead,0,3,0.0,0.0,0.0
406
+ 404,Place bead on table,0,3,0.0,0.0,0.0
407
+ 405,Reposition sign and organize beads,0,0,0.0,0.0,0.0
408
+ 406,Reposition ruler and pen,0,0,0.0,0.0,0.0
409
+ 407,Reposition pen and prepare for next line,0,0,0.0,0.0,0.0
410
+ 408,Draw straight lines on cardboard,0,0,0.0,0.0,0.0
411
+ 409,Draw lines with ruler,0,0,0.0,0.0,0.0
412
+ 410,Sort origami stars,0,0,0.0,0.0,0.0
413
+ 411,Walk in hallway,0,2,0.0,0.0,0.0
414
+ 412,Reach for stars,0,0,0.0,0.0,0.0
415
+ 413,Walk towards desk,0,0,0.0,0.0,0.0
416
+ 414,Grasp origami stars,0,0,0.0,0.0,0.0
417
+ 415,Place stars in container,0,0,0.0,0.0,0.0
418
+ 416,Sort light blue origami stars,0,0,0.0,0.0,0.0
419
+ 417,Sort origami stars by color,0,0,0.0,0.0,0.0
420
+ 418,Move origami stars,0,0,0.0,0.0,0.0
421
+ 419,Put down scissors,0,0,0.0,0.0,0.0
422
+ 420,Use smartphone,70,2,0.0,0.0,0.0
423
+ 421,Pick up water bottle,0,0,0.0,0.0,0.0
424
+ 422,Hold water bottle,0,0,0.0,0.0,0.0
425
+ 423,Place water bottle on table,0,0,0.0,0.0,0.0
426
+ 424,Hold phone,0,0,0.0,0.0,0.0
427
+ 425,Hold and view phone,0,0,0.0,0.0,0.0
428
+ 426,Cut cardboard pieces with scissors,0,0,0.0,0.0,0.0
429
+ 427,Vacuum the carpet,0,0,0.0,0.0,0.0
430
+ 428,Push vacuum cleaner,0,0,0.0,0.0,0.0
431
+ 429,Adjust vacuum cleaner position,0,0,0.0,0.0,0.0
432
+ 430,Vacuuming carpet edge,0,0,0.0,0.0,0.0
433
+ 431,Vacuum edge of carpet,0,0,0.0,0.0,0.0
434
+ 432,Vacuuming carpet corner,0,0,0.0,0.0,0.0
435
+ 433,Vacuuming the carpet edge,0,0,0.0,0.0,0.0
436
+ 434,Move vacuum cleaner,0,0,0.0,0.0,0.0
437
+ 435,Vacuuming along the wall edge,0,0,0.0,0.0,0.0
438
+ 436,Fold paper strip into star,0,0,0.0,0.0,0.0
439
+ 437,Arrange Mahjong tiles,0,0,0.0,0.0,0.0
440
+ 438,Rearrange Mahjong tiles,0,203,0.0,0.0,0.0
441
+ 439,Adjust Mahjong tiles,0,0,0.0,0.0,0.0
442
+ 440,Reach for Mahjong tiles,0,73,0.0,0.0,0.0
443
+ 441,Rearrange Mahjong tile,0,2,0.0,0.0,0.0
444
+ 442,Adjust Mahjong tile,0,0,0.0,0.0,0.0
445
+ 443,Align Mahjong tiles,0,0,0.0,0.0,0.0
446
+ 444,Move Mahjong tile,0,0,0.0,0.0,0.0
447
+ 445,Realign Mahjong tiles,0,0,0.0,0.0,0.0
448
+ 446,Cut cardboard square,0,0,0.0,0.0,0.0
449
+ 447,Trim cardboard piece,0,0,0.0,0.0,0.0
450
+ 448,Pick up cereal boxes,0,0,0.0,0.0,0.0
451
+ 449,Carry cereal boxes,0,0,0.0,0.0,0.0
452
+ 450,Carry cereal towards aisle,0,1,0.0,0.0,0.0
453
+ 451,Carry pasta box towards aisle,0,0,0.0,0.0,0.0
454
+ 452,Pick up container from box,0,0,0.0,0.0,0.0
455
+ 453,Hold container,0,1,0.0,0.0,0.0
456
+ 454,Reach for items in box,0,1,0.0,0.0,0.0
457
+ 455,Pick up grocery item,0,0,0.0,0.0,0.0
458
+ 456,Carry item to shelf,0,0,0.0,0.0,0.0
459
+ 457,Move to stock products,0,0,0.0,0.0,0.0
460
+ 458,Wipe shelf surface,0,0,0.0,0.0,0.0
461
+ 459,Move cardboard box,0,5,0.0,0.0,0.0
462
+ 460,Place snack on shelf,0,0,0.0,0.0,0.0
463
+ 461,Retrieve snack from container,0,0,0.0,0.0,0.0
464
+ 462,Pick up gift box,0,0,0.0,0.0,0.0
465
+ 463,Pick up next gift box,0,0,0.0,0.0,0.0
466
+ 464,Pick up snack pouch,0,0,0.0,0.0,0.0
467
+ 465,Place snack pouch in container,0,0,0.0,0.0,0.0
468
+ 466,Reach for snack pouch,0,0,0.0,0.0,0.0
469
+ 467,Move storage bin,0,0,0.0,0.0,0.0
470
+ 468,Reach for shelf,0,1,0.0,0.0,0.0
471
+ 469,Hold bin and move through aisle,0,0,0.0,0.0,0.0
472
+ 470,Remove storage bin from shelf,0,0,0.0,0.0,0.0
473
+ 471,Reach for empty shelf space,0,0,0.0,0.0,0.0
474
+ 472,Grasp plastic bag on shelf,0,0,0.0,0.0,0.0
475
+ 473,Remove plastic container from shelf,0,0,0.0,0.0,0.0
476
+ 474,Arrange plastic containers,0,0,0.0,0.0,0.0
477
+ 475,Retrieve another container,0,0,0.0,0.0,0.0
478
+ 476,Arrange container on shelf,0,0,0.0,0.0,0.0
479
+ 477,Hold smartphone,42,4,0.0,0.0,0.0
480
+ 478,Place smartphone on desk,0,0,0.0,0.0,0.0
481
+ 479,Reach for water bottle,0,1,0.0,0.0,0.0
482
+ 480,Hold scissors,0,2,0.0,0.0,0.0
483
+ 481,Cut newspaper,0,10,0.0,0.0,0.0
484
+ 482,Continue cutting newspaper,0,0,0.0,0.0,0.0
485
+ 483,Place scissors on table,0,18,0.0,0.0,0.0
486
+ 484,Move scissors away,0,1,0.0,0.0,0.0
487
+ 485,Place scissors down,0,1,0.0,0.0,0.0
488
+ 486,Arrange tiles into row,0,0,0.0,0.0,0.0
489
+ 487,Adjust tile row alignment,0,0,0.0,0.0,0.0
490
+ 488,Adjust Mahjong tile alignment,0,0,0.0,0.0,0.0
491
+ 489,Adjust Mahjong tile on the stack,0,47,0.0,0.0,0.0
492
+ 490,Pick up Mahjong tile,0,0,0.0,0.0,0.0
493
+ 491,Place Mahjong tile on the stack,0,0,0.0,0.0,0.0
494
+ 492,Place Mahjong tile on stack,0,0,0.0,0.0,0.0
495
+ 493,Hold ruler and draw line,0,0,0.0,0.0,0.0
496
+ 494,Draw line,0,0,0.0,0.0,0.0
497
+ 495,Mark lines with pen along ruler,0,0,0.0,0.0,0.0
498
+ 496,Hold ruler and mark cardboard,0,0,0.0,0.0,0.0
499
+ 497,Hold ruler and marker,0,0,0.0,0.0,0.0
500
+ 498,Inspect charging case,0,0,0.0,0.0,0.0
501
+ 499,Place charging case down,0,0,0.0,0.0,0.0
502
+ 500,Hold paper strip,0,0,0.0,0.0,0.0
503
+ 501,Measure and mark cardboard,0,0,0.0,0.0,0.0
504
+ 502,Hold and align cardboard,0,0,0.0,0.0,0.0
505
+ 503,Position cardboard tube,0,0,0.0,0.0,0.0
506
+ 504,Cut cardboard strip with scissors,0,0,0.0,0.0,0.0
507
+ 505,Scroll on smartphone,0,0,0.0,0.0,0.0
508
+ 506,Tap smartphone screen,0,0,0.0,0.0,0.0
509
+ 507,Scroll through photo gallery,0,0,0.0,0.0,0.0
510
+ 508,Typing message on smartphone,0,0,0.0,0.0,0.0
511
+ 509,Typing on smartphone,0,0,0.0,0.0,0.0
512
+ 510,Tapping smartphone screen,0,0,0.0,0.0,0.0
513
+ 511,Putting away smartphone,0,0,0.0,0.0,0.0
514
+ 512,Stop measuring and put down tools,0,0,0.0,0.0,0.0
515
+ 513,Draw line with pen,0,0,0.0,0.0,0.0
516
+ 514,Prepare to draw lines,0,0,0.0,0.0,0.0
517
+ 515,Remove ruler and marker,0,0,0.0,0.0,0.0
518
+ 516,Align ruler and mark cardboard,0,0,0.0,0.0,0.0
519
+ 517,Walking through classroom,0,1,0.0,0.0,0.0
520
+ 518,Assemble cardboard pieces,0,1,0.0,0.0,0.0
521
+ 519,Move marker away,0,2,0.0,0.0,0.0
522
+ 520,Arrange cardboard piece,0,0,0.0,0.0,0.0
523
+ 521,Position ruler and mark cardboard,0,0,0.0,0.0,0.0
524
+ 522,Place canned good on shelf,0,0,0.0,0.0,0.0
525
+ 523,Move canned goods container,0,0,0.0,0.0,0.0
526
+ 524,Position container near shelf,0,0,0.0,0.0,0.0
527
+ 525,Adjust container on shelf,0,0,0.0,0.0,0.0
528
+ 526,Pick up canned food,13,0,0.0,0.0,0.0
529
+ 527,Place canned food in container,0,0,0.0,0.0,0.0
530
+ 528,Adjust cans in container,0,0,0.0,0.0,0.0
531
+ 529,Adjust cans in tray,0,0,0.0,0.0,0.0
532
+ 530,Adjusting canned goods on shelf,0,0,0.0,0.0,0.0
533
+ 531,Align canned goods on shelf,0,0,0.0,0.0,0.0
534
+ 532,Place canned food on shelf,69,0,0.0,0.0,0.0
535
+ 533,Reach for next canned food item,0,0,0.0,0.0,0.0
536
+ 534,Wipe the plastic jar,0,5,0.0,0.0,0.0
537
+ 535,Finish wiping and inspect jar,0,8,0.0,0.0,0.0
538
+ 536,Inspect jar,0,109,0.0,0.0,0.0
539
+ 537,Pick up tin can,0,0,0.0,0.0,0.0
540
+ 538,Hold items and inspect shelf,0,72,0.0,0.0,0.0
541
+ 539,Move cardboard,0,0,0.0,0.0,0.0
542
+ 540,Stabilize cardboard,0,0,0.0,0.0,0.0
543
+ 541,Stabilize ruler,0,0,0.0,0.0,0.0
544
+ 542,Labeling cardboard squares,0,0,0.0,0.0,0.0
545
+ 543,Moving cardboard square,0,0,0.0,0.0,0.0
546
+ 544,Labeling cardboard square,0,0,0.0,0.0,0.0
547
+ 545,Starting to label next square,0,0,0.0,0.0,0.0
548
+ 546,Placing labeled cardboard square,0,0,0.0,0.0,0.0
549
+ 547,Labeling cardboard piece,0,2,0.0,0.0,0.0
550
+ 548,Move cardboard piece,0,0,0.0,0.0,0.0
551
+ 549,Reach for next piece,0,0,0.0,0.0,0.0
552
+ 550,Marking cardboard with pen,0,0,0.0,0.0,0.0
553
+ 551,Repositioning ruler and cardboard,0,2,0.0,0.0,0.0
554
+ 552,Folding cardboard,0,0,0.0,0.0,0.0
555
+ 553,Place cardboard piece on stack,0,2,0.0,0.0,0.0
556
+ 554,Arrange buttons on the table,0,14,0.0,0.0,0.0
557
+ 555,Arrange buttons,33,2,0.5,0.030303030303030304,0.05714285714285715
558
+ 556,Sorting buttons,0,0,0.0,0.0,0.0
559
+ 557,Sort orange buttons,0,16,0.0,0.0,0.0
560
+ 558,Sort orange button,0,0,0.0,0.0,0.0
561
+ 559,Move hand over button pile,0,0,0.0,0.0,0.0
562
+ 560,Move orange buttons,0,11,0.0,0.0,0.0
563
+ 561,Arrange orange buttons,0,1,0.0,0.0,0.0
564
+ 562,Pick up stapler,0,0,0.0,0.0,0.0
565
+ 563,Drawing grid line with ruler,0,0,0.0,0.0,0.0
566
+ 564,Drawing grid line with pen and ruler,0,0,0.0,0.0,0.0
567
+ 565,Draw grid line with pen,0,0,0.0,0.0,0.0
568
+ 566,Pick up cardboard,0,0,0.0,0.0,0.0
569
+ 567,Draw grid line,0,2,0.0,0.0,0.0
570
+ 568,Drawing grid line,0,0,0.0,0.0,0.0
571
+ 569,Manipulate paper star,0,10,0.0,0.0,0.0
572
+ 570,Fold paper star,0,6,0.0,0.0,0.0
573
+ 571,Reach for beads,0,45,0.0,0.0,0.0
574
+ 572,Sort purple beads,0,0,0.0,0.0,0.0
575
+ 573,Write on paper,13,1,0.0,0.0,0.0
576
+ 574,Gathering star beads,0,0,0.0,0.0,0.0
577
+ 575,Sort beads by hand,0,0,0.0,0.0,0.0
578
+ 576,Pick up can,3,0,0.0,0.0,0.0
579
+ 577,Hold tray of canned goods,0,0,0.0,0.0,0.0
580
+ 578,Position tray,0,0,0.0,0.0,0.0
581
+ 579,Sort canned goods in tray,0,0,0.0,0.0,0.0
582
+ 580,Carry crate of cans,0,0,0.0,0.0,0.0
583
+ 581,Move can towards shelf,0,0,0.0,0.0,0.0
584
+ 582,Wipe item,0,0,0.0,0.0,0.0
585
+ 583,Place item back,0,0,0.0,0.0,0.0
586
+ 584,Wipe retail item,0,0,0.0,0.0,0.0
587
+ 585,Reach for retail item,0,0,0.0,0.0,0.0
588
+ 586,Grasp retail item,0,0,0.0,0.0,0.0
589
+ 587,Adjust retail items on shelf,0,0,0.0,0.0,0.0
590
+ 588,Align and place retail item,0,22,0.0,0.0,0.0
591
+ 589,Arrange items on shelf,0,0,0.0,0.0,0.0
592
+ 590,Pick up pink water bottle,0,0,0.0,0.0,0.0
593
+ 591,Place down pink water bottle,0,0,0.0,0.0,0.0
594
+ 592,Place star in row,0,0,0.0,0.0,0.0
595
+ 593,Reach for star,0,0,0.0,0.0,0.0
596
+ 594,Retrieve star,0,0,0.0,0.0,0.0
597
+ 595,Pick up star,0,0,0.0,0.0,0.0
598
+ 596,Hold recording sheet and pen,0,0,0.0,0.0,0.0
599
+ 597,Record star count,0,0,0.0,0.0,0.0
600
+ 598,Hold pen and paper,0,0,0.0,0.0,0.0
601
+ 599,Observe paper and count objects,0,0,0.0,0.0,0.0
602
+ 600,Write count on paper,17,0,0.0,0.0,0.0
603
+ 601,Place pen on table,0,0,0.0,0.0,0.0
604
+ 602,View content on smartphone,0,0,0.0,0.0,0.0
605
+ 603,Resume writing on paper,0,0,0.0,0.0,0.0
606
+ 604,Pick up paper star,0,0,0.0,0.0,0.0
607
+ 605,Place paper star in row,0,0,0.0,0.0,0.0
608
+ 606,Manipulate star,0,0,0.0,0.0,0.0
609
+ 607,Arrange paper stars,0,0,0.0,0.0,0.0
610
+ 608,Cut cardboard grid,0,0,0.0,0.0,0.0
611
+ 609,Pick up small item,0,0,0.0,0.0,0.0
612
+ 610,Walking to sink,0,0,0.0,0.0,0.0
613
+ 611,Washing hands,0,0,0.0,0.0,0.0
614
+ 612,Finish washing hands,0,0,0.0,0.0,0.0
615
+ 613,Pick up paper towel,0,0,0.0,0.0,0.0
616
+ 614,Dry hands,0,0,0.0,0.0,0.0
617
+ 615,Begin folding paper strip,0,2,0.0,0.0,0.0
618
+ 616,Fold paper strip into a star,0,0,0.0,0.0,0.0
619
+ 617,Prepare paper strip,0,3,0.0,0.0,0.0
620
+ 618,Continue folding paper strip,0,50,0.0,0.0,0.0
621
+ 619,Fold lucky star,0,0,0.0,0.0,0.0
622
+ 620,Manipulate folded paper star,0,1,0.0,0.0,0.0
623
+ 621,Grasp paper strip,0,0,0.0,0.0,0.0
624
+ 622,Sort colored tiles,0,16,0.0,0.0,0.0
625
+ 623,Pick up colored tile,0,6,0.0,0.0,0.0
626
+ 624,Place colored tile,0,0,0.0,0.0,0.0
627
+ 625,Sort tiles,0,24,0.0,0.0,0.0
628
+ 626,Sort tiles by color,0,8,0.0,0.0,0.0
629
+ 627,Write on notepad,0,32,0.0,0.0,0.0
630
+ 628,Writing on notepad,0,1,0.0,0.0,0.0
631
+ 629,Reaching for beads,0,1,0.0,0.0,0.0
632
+ 630,Cut section from newspaper,0,0,0.0,0.0,0.0
633
+ 631,Tear newspaper,0,0,0.0,0.0,0.0
634
+ 632,Hold newspaper,0,1,0.0,0.0,0.0
635
+ 633,Hold and align newspaper,0,0,0.0,0.0,0.0
636
+ 634,Fold newspaper,0,7,0.0,0.0,0.0
637
+ 635,Reposition newspaper,0,0,0.0,0.0,0.0
638
+ 636,Cut along the edge of the newspaper,0,0,0.0,0.0,0.0
639
+ 637,Cut along the newspaper edge,0,1,0.0,0.0,0.0
640
+ 638,Browsing mobile phone,0,0,0.0,0.0,0.0
641
+ 639,Browse mobile phone,0,0,0.0,0.0,0.0
642
+ 640,Cut newspaper with scissors,0,0,0.0,0.0,0.0
643
+ 641,Sort blue star-shaped pieces,0,0,0.0,0.0,0.0
644
+ 642,Sort small plastic pieces,0,0,0.0,0.0,0.0
645
+ 643,Reach for more pieces,0,0,0.0,0.0,0.0
646
+ 644,Sort plastic pieces,0,0,0.0,0.0,0.0
647
+ 645,Move pieces into box,0,0,0.0,0.0,0.0
648
+ 646,Gather pieces into box,0,0,0.0,0.0,0.0
649
+ 647,Typing on phone,0,0,0.0,0.0,0.0
650
+ 648,Scrolling and viewing content on phone,0,0,0.0,0.0,0.0
651
+ 649,Pick up item from shelf,0,2,0.0,0.0,0.0
652
+ 650,Pick up charging cable,0,0,0.0,0.0,0.0
653
+ 651,Pick up electronic item,0,99,0.0,0.0,0.0
654
+ 652,Wipe electronic item,0,0,0.0,0.0,0.0
655
+ 653,Place item in bag,0,0,0.0,0.0,0.0
656
+ 654,Inspect smartphone box,0,0,0.0,0.0,0.0
657
+ 655,Hold smartphone box,0,0,0.0,0.0,0.0
658
+ 656,Examine product,0,1,0.0,0.0,0.0
659
+ 657,Pick up another canned item,0,0,0.0,0.0,0.0
660
+ 658,Carry plastic container,0,1,0.0,0.0,0.0
661
+ 659,Reach for another container,0,33,0.0,0.0,0.0
662
+ 660,Release container,0,3,0.0,0.0,0.0
663
+ 661,Pick up storage container,0,0,0.0,0.0,0.0
664
+ 662,Move container toward shelf,0,1,0.0,0.0,0.0
665
+ 663,Position container on shelf,0,0,0.0,0.0,0.0
666
+ 664,Remove lid from container,0,1,0.0,0.0,0.0
667
+ 665,Pick up canned goods,0,0,0.0,0.0,0.0
668
+ 666,Place canned goods in container,0,5,0.0,0.0,0.0
669
+ 667,Pick up next product from bin,0,0,0.0,0.0,0.0
670
+ 668,Move bin,0,13,0.0,0.0,0.0
671
+ 669,Walking along the aisle,0,0,0.0,0.0,0.0
672
+ 670,Move plastic storage bin,0,0,0.0,0.0,0.0
673
+ 671,Place canned food in bin,0,0,0.0,0.0,0.0
674
+ 672,Hold container of canned food,0,0,0.0,0.0,0.0
675
+ 673,Move towards aisle,0,0,0.0,0.0,0.0
676
+ 674,Approach restocking supplies,0,0,0.0,0.0,0.0
677
+ 675,Pick up plastic container,0,1,0.0,0.0,0.0
678
+ 676,Move along the shelves,0,0,0.0,0.0,0.0
679
+ 677,Forming quilled paper shape,0,0,0.0,0.0,0.0
680
+ 678,Manipulate quilled paper shape,0,0,0.0,0.0,0.0
681
+ 679,Place quilled paper shape,0,0,0.0,0.0,0.0
682
+ 680,Retrieve paper strip,0,0,0.0,0.0,0.0
683
+ 681,Select paper strip,0,0,0.0,0.0,0.0
684
+ 682,Transition to standing position,0,0,0.0,0.0,0.0
685
+ 683,Observe paper quilling station,0,0,0.0,0.0,0.0
686
+ 684,Sort quilled paper pieces,0,0,0.0,0.0,0.0
687
+ 685,Walk towards storage area,0,0,0.0,0.0,0.0
688
+ 686,Hold device and cable,0,0,0.0,0.0,0.0
689
+ 687,Move piece to pile,0,0,0.0,0.0,0.0
690
+ 688,Manipulate quilled paper,0,0,0.0,0.0,0.0
691
+ 689,Pick up and sort cardboard,0,1,0.0,0.0,0.0
692
+ 690,Sort and arrange cardboard pieces,0,5,0.0,0.0,0.0
693
+ 691,Move camera over surface,0,9,0.0,0.0,0.0
694
+ 692,Observe sorting progress,0,0,0.0,0.0,0.0
695
+ 693,Reach for cardboard piece,0,0,0.0,0.0,0.0
696
+ 694,Lock phone,0,4,0.0,0.0,0.0
697
+ 695,Sort and stack cardboard pieces,0,0,0.0,0.0,0.0
698
+ 696,Mark list with pen,0,0,0.0,0.0,0.0
699
+ 697,Adjust bead piles,0,0,0.0,0.0,0.0
700
+ 698,Sort blue beads,0,1,0.0,0.0,0.0
701
+ 699,Place down pen,0,6,0.0,0.0,0.0
702
+ 700,Move away from desk,0,0,0.0,0.0,0.0
703
+ 701,Walking through the office,0,0,0.0,0.0,0.0
704
+ 702,Resume sorting blue beads,0,2,0.0,0.0,0.0
705
+ 703,Fold cardboard shape,0,0,0.0,0.0,0.0
706
+ 704,Reach for cardboard box,0,5,0.0,0.0,0.0
707
+ 705,Reach for object,0,26,0.0,0.0,0.0
708
+ 706,Release cardboard shape,0,20,0.0,0.0,0.0
709
+ 707,Reposition hands,0,0,0.0,0.0,0.0
710
+ 708,Rolling paper strip,0,0,0.0,0.0,0.0
711
+ 709,Finishing coil,0,0,0.0,0.0,0.0
712
+ 710,Start folding paper strip,0,0,0.0,0.0,0.0
713
+ 711,Folding paper strip,0,0,0.0,0.0,0.0
714
+ 712,Positioning paper strip,0,0,0.0,0.0,0.0
715
+ 713,Manipulate quilling paper,0,2,0.0,0.0,0.0
716
+ 714,Walk towards workspace,0,0,0.0,0.0,0.0
717
+ 715,Interaction with coworker,0,22,0.0,0.0,0.0
718
+ 716,Walk through workspace,0,9,0.0,0.0,0.0
719
+ 717,Manipulate small object,0,0,0.0,0.0,0.0
720
+ 718,Manipulate paper quilling piece,0,0,0.0,0.0,0.0
721
+ 719,Hold quilled paper piece,0,0,0.0,0.0,0.0
722
+ 720,Pull paper strip,0,0,0.0,0.0,0.0
723
+ 721,Hold and align paper strip,0,9,0.0,0.0,0.0
724
+ 722,Hold and rotate paper strip,0,0,0.0,0.0,0.0
725
+ 723,Marking cardboard piece,30,1,0.0,0.0,0.0
726
+ 724,Hold and mark cardboard piece,0,0,0.0,0.0,0.0
727
+ 725,Organize cardboard pieces,15,0,0.0,0.0,0.0
728
+ 726,Walking towards workstation,0,0,0.0,0.0,0.0
729
+ 727,Move to desk,0,0,0.0,0.0,0.0
730
+ 728,Sort small objects,0,0,0.0,0.0,0.0
731
+ 729,Gathering items,0,0,0.0,0.0,0.0
732
+ 730,Place items on table,0,0,0.0,0.0,0.0
733
+ 731,Gathering colored beads,0,0,0.0,0.0,0.0
734
+ 732,Arrange beads by color,0,0,0.0,0.0,0.0
735
+ 733,Sort star-shaped objects by color,0,1,0.0,0.0,0.0
736
+ 734,Sort star-shaped objects,0,0,0.0,0.0,0.0
737
+ 735,Sort yellow star-shaped objects,0,0,0.0,0.0,0.0
738
+ 736,Sort purple star-shaped objects,0,0,0.0,0.0,0.0
739
+ 737,View phone screen,0,0,0.0,0.0,0.0
740
+ 738,Viewing phone screen,0,0,0.0,0.0,0.0
741
+ 739,Initiate star folding,0,0,0.0,0.0,0.0
742
+ 740,Reach for next canned product,0,0,0.0,0.0,0.0
743
+ 741,Place jar in box,0,0,0.0,0.0,0.0
744
+ 742,Place pickle jar in box,0,0,0.0,0.0,0.0
745
+ 743,Grasp product from shelf,0,0,0.0,0.0,0.0
746
+ 744,Place red button,0,0,0.0,0.0,0.0
747
+ 745,Move and place black buttons,0,0,0.0,0.0,0.0
748
+ 746,Arrange red buttons,0,0,0.0,0.0,0.0
749
+ 747,Adjust red button position,0,0,0.0,0.0,0.0
750
+ 748,Withdraw hand from buttons,0,0,0.0,0.0,0.0
751
+ 749,Arrive at a different workstation,0,0,0.0,0.0,0.0
752
+ 750,Move vacuum cleaner hose,0,1,0.0,0.0,0.0
753
+ 751,Place smartphone on cardboard,0,0,0.0,0.0,0.0
754
+ 752,Reach into bag,0,0,0.0,0.0,0.0
755
+ 753,Organize products,0,0,0.0,0.0,0.0
756
+ 754,Close cardboard box,0,0,0.0,0.0,0.0
757
+ 755,Pick up item,0,3,0.0,0.0,0.0
758
+ 756,Stand up and walk away,0,0,0.0,0.0,0.0
759
+ 757,Interact with colleagues,0,0,0.0,0.0,0.0
760
+ 758,Moving hand towards cardboard stack,0,0,0.0,0.0,0.0
761
+ 759,Put down water bottle,0,0,0.0,0.0,0.0
762
+ 760,Placing piece on stack,0,0,0.0,0.0,0.0
763
+ 761,Reach for and pick up smartphone,0,0,0.0,0.0,0.0
764
+ 762,Move cardboard to pile,0,0,0.0,0.0,0.0
765
+ 763,Fold cardboard sheet,0,0,0.0,0.0,0.0
766
+ 764,Reach for shelving divider,0,0,0.0,0.0,0.0
767
+ 765,Rearrange shelf item,0,0,0.0,0.0,0.0
768
+ 766,Arrange paper strips,0,0,0.0,0.0,0.0
769
+ 767,Place down strip,0,9,0.0,0.0,0.0
770
+ 768,Move puzzle piece,0,2,0.0,0.0,0.0
771
+ 769,Cap marker,0,0,0.0,0.0,0.0
772
+ 770,Combine bead piles,0,0,0.0,0.0,0.0
773
+ 771,Draw lines with pen and ruler,0,0,0.0,0.0,0.0
774
+ 772,Put down phone,0,0,0.0,0.0,0.0
775
+ 773,Pick up pasta box,0,0,0.0,0.0,0.0
776
+ 774,Place gift box into bin,0,0,0.0,0.0,0.0
777
+ 775,Remove plastic container from storage box,0,0,0.0,0.0,0.0
778
+ 776,Hold ruler,0,0,0.0,0.0,0.0
779
+ 777,Move pen away,0,0,0.0,0.0,0.0
780
+ 778,Place crate on floor,0,0,0.0,0.0,0.0
781
+ 779,Place smartphone on table,0,0,0.0,0.0,0.0
782
+ 780,Discard paper towel,0,0,0.0,0.0,0.0
783
+ 781,Release paper star,0,0,0.0,0.0,0.0
784
+ 782,Place phone on table,0,0,0.0,0.0,0.0
785
+ 783,Scrolling or navigating on phone,0,0,0.0,0.0,0.0
786
+ 784,Hold electronic item,0,0,0.0,0.0,0.0
787
+ 785,Inspect electronic item,0,0,0.0,0.0,0.0
788
+ 786,Move pineapple chips,0,0,0.0,0.0,0.0
789
+ 787,Mark paper list,0,0,0.0,0.0,0.0
790
+ 788,Placing phone down,0,0,0.0,0.0,0.0
791
+ 789,Pick up nut bar box,0,0,0.0,0.0,0.0
792
+ 790,Pick up plastic bin,0,0,0.0,0.0,0.0
793
+ 791,Pick up pickle jar,0,0,0.0,0.0,0.0
794
+ 792,Pick up product from shelf,0,0,0.0,0.0,0.0
795
+ 793,Place jar into shelf box,0,0,0.0,0.0,0.0
796
+ 794,Wipe grocery shelf,0,0,0.0,0.0,0.0
797
+ 795,Rearrange buttons,0,0,0.0,0.0,0.0
798
+ 796,Release button,0,0,0.0,0.0,0.0
799
+ 797,Pick up orange button,0,0,0.0,0.0,0.0
800
+ 798,Arrange small buttons,0,0,0.0,0.0,0.0
801
+ 799,Align buttons,0,0,0.0,0.0,0.0
802
+ 800,Look around the table,0,0,0.0,0.0,0.0
803
+ 801,Align red buttons,0,0,0.0,0.0,0.0
804
+ 802,Reach for black button,0,0,0.0,0.0,0.0
805
+ 803,Reach for buttons,0,0,0.0,0.0,0.0
806
+ 804,Place and align button,0,0,0.0,0.0,0.0
807
+ 805,Move hand,0,0,0.0,0.0,0.0
808
+ 806,Move button to line,0,0,0.0,0.0,0.0
809
+ 807,Reach for utility knife,0,0,0.0,0.0,0.0
810
+ 808,Place down paper pieces,0,0,0.0,0.0,0.0
811
+ 809,Switch to scissors,0,0,0.0,0.0,0.0
812
+ 810,Place phone on shelf,0,10,0.0,0.0,0.0
813
+ 811,Inspect product lid,0,0,0.0,0.0,0.0
814
+ 812,Sweep floor debris,0,0,0.0,0.0,0.0
815
+ 813,Adjust grip on container,0,1,0.0,0.0,0.0
816
+ 814,Manipulate paper piece,0,0,0.0,0.0,0.0
817
+ 815,Hold quilled paper coil,0,0,0.0,0.0,0.0
818
+ 816,Place scissors aside,0,0,0.0,0.0,0.0
819
+ 817,Finish placing cardboard cutouts,0,0,0.0,0.0,0.0
820
+ 818,Fold cut cardboard,0,0,0.0,0.0,0.0
821
+ 819,Look away,0,0,0.0,0.0,0.0
822
+ 820,Pick up cut cardboard piece,0,0,0.0,0.0,0.0
823
+ 821,Reposition scissors,0,1,0.0,0.0,0.0
824
+ 822,Hold cardboard piece,7,0,0.0,0.0,0.0
825
+ 823,Picking up stock,0,0,0.0,0.0,0.0
826
+ 824,Carry container,0,0,0.0,0.0,0.0
827
+ 825,Positioning cardboard on workspace,0,0,0.0,0.0,0.0
828
+ 826,Stop sorting stars,0,0,0.0,0.0,0.0
829
+ 827,Place knife down,0,0,0.0,0.0,0.0
830
+ 828,Search for puzzle piece,20,3,0.0,0.0,0.0
831
+ 829,Lift pen and shift ruler,0,0,0.0,0.0,0.0
832
+ 830,Moving ruler,0,0,0.0,0.0,0.0
833
+ 831,Hold beads,19,0,0.0,0.0,0.0
834
+ 832,Adjusting fabric position,0,0,0.0,0.0,0.0
835
+ 833,Pick up new cardboard piece,24,5,0.0,0.0,0.0
836
+ 834,Gather cardboard pieces,0,0,0.0,0.0,0.0
837
+ 835,Hold electronic accessory,0,1,0.0,0.0,0.0
838
+ 836,Pick up electronic accessory,0,0,0.0,0.0,0.0
839
+ 837,Place accessory box,0,0,0.0,0.0,0.0
840
+ 838,Release product on shelf,0,0,0.0,0.0,0.0
841
+ 839,Pick up new product from box,0,0,0.0,0.0,0.0
842
+ 840,Pick up shopping bag,0,0,0.0,0.0,0.0
843
+ 841,Move to shelf,3,0,0.0,0.0,0.0
844
+ 842,Grasp snack package,0,0,0.0,0.0,0.0
845
+ 843,Place snack in box,0,0,0.0,0.0,0.0
846
+ 844,Place snack packages on shelf,0,0,0.0,0.0,0.0
847
+ 845,Reach for snack package,0,0,0.0,0.0,0.0
848
+ 846,Reach for item,0,0,0.0,0.0,0.0
849
+ 847,Organize item on shelf,0,0,0.0,0.0,0.0
850
+ 848,Place pen on cardboard,0,0,0.0,0.0,0.0
851
+ 849,Adjust cardboard divider,0,0,0.0,0.0,0.0
852
+ 850,Place finished star on table,0,0,0.0,0.0,0.0
853
+ 851,Inspect shelf,0,0,0.0,0.0,0.0
854
+ 852,Pick up snack packs,0,0,0.0,0.0,0.0
855
+ 853,Move to shelf base,0,0,0.0,0.0,0.0
856
+ 854,Place gift box on shelf,0,0,0.0,0.0,0.0
857
+ 855,Place snack pouch on shelf,0,0,0.0,0.0,0.0
858
+ 856,Sort Mahjong tiles,0,0,0.0,0.0,0.0
859
+ 857,Pick up charging case,0,0,0.0,0.0,0.0
860
+ 858,Place ruler on cardboard,0,0,0.0,0.0,0.0
861
+ 859,Reposition tools,0,0,0.0,0.0,0.0
862
+ 860,Position scissors for next cut,0,0,0.0,0.0,0.0
863
+ 861,Tapping on smartphone screen,0,0,0.0,0.0,0.0
864
+ 862,Positioning ruler on cardboard,0,0,0.0,0.0,0.0
865
+ 863,Placing labeled square,0,0,0.0,0.0,0.0
866
+ 864,Switching marker,0,0,0.0,0.0,0.0
867
+ 865,Placing pen on table,0,0,0.0,0.0,0.0
868
+ 866,Manipulate cardboard sheet,0,1,0.0,0.0,0.0
869
+ 867,Interact with smartphone,21,10,0.0,0.0,0.0
870
+ 868,Pick up retail item,0,0,0.0,0.0,0.0
871
+ 869,Adjust retail item position,0,0,0.0,0.0,0.0
872
+ 870,Observe surroundings,0,0,0.0,0.0,0.0
873
+ 871,Manipulate paper stars,0,0,0.0,0.0,0.0
874
+ 872,Pick up power bank,0,0,0.0,0.0,0.0
875
+ 873,Rub hands together,0,0,0.0,0.0,0.0
876
+ 874,Place star on table,0,3,0.0,0.0,0.0
877
+ 875,Gather pieces,0,0,0.0,0.0,0.0
878
+ 876,Select another item,0,8,0.0,0.0,0.0
879
+ 877,Place container on floor,0,0,0.0,0.0,0.0
880
+ 878,Place storage container on floor,0,2,0.0,0.0,0.0
881
+ 879,Reorganize bin contents,0,0,0.0,0.0,0.0
882
+ 880,Observe stocking,0,0,0.0,0.0,0.0
883
+ 881,Manipulate quilled paper strips,0,0,0.0,0.0,0.0
884
+ 882,Move blue beads,0,1,0.0,0.0,0.0
885
+ 883,Place controller on table,0,0,0.0,0.0,0.0
886
+ 884,Selecting new paper strip,0,0,0.0,0.0,0.0
887
+ 885,Grasp electronic object,0,0,0.0,0.0,0.0
888
+ 886,Reach for paper strip,0,0,0.0,0.0,0.0
889
+ 887,Reach for canned food,0,0,0.0,0.0,0.0
890
+ 888,Hold blue product box,0,0,0.0,0.0,0.0
891
+ 889,Inspect product,0,0,0.0,0.0,0.0
892
+ 890,Clean shelf,0,0,0.0,0.0,0.0
893
+ 891,Walk towards shelf,0,0,0.0,0.0,0.0
894
+ 892,Select product from box,0,0,0.0,0.0,0.0
895
+ 893,Wipe ketchup bottle,0,0,0.0,0.0,0.0
896
+ 894,Place ketchup bottle on shelf,0,0,0.0,0.0,0.0
897
+ 895,Draw line with marker,0,0,0.0,0.0,0.0
898
+ 896,Draw straight line,0,0,0.0,0.0,0.0
899
+ 897,Mark straight line,0,0,0.0,0.0,0.0
900
+ 898,Pick up small cardboard piece,0,0,0.0,0.0,0.0
901
+ 899,Walk through office,0,0,0.0,0.0,0.0
902
+ 900,Cut cardboard along line,0,0,0.0,0.0,0.0
903
+ 901,Reposition hands and ruler,0,0,0.0,0.0,0.0
904
+ 902,Align ruler with crease,0,0,0.0,0.0,0.0
905
+ 903,Press fold,0,0,0.0,0.0,0.0
906
+ 904,Cut cardboard strip with utility knife,0,0,0.0,0.0,0.0
907
+ 905,Pick up dustpan,17,0,0.0,0.0,0.0
908
+ 906,Hold container lid,25,0,0.0,0.0,0.0
909
+ 907,Move towards the stove,9,0,0.0,0.0,0.0
910
+ 908,Open stove pot lid,20,0,0.0,0.0,0.0
911
+ 909,Closing the door,8,0,0.0,0.0,0.0
912
+ 910,Picking up bottle,11,0,0.0,0.0,0.0
913
+ 911,Wipe kitchen counter,16,0,0.0,0.0,0.0
914
+ 912,Move towards kitchen area,15,0,0.0,0.0,0.0
915
+ 913,Place cloth on floor,6,0,0.0,0.0,0.0
916
+ 914,Reach for cleaning supplies,18,0,0.0,0.0,0.0
917
+ 915,Remove cleaning bottle,11,0,0.0,0.0,0.0
918
+ 916,Washing hands in sink,10,0,0.0,0.0,0.0
919
+ 917,Grasping cleaning cloth,7,0,0.0,0.0,0.0
920
+ 918,Wiping countertop,11,0,0.0,0.0,0.0
921
+ 919,Lift pot lid,9,0,0.0,0.0,0.0
922
+ 920,Stir contents,8,0,0.0,0.0,0.0
923
+ 921,Place lid back,9,0,0.0,0.0,0.0
924
+ 922,Adjust pot position,6,0,0.0,0.0,0.0
925
+ 923,Move pot,7,0,0.0,0.0,0.0
926
+ 924,Place towel,16,0,0.0,0.0,0.0
927
+ 925,Start cutting,7,0,0.0,0.0,0.0
928
+ 926,Cut along the marked line,51,0,0.0,0.0,0.0
929
+ 927,Pick up item from bin,0,0,0.0,0.0,0.0
930
+ 928,Hold item,0,0,0.0,0.0,0.0
931
+ 929,Check smart watch,0,0,0.0,0.0,0.0
932
+ 930,Pick up jar,0,0,0.0,0.0,0.0
933
+ 931,Pick up sauce bottle,0,0,0.0,0.0,0.0
934
+ 932,Place sauce bottle on shelf,0,0,0.0,0.0,0.0
935
+ 933,Hold empty container,0,0,0.0,0.0,0.0
936
+ 934,Assess shelf arrangement,0,0,0.0,0.0,0.0
937
+ 935,Pick up bottle,0,0,0.0,0.0,0.0
938
+ 936,Release foam strip,0,0,0.0,0.0,0.0
939
+ 937,Observe craft layout,0,0,0.0,0.0,0.0
940
+ 938,Reach for foam strips,0,0,0.0,0.0,0.0
941
+ 939,Adjust foam strip,0,0,0.0,0.0,0.0
942
+ 940,Align foam strip,0,0,0.0,0.0,0.0
943
+ 941,Attach foam strip,0,0,0.0,0.0,0.0
944
+ 942,Curve foam strip into loop,0,0,0.0,0.0,0.0
945
+ 943,Press ends of foam strip together,0,0,0.0,0.0,0.0
946
+ 944,Position yellow foam piece on strip,0,0,0.0,0.0,0.0
947
+ 945,Press foam strip,0,0,0.0,0.0,0.0
948
+ 946,Fold foam piece,0,0,0.0,0.0,0.0
949
+ 947,Pinch foam strips,0,0,0.0,0.0,0.0
950
+ 948,Pull blue foam strip,0,0,0.0,0.0,0.0
951
+ 949,Tear blue foam strip,0,0,0.0,0.0,0.0
952
+ 950,Pick up blue foam piece,0,0,0.0,0.0,0.0
953
+ 951,Tear blue foam piece,0,0,0.0,0.0,0.0
954
+ 952,Tear off blue foam piece,0,0,0.0,0.0,0.0
955
+ 953,Peel foam strip,0,0,0.0,0.0,0.0
956
+ 954,Move small blue foam piece towards the strip,0,0,0.0,0.0,0.0
957
+ 955,Align blue strip,0,0,0.0,0.0,0.0
958
+ 956,Press blue strip,0,0,0.0,0.0,0.0
959
+ 957,Position blue strip,0,0,0.0,0.0,0.0
960
+ 958,Lift blue strip,0,0,0.0,0.0,0.0
961
+ 959,Hold blue strip,0,0,0.0,0.0,0.0
962
+ 960,Peel blue strip,0,0,0.0,0.0,0.0
963
+ 961,Align paper strip,0,0,0.0,0.0,0.0
964
+ 962,Interlock paper strips,0,0,0.0,0.0,0.0
965
+ 963,Turn away from table,0,0,0.0,0.0,0.0
966
+ 964,Touch phone and paper strip,0,0,0.0,0.0,0.0
967
+ 965,Attach material to paper strip,0,0,0.0,0.0,0.0
968
+ 966,Pick up tool,0,0,0.0,0.0,0.0
969
+ 967,Walk through the room,0,0,0.0,0.0,0.0
970
+ 968,Walk down hallway,0,0,0.0,0.0,0.0
971
+ 969,Reach for door handle,0,0,0.0,0.0,0.0
972
+ 970,Grasp door handle,0,0,0.0,0.0,0.0
973
+ 971,Walk to table,0,0,0.0,0.0,0.0
974
+ 972,Pick up supplies from box,0,0,0.0,0.0,0.0
975
+ 973,Approach work table,0,0,0.0,0.0,0.0
976
+ 974,Touch colleague's back,0,0,0.0,0.0,0.0
977
+ 975,Position the chair,0,0,0.0,0.0,0.0
978
+ 976,Observe and walk through store,15,0,0.0,0.0,0.0
979
+ 977,Inspect shelf condition,27,0,0.0,0.0,0.0
980
+ 978,Approach boxes,12,0,0.0,0.0,0.0
981
+ 979,Reach for wire hangers,13,0,0.0,0.0,0.0
982
+ 980,Extract wire hangers from box,30,0,0.0,0.0,0.0
983
+ 981,Bundle display hooks,22,0,0.0,0.0,0.0
984
+ 982,Release hook,14,0,0.0,0.0,0.0
985
+ 983,Move through aisle,10,0,0.0,0.0,0.0
986
+ 984,Pick up items from the shopping bag,23,0,0.0,0.0,0.0
987
+ 985,Place items on the shelf,6,0,0.0,0.0,0.0
988
+ 986,Release cardboard piece and gesture,16,0,0.0,0.0,0.0
989
+ 987,Move marker and adjust hand,8,0,0.0,0.0,0.0
990
+ 988,Identify next cardboard piece,21,0,0.0,0.0,0.0
991
+ 989,Observe and pause,11,0,0.0,0.0,0.0
992
+ 990,Resume observation,4,0,0.0,0.0,0.0
993
+ 991,Reach for and examine canned goods,0,0,0.0,0.0,0.0
994
+ 992,Examine canned goods,0,0,0.0,0.0,0.0
995
+ 993,Select and pick up a canned item,0,0,0.0,0.0,0.0
996
+ 994,Place item back on shelf,0,0,0.0,0.0,0.0
997
+ 995,Inspect Dior gift box,0,0,0.0,0.0,0.0
998
+ 996,Move along the shelf,0,0,0.0,0.0,0.0
999
+ 997,Select a bottle,0,0,0.0,0.0,0.0
1000
+ 998,Place bottle back on shelf,0,0,0.0,0.0,0.0
1001
+ 999,Pick up another bottle,0,0,0.0,0.0,0.0
1002
+ 1000,Release bottle,0,0,0.0,0.0,0.0
1003
+ 1001,Inspect bottle,0,0,0.0,0.0,0.0
1004
+ 1002,Inspect almond package,0,0,0.0,0.0,0.0
1005
+ 1003,Scan supermarket shelves,0,0,0.0,0.0,0.0
1006
+ 1004,Move along the supermarket aisle,0,0,0.0,0.0,0.0
1007
+ 1005,Reach for canned goods,0,0,0.0,0.0,0.0
1008
+ 1006,Touch canned goods,0,0,0.0,0.0,0.0
1009
+ 1007,Manipulate cardboard shape,0,0,0.0,0.0,0.0
1010
+ 1008,Hold small cardboard pieces,0,0,0.0,0.0,0.0
1011
+ 1009,Prepare to place cardboard,0,0,0.0,0.0,0.0
1012
+ 1010,Reach for next can,18,0,0.0,0.0,0.0
1013
+ 1011,Hold canned food,24,0,0.0,0.0,0.0
1014
+ 1012,Retrieve next canned food item,17,0,0.0,0.0,0.0
1015
+ 1013,Align canned food on shelf,9,0,0.0,0.0,0.0
1016
+ 1014,Retrieve canned food from box,12,0,0.0,0.0,0.0
1017
+ 1015,Place another canned food on shelf,11,0,0.0,0.0,0.0
1018
+ 1016,Adjust canned food on shelf,9,0,0.0,0.0,0.0
1019
+ 1017,Move hand away from shelf,8,0,0.0,0.0,0.0
1020
+ 1018,Hold earbud case,21,0,0.0,0.0,0.0
1021
+ 1019,sort craft materials,36,0,0.0,0.0,0.0
1022
+ 1020,Manipulate craft piece,38,0,0.0,0.0,0.0
1023
+ 1021,Manipulate craft paper strips,33,0,0.0,0.0,0.0
1024
+ 1022,Operate smartphone,40,0,0.0,0.0,0.0
1025
+ 1023,Release smartphone,7,0,0.0,0.0,0.0
1026
+ 1024,Sort small craft pieces,39,0,0.0,0.0,0.0
1027
+ 1025,Hold product package,0,0,0.0,0.0,0.0
1028
+ 1026,Check phone,0,0,0.0,0.0,0.0
1029
+ 1027,Hold charging cable,0,0,0.0,0.0,0.0
1030
+ 1028,Hold items in hand,0,0,0.0,0.0,0.0
1031
+ 1029,Hold and examine item,0,0,0.0,0.0,0.0
1032
+ 1030,Remove item from bag,0,0,0.0,0.0,0.0
1033
+ 1031,Pick up pack from shelf,0,0,0.0,0.0,0.0
1034
+ 1032,fold purple ribbon,0,0,0.0,0.0,0.0
1035
+ 1033,Fold ribbon,0,0,0.0,0.0,0.0
1036
+ 1034,Hold small piece of ribbon,0,0,0.0,0.0,0.0
1037
+ 1035,Position ribbon piece,0,0,0.0,0.0,0.0
1038
+ 1036,Manipulate ribbon piece,0,0,0.0,0.0,0.0
1039
+ 1037,Place ribbon onto project,0,0,0.0,0.0,0.0
1040
+ 1038,Fold and manipulate ribbon,0,0,0.0,0.0,0.0
1041
+ 1039,Manipulate ribbon knot,0,0,0.0,0.0,0.0
1042
+ 1040,Secure ribbon with needle,0,0,0.0,0.0,0.0
1043
+ 1041,Open paper lantern,29,0,0.0,0.0,0.0
1044
+ 1042,Fold paper lantern,9,0,0.0,0.0,0.0
1045
+ 1043,Grasp lantern,15,0,0.0,0.0,0.0
1046
+ 1044,Grasp lantern component,15,0,0.0,0.0,0.0
1047
+ 1045,Align paper lantern edges,29,0,0.0,0.0,0.0
1048
+ 1046,Release lantern,13,0,0.0,0.0,0.0
1049
+ 1047,Pick up packaged paper lantern component,12,0,0.0,0.0,0.0
1050
+ 1048,Handle paper lantern component,19,0,0.0,0.0,0.0
1051
+ 1049,Open folded paper lantern,21,0,0.0,0.0,0.0
1052
+ 1050,Hold paper lantern,19,0,0.0,0.0,0.0
1053
+ 1051,Apply adhesive tape to lantern,14,0,0.0,0.0,0.0
1054
+ 1052,Remove paper lantern part from packaging,16,0,0.0,0.0,0.0
1055
+ 1053,Remove plastic packaging,8,0,0.0,0.0,0.0
1056
+ 1054,Open paper lantern component,24,0,0.0,0.0,0.0
1057
+ 1055,Expand paper lantern,22,0,0.0,0.0,0.0
1058
+ 1056,Align edges of paper lantern,6,0,0.0,0.0,0.0
1059
+ 1057,Mark cardboard with ruler,0,0,0.0,0.0,0.0
1060
+ 1058,Cut along the line,0,0,0.0,0.0,0.0
1061
+ 1059,Release cardboard,0,0,0.0,0.0,0.0
1062
+ 1060,Reposition utility knife,0,0,0.0,0.0,0.0
1063
+ 1061,Tear off cardboard segment,0,0,0.0,0.0,0.0
1064
+ 1062,Browsing smartphone content,0,0,0.0,0.0,0.0
1065
+ 1063,Manipulate small component,0,0,0.0,0.0,0.0
1066
+ 1064,Manipulate component on strip,0,0,0.0,0.0,0.0
1067
+ 1065,Place strip on table,0,0,0.0,0.0,0.0
1068
+ 1066,Manipulate component,0,0,0.0,0.0,0.0
1069
+ 1067,Reach for craft items,18,0,0.0,0.0,0.0
1070
+ 1068,Place hand on table,33,0,0.0,0.0,0.0
1071
+ 1069,Browse smartphone screen,33,0,0.0,0.0,0.0
1072
+ 1070,Scroll smartphone screen,31,0,0.0,0.0,0.0
1073
+ 1071,Put down smartphone,26,0,0.0,0.0,0.0
1074
+ 1072,Place smartphone down,24,0,0.0,0.0,0.0
1075
+ 1073,Record count on notepad,0,0,0.0,0.0,0.0
1076
+ 1074,Count and record paper stars,0,0,0.0,0.0,0.0
1077
+ 1075,Record star count on paper,0,0,0.0,0.0,0.0
1078
+ 1076,Connect cable to device,0,0,0.0,0.0,0.0
1079
+ 1077,Place device on lap,0,0,0.0,0.0,0.0
1080
+ 1078,Count and arrange paper stars,0,0,0.0,0.0,0.0
1081
+ 1079,Count paper stars,0,0,0.0,0.0,0.0
1082
+ 1080,Move hand to paper stars,0,0,0.0,0.0,0.0
1083
+ 1081,Resume counting stars,0,0,0.0,0.0,0.0
1084
+ 1082,Reviewing count record,0,0,0.0,0.0,0.0
1085
+ 1083,Write on paper record,0,0,0.0,0.0,0.0
1086
+ 1084,Update paper record,0,0,0.0,0.0,0.0
1087
+ 1085,Adjust cardboard,0,0,0.0,0.0,0.0
1088
+ 1086,Set down scissors and pick up power bank,0,0,0.0,0.0,0.0
1089
+ 1087,Reposition cardboard for cutting,0,0,0.0,0.0,0.0
1090
+ 1088,Arrange cardboard pieces,0,0,0.0,0.0,0.0
1091
+ 1089,Mark cardboard strip with pen,0,0,0.0,0.0,0.0
1092
+ 1090,Pick up puzzle piece,18,0,0.0,0.0,0.0
1093
+ 1091,Place piece into puzzle,25,0,0.0,0.0,0.0
1094
+ 1092,Manipulate puzzle piece,38,0,0.0,0.0,0.0
1095
+ 1093,Observe puzzle progress,32,0,0.0,0.0,0.0
1096
+ 1094,Reach for puzzle piece,16,0,0.0,0.0,0.0
1097
+ 1095,Attempt to fit puzzle piece,31,0,0.0,0.0,0.0
1098
+ 1096,Sort puzzle pieces,34,0,0.0,0.0,0.0
1099
+ 1097,Walking across the room,17,0,0.0,0.0,0.0
1100
+ 1098,Approaching the table,9,0,0.0,0.0,0.0
1101
+ 1099,Preparing to craft,10,0,0.0,0.0,0.0
1102
+ 1100,Picking up crafting material,12,0,0.0,0.0,0.0
1103
+ 1101,Manipulate material,16,0,0.0,0.0,0.0
1104
+ 1102,Place material,13,0,0.0,0.0,0.0
1105
+ 1103,Manipulate yellow strip,31,0,0.0,0.0,0.0
1106
+ 1104,Manipulating paper strips,22,0,0.0,0.0,0.0
1107
+ 1105,Manipulate bead,23,0,0.0,0.0,0.0
1108
+ 1106,Manipulate beads,22,0,0.0,0.0,0.0
1109
+ 1107,Hold and manipulate paper strip,31,0,0.0,0.0,0.0
1110
+ 1108,Repositioning ruler,0,0,0.0,0.0,0.0
1111
+ 1109,Place down ruler and pen,0,0,0.0,0.0,0.0
1112
+ 1110,Walk through hallway,0,0,0.0,0.0,0.0
1113
+ 1111,Fold cardboard edge,0,0,0.0,0.0,0.0
1114
+ 1112,Pick up marker,0,0,0.0,0.0,0.0
1115
+ 1113,Drop cardboard square into box,0,0,0.0,0.0,0.0
1116
+ 1114,Retrieve hand to table,0,0,0.0,0.0,0.0
1117
+ 1115,Pick up cardboard stack,0,0,0.0,0.0,0.0
1118
+ 1116,Walk with cardboard,0,0,0.0,0.0,0.0
1119
+ 1117,Deposit cardboard squares,0,0,0.0,0.0,0.0
1120
+ 1118,Move away from collection box,0,0,0.0,0.0,0.0
1121
+ 1119,Walking through office hallway,0,0,0.0,0.0,0.0
1122
+ 1120,Grasp cardboard sheet,0,0,0.0,0.0,0.0
1123
+ 1121,Cut cardboard sheet with scissors,0,0,0.0,0.0,0.0
1124
+ 1122,Sort cut cardboard,0,0,0.0,0.0,0.0
1125
+ 1123,Cut cardboard sheet,0,0,0.0,0.0,0.0
1126
+ 1124,Place cardboard square,0,0,0.0,0.0,0.0
1127
+ 1125,Sort buttons,25,0,0.0,0.0,0.0
1128
+ 1126,Arrange buttons in a line,29,0,0.0,0.0,0.0
1129
+ 1127,Sort and arrange buttons,32,0,0.0,0.0,0.0
1130
+ 1128,Sort button,36,0,0.0,0.0,0.0
1131
+ 1129,Sort and adjust button line,29,0,0.0,0.0,0.0
1132
+ 1130,Sort and place buttons,31,0,0.0,0.0,0.0
1133
+ 1131,Walking in the hallway,13,0,0.0,0.0,0.0
1134
+ 1132,Approaching and pressing the door switch,22,0,0.0,0.0,0.0
1135
+ 1133,Entering the VR training room,16,0,0.0,0.0,0.0
1136
+ 1134,Greeting/acknowledging participants,33,0,0.0,0.0,0.0
1137
+ 1135,Move through the training room,20,0,0.0,0.0,0.0
1138
+ 1136,Manipulate plastic strips,34,0,0.0,0.0,0.0
1139
+ 1137,Manipulate plastic strip,37,0,0.0,0.0,0.0
1140
+ 1138,Hold and bend plastic strip,16,0,0.0,0.0,0.0
1141
+ 1139,Bend and manipulate plastic strip,37,0,0.0,0.0,0.0
1142
+ 1140,Fold plastic strip,57,0,0.0,0.0,0.0
1143
+ 1141,Sort buttons by color,0,0,0.0,0.0,0.0
1144
+ 1142,Sort button by color,0,0,0.0,0.0,0.0
1145
+ 1143,Place button in group,0,0,0.0,0.0,0.0
1146
+ 1144,Move away from table,0,0,0.0,0.0,0.0
1147
+ 1145,Return to sorting,0,0,0.0,0.0,0.0
1148
+ 1146,Manipulate paper decoration,41,0,0.0,0.0,0.0
1149
+ 1147,Manipulate paper edge,35,0,0.0,0.0,0.0
1150
+ 1148,Placing paper strip,44,0,0.0,0.0,0.0
1151
+ 1149,Securing paper structure,37,0,0.0,0.0,0.0
1152
+ 1150,Manipulate adhesive strip,44,0,0.0,0.0,0.0
1153
+ 1151,Secure paper edges with adhesive,40,0,0.0,0.0,0.0
1154
+ 1152,Record count,18,0,0.0,0.0,0.0
1155
+ 1153,Sort beads and write count,18,0,0.0,0.0,0.0
1156
+ 1154,Counting and organizing beads,19,0,0.0,0.0,0.0
1157
+ 1155,Pick up star bead,6,0,0.0,0.0,0.0
1158
+ 1156,Place and count bead,27,0,0.0,0.0,0.0
1159
+ 1157,Arrange star beads,15,0,0.0,0.0,0.0
1160
+ 1158,Counting star beads,23,0,0.0,0.0,0.0
1161
+ 1159,Adjust paper,9,0,0.0,0.0,0.0
1162
+ 1160,Gather star beads,13,0,0.0,0.0,0.0
1163
+ 1161,Arrange star beads for counting,16,0,0.0,0.0,0.0
1164
+ 1162,Sort and count beads,27,0,0.0,0.0,0.0
1165
+ 1163,Rinse cloth in sink,4,0,0.0,0.0,0.0
1166
+ 1164,Reposition hand,7,0,0.0,0.0,0.0
1167
+ 1165,Touch foam strip,0,0,0.0,0.0,0.0
1168
+ 1166,Assemble foam strips,0,0,0.0,0.0,0.0
1169
+ 1167,Press foam piece to strip,0,0,0.0,0.0,0.0
1170
+ 1168,Walk towards other aisles,4,0,0.0,0.0,0.0
1171
+ 1169,Place marked piece down,14,0,0.0,0.0,0.0
1172
+ 1170,Gesturing,2,0,0.0,0.0,0.0
1173
+ 1171,Prepare to resume cutting,0,0,0.0,0.0,0.0
1174
+ 1172,Reach for next canned food,3,0,0.0,0.0,0.0
1175
+ 1173,Move hand away,5,0,0.0,0.0,0.0
1176
+ 1174,Sort craft items,6,0,0.0,0.0,0.0
1177
+ 1175,Retrieving more beads,5,0,0.0,0.0,0.0
1178
+ 1176,Pick up yellow item,0,0,0.0,0.0,0.0
1179
+ 1177,Prepare to place bottle on shelf,0,0,0.0,0.0,0.0
1180
+ 1178,Move ruler and tools,0,0,0.0,0.0,0.0
1181
+ 1179,Transition to cutting,0,0,0.0,0.0,0.0
1182
+ 1180,Position utility knife on cardboard,0,0,0.0,0.0,0.0
1183
+ 1181,Place smartphone on stand,3,0,0.0,0.0,0.0
1184
+ 1182,Move dustpan to side,3,0,0.0,0.0,0.0
1185
+ 1183,Walking towards door,3,0,0.0,0.0,0.0
1186
+ 1184,Grasp cleaning bottle,3,0,0.0,0.0,0.0
1187
+ 1185,Pick up next item from bin,0,0,0.0,0.0,0.0
1188
+ 1186,Inspect and place item on shelf,0,0,0.0,0.0,0.0
1189
+ 1187,Place blue foam piece,0,0,0.0,0.0,0.0
1190
+ 1188,Hold foam pieces,0,0,0.0,0.0,0.0
1191
+ 1189,Fold blue strip,0,0,0.0,0.0,0.0
1192
+ 1190,Pick up craft material,0,0,0.0,0.0,0.0
1193
+ 1191,Enter workspace,0,0,0.0,0.0,0.0
1194
+ 1192,Enter the room,0,0,0.0,0.0,0.0
1195
+ 1193,Pull chair,0,0,0.0,0.0,0.0
1196
+ 1194,Observe colleague and workspace,3,0,0.0,0.0,0.0
1197
+ 1195,Pick up Dior gift box,0,0,0.0,0.0,0.0
1198
+ 1196,Place back Dior gift box,0,0,0.0,0.0,0.0
1199
+ 1197,Pick up canned good,0,0,0.0,0.0,0.0
1200
+ 1198,Open earbud case,3,0,0.0,0.0,0.0
1201
+ 1199,Retrieve items from bag,0,0,0.0,0.0,0.0
1202
+ 1200,Adjust lantern string,3,0,0.0,0.0,0.0
1203
+ 1201,Adjust lantern shape,3,0,0.0,0.0,0.0
1204
+ 1202,Pick up electronic device,0,0,0.0,0.0,0.0
1205
+ 1203,Pick up small piece of material,3,0,0.0,0.0,0.0
1206
+ 1204,Use phone while crafting,3,0,0.0,0.0,0.0
1207
+ 1205,Approaching work table,0,0,0.0,0.0,0.0
1208
+ 1206,Set down utility knife,0,0,0.0,0.0,0.0
1209
+ 1207,Prepare to cut cardboard,0,0,0.0,0.0,0.0
1210
+ 1208,Score cardboard,0,0,0.0,0.0,0.0
1211
+ 1209,Move cardboard sheet,0,0,0.0,0.0,0.0
1212
+ 1210,Trim cardboard,0,0,0.0,0.0,0.0
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/long_horizon_next_action/predictions.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/confusion_matrix.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/metrics.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "status": "pass",
3
+ "task": "next_subtask_forecast",
4
+ "task_display_name": "Next Subtask Forecast",
5
+ "model_family": "neural_mlp_metadata",
6
+ "source": "128_episode_qwen_jsonl_metadata",
7
+ "input_features": "frame/context metadata plus hashed prompt/options/main_task text; answer_json fields are excluded from inputs",
8
+ "split_policy": "train on train split, report held-out test split",
9
+ "num_train_windows": 25068,
10
+ "num_val_windows": 4496,
11
+ "num_test_windows": 3951,
12
+ "num_classes": 891,
13
+ "num_train_classes": 651,
14
+ "history": [
15
+ {
16
+ "epoch": 1,
17
+ "loss": 5.732333925288888,
18
+ "train_accuracy": 0.08217647997446945
19
+ },
20
+ {
21
+ "epoch": 7,
22
+ "loss": 0.5706693253221172,
23
+ "train_accuracy": 0.751515876815063
24
+ },
25
+ {
26
+ "epoch": 14,
27
+ "loss": 0.293251357785971,
28
+ "train_accuracy": 0.8487314504547631
29
+ },
30
+ {
31
+ "epoch": 21,
32
+ "loss": 0.2165332631249167,
33
+ "train_accuracy": 0.8814424764640179
34
+ },
35
+ {
36
+ "epoch": 28,
37
+ "loss": 0.1831686576697289,
38
+ "train_accuracy": 0.8973990745173129
39
+ },
40
+ {
41
+ "epoch": 35,
42
+ "loss": 0.1616299121035939,
43
+ "train_accuracy": 0.9086883676400191
44
+ }
45
+ ],
46
+ "device": "cuda",
47
+ "splits": {
48
+ "test": {
49
+ "accuracy": 0.0005062009617818274,
50
+ "balanced_accuracy": 0.0008888888888888888,
51
+ "macro_f1": 2.086049543676662e-05,
52
+ "weighted_f1": 1.1879553210003771e-05,
53
+ "num_eval_windows": 3951,
54
+ "num_classes": 891
55
+ }
56
+ },
57
+ "primary_metric": "macro_f1",
58
+ "primary_score": 2.086049543676662e-05
59
+ }
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/per_class_metrics.csv ADDED
@@ -0,0 +1,892 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ class_id,class_name,support,predicted,precision,recall,f1
2
+ 0,Retrieving items from boxes,0,0,0.0,0.0,0.0
3
+ 1,Placing products on shelf,0,0,0.0,0.0,0.0
4
+ 2,Stocking canned goods,0,4,0.0,0.0,0.0
5
+ 3,Arranging products on shelf,0,0,0.0,0.0,0.0
6
+ 4,Organizing shelf workspace,0,0,0.0,0.0,0.0
7
+ 5,Finalizing shelf placement,0,0,0.0,0.0,0.0
8
+ 6,Wiping and organizing grocery products,0,0,0.0,0.0,0.0
9
+ 7,Wiping shelves and cleaning product items,0,0,0.0,0.0,0.0
10
+ 8,Cleaning products and retrieving items from boxes,0,0,0.0,0.0,0.0
11
+ 9,Cleaning and organizing products in boxes,0,0,0.0,0.0,0.0
12
+ 10,Cleaning and boxing jars,0,0,0.0,0.0,0.0
13
+ 11,Wiping shelves,0,0,0.0,0.0,0.0
14
+ 12,Cleaning shelves and handling pickle jars,0,0,0.0,0.0,0.0
15
+ 13,Organizing pickle jars and maintaining cleaning tools,0,0,0.0,0.0,0.0
16
+ 14,Wiping shelves and picking up products,0,3,0.0,0.0,0.0
17
+ 15,Wiping product jars and shelf stocking,0,0,0.0,0.0,0.0
18
+ 16,Wiping product jars,0,0,0.0,0.0,0.0
19
+ 17,Cleaning shelves and relocating,0,0,0.0,0.0,0.0
20
+ 18,Picking up and packing products,0,11,0.0,0.0,0.0
21
+ 19,Holding and inspecting product,0,22,0.0,0.0,0.0
22
+ 20,Transferring products from box to shelf,0,52,0.0,0.0,0.0
23
+ 21,Adjusting posture while holding item,0,19,0.0,0.0,0.0
24
+ 22,Grasping and placing products on shelf,0,0,0.0,0.0,0.0
25
+ 23,Rearranging items from shelf back to box,0,0,0.0,0.0,0.0
26
+ 24,Sorting and grouping buttons,0,0,0.0,0.0,0.0
27
+ 25,Organizing buttons into patterns,0,0,0.0,0.0,0.0
28
+ 26,Refining button layout,0,0,0.0,0.0,0.0
29
+ 27,Sorting buttons by color,0,11,0.0,0.0,0.0
30
+ 28,Aligning button rows,0,0,0.0,0.0,0.0
31
+ 29,Final button arrangement,0,0,0.0,0.0,0.0
32
+ 30,Cutting cardboard pieces with scissors,0,2,0.0,0.0,0.0
33
+ 31,Manipulating cardboard piece,0,0,0.0,0.0,0.0
34
+ 32,Positioning scissors to cut cardboard,0,2,0.0,0.0,0.0
35
+ 33,Releasing scissors,0,3,0.0,0.0,0.0
36
+ 34,Manipulating cardboard and picking up scissors,0,0,0.0,0.0,0.0
37
+ 35,Cutting cardboard,0,11,0.0,0.0,0.0
38
+ 36,Picking up and positioning scissors,0,8,0.0,0.0,0.0
39
+ 37,Transitioning to new workstation,0,0,0.0,0.0,0.0
40
+ 38,Marking dimensions on cardboard,0,0,0.0,0.0,0.0
41
+ 39,Cutting cardboard with scissors,0,333,0.0,0.0,0.0
42
+ 40,Finishing cut and placing scissors down,0,0,0.0,0.0,0.0
43
+ 41,unknown,18,1516,0.0013192612137203166,0.1111111111111111,0.002607561929595828
44
+ 42,Positioning ruler for measurement,0,2,0.0,0.0,0.0
45
+ 43,Marking lines with pen,0,4,0.0,0.0,0.0
46
+ 44,Adjusting marker and ruler,0,0,0.0,0.0,0.0
47
+ 45,Drawing lines and checking smartphone,0,0,0.0,0.0,0.0
48
+ 46,Marking and cutting cardboard,0,0,0.0,0.0,0.0
49
+ 47,Aligning ruler for final measurements,0,0,0.0,0.0,0.0
50
+ 48,Finalizing marks on cardboard,0,0,0.0,0.0,0.0
51
+ 49,Connecting power cables to a portable charger,0,3,0.0,0.0,0.0
52
+ 50,Preparing workspace for craft activity,0,0,0.0,0.0,0.0
53
+ 51,Crafting with paper strips,0,1,0.0,0.0,0.0
54
+ 52,Folding paper strips,0,63,0.0,0.0,0.0
55
+ 53,Folding and organizing paper strips,0,0,0.0,0.0,0.0
56
+ 54,Folding paper strips while using phone,0,4,0.0,0.0,0.0
57
+ 55,Manipulating paper strips,0,17,0.0,0.0,0.0
58
+ 56,Adjusting and placing down paper pieces,0,0,0.0,0.0,0.0
59
+ 57,Finalizing and releasing folded paper,0,0,0.0,0.0,0.0
60
+ 58,Fold paper strip,0,0,0.0,0.0,0.0
61
+ 59,Fold paper strip into knot,0,0,0.0,0.0,0.0
62
+ 60,Fold paper strip into lucky star,0,0,0.0,0.0,0.0
63
+ 61,Inflate paper star,0,0,0.0,0.0,0.0
64
+ 62,Folding purple paper strip,0,0,0.0,0.0,0.0
65
+ 63,Holding and creasing purple paper,0,0,0.0,0.0,0.0
66
+ 64,Releasing paper and reaching for phone,0,0,0.0,0.0,0.0
67
+ 65,Organizing paper strips,0,0,0.0,0.0,0.0
68
+ 66,Preparation and initial cutting with utility knife,0,0,0.0,0.0,0.0
69
+ 67,Cutting and separating cardboard pieces,0,0,0.0,0.0,0.0
70
+ 68,Cutting cardboard with utility knife,0,0,0.0,0.0,0.0
71
+ 69,Marking cardboard piece,0,66,0.0,0.0,0.0
72
+ 70,Marking cardboard piece and preparing workspace,0,0,0.0,0.0,0.0
73
+ 71,Retrieving tools from bag,0,0,0.0,0.0,0.0
74
+ 72,Retrieving smartphone,0,2,0.0,0.0,0.0
75
+ 73,Measuring and adjusting workspace,0,4,0.0,0.0,0.0
76
+ 74,Marking cardboard with pen,0,0,0.0,0.0,0.0
77
+ 75,Drawing guide lines on cardboard,0,0,0.0,0.0,0.0
78
+ 76,Walking towards the desk,0,0,0.0,0.0,0.0
79
+ 77,Preparing materials and positioning hands for quilling,0,0,0.0,0.0,0.0
80
+ 78,Manipulating and releasing quilling strips,0,0,0.0,0.0,0.0
81
+ 79,Manipulating and releasing paper strips,0,10,0.0,0.0,0.0
82
+ 80,Beginning to roll the quilling strip,0,7,0.0,0.0,0.0
83
+ 81,Typing on smartphone while working with paper,0,0,0.0,0.0,0.0
84
+ 82,Checking stock information,0,1,0.0,0.0,0.0
85
+ 83,Organizing products on shelf,0,0,0.0,0.0,0.0
86
+ 84,Handling shipping box,0,2,0.0,0.0,0.0
87
+ 85,Placing items on shelf,0,8,0.0,0.0,0.0
88
+ 86,Assessing shelf status and relocating,0,0,0.0,0.0,0.0
89
+ 87,Inspecting and approaching shelf,0,0,0.0,0.0,0.0
90
+ 88,Adjusting items on shelf,0,1,0.0,0.0,0.0
91
+ 89,Removing and discarding damaged items,0,0,0.0,0.0,0.0
92
+ 90,Sweeping floor debris,0,0,0.0,0.0,0.0
93
+ 91,Reaching for products and preparing the shelf,0,0,0.0,0.0,0.0
94
+ 92,Peeling and disposing of shelf labels,0,0,0.0,0.0,0.0
95
+ 93,Removing old shelf labels,0,73,0.0,0.0,0.0
96
+ 94,Removing labels and moving along the shelf,0,0,0.0,0.0,0.0
97
+ 95,Repositioning stool for the next section,0,0,0.0,0.0,0.0
98
+ 96,Observing shelf and relocating to next aisle,0,3,0.0,0.0,0.0
99
+ 97,Reviewing new product labels,0,0,0.0,0.0,0.0
100
+ 98,Selecting and collecting bottled goods,0,0,0.0,0.0,0.0
101
+ 99,Inspecting supplement bottle,0,0,0.0,0.0,0.0
102
+ 100,Handling miscellaneous items,0,0,0.0,0.0,0.0
103
+ 101,Navigating store and browsing shelves,0,0,0.0,0.0,0.0
104
+ 102,Selecting and evaluating items,0,0,0.0,0.0,0.0
105
+ 103,Adding items to shopping container,0,0,0.0,0.0,0.0
106
+ 104,Selecting and collecting oil bottles,0,0,0.0,0.0,0.0
107
+ 105,Inspecting and packing supplement bottle,0,0,0.0,0.0,0.0
108
+ 106,Selecting and packing spice jars,0,0,0.0,0.0,0.0
109
+ 107,Finalizing spice jar selection,0,0,0.0,0.0,0.0
110
+ 108,Sorting beads by color,0,2,0.0,0.0,0.0
111
+ 109,Refining bead arrangement with marker,0,0,0.0,0.0,0.0
112
+ 110,Arranging star-shaped beads,0,1,0.0,0.0,0.0
113
+ 111,Fine-tuning bead placement,0,145,0.0,0.0,0.0
114
+ 112,Moving away from workstation,0,3,0.0,0.0,0.0
115
+ 113,Walking through the room,0,0,0.0,0.0,0.0
116
+ 114,Returning to desk and setting up smartphone,0,1,0.0,0.0,0.0
117
+ 115,Preparing for further sorting,0,2,0.0,0.0,0.0
118
+ 116,Sorting beads,0,0,0.0,0.0,0.0
119
+ 117,Cutting cardboard with a utility knife,0,0,0.0,0.0,0.0
120
+ 118,Adjusting ruler position,0,0,0.0,0.0,0.0
121
+ 119,Marking guidelines on cardboard,0,0,0.0,0.0,0.0
122
+ 120,Drawing lines on cardboard,0,0,0.0,0.0,0.0
123
+ 121,Drawing lines along ruler,0,0,0.0,0.0,0.0
124
+ 122,Marking grid lines,0,0,0.0,0.0,0.0
125
+ 123,Capping marker and marking lines,0,0,0.0,0.0,0.0
126
+ 124,Capping marker and positioning ruler,0,0,0.0,0.0,0.0
127
+ 125,Organizing cardboard pieces,15,0,0.0,0.0,0.0
128
+ 126,Walking in workspace,0,0,0.0,0.0,0.0
129
+ 127,Charging power bank,0,0,0.0,0.0,0.0
130
+ 128,Manipulate and inspect colorful pieces,0,0,0.0,0.0,0.0
131
+ 129,Manipulate colorful pieces,0,0,0.0,0.0,0.0
132
+ 130,Sort colorful pieces,0,0,0.0,0.0,0.0
133
+ 131,Handle power bank and cable,0,0,0.0,0.0,0.0
134
+ 132,Touch pieces in box and interact with colleagues,0,0,0.0,0.0,0.0
135
+ 133,Hold small white box,0,0,0.0,0.0,0.0
136
+ 134,"Place white box, adjust smartphone, and resume sorting",0,0,0.0,0.0,0.0
137
+ 135,Sort small colorful pieces,0,0,0.0,0.0,0.0
138
+ 136,Sort colorful paper pieces,0,0,0.0,0.0,0.0
139
+ 137,Check instructions on phone,0,2,0.0,0.0,0.0
140
+ 138,Draw lines and patterns on cardboard,0,0,0.0,0.0,0.0
141
+ 139,Trace and remove pattern,0,0,0.0,0.0,0.0
142
+ 140,Trace and remove pattern piece,0,2,0.0,0.0,0.0
143
+ 141,Cut out cardboard pattern,0,0,0.0,0.0,0.0
144
+ 142,Cut cardboard pattern,0,16,0.0,0.0,0.0
145
+ 143,Adjust position and check phone,0,0,0.0,0.0,0.0
146
+ 144,Prepare tools for marking,0,0,0.0,0.0,0.0
147
+ 145,Mark cardboard for cutting,0,12,0.0,0.0,0.0
148
+ 146,Cut cardboard with utility knife,0,0,0.0,0.0,0.0
149
+ 147,Reposition and cut cardboard,0,0,0.0,0.0,0.0
150
+ 148,Hold quilling paper,0,0,0.0,0.0,0.0
151
+ 149,Release paper strip,0,0,0.0,0.0,0.0
152
+ 150,Roll quilling paper,0,0,0.0,0.0,0.0
153
+ 151,Handle and prepare paper coil,0,0,0.0,0.0,0.0
154
+ 152,Manipulate quilled paper strip,0,0,0.0,0.0,0.0
155
+ 153,Release and prepare new strip,0,0,0.0,0.0,0.0
156
+ 154,Manipulate small paper segment,0,0,0.0,0.0,0.0
157
+ 155,Place down paper segment and reach for supplies,0,0,0.0,0.0,0.0
158
+ 156,Clear workspace and pick up phone,0,0,0.0,0.0,0.0
159
+ 157,Browse and interact with phone interface,0,0,0.0,0.0,0.0
160
+ 158,Browsing phone interface,0,0,0.0,0.0,0.0
161
+ 159,Pick up and inspect light blue strip,0,0,0.0,0.0,0.0
162
+ 160,Manipulate light blue strip,0,0,0.0,0.0,0.0
163
+ 161,Cutting cardboard tube,0,0,0.0,0.0,0.0
164
+ 162,Stacking cardboard pieces,0,0,0.0,0.0,0.0
165
+ 163,Positioning and cutting cardboard,0,1,0.0,0.0,0.0
166
+ 164,Cutting and placing cardboard piece,0,4,0.0,0.0,0.0
167
+ 165,Preparing tools and materials,0,0,0.0,0.0,0.0
168
+ 166,Cutting cardboard piece,0,0,0.0,0.0,0.0
169
+ 167,Cutting and releasing cardboard piece,0,0,0.0,0.0,0.0
170
+ 168,Approaching workstation,0,1,0.0,0.0,0.0
171
+ 169,Cutting cardboard into triangles,0,0,0.0,0.0,0.0
172
+ 170,Cutting cardboard and picking up smartphone,0,0,0.0,0.0,0.0
173
+ 171,Moving cardboard cutouts,0,0,0.0,0.0,0.0
174
+ 172,Cleaning up workstation,0,0,0.0,0.0,0.0
175
+ 173,Organizing tools and materials,0,0,0.0,0.0,0.0
176
+ 174,Cutting cardboard shapes,0,3,0.0,0.0,0.0
177
+ 175,Cutting cardboard triangles,0,0,0.0,0.0,0.0
178
+ 176,Sorting cardboard shapes and holding marker,0,0,0.0,0.0,0.0
179
+ 177,Marking cardboard,0,0,0.0,0.0,0.0
180
+ 178,Inspecting pieces and switching to scissors,0,0,0.0,0.0,0.0
181
+ 179,Cutting cardboard strips with scissors,0,0,0.0,0.0,0.0
182
+ 180,Cutting and organizing cardboard pieces,0,7,0.0,0.0,0.0
183
+ 181,Inspecting and folding cardboard pieces,0,0,0.0,0.0,0.0
184
+ 182,Performing precision cuts on cardboard,0,0,0.0,0.0,0.0
185
+ 183,Folding cardboard and handling utility knife,0,0,0.0,0.0,0.0
186
+ 184,Inspecting and placing cardboard strips,0,0,0.0,0.0,0.0
187
+ 185,Gathering and boxing inventory,0,0,0.0,0.0,0.0
188
+ 186,Organizing cans into a storage box,0,0,0.0,0.0,0.0
189
+ 187,Placing cans onto the shelf,0,0,0.0,0.0,0.0
190
+ 188,Arranging shelf display,0,0,0.0,0.0,0.0
191
+ 189,Preparing to organize container bin,0,0,0.0,0.0,0.0
192
+ 190,Placing items into the bin,0,0,0.0,0.0,0.0
193
+ 191,Inspecting and placing cans on shelf,0,0,0.0,0.0,0.0
194
+ 192,Moving along the aisle to assess stock,0,0,0.0,0.0,0.0
195
+ 193,Organizing shelf display,0,0,0.0,0.0,0.0
196
+ 194,Replenishing shelf stock,0,0,0.0,0.0,0.0
197
+ 195,Holding and grasping product bags,0,1,0.0,0.0,0.0
198
+ 196,Placing product bags on shelf,0,0,0.0,0.0,0.0
199
+ 197,Positioning shelving dividers,0,0,0.0,0.0,0.0
200
+ 198,Placing items and transitioning,0,2,0.0,0.0,0.0
201
+ 199,Retrieving and carrying containers,0,0,0.0,0.0,0.0
202
+ 200,"Placing, rearranging, and cleaning shelf items",0,0,0.0,0.0,0.0
203
+ 201,Stocking product boxes,0,10,0.0,0.0,0.0
204
+ 202,Retrieving next product,0,7,0.0,0.0,0.0
205
+ 203,Placing and adjusting plush toys,0,0,0.0,0.0,0.0
206
+ 204,Organizing shelf products,0,0,0.0,0.0,0.0
207
+ 205,Reaching for products,0,0,0.0,0.0,0.0
208
+ 206,Marking cardboard squares,0,0,0.0,0.0,0.0
209
+ 207,Walking to workspace,0,0,0.0,0.0,0.0
210
+ 208,Moving to workspace,0,0,0.0,0.0,0.0
211
+ 209,Arranging cardboard squares,0,0,0.0,0.0,0.0
212
+ 210,Stacking cardboard squares,0,0,0.0,0.0,0.0
213
+ 211,Sorting colorful paper stars,0,0,0.0,0.0,0.0
214
+ 212,Stopping sorting activity,0,0,0.0,0.0,0.0
215
+ 213,Leaving the room and retrieving object,0,0,0.0,0.0,0.0
216
+ 214,Returning to the workspace,0,0,0.0,0.0,0.0
217
+ 215,Resuming sorting paper stars,0,0,0.0,0.0,0.0
218
+ 216,Holding cardboard pieces,0,0,0.0,0.0,0.0
219
+ 217,Cutting cardboard and placing knife down,0,0,0.0,0.0,0.0
220
+ 218,Cutting cardboard strip,0,0,0.0,0.0,0.0
221
+ 219,Positioning cardboard piece,0,0,0.0,0.0,0.0
222
+ 220,Aligning cardboard strip,0,0,0.0,0.0,0.0
223
+ 221,Holding cardboard with ruler,0,0,0.0,0.0,0.0
224
+ 222,Moving utility knife along ruler,0,0,0.0,0.0,0.0
225
+ 223,Sliding utility knife along ruler,0,0,0.0,0.0,0.0
226
+ 224,Guiding utility knife along ruler,0,0,0.0,0.0,0.0
227
+ 225,Marking lines on cardboard,0,3,0.0,0.0,0.0
228
+ 226,Prepare workspace and tools,0,8,0.0,0.0,0.0
229
+ 227,Arrange paper strips,0,0,0.0,0.0,0.0
230
+ 228,Manipulate paper strips,86,3,0.0,0.0,0.0
231
+ 229,Select and handle paper strips,0,60,0.0,0.0,0.0
232
+ 230,Bend and shape paper strips,0,0,0.0,0.0,0.0
233
+ 231,Hold small object,0,2,0.0,0.0,0.0
234
+ 232,Observe workspace,0,2,0.0,0.0,0.0
235
+ 233,Placing and scanning for puzzle pieces,0,8,0.0,0.0,0.0
236
+ 234,Scanning for and placing puzzle pieces,0,20,0.0,0.0,0.0
237
+ 235,Placing puzzle piece,0,0,0.0,0.0,0.0
238
+ 236,Positioning puzzle piece,0,20,0.0,0.0,0.0
239
+ 237,Manipulating puzzle pieces,0,1,0.0,0.0,0.0
240
+ 238,"Moving, placing, and adjusting puzzle pieces",0,11,0.0,0.0,0.0
241
+ 239,Adjusting puzzle piece,0,29,0.0,0.0,0.0
242
+ 240,Adjusting a puzzle piece,0,0,0.0,0.0,0.0
243
+ 241,Draw lines using a ruler,0,0,0.0,0.0,0.0
244
+ 242,Reposition ruler and draw lines,0,0,0.0,0.0,0.0
245
+ 243,Steadying the ruler and pen,0,0,0.0,0.0,0.0
246
+ 244,Cap marker and place down,0,0,0.0,0.0,0.0
247
+ 245,Walk to and approach packing area,0,0,0.0,0.0,0.0
248
+ 246,Pack beads into box,0,0,0.0,0.0,0.0
249
+ 247,Pick up and deposit beads into box,0,0,0.0,0.0,0.0
250
+ 248,Retrieve and move cardboard tray,0,0,0.0,0.0,0.0
251
+ 249,Sort beads and adjust tray position,0,0,0.0,0.0,0.0
252
+ 250,Sort beads by color,0,41,0.0,0.0,0.0
253
+ 251,Sort beads and adjust phone,0,0,0.0,0.0,0.0
254
+ 252,Sort beads,0,16,0.0,0.0,0.0
255
+ 253,Cut light green fabric,0,0,0.0,0.0,0.0
256
+ 254,Cut fabric with scissors,0,0,0.0,0.0,0.0
257
+ 255,Adjust fabric for cutting,0,0,0.0,0.0,0.0
258
+ 256,Adjust and cut fabric,0,0,0.0,0.0,0.0
259
+ 257,Mark fabric with pen and remove ruler,0,0,0.0,0.0,0.0
260
+ 258,Mark fabric with pen,0,0,0.0,0.0,0.0
261
+ 259,Mark fabric,0,0,0.0,0.0,0.0
262
+ 260,Mark fabric with pen and ruler,0,0,0.0,0.0,0.0
263
+ 261,Cutting and placing cardboard,0,0,0.0,0.0,0.0
264
+ 262,Gathering cardboard pieces,0,1,0.0,0.0,0.0
265
+ 263,Pick up electronic accessory from box,0,0,0.0,0.0,0.0
266
+ 264,Place accessory on shelf,0,3,0.0,0.0,0.0
267
+ 265,Pick up and place accessory on shelf,0,0,0.0,0.0,0.0
268
+ 266,Reorganize items in box,0,0,0.0,0.0,0.0
269
+ 267,Transition to new product selection,0,0,0.0,0.0,0.0
270
+ 268,Pick up and place product on shelf,0,1,0.0,0.0,0.0
271
+ 269,Pick up product from box,0,7,0.0,0.0,0.0
272
+ 270,Stock multiple products on shelf,0,0,0.0,0.0,0.0
273
+ 271,Move product towards shelf,0,0,0.0,0.0,0.0
274
+ 272,Carry shopping bag,0,0,0.0,0.0,0.0
275
+ 273,Hang shopping bag and exit area,0,0,0.0,0.0,0.0
276
+ 274,Picking and moving items to shelf,0,0,0.0,0.0,0.0
277
+ 275,Placing items and returning to box,0,0,0.0,0.0,0.0
278
+ 276,Stocking and repositioning box,0,0,0.0,0.0,0.0
279
+ 277,Stocking products on shelf,0,0,0.0,0.0,0.0
280
+ 278,Handling snack packages,0,0,0.0,0.0,0.0
281
+ 279,Placing snack packages on shelf,0,0,0.0,0.0,0.0
282
+ 280,Organizing snack packages in box,0,0,0.0,0.0,0.0
283
+ 281,Transporting snack packages to shelf,0,0,0.0,0.0,0.0
284
+ 282,Organizing snacks in box,0,0,0.0,0.0,0.0
285
+ 283,Adjusting snack package,0,6,0.0,0.0,0.0
286
+ 284,Opening cardboard box,0,4,0.0,0.0,0.0
287
+ 285,Aligning plastic containers on the shelf,0,0,0.0,0.0,0.0
288
+ 286,Reaching for and adjusting shelf containers,0,0,0.0,0.0,0.0
289
+ 287,Adjusting container positions,0,0,0.0,0.0,0.0
290
+ 288,Placing containers on the shelf,0,1,0.0,0.0,0.0
291
+ 289,Adjusting items on the shelf,0,0,0.0,0.0,0.0
292
+ 290,Transferring items from shelf to shopping bag,0,5,0.0,0.0,0.0
293
+ 291,Retrieving items for bag placement,0,0,0.0,0.0,0.0
294
+ 292,Organizing shelf and placing plush toy into bag,0,0,0.0,0.0,0.0
295
+ 293,Preparing shopping bag,0,2,0.0,0.0,0.0
296
+ 294,Placing items into and organizing shopping bag,0,0,0.0,0.0,0.0
297
+ 295,Retrieving items from shelf,0,1,0.0,0.0,0.0
298
+ 296,Finalizing item placement into shopping bag,0,0,0.0,0.0,0.0
299
+ 297,Sort star-shaped beads,0,1,0.0,0.0,0.0
300
+ 298,Sort beads on the table,0,5,0.0,0.0,0.0
301
+ 299,Sort beads on table,0,0,0.0,0.0,0.0
302
+ 300,Hold instructional sign,0,0,0.0,0.0,0.0
303
+ 301,Pick up and place star-shaped beads,0,13,0.0,0.0,0.0
304
+ 302,Reposition sign and organize beads,0,5,0.0,0.0,0.0
305
+ 303,Measuring and marking cardboard for crafting,0,0,0.0,0.0,0.0
306
+ 304,Sorting origami stars,0,0,0.0,0.0,0.0
307
+ 305,Fetching materials,0,0,0.0,0.0,0.0
308
+ 306,Returning to desk,0,0,0.0,0.0,0.0
309
+ 307,Collecting origami stars,0,0,0.0,0.0,0.0
310
+ 308,Sorting light blue origami stars,0,0,0.0,0.0,0.0
311
+ 309,Sorting origami stars by color,0,0,0.0,0.0,0.0
312
+ 310,Moving origami stars,0,0,0.0,0.0,0.0
313
+ 311,Cutting cardboard with scissors and checking phone,0,0,0.0,0.0,0.0
314
+ 312,Resuming cardboard cutting,0,0,0.0,0.0,0.0
315
+ 313,Finishing cutting and switching to phone,0,0,0.0,0.0,0.0
316
+ 314,Using phone and drinking water,0,1,0.0,0.0,0.0
317
+ 315,Setting down water and picking up phone,0,0,0.0,0.0,0.0
318
+ 316,Using phone,31,0,0.0,0.0,0.0
319
+ 317,Resuming cutting cardboard,0,0,0.0,0.0,0.0
320
+ 318,Vacuuming the carpet,0,5,0.0,0.0,0.0
321
+ 319,Pushing the vacuum cleaner,0,0,0.0,0.0,0.0
322
+ 320,Moving and adjusting the vacuum cleaner,0,0,0.0,0.0,0.0
323
+ 321,Vacuuming the carpet edge,0,0,0.0,0.0,0.0
324
+ 322,Vacuuming the carpet corner,0,0,0.0,0.0,0.0
325
+ 323,Moving the vacuum cleaner,0,0,0.0,0.0,0.0
326
+ 324,Vacuuming along the wall edge,0,0,0.0,0.0,0.0
327
+ 325,Fold paper strip into star,0,0,0.0,0.0,0.0
328
+ 326,Arrange Mahjong tiles,0,0,0.0,0.0,0.0
329
+ 327,Rearrange Mahjong tiles,0,76,0.0,0.0,0.0
330
+ 328,Adjust Mahjong tiles,0,5,0.0,0.0,0.0
331
+ 329,Reach for and adjust Mahjong tiles,0,107,0.0,0.0,0.0
332
+ 330,Rearrange Mahjong tile,0,3,0.0,0.0,0.0
333
+ 331,Adjust and align Mahjong tiles,0,9,0.0,0.0,0.0
334
+ 332,"Adjust, move, and realign Mahjong tiles",0,2,0.0,0.0,0.0
335
+ 333,Cutting cardboard square,0,0,0.0,0.0,0.0
336
+ 334,Trimming cardboard piece,0,0,0.0,0.0,0.0
337
+ 335,Retrieve food items from boxes,0,0,0.0,0.0,0.0
338
+ 336,Pick up cereal boxes,0,7,0.0,0.0,0.0
339
+ 337,Carry cereal boxes to aisle,0,0,0.0,0.0,0.0
340
+ 338,Retrieve and transport cereal,0,0,0.0,0.0,0.0
341
+ 339,Transport cereal to shelf,0,15,0.0,0.0,0.0
342
+ 340,Transport pasta to shelf,0,1,0.0,0.0,0.0
343
+ 341,Handle container from box,0,0,0.0,0.0,0.0
344
+ 342,Reach for items in box,0,6,0.0,0.0,0.0
345
+ 343,Pick up grocery item,0,6,0.0,0.0,0.0
346
+ 344,Place grocery item on shelf,0,10,0.0,0.0,0.0
347
+ 345,Clean and inspect shelf,0,10,0.0,0.0,0.0
348
+ 346,Prepare shelf for stocking,0,3,0.0,0.0,0.0
349
+ 347,Retrieve snack packs,0,0,0.0,0.0,0.0
350
+ 348,Clean shelf surface,0,0,0.0,0.0,0.0
351
+ 349,Organize inventory and reach for products,0,0,0.0,0.0,0.0
352
+ 350,Stocking snack packets on the shelf,0,0,0.0,0.0,0.0
353
+ 351,Sorting gift boxes into a bin,0,0,0.0,0.0,0.0
354
+ 352,Stocking gift boxes on the shelf,0,0,0.0,0.0,0.0
355
+ 353,Organizing snack pouches into containers,0,0,0.0,0.0,0.0
356
+ 354,Stocking snack pouches on the shelf,0,0,0.0,0.0,0.0
357
+ 355,Relocating storage bins along the aisle,0,0,0.0,0.0,0.0
358
+ 356,Clearing space on the shelf,0,0,0.0,0.0,0.0
359
+ 357,Rearranging containers on the shelf,0,0,0.0,0.0,0.0
360
+ 358,Stocking containers on the shelf,0,0,0.0,0.0,0.0
361
+ 359,Holding smartphone,28,3,0.0,0.0,0.0
362
+ 360,Preparing workspace,0,0,0.0,0.0,0.0
363
+ 361,Positioning newspaper and scissors,0,0,0.0,0.0,0.0
364
+ 362,Cutting newspaper,0,0,0.0,0.0,0.0
365
+ 363,Cutting and pausing,0,0,0.0,0.0,0.0
366
+ 364,Resuming cutting,0,0,0.0,0.0,0.0
367
+ 365,Cutting and adjusting scissors,0,0,0.0,0.0,0.0
368
+ 366,Resuming cutting newspaper,0,0,0.0,0.0,0.0
369
+ 367,Arrange tiles into row,0,0,0.0,0.0,0.0
370
+ 368,Adjust tile alignment,0,0,0.0,0.0,0.0
371
+ 369,Adjust tiles on stack,0,0,0.0,0.0,0.0
372
+ 370,Pick up and place tiles on stack,0,17,0.0,0.0,0.0
373
+ 371,Measuring and marking cardboard for cutting,0,7,0.0,0.0,0.0
374
+ 372,Interruption: handling charging case,0,0,0.0,0.0,0.0
375
+ 373,Hold paper strip,0,0,0.0,0.0,0.0
376
+ 374,Manipulate paper strip,44,0,0.0,0.0,0.0
377
+ 375,Cutting initial cardboard pieces,0,0,0.0,0.0,0.0
378
+ 376,Measuring and marking cardboard with ruler,0,0,0.0,0.0,0.0
379
+ 377,Marking cardboard and preparing to cut,0,0,0.0,0.0,0.0
380
+ 378,Aligning cardboard for cutting,0,0,0.0,0.0,0.0
381
+ 379,Cutting cardboard sheet,0,0,0.0,0.0,0.0
382
+ 380,Measuring and marking cardboard,0,0,0.0,0.0,0.0
383
+ 381,Cutting cardboard strips,0,0,0.0,0.0,0.0
384
+ 382,Final cardboard cutting,0,0,0.0,0.0,0.0
385
+ 383,Typing on smartphone,0,0,0.0,0.0,0.0
386
+ 384,Scrolling on smartphone,0,50,0.0,0.0,0.0
387
+ 385,Scrolling and tapping on smartphone,0,0,0.0,0.0,0.0
388
+ 386,Browsing photo gallery,0,0,0.0,0.0,0.0
389
+ 387,Typing message on smartphone,0,0,0.0,0.0,0.0
390
+ 388,Tapping and putting away smartphone,0,0,0.0,0.0,0.0
391
+ 389,Stop measuring and transition to smartphone usage,0,0,0.0,0.0,0.0
392
+ 390,Position ruler and draw initial lines,0,0,0.0,0.0,0.0
393
+ 391,Draw and reposition ruler,0,0,0.0,0.0,0.0
394
+ 392,Prepare for further marking,0,0,0.0,0.0,0.0
395
+ 393,Align ruler and draw line,0,0,0.0,0.0,0.0
396
+ 394,Mark lines on cardboard,0,0,0.0,0.0,0.0
397
+ 395,Remove tools after marking,0,0,0.0,0.0,0.0
398
+ 396,Final alignment and marking,0,0,0.0,0.0,0.0
399
+ 397,Walking to workstation,0,0,0.0,0.0,0.0
400
+ 398,Assembling cardboard base,0,0,0.0,0.0,0.0
401
+ 399,Marking cardboard measurements,0,0,0.0,0.0,0.0
402
+ 400,Marking and positioning cardboard,0,0,0.0,0.0,0.0
403
+ 401,Adjusting cardboard layout,0,0,0.0,0.0,0.0
404
+ 402,Cutting cardboard pieces,49,50,0.0,0.0,0.0
405
+ 403,Placing canned goods onto the shelf,0,0,0.0,0.0,0.0
406
+ 404,Placing canned goods onto the shelf and retrieving container,0,0,0.0,0.0,0.0
407
+ 405,Positioning container near the shelf,0,0,0.0,0.0,0.0
408
+ 406,Organizing container contents,0,0,0.0,0.0,0.0
409
+ 407,Placing and arranging cans on the shelf,0,0,0.0,0.0,0.0
410
+ 408,Adjusting canned goods on the shelf,0,0,0.0,0.0,0.0
411
+ 409,Stocking canned goods onto the shelf,0,0,0.0,0.0,0.0
412
+ 410,Wiping the plastic jar,0,0,0.0,0.0,0.0
413
+ 411,Inspecting the cleaned plastic jar,0,0,0.0,0.0,0.0
414
+ 412,Transitioning from jar to tin can,0,0,0.0,0.0,0.0
415
+ 413,Inspecting shelf contents,0,1,0.0,0.0,0.0
416
+ 414,Cutting and measuring cardboard,0,0,0.0,0.0,0.0
417
+ 415,Stabilizing cardboard for cutting,0,0,0.0,0.0,0.0
418
+ 416,Positioning ruler for cutting,0,0,0.0,0.0,0.0
419
+ 417,Labeling cardboard squares,0,0,0.0,0.0,0.0
420
+ 418,Labeling and organizing cardboard squares,0,0,0.0,0.0,0.0
421
+ 419,Labeling cardboard square,0,0,0.0,0.0,0.0
422
+ 420,Labeling and placing cardboard squares,0,2,0.0,0.0,0.0
423
+ 421,Labeling and placing cardboard square,0,5,0.0,0.0,0.0
424
+ 422,Labeling and switching markers,0,4,0.0,0.0,0.0
425
+ 423,Labeling and retrieving cardboard pieces,0,4,0.0,0.0,0.0
426
+ 424,Positioning cardboard pieces,0,0,0.0,0.0,0.0
427
+ 425,Adjusting and folding cardboard,0,4,0.0,0.0,0.0
428
+ 426,Observing workspace,0,1,0.0,0.0,0.0
429
+ 427,Arranging buttons on the table,0,0,0.0,0.0,0.0
430
+ 428,Arranging buttons,0,4,0.0,0.0,0.0
431
+ 429,Sorting buttons,0,0,0.0,0.0,0.0
432
+ 430,Sorting orange buttons,0,0,0.0,0.0,0.0
433
+ 431,Sorting orange button,0,0,0.0,0.0,0.0
434
+ 432,Moving hand over pile and sorting orange buttons,0,1,0.0,0.0,0.0
435
+ 433,Moving orange buttons,0,2,0.0,0.0,0.0
436
+ 434,Arranging orange buttons,0,9,0.0,0.0,0.0
437
+ 435,folding paper strips into stars,0,0,0.0,0.0,0.0
438
+ 436,folding paper strips and retrieving stapler,0,0,0.0,0.0,0.0
439
+ 437,Drawing grid lines with a ruler and pen,0,0,0.0,0.0,0.0
440
+ 438,Drawing grid lines with a pen,0,0,0.0,0.0,0.0
441
+ 439,Drawing grid lines and repositioning cardboard,0,0,0.0,0.0,0.0
442
+ 440,Drawing grid lines with a ruler,0,0,0.0,0.0,0.0
443
+ 441,Drawing grid lines,0,0,0.0,0.0,0.0
444
+ 442,folding paper stars and typing on smartphone,0,0,0.0,0.0,0.0
445
+ 443,typing on smartphone,0,0,0.0,0.0,0.0
446
+ 444,folding paper strip,0,0,0.0,0.0,0.0
447
+ 445,folding paper star,0,0,0.0,0.0,0.0
448
+ 446,manipulating paper star and using smartphone,0,0,0.0,0.0,0.0
449
+ 447,manipulating paper strip,0,0,0.0,0.0,0.0
450
+ 448,Observe workspace and reach for beads,0,0,0.0,0.0,0.0
451
+ 449,Sort purple beads,0,1,0.0,0.0,0.0
452
+ 450,Write on paper,0,0,0.0,0.0,0.0
453
+ 451,Gathering star beads,0,0,0.0,0.0,0.0
454
+ 452,Sort beads by hand,0,8,0.0,0.0,0.0
455
+ 453,Pick up and place canned goods on shelf,0,0,0.0,0.0,0.0
456
+ 454,Hold and carry tray of canned goods,0,0,0.0,0.0,0.0
457
+ 455,Move towards shelf and position tray,0,0,0.0,0.0,0.0
458
+ 456,Sort canned goods in tray,0,0,0.0,0.0,0.0
459
+ 457,Align canned goods on shelf,0,1,0.0,0.0,0.0
460
+ 458,Handle crate of cans,0,0,0.0,0.0,0.0
461
+ 459,Wiping and cleaning retail items,0,0,0.0,0.0,0.0
462
+ 460,Adjusting retail items on shelf,0,0,0.0,0.0,0.0
463
+ 461,Arranging items on shelf,0,0,0.0,0.0,0.0
464
+ 462,Folding and sorting paper stars while handling a water bottle,0,0,0.0,0.0,0.0
465
+ 463,Folding and sorting paper stars,0,4,0.0,0.0,0.0
466
+ 464,Organizing stars into a row,0,0,0.0,0.0,0.0
467
+ 465,Reaching for and picking up stars,0,0,0.0,0.0,0.0
468
+ 466,Manipulating paper stars,0,0,0.0,0.0,0.0
469
+ 467,Holding writing materials,0,0,0.0,0.0,0.0
470
+ 468,Recording star count,0,0,0.0,0.0,0.0
471
+ 469,Holding pen and paper,0,0,0.0,0.0,0.0
472
+ 470,Counting and recording stars,0,0,0.0,0.0,0.0
473
+ 471,Using smartphone,26,0,0.0,0.0,0.0
474
+ 472,Resuming recording star count,0,0,0.0,0.0,0.0
475
+ 473,Arranging paper stars,0,0,0.0,0.0,0.0
476
+ 474,Cutting and preparing cardboard pieces,0,0,0.0,0.0,0.0
477
+ 475,Cleaning up workspace and moving items,0,0,0.0,0.0,0.0
478
+ 476,Washing hands,0,1,0.0,0.0,0.0
479
+ 477,Assembling cardboard boxes,0,0,0.0,0.0,0.0
480
+ 478,Pick up and begin folding paper strip,0,0,0.0,0.0,0.0
481
+ 479,Form paper strip into a star,0,0,0.0,0.0,0.0
482
+ 480,Manipulate star and prepare next strip,0,0,0.0,0.0,0.0
483
+ 481,Folding paper strip into lucky star,0,0,0.0,0.0,0.0
484
+ 482,Folding paper strip,0,0,0.0,0.0,0.0
485
+ 483,Folding lucky star,0,0,0.0,0.0,0.0
486
+ 484,Manipulating paper star,0,0,0.0,0.0,0.0
487
+ 485,Manipulating paper strip,0,0,0.0,0.0,0.0
488
+ 486,Finalizing paper star,0,0,0.0,0.0,0.0
489
+ 487,Sorting small colored tiles,0,0,0.0,0.0,0.0
490
+ 488,Picking up and placing tiles,0,0,0.0,0.0,0.0
491
+ 489,Sorting tiles by color,0,0,0.0,0.0,0.0
492
+ 490,Sorting and counting beads,0,66,0.0,0.0,0.0
493
+ 491,Cut section from newspaper,0,0,0.0,0.0,0.0
494
+ 492,Tear newspaper,0,0,0.0,0.0,0.0
495
+ 493,Hold newspaper,0,0,0.0,0.0,0.0
496
+ 494,Align and fold newspaper,0,0,0.0,0.0,0.0
497
+ 495,Reposition and cut newspaper,0,0,0.0,0.0,0.0
498
+ 496,Cut newspaper,0,0,0.0,0.0,0.0
499
+ 497,Cut along the newspaper edge,0,9,0.0,0.0,0.0
500
+ 498,Browse mobile phone,0,30,0.0,0.0,0.0
501
+ 499,Browse mobile phone and cut newspaper,0,0,0.0,0.0,0.0
502
+ 500,Hold and cut newspaper with scissors,0,0,0.0,0.0,0.0
503
+ 501,Cut newspaper and place scissors on table,0,0,0.0,0.0,0.0
504
+ 502,Sorting small star-shaped plastic pieces,0,0,0.0,0.0,0.0
505
+ 503,Sorting and reaching for plastic pieces,0,0,0.0,0.0,0.0
506
+ 504,Sorting small plastic pieces,0,0,0.0,0.0,0.0
507
+ 505,Sorting pieces and writing on paper,0,0,0.0,0.0,0.0
508
+ 506,Sorting plastic pieces,0,0,0.0,0.0,0.0
509
+ 507,Gathering and boxing plastic pieces,0,0,0.0,0.0,0.0
510
+ 508,Boxing pieces and picking up phone,0,0,0.0,0.0,0.0
511
+ 509,Typing and navigating on phone,0,0,0.0,0.0,0.0
512
+ 510,Scrolling and viewing content on phone,0,0,0.0,0.0,0.0
513
+ 511,Selecting and bagging electronic accessories,0,0,0.0,0.0,0.0
514
+ 512,Picking up and bagging a charging cable,0,0,0.0,0.0,0.0
515
+ 513,Bagging a held electronic item,0,0,0.0,0.0,0.0
516
+ 514,Picking up and bagging an electronic item,0,2,0.0,0.0,0.0
517
+ 515,"Cleaning, inspecting, and bagging an electronic item",0,3,0.0,0.0,0.0
518
+ 516,Cleaning and bagging an electronic item,0,0,0.0,0.0,0.0
519
+ 517,Inspecting and bagging a smartphone box,0,0,0.0,0.0,0.0
520
+ 518,Holding and bagging a smartphone box,0,0,0.0,0.0,0.0
521
+ 519,Examining a product,0,0,0.0,0.0,0.0
522
+ 520,Picking and stocking items from container,0,20,0.0,0.0,0.0
523
+ 521,Stocking items and carrying container,0,12,0.0,0.0,0.0
524
+ 522,Moving container and stocking items,0,6,0.0,0.0,0.0
525
+ 523,Retrieving and shelving container,0,13,0.0,0.0,0.0
526
+ 524,Stocking items and reaching for container,0,151,0.0,0.0,0.0
527
+ 525,Placing container on floor,0,35,0.0,0.0,0.0
528
+ 526,Preparing container on shelf,0,47,0.0,0.0,0.0
529
+ 527,Organizing canned goods in container,0,17,0.0,0.0,0.0
530
+ 528,Stocking products from bin onto shelf,0,1,0.0,0.0,0.0
531
+ 529,Finalizing shelf placement and moving bin,0,55,0.0,0.0,0.0
532
+ 530,Walking along the aisle,0,0,0.0,0.0,0.0
533
+ 531,Moving plastic storage bin,0,0,0.0,0.0,0.0
534
+ 532,Collecting canned food into bin,0,0,0.0,0.0,0.0
535
+ 533,Placing canned food onto shelf,0,0,0.0,0.0,0.0
536
+ 534,Holding container of canned food,0,0,0.0,0.0,0.0
537
+ 535,Moving towards aisle,0,0,0.0,0.0,0.0
538
+ 536,Approaching restocking supplies,0,0,0.0,0.0,0.0
539
+ 537,Retrieving plastic container and moving to aisle,0,0,0.0,0.0,0.0
540
+ 538,Forming quilled paper shapes,0,0,0.0,0.0,0.0
541
+ 539,Manipulating and placing paper shapes,0,0,0.0,0.0,0.0
542
+ 540,Selecting and manipulating paper strips,0,0,0.0,0.0,0.0
543
+ 541,Transitioning and observing workspace,0,0,0.0,0.0,0.0
544
+ 542,Sorting quilled paper pieces,0,0,0.0,0.0,0.0
545
+ 543,Walking to storage area,0,0,0.0,0.0,0.0
546
+ 544,Returning to workspace,0,0,0.0,0.0,0.0
547
+ 545,Organizing paper pieces into piles,0,0,0.0,0.0,0.0
548
+ 546,Manipulating quilled paper,0,0,0.0,0.0,0.0
549
+ 547,Sorting pieces and retrieving smartphone,0,1,0.0,0.0,0.0
550
+ 548,Sorting cardboard pieces,0,0,0.0,0.0,0.0
551
+ 549,Scanning workspace,0,0,0.0,0.0,0.0
552
+ 550,Sorting and reaching for pieces,0,0,0.0,0.0,0.0
553
+ 551,Interacting with phone,0,12,0.0,0.0,0.0
554
+ 552,Sorting and stacking cardboard pieces,0,0,0.0,0.0,0.0
555
+ 553,Marking list and sorting beads,0,0,0.0,0.0,0.0
556
+ 554,Marking paper and adjusting piles,0,0,0.0,0.0,0.0
557
+ 555,Sorting blue beads and marking list,0,0,0.0,0.0,0.0
558
+ 556,Completing list and moving away from desk,0,0,0.0,0.0,0.0
559
+ 557,Walking through office,0,0,0.0,0.0,0.0
560
+ 558,Returning to table and placing controller,0,0,0.0,0.0,0.0
561
+ 559,Sorting blue beads and writing on paper,0,7,0.0,0.0,0.0
562
+ 560,Sorting beads and writing on paper,0,0,0.0,0.0,0.0
563
+ 561,Cutting and folding cardboard shapes,0,1,0.0,0.0,0.0
564
+ 562,Retrieving materials,0,0,0.0,0.0,0.0
565
+ 563,Cutting and adjusting cardboard pieces,0,17,0.0,0.0,0.0
566
+ 564,Repositioning and cutting cardboard pieces,0,2,0.0,0.0,0.0
567
+ 565,Rolling paper strips into coils,0,0,0.0,0.0,0.0
568
+ 566,Finishing coil and selecting new strip,0,4,0.0,0.0,0.0
569
+ 567,Positioning paper strip,0,24,0.0,0.0,0.0
570
+ 568,Manipulating tools and quilling materials,0,3,0.0,0.0,0.0
571
+ 569,Walking to and approaching workspace,0,0,0.0,0.0,0.0
572
+ 570,Interacting with coworker,0,1,0.0,0.0,0.0
573
+ 571,Walking through workspace,0,0,0.0,0.0,0.0
574
+ 572,Manipulating small object,0,45,0.0,0.0,0.0
575
+ 573,Manipulating paper quilling piece,0,4,0.0,0.0,0.0
576
+ 574,Holding and retrieving paper strips,0,0,0.0,0.0,0.0
577
+ 575,Holding quilled paper piece,0,0,0.0,0.0,0.0
578
+ 576,Holding and aligning paper strip,0,2,0.0,0.0,0.0
579
+ 577,Holding and rotating paper strip,0,23,0.0,0.0,0.0
580
+ 578,Hold and mark cardboard piece,0,0,0.0,0.0,0.0
581
+ 579,Marking cardboard pieces,0,0,0.0,0.0,0.0
582
+ 580,Placing cardboard piece,0,30,0.0,0.0,0.0
583
+ 581,Walking towards workstation,0,0,0.0,0.0,0.0
584
+ 582,Preparing workstation,0,0,0.0,0.0,0.0
585
+ 583,Gathering items,0,0,0.0,0.0,0.0
586
+ 584,Gathering colored beads,0,0,0.0,0.0,0.0
587
+ 585,Sorting star-shaped objects by color,0,18,0.0,0.0,0.0
588
+ 586,Sorting star-shaped objects,0,0,0.0,0.0,0.0
589
+ 587,Sorting yellow star-shaped objects,0,32,0.0,0.0,0.0
590
+ 588,Sorting purple star-shaped objects,0,0,0.0,0.0,0.0
591
+ 589,Grasp paper strip,0,0,0.0,0.0,0.0
592
+ 590,Fold paper star,0,0,0.0,0.0,0.0
593
+ 591,Cleaning workspace,0,9,0.0,0.0,0.0
594
+ 592,Reaching for utility knife,0,0,0.0,0.0,0.0
595
+ 593,Using smartphone as a guide on cardboard,0,1,0.0,0.0,0.0
596
+ 594,Manipulating quilling strips,0,0,0.0,0.0,0.0
597
+ 595,Collecting oil and supplement bottles,0,0,0.0,0.0,0.0
598
+ 596,Observing and reaching for products,0,0,0.0,0.0,0.0
599
+ 597,Place down paper strip,0,8,0.0,0.0,0.0
600
+ 598,Sort and combine bead piles,0,0,0.0,0.0,0.0
601
+ 599,Mark fabric and reposition ruler,0,0,0.0,0.0,0.0
602
+ 600,Move empty cardboard boxes,0,10,0.0,0.0,0.0
603
+ 601,Finishing segment and placing scissors,0,0,0.0,0.0,0.0
604
+ 602,Draw line on cardboard,0,0,0.0,0.0,0.0
605
+ 603,manipulating paper star,0,0,0.0,0.0,0.0
606
+ 604,Using and placing phone,0,0,0.0,0.0,0.0
607
+ 605,Holding an electronic item,0,0,0.0,0.0,0.0
608
+ 606,Restocking pineapple chips,0,0,0.0,0.0,0.0
609
+ 607,Sorting and arranging cardboard pieces,0,0,0.0,0.0,0.0
610
+ 608,Marking list and moving blue beads,0,0,0.0,0.0,0.0
611
+ 609,Sorting blue beads,0,2,0.0,0.0,0.0
612
+ 610,Cleaning shelves and rearranging products,0,0,0.0,0.0,0.0
613
+ 611,Holding and positioning item on shelf,0,3,0.0,0.0,0.0
614
+ 612,Placing down scissors,0,0,0.0,0.0,0.0
615
+ 613,Cutting cardboard and putting down scissors,0,0,0.0,0.0,0.0
616
+ 614,Cutting and stacking cardboard pieces,0,0,0.0,0.0,0.0
617
+ 615,Transitioning from utility knife to scissors,0,0,0.0,0.0,0.0
618
+ 616,Placing items on shelf and moving box,0,1,0.0,0.0,0.0
619
+ 617,Managing shopping container,0,0,0.0,0.0,0.0
620
+ 618,Gathering materials and walking,0,0,0.0,0.0,0.0
621
+ 619,Manipulate paper piece,0,0,0.0,0.0,0.0
622
+ 620,Cutting cardboard tube and setting aside scissors,0,0,0.0,0.0,0.0
623
+ 621,Placing and searching for puzzle pieces,0,5,0.0,0.0,0.0
624
+ 622,Marking lines and repositioning ruler,0,0,0.0,0.0,0.0
625
+ 623,Position tray and reach for beads,0,0,0.0,0.0,0.0
626
+ 624,Gather and hold beads,0,0,0.0,0.0,0.0
627
+ 625,Adjust phone and reach for beads,0,0,0.0,0.0,0.0
628
+ 626,Cut light green fabric and reposition scissors,0,0,0.0,0.0,0.0
629
+ 627,Cutting and picking up new cardboard,0,0,0.0,0.0,0.0
630
+ 628,Cutting and gathering cardboard,0,0,0.0,0.0,0.0
631
+ 629,Place product on shelf,0,1,0.0,0.0,0.0
632
+ 630,Placing snack packages in box,0,0,0.0,0.0,0.0
633
+ 631,Placing and adjusting items on the shelf,0,0,0.0,0.0,0.0
634
+ 632,Transferring items from shelf to bag,0,1,0.0,0.0,0.0
635
+ 633,Adjusting cardboard divider,0,0,0.0,0.0,0.0
636
+ 634,Finish and place origami star,0,0,0.0,0.0,0.0
637
+ 635,Cutting cardboard and placing down scissors,0,0,0.0,0.0,0.0
638
+ 636,Sort Mahjong tiles,0,0,0.0,0.0,0.0
639
+ 637,Draw grid lines,0,0,0.0,0.0,0.0
640
+ 638,Stacking and organizing cardboard,0,0,0.0,0.0,0.0
641
+ 639,Moving orange buttons and interacting with smartphone,0,0,0.0,0.0,0.0
642
+ 640,typing on smartphone and picking up paper strip,0,0,0.0,0.0,0.0
643
+ 641,folding paper strip and typing on smartphone,0,0,0.0,0.0,0.0
644
+ 642,using smartphone and picking up paper strip,0,0,0.0,0.0,0.0
645
+ 643,Sorting paper stars,0,0,0.0,0.0,0.0
646
+ 644,Complete folding and place star on table,0,0,0.0,0.0,0.0
647
+ 645,"Cleaning, bagging, and selecting another item",0,0,0.0,0.0,0.0
648
+ 646,Positioning the container on the floor,0,3,0.0,0.0,0.0
649
+ 647,Stocking products and reorganizing bin,0,0,0.0,0.0,0.0
650
+ 648,Observing restocking needs,0,0,0.0,0.0,0.0
651
+ 649,Writing on paper and sorting blue beads,0,0,0.0,0.0,0.0
652
+ 650,Writing on paper and reaching for beads,0,0,0.0,0.0,0.0
653
+ 651,Wipe shelf and retrieve canned food,0,0,0.0,0.0,0.0
654
+ 652,Place canned food on shelf,0,0,0.0,0.0,0.0
655
+ 653,Examine yellow item,0,0,0.0,0.0,0.0
656
+ 654,Hold blue product box,0,0,0.0,0.0,0.0
657
+ 655,Wipe shelf and pick up product,0,0,0.0,0.0,0.0
658
+ 656,Inspect product,0,0,0.0,0.0,0.0
659
+ 657,Clean shelf and stock product,0,0,0.0,0.0,0.0
660
+ 658,Retrieve product from box,0,0,0.0,0.0,0.0
661
+ 659,Wipe and transport product to shelf,0,0,0.0,0.0,0.0
662
+ 660,Wipe and select ketchup bottle,0,0,0.0,0.0,0.0
663
+ 661,Wipe and place ketchup bottle on shelf,0,0,0.0,0.0,0.0
664
+ 662,Inspecting cardboard,0,0,0.0,0.0,0.0
665
+ 663,Moving through workspace,0,0,0.0,0.0,0.0
666
+ 664,Transitioning to cutting,0,0,0.0,0.0,0.0
667
+ 665,Cutting and folding cardboard,0,0,0.0,0.0,0.0
668
+ 666,Folding cardboard,0,0,0.0,0.0,0.0
669
+ 667,Using a smartphone,28,0,0.0,0.0,0.0
670
+ 668,Tidying workspace,20,0,0.0,0.0,0.0
671
+ 669,Handling container lid,25,0,0.0,0.0,0.0
672
+ 670,Approaching the stove,9,0,0.0,0.0,0.0
673
+ 671,Checking cooking pot,20,0,0.0,0.0,0.0
674
+ 672,Moving around the kitchen,22,0,0.0,0.0,0.0
675
+ 673,Wiping counter,16,0,0.0,0.0,0.0
676
+ 674,Cleaning cloth maintenance,25,0,0.0,0.0,0.0
677
+ 675,Retrieving cleaning supplies,32,0,0.0,0.0,0.0
678
+ 676,Cleaning kitchen surfaces,28,0,0.0,0.0,0.0
679
+ 677,Cooking at the stove,32,0,0.0,0.0,0.0
680
+ 678,Adjusting cookware,23,0,0.0,0.0,0.0
681
+ 679,Prepare to cut cardboard,49,0,0.0,0.0,0.0
682
+ 680,Cut cardboard,172,0,0.0,0.0,0.0
683
+ 681,Cut along the marked line,51,0,0.0,0.0,0.0
684
+ 682,Picking and placing items from bin onto shelves,0,0,0.0,0.0,0.0
685
+ 683,Retrieving items from bin,0,0,0.0,0.0,0.0
686
+ 684,Placing held items on shelf,0,0,0.0,0.0,0.0
687
+ 685,Monitoring task progress via smartwatch,0,0,0.0,0.0,0.0
688
+ 686,Placing jar on shelf,0,0,0.0,0.0,0.0
689
+ 687,Stocking jars on shelf,0,0,0.0,0.0,0.0
690
+ 688,Stocking sauce bottles on shelf,0,0,0.0,0.0,0.0
691
+ 689,Stocking miscellaneous products on shelf,0,0,0.0,0.0,0.0
692
+ 690,Handling and organizing containers,0,0,0.0,0.0,0.0
693
+ 691,Assessing shelf arrangement,0,0,0.0,0.0,0.0
694
+ 692,Assembling the foam base loop,0,0,0.0,0.0,0.0
695
+ 693,Attaching decorative yellow pieces,0,0,0.0,0.0,0.0
696
+ 694,Tearing and preparing blue foam pieces,0,0,0.0,0.0,0.0
697
+ 695,Attaching and folding blue foam strips,0,0,0.0,0.0,0.0
698
+ 696,Interlocking the craft strips,0,0,0.0,0.0,0.0
699
+ 697,Reviewing craft instructions,0,0,0.0,0.0,0.0
700
+ 698,Finalizing the craft assembly,0,0,0.0,0.0,0.0
701
+ 699,Retrieving additional supplies,0,0,0.0,0.0,0.0
702
+ 700,Walk to shelf location,18,0,0.0,0.0,0.0
703
+ 701,Inspect shelf condition,27,0,0.0,0.0,0.0
704
+ 702,Approach inventory boxes,34,0,0.0,0.0,0.0
705
+ 703,Extract wire hangers from inventory,43,0,0.0,0.0,0.0
706
+ 704,Bundle display hooks,22,0,0.0,0.0,0.0
707
+ 705,Install display hooks,18,0,0.0,0.0,0.0
708
+ 706,Move to stocking area,10,0,0.0,0.0,0.0
709
+ 707,Unpack and place items on shelf,29,0,0.0,0.0,0.0
710
+ 708,Place and adjust items on shelf,42,0,0.0,0.0,0.0
711
+ 709,Place and retrieve items for stocking,27,0,0.0,0.0,0.0
712
+ 710,Labeling cardboard pieces,196,0,0.0,0.0,0.0
713
+ 711,Observing and pausing,17,0,0.0,0.0,0.0
714
+ 712,Browsing and selecting canned goods,0,0,0.0,0.0,0.0
715
+ 713,Selecting and picking up canned goods,0,0,0.0,0.0,0.0
716
+ 714,Returning canned goods to shelf,0,0,0.0,0.0,0.0
717
+ 715,Moving along the shelves,0,0,0.0,0.0,0.0
718
+ 716,Inspecting Dior gift box,0,0,0.0,0.0,0.0
719
+ 717,Selecting a bottle,0,0,0.0,0.0,0.0
720
+ 718,Comparing and replacing bottles,0,0,0.0,0.0,0.0
721
+ 719,Inspecting bottle,0,0,0.0,0.0,0.0
722
+ 720,Inspecting almond package and scanning shelves,0,0,0.0,0.0,0.0
723
+ 721,Moving along the aisle,0,0,0.0,0.0,0.0
724
+ 722,Touching canned goods,0,0,0.0,0.0,0.0
725
+ 723,Cutting triangular cardboard pieces,0,0,0.0,0.0,0.0
726
+ 724,Manipulating cardboard shapes,0,0,0.0,0.0,0.0
727
+ 725,Folding cardboard and checking phone,0,0,0.0,0.0,0.0
728
+ 726,Moving cardboard pieces across the workspace,0,0,0.0,0.0,0.0
729
+ 727,Trimming and placing cardboard pieces,0,0,0.0,0.0,0.0
730
+ 728,Placing canned goods on the shelf,63,0,0.0,0.0,0.0
731
+ 729,Reaching into the box for more stock,33,0,0.0,0.0,0.0
732
+ 730,Picking up and placing canned goods,17,0,0.0,0.0,0.0
733
+ 731,Placing and retrieving canned goods,19,0,0.0,0.0,0.0
734
+ 732,Aligning canned goods on the shelf,9,0,0.0,0.0,0.0
735
+ 733,Stocking multiple cans on the shelf,35,0,0.0,0.0,0.0
736
+ 734,Adjusting stock and finishing placement,17,0,0.0,0.0,0.0
737
+ 735,Finalizing shelf organization,27,0,0.0,0.0,0.0
738
+ 736,Handling earbud case,24,0,0.0,0.0,0.0
739
+ 737,Checking smartphone,24,0,0.0,0.0,0.0
740
+ 738,Sort craft materials into piles,36,0,0.0,0.0,0.0
741
+ 739,Manipulate craft pieces,38,0,0.0,0.0,0.0
742
+ 740,Release scissors and operate smartphone,44,0,0.0,0.0,0.0
743
+ 741,Operate and release smartphone,7,0,0.0,0.0,0.0
744
+ 742,Manipulate and release paper strips,34,0,0.0,0.0,0.0
745
+ 743,Sort small craft pieces,39,0,0.0,0.0,0.0
746
+ 744,Placing items on the shelf,0,0,0.0,0.0,0.0
747
+ 745,Holding and organizing product packages,0,0,0.0,0.0,0.0
748
+ 746,Placing items and checking phone,0,0,0.0,0.0,0.0
749
+ 747,Handling charging cables,0,0,0.0,0.0,0.0
750
+ 748,Retrieving items from bag,0,0,0.0,0.0,0.0
751
+ 749,Adjusting items and reaching for stock,0,0,0.0,0.0,0.0
752
+ 750,Retrieving and examining items from bag,0,0,0.0,0.0,0.0
753
+ 751,Removing items from bag and stocking them,0,0,0.0,0.0,0.0
754
+ 752,Reorganizing stock on shelf,0,0,0.0,0.0,0.0
755
+ 753,Folding purple ribbon,0,0,0.0,0.0,0.0
756
+ 754,Positioning ribbon piece,0,0,0.0,0.0,0.0
757
+ 755,Manipulating ribbon piece,0,0,0.0,0.0,0.0
758
+ 756,Placing ribbon onto project,0,0,0.0,0.0,0.0
759
+ 757,Folding and shaping ribbon,0,0,0.0,0.0,0.0
760
+ 758,Forming ribbon knot,0,0,0.0,0.0,0.0
761
+ 759,Securing ribbon with needle,0,0,0.0,0.0,0.0
762
+ 760,Unfold paper lantern,16,0,0.0,0.0,0.0
763
+ 761,Fold and grasp lantern,24,0,0.0,0.0,0.0
764
+ 762,Grasp lantern component,15,0,0.0,0.0,0.0
765
+ 763,Open paper lantern,13,0,0.0,0.0,0.0
766
+ 764,Align paper lantern edges,29,0,0.0,0.0,0.0
767
+ 765,Adjust lantern string and handle components,28,0,0.0,0.0,0.0
768
+ 766,Handle paper lantern component,19,0,0.0,0.0,0.0
769
+ 767,Expand and adjust lantern shape,24,0,0.0,0.0,0.0
770
+ 768,Secure lantern with adhesive,33,0,0.0,0.0,0.0
771
+ 769,Unpack additional lantern component,16,0,0.0,0.0,0.0
772
+ 770,Remove packaging and prepare component,32,0,0.0,0.0,0.0
773
+ 771,Expand paper lantern,22,0,0.0,0.0,0.0
774
+ 772,Final alignment of lantern edges,6,0,0.0,0.0,0.0
775
+ 773,Measure and mark cardboard with ruler,0,0,0.0,0.0,0.0
776
+ 774,Mark cardboard with marker,0,0,0.0,0.0,0.0
777
+ 775,Position utility knife and begin cutting,0,0,0.0,0.0,0.0
778
+ 776,Cut and release cardboard,0,0,0.0,0.0,0.0
779
+ 777,Cut and fold cardboard,0,0,0.0,0.0,0.0
780
+ 778,Cut and reposition utility knife,0,0,0.0,0.0,0.0
781
+ 779,Cut and tear off cardboard segment,0,0,0.0,0.0,0.0
782
+ 780,Holding a smartphone,0,0,0.0,0.0,0.0
783
+ 781,Browsing smartphone content,0,0,0.0,0.0,0.0
784
+ 782,Assembling small decorative components,0,0,0.0,0.0,0.0
785
+ 783,Manipulating components on a strip,0,0,0.0,0.0,0.0
786
+ 784,Manipulating and placing strip components,0,0,0.0,0.0,0.0
787
+ 785,Final component assembly,0,0,0.0,0.0,0.0
788
+ 786,Preparing craft area,24,0,0.0,0.0,0.0
789
+ 787,Transitioning to smartphone usage,59,0,0.0,0.0,0.0
790
+ 788,Browsing smartphone,32,0,0.0,0.0,0.0
791
+ 789,Scrolling smartphone screen,31,0,0.0,0.0,0.0
792
+ 790,Scrolling and setting down smartphone,24,0,0.0,0.0,0.0
793
+ 791,Scrolling and placing smartphone down,21,0,0.0,0.0,0.0
794
+ 792,Paper quilling craft,32,0,0.0,0.0,0.0
795
+ 793,Counting and recording paper stars,0,0,0.0,0.0,0.0
796
+ 794,Handling electronic device,0,0,0.0,0.0,0.0
797
+ 795,Reviewing and organizing records,0,0,0.0,0.0,0.0
798
+ 796,Cutting and adjusting cardboard,0,0,0.0,0.0,0.0
799
+ 797,Cutting cardboard and retrieving power bank,0,0,0.0,0.0,0.0
800
+ 798,Cutting and repositioning cardboard,0,0,0.0,0.0,0.0
801
+ 799,Arranging and marking cardboard strips,0,0,0.0,0.0,0.0
802
+ 800,Pick up and place puzzle piece,31,0,0.0,0.0,0.0
803
+ 801,Manipulate puzzle piece,38,0,0.0,0.0,0.0
804
+ 802,Manipulate puzzle pieces,35,0,0.0,0.0,0.0
805
+ 803,Place puzzle piece,16,0,0.0,0.0,0.0
806
+ 804,Observe puzzle progress,32,0,0.0,0.0,0.0
807
+ 805,Search for and pick up puzzle piece,26,0,0.0,0.0,0.0
808
+ 806,Release and adjust puzzle piece,15,0,0.0,0.0,0.0
809
+ 807,"Reach for, pick up, and attempt to fit puzzle piece",53,0,0.0,0.0,0.0
810
+ 808,Sort puzzle pieces,34,0,0.0,0.0,0.0
811
+ 809,Walking to the crafting area,17,0,0.0,0.0,0.0
812
+ 810,Preparing materials,31,0,0.0,0.0,0.0
813
+ 811,Assembling material pieces,32,0,0.0,0.0,0.0
814
+ 812,Manipulating yellow strips,31,0,0.0,0.0,0.0
815
+ 813,Working with paper strips,76,0,0.0,0.0,0.0
816
+ 814,Manipulating beads,45,0,0.0,0.0,0.0
817
+ 815,Using phone and resuming work,14,0,0.0,0.0,0.0
818
+ 816,Holding and adjusting cardboard,0,0,0.0,0.0,0.0
819
+ 817,Folding and positioning cardboard,0,0,0.0,0.0,0.0
820
+ 818,Completing marking and stepping away,0,0,0.0,0.0,0.0
821
+ 819,Returning to workstation,0,0,0.0,0.0,0.0
822
+ 820,Folding cardboard edge,0,0,0.0,0.0,0.0
823
+ 821,Folding cardboard and preparing marker,0,0,0.0,0.0,0.0
824
+ 822,Adjusting edge and marking cardboard,0,0,0.0,0.0,0.0
825
+ 823,Sorting and collecting cardboard squares,0,0,0.0,0.0,0.0
826
+ 824,Transporting cardboard to collection area,0,0,0.0,0.0,0.0
827
+ 825,Depositing cardboard squares,0,0,0.0,0.0,0.0
828
+ 826,Returning to work table,0,0,0.0,0.0,0.0
829
+ 827,Preparing and folding cardboard,0,0,0.0,0.0,0.0
830
+ 828,Cutting cardboard into strips,0,0,0.0,0.0,0.0
831
+ 829,Cutting and sorting cardboard squares,0,0,0.0,0.0,0.0
832
+ 830,Scoring and cutting cardboard,0,0,0.0,0.0,0.0
833
+ 831,Refining cardboard cuts,0,0,0.0,0.0,0.0
834
+ 832,Cutting and placing cardboard squares,0,0,0.0,0.0,0.0
835
+ 833,Sort buttons by color,25,0,0.0,0.0,0.0
836
+ 834,Arrange buttons,33,0,0.0,0.0,0.0
837
+ 835,Arrange buttons in a line,29,0,0.0,0.0,0.0
838
+ 836,Sort and arrange buttons,32,0,0.0,0.0,0.0
839
+ 837,Pick up and place buttons,40,0,0.0,0.0,0.0
840
+ 838,Sort buttons,36,0,0.0,0.0,0.0
841
+ 839,Sort and adjust button line,29,0,0.0,0.0,0.0
842
+ 840,Use smartphone,30,0,0.0,0.0,0.0
843
+ 841,Sort and place buttons,31,0,0.0,0.0,0.0
844
+ 842,Entering the training area,35,0,0.0,0.0,0.0
845
+ 843,Greeting participants,49,0,0.0,0.0,0.0
846
+ 844,Moving through the room,20,0,0.0,0.0,0.0
847
+ 845,Initiating assembly,34,0,0.0,0.0,0.0
848
+ 846,Handling plastic strip,28,0,0.0,0.0,0.0
849
+ 847,Bending plastic strip,16,0,0.0,0.0,0.0
850
+ 848,Manipulating plastic strip,37,0,0.0,0.0,0.0
851
+ 849,Folding plastic strip,57,0,0.0,0.0,0.0
852
+ 850,Trimming and stacking cardboard pieces,0,0,0.0,0.0,0.0
853
+ 851,Cutting and adjusting cardboard sheet,0,0,0.0,0.0,0.0
854
+ 852,Positioning and cutting cardboard piece,0,0,0.0,0.0,0.0
855
+ 853,Taking a break to drink water and check phone,0,0,0.0,0.0,0.0
856
+ 854,Moving away from and returning to the table,0,0,0.0,0.0,0.0
857
+ 855,Manipulate paper decoration,41,0,0.0,0.0,0.0
858
+ 856,Manipulate paper edge,35,0,0.0,0.0,0.0
859
+ 857,Adjusting paper edge and placing strip,43,0,0.0,0.0,0.0
860
+ 858,Adjusting and securing paper structure,36,0,0.0,0.0,0.0
861
+ 859,Manipulate adhesive strip,44,0,0.0,0.0,0.0
862
+ 860,Secure paper edges with adhesive,39,0,0.0,0.0,0.0
863
+ 861,Sort and record bead counts,49,0,0.0,0.0,0.0
864
+ 862,Sort and group beads,22,0,0.0,0.0,0.0
865
+ 863,"Pick, place, and count beads",33,0,0.0,0.0,0.0
866
+ 864,Sort beads and record count,75,0,0.0,0.0,0.0
867
+ 865,Count beads and retrieve more,28,0,0.0,0.0,0.0
868
+ 866,Document bead counts,22,0,0.0,0.0,0.0
869
+ 867,Organize and count beads,56,0,0.0,0.0,0.0
870
+ 868,Finish cutting along the marked line and reposition,7,0,0.0,0.0,0.0
871
+ 869,Checking smartwatch while reaching for product,0,0,0.0,0.0,0.0
872
+ 870,Positioning ruler and drawing lines,0,0,0.0,0.0,0.0
873
+ 871,Finalize wiping and placement of ketchup bottle,0,0,0.0,0.0,0.0
874
+ 872,Positioning utility knife,0,0,0.0,0.0,0.0
875
+ 873,Setting up smartphone,3,0,0.0,0.0,0.0
876
+ 874,Placing items on shelf and retrieving new items,0,0,0.0,0.0,0.0
877
+ 875,Inspecting and stocking shelf items,0,0,0.0,0.0,0.0
878
+ 876,Organizing products into box,0,0,0.0,0.0,0.0
879
+ 877,Inspect shelf condition and observe surroundings,3,0,0.0,0.0,0.0
880
+ 878,Place items on shelf,6,0,0.0,0.0,0.0
881
+ 879,Returning Dior gift box to shelf,0,0,0.0,0.0,0.0
882
+ 880,Picking up canned goods,0,0,0.0,0.0,0.0
883
+ 881,Cutting and releasing cardboard shapes,0,0,0.0,0.0,0.0
884
+ 882,Retrieving and placing canned goods,7,0,0.0,0.0,0.0
885
+ 883,Placing remaining items on the shelf,0,0,0.0,0.0,0.0
886
+ 884,Attempt to fit and place puzzle piece,5,0,0.0,0.0,0.0
887
+ 885,Working with paper strips and using phone,3,0,0.0,0.0,0.0
888
+ 886,Finalizing cuts and storing tools,0,0,0.0,0.0,0.0
889
+ 887,Sorting squares and preparing for next cut,0,0,0.0,0.0,0.0
890
+ 888,Sorting and processing cardboard strips,0,0,0.0,0.0,0.0
891
+ 889,Preparing plastic strip for folding,9,0,0.0,0.0,0.0
892
+ 890,Final trimming of cardboard,0,0,0.0,0.0,0.0
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/next_subtask_forecast/predictions.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_relevance/metrics.json CHANGED
@@ -4,6 +4,7 @@
4
  "task_display_name": "Object Relevance Prediction",
5
  "model_family": "neural_mlp_metadata_multilabel",
6
  "source": "128_episode_qwen_jsonl_metadata",
 
7
  "num_train_windows": 25629,
8
  "num_val_windows": 4608,
9
  "num_test_windows": 4032,
 
4
  "task_display_name": "Object Relevance Prediction",
5
  "model_family": "neural_mlp_metadata_multilabel",
6
  "source": "128_episode_qwen_jsonl_metadata",
7
+ "input_features": "frame/context metadata plus hashed prompt/options/main_task text; answer_json fields are excluded from inputs",
8
  "num_train_windows": 25629,
9
  "num_val_windows": 4608,
10
  "num_test_windows": 4032,
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/metrics.json ADDED
@@ -0,0 +1,50 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "status": "pass",
3
+ "task": "object_set_forecast",
4
+ "task_display_name": "Object Set Forecast",
5
+ "model_family": "neural_mlp_metadata_multilabel",
6
+ "source": "128_episode_qwen_jsonl_metadata",
7
+ "input_features": "current-window frame/context metadata plus hashed prompt/options/main_task text; target object set +100 frames",
8
+ "num_train_windows": 25068,
9
+ "num_val_windows": 4496,
10
+ "num_test_windows": 3951,
11
+ "num_objects": 256,
12
+ "history": [
13
+ {
14
+ "epoch": 1,
15
+ "loss": 0.40618257693683624
16
+ },
17
+ {
18
+ "epoch": 7,
19
+ "loss": 0.09393147091218229
20
+ },
21
+ {
22
+ "epoch": 14,
23
+ "loss": 0.06462562757330076
24
+ },
25
+ {
26
+ "epoch": 21,
27
+ "loss": 0.051652751573515565
28
+ },
29
+ {
30
+ "epoch": 28,
31
+ "loss": 0.04472079514406202
32
+ },
33
+ {
34
+ "epoch": 35,
35
+ "loss": 0.040127100790633106
36
+ }
37
+ ],
38
+ "device": "cuda",
39
+ "splits": {
40
+ "test": {
41
+ "precision": 0.12744893705710714,
42
+ "recall": 0.27504779039694144,
43
+ "micro_f1": 0.17418550827844048,
44
+ "macro_f1": 0.02176780675052154,
45
+ "exact_match": 0.019994937990382183
46
+ }
47
+ },
48
+ "primary_metric": "micro_f1",
49
+ "primary_score": 0.17418550827844048
50
+ }
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/object_set_forecast/predictions.csv ADDED
The diff for this file is too large to render. See raw diff
 
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/metrics.json ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "status": "pass",
3
+ "task": "time_to_transition",
4
+ "task_display_name": "Time To Transition",
5
+ "model_family": "neural_mlp_metadata_regressor",
6
+ "source": "128_episode_qwen_jsonl_metadata",
7
+ "input_features": "frame/context metadata plus hashed prompt/options/main_task text; target is frames to next action boundary capped at 200",
8
+ "split_policy": "train neural regressor on train split, report held-out test timing error",
9
+ "num_train_windows": 25629,
10
+ "num_test_windows": 4032,
11
+ "history": [
12
+ {
13
+ "epoch": 1,
14
+ "loss": 0.8395254544245428
15
+ },
16
+ {
17
+ "epoch": 7,
18
+ "loss": 0.4669298557358576
19
+ },
20
+ {
21
+ "epoch": 14,
22
+ "loss": 0.2693304531382335
23
+ },
24
+ {
25
+ "epoch": 21,
26
+ "loss": 0.19281703676294232
27
+ },
28
+ {
29
+ "epoch": 28,
30
+ "loss": 0.15983608896562138
31
+ },
32
+ {
33
+ "epoch": 35,
34
+ "loss": 0.13736773695386975
35
+ }
36
+ ],
37
+ "device": "cuda",
38
+ "splits": {
39
+ "test": {
40
+ "mae": 41.4664421081543,
41
+ "rmse": 55.98573684692383,
42
+ "r2": -0.24839289780206042
43
+ }
44
+ },
45
+ "primary_metric": "mae",
46
+ "metric_direction": "lower",
47
+ "primary_score": 41.4664421081543
48
+ }
results/omni_finetune/a100_128_metadata_task_baselines_20260616_v2/neural_mlp/time_to_transition/predictions.csv ADDED
The diff for this file is too large to render. See raw diff
 
scripts/omni/run_128_task_baselines.py CHANGED
@@ -7,9 +7,9 @@ The 128-episode public package intentionally does not redistribute raw sensor
7
  feature NPZ files, so this runner uses only public-safe JSONL metadata and
8
  strictly avoids answer fields as input features.
9
 
10
- The output keeps the same twelve task IDs. Tasks with enough JSONL signal get
11
  simple and, where appropriate, neural baselines over the same train/val/test
12
- episode split used by the Qwen3-Omni pilot. Tasks whose original target
13
  requires missing raw motion/depth/audio feature blocks are emitted as explicit
14
  unsupported records instead of fabricated scores.
15
  """
@@ -49,6 +49,14 @@ TASKS = [
49
  "modality_reconstruction",
50
  "temporal_order",
51
  "misalignment_detection",
 
 
 
 
 
 
 
 
52
  ]
53
 
54
  CLASSIFICATION_TASKS = {
@@ -76,6 +84,18 @@ UNSUPPORTED_TASKS = {
76
  "primary_metric": "f1",
77
  "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
78
  },
 
 
 
 
 
 
 
 
 
 
 
 
79
  }
80
 
81
  DEFAULT_DATASET = ROOT / "tmp/omni_128_dataset_fetch/dataset.jsonl"
@@ -122,6 +142,14 @@ def parse_args() -> argparse.Namespace:
122
  parser.add_argument("--neural-dropout", type=float, default=0.10)
123
  parser.add_argument("--neural-device", default="auto", choices=["auto", "cpu", "cuda"])
124
  parser.add_argument("--max-object-vocab", type=int, default=256)
 
 
 
 
 
 
 
 
125
  return parser.parse_args()
126
 
127
 
@@ -378,6 +406,62 @@ def split_indices(rows: list[dict[str, Any]]) -> dict[str, np.ndarray]:
378
  return {key: np.asarray(value, dtype=np.int64) for key, value in indices.items()}
379
 
380
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
381
  def encode_labels(values: list[str]) -> tuple[np.ndarray, list[str]]:
382
  seen: OrderedDict[str, int] = OrderedDict()
383
  for value in values:
@@ -654,6 +738,19 @@ def neural_classification(
654
  from neural_task_models import train_classifier
655
 
656
  out_dir.mkdir(parents=True, exist_ok=True)
 
 
 
 
 
 
 
 
 
 
 
 
 
657
  result = train_classifier(
658
  X.astype(np.float32),
659
  y,
@@ -745,21 +842,24 @@ def multilabel_metrics(y_true: np.ndarray, y_pred: np.ndarray) -> dict[str, floa
745
 
746
 
747
  def simple_multilabel(
 
748
  rows: list[dict[str, Any]],
 
749
  feature_rows: list[dict[str, Any]],
750
  X: np.ndarray,
751
  splits: dict[str, np.ndarray],
752
  out_root: Path,
753
  args: argparse.Namespace,
 
754
  ) -> dict[str, Any]:
755
  train_objects = Counter()
756
  for idx in splits["train"]:
757
- for obj in answer(rows[int(idx)]).get("objects", []) or []:
758
  key = norm(obj).lower()
759
  if key:
760
  train_objects[key] += 1
761
  vocab = [name for name, _count in train_objects.most_common(args.max_object_vocab)]
762
- Y = object_matrix(rows, vocab)
763
  train_idx, val_idx, test_idx = splits["train"], splits["val"], splits["test"]
764
  freq = Y[train_idx].mean(axis=0)
765
  # Public-safe simple baseline: predict object labels seen in at least 10% of
@@ -768,7 +868,7 @@ def simple_multilabel(
768
  pred_all = np.tile((freq >= 0.10).astype(np.float32), (len(rows), 1))
769
  test_metrics = multilabel_metrics(Y[test_idx], pred_all[test_idx])
770
  val_metrics = multilabel_metrics(Y[val_idx], pred_all[val_idx]) if len(val_idx) else {}
771
- out_dir = out_root / "object_relevance"
772
  out_dir.mkdir(parents=True, exist_ok=True)
773
  pred_rows = []
774
  for idx in test_idx:
@@ -778,10 +878,11 @@ def simple_multilabel(
778
  pred_rows.append({**feature_rows[idx], "true_objects": ";".join(true_objs), "predicted_objects": ";".join(pred_objs)})
779
  metrics = {
780
  "status": "pass",
781
- "task": "object_relevance",
782
- "task_display_name": task_display_name("object_relevance"),
783
  "model_family": "simple_train_object_frequency",
784
  "source": "128_episode_qwen_jsonl_metadata",
 
785
  "split_policy": "object vocabulary and frequencies are learned from train split only",
786
  "num_train_windows": int(len(train_idx)),
787
  "num_val_windows": int(len(val_idx)),
@@ -797,14 +898,14 @@ def simple_multilabel(
797
 
798
  neural_result = None
799
  if args.include_neural:
800
- neural_dir = out_root / "neural_mlp" / "object_relevance"
801
  try:
802
- neural_result = neural_multilabel(rows, feature_rows, X, Y, splits, neural_dir, args, vocab)
803
  except Exception as exc: # pragma: no cover - protects long batch runs from optional NN environment failures.
804
  neural_result = {
805
  "status": "failed",
806
- "task": "object_relevance",
807
- "task_display_name": task_display_name("object_relevance"),
808
  "model_family": "neural_mlp_metadata_multilabel",
809
  "source": "128_episode_qwen_jsonl_metadata",
810
  "primary_metric": "micro_f1",
@@ -816,6 +917,7 @@ def simple_multilabel(
816
 
817
 
818
  def neural_multilabel(
 
819
  rows: list[dict[str, Any]],
820
  feature_rows: list[dict[str, Any]],
821
  X: np.ndarray,
@@ -824,6 +926,7 @@ def neural_multilabel(
824
  out_dir: Path,
825
  args: argparse.Namespace,
826
  vocab: list[str],
 
827
  ) -> dict[str, Any]:
828
  from neural_task_models import train_multilabel
829
 
@@ -838,10 +941,11 @@ def neural_multilabel(
838
  pred_rows.append({**feature_rows[idx], "true_objects": ";".join(true_objs), "predicted_objects": ";".join(pred_objs)})
839
  metrics = {
840
  "status": "pass",
841
- "task": "object_relevance",
842
- "task_display_name": task_display_name("object_relevance"),
843
  "model_family": "neural_mlp_metadata_multilabel",
844
  "source": "128_episode_qwen_jsonl_metadata",
 
845
  "num_train_windows": int(len(splits["train"])),
846
  "num_val_windows": int(len(splits["val"])),
847
  "num_test_windows": int(len(splits["test"])),
@@ -868,6 +972,15 @@ def caption_query_text(row: dict[str, Any]) -> str:
868
  )
869
 
870
 
 
 
 
 
 
 
 
 
 
871
  def retrieval_metrics_from_scores(scores: np.ndarray, rows: list[dict[str, Any]], test_idx: np.ndarray) -> tuple[dict[str, Any], list[dict[str, Any]]]:
872
  ranks = []
873
  rank_rows = []
@@ -1167,6 +1280,174 @@ def temporal_order(rows: list[dict[str, Any]], X: np.ndarray, splits: dict[str,
1167
  return {"simple": payload, "neural": neural_result}
1168
 
1169
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1170
  def unsupported_record(task_id: str, out_root: Path, reason: str, primary_metric: str) -> dict[str, Any]:
1171
  payload = {
1172
  "status": "unsupported_without_raw_128_feature_blocks",
@@ -1267,11 +1548,129 @@ def main() -> int:
1267
  result = classification_baseline(task_id, rows, feature_rows, X, splits, lambda row, k=key: norm(answer(row).get(k)), out, args)
1268
  task_results.append({"task": task_id, "task_display_name": task_display_name(task_id), **result})
1269
  log("object_relevance: start")
1270
- task_results.append({"task": "object_relevance", "task_display_name": task_display_name("object_relevance"), **simple_multilabel(rows, feature_rows, X, splits, out, args)})
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1271
  log("caption_grounding: start")
1272
  task_results.append({"task": "caption_grounding", "task_display_name": task_display_name("caption_grounding"), **caption_grounding(rows, episodes, splits, out, args)})
1273
  log("temporal_order: start")
1274
  task_results.append({"task": "temporal_order", "task_display_name": task_display_name("temporal_order"), **temporal_order(rows, X, splits, out, args)})
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1275
  for task_id, spec in UNSUPPORTED_TASKS.items():
1276
  log(f"{task_id}: recording unsupported status")
1277
  task_results.append({"task": task_id, "task_display_name": task_display_name(task_id), **unsupported_record(task_id, out, spec["reason"], spec["primary_metric"])})
 
7
  feature NPZ files, so this runner uses only public-safe JSONL metadata and
8
  strictly avoids answer fields as input features.
9
 
10
+ The output keeps the unified 20 task IDs. Tasks with enough JSONL signal get
11
  simple and, where appropriate, neural baselines over the same train/val/test
12
+ episode split used by the Qwen3-Omni pilot. Tasks whose original target
13
  requires missing raw motion/depth/audio feature blocks are emitted as explicit
14
  unsupported records instead of fabricated scores.
15
  """
 
49
  "modality_reconstruction",
50
  "temporal_order",
51
  "misalignment_detection",
52
+ "long_horizon_next_action",
53
+ "next_subtask_forecast",
54
+ "interaction_text_prediction",
55
+ "action_object_relation",
56
+ "object_set_forecast",
57
+ "imu_to_hand_pose",
58
+ "camera_view_sync_retrieval",
59
+ "time_to_transition",
60
  ]
61
 
62
  CLASSIFICATION_TASKS = {
 
84
  "primary_metric": "f1",
85
  "reason": "requires deliberately shifted cross-modal feature pairs, which cannot be reconstructed from the public JSONL labels alone",
86
  },
87
+ "interaction_text_prediction": {
88
+ "primary_metric": "macro_f1",
89
+ "reason": "requires raw annotation.hdf5 caption interaction text; the public 128 JSONL keeps only structured labels and derived metadata",
90
+ },
91
+ "imu_to_hand_pose": {
92
+ "primary_metric": "mae",
93
+ "reason": "requires raw IMU and hand-joint feature blocks, which are not in the public 128 JSONL metadata package",
94
+ },
95
+ "camera_view_sync_retrieval": {
96
+ "primary_metric": "mrr",
97
+ "reason": "requires paired camera-view feature blocks, which are not in the public 128 JSONL metadata package",
98
+ },
99
  }
100
 
101
  DEFAULT_DATASET = ROOT / "tmp/omni_128_dataset_fetch/dataset.jsonl"
 
142
  parser.add_argument("--neural-dropout", type=float, default=0.10)
143
  parser.add_argument("--neural-device", default="auto", choices=["auto", "cpu", "cuda"])
144
  parser.add_argument("--max-object-vocab", type=int, default=256)
145
+ parser.add_argument("--future-frames", type=int, default=100)
146
+ parser.add_argument("--transition-cap-frames", type=int, default=200)
147
+ parser.add_argument(
148
+ "--max-neural-classes",
149
+ type=int,
150
+ default=4096,
151
+ help="Emit an explicit unsupported NN record instead of fitting very large classification heads.",
152
+ )
153
  return parser.parse_args()
154
 
155
 
 
406
  return {key: np.asarray(value, dtype=np.int64) for key, value in indices.items()}
407
 
408
 
409
+ def row_start(row: dict[str, Any]) -> int:
410
+ return int((row.get("center_window") or {}).get("start_frame", 0) or 0)
411
+
412
+
413
+ def by_episode_sorted(rows: list[dict[str, Any]]) -> dict[str, list[int]]:
414
+ grouped: dict[str, list[int]] = {}
415
+ for idx, row in enumerate(rows):
416
+ grouped.setdefault(str(row.get("episode_id")), []).append(idx)
417
+ for episode_id in grouped:
418
+ grouped[episode_id].sort(key=lambda idx: row_start(rows[idx]))
419
+ return grouped
420
+
421
+
422
+ def future_index_map(rows: list[dict[str, Any]], frame_offset: int) -> dict[int, int]:
423
+ mapping: dict[int, int] = {}
424
+ for indices in by_episode_sorted(rows).values():
425
+ starts = np.asarray([row_start(rows[idx]) for idx in indices], dtype=np.int64)
426
+ for idx in indices:
427
+ target_start = row_start(rows[idx]) + frame_offset
428
+ future_pos = int(np.searchsorted(starts, target_start, side="left"))
429
+ if future_pos < len(indices):
430
+ mapping[idx] = indices[future_pos]
431
+ return mapping
432
+
433
+
434
+ def make_future_subset(rows: list[dict[str, Any]], frame_offset: int) -> tuple[np.ndarray, np.ndarray]:
435
+ mapping = future_index_map(rows, frame_offset)
436
+ current = np.asarray(sorted(mapping.keys()), dtype=np.int64)
437
+ future = np.asarray([mapping[int(idx)] for idx in current], dtype=np.int64)
438
+ return current, future
439
+
440
+
441
+ def subset_splits(splits: dict[str, np.ndarray], keep: np.ndarray) -> dict[str, np.ndarray]:
442
+ local = {int(global_idx): local_idx for local_idx, global_idx in enumerate(keep)}
443
+ return {
444
+ split: np.asarray([local[int(idx)] for idx in values if int(idx) in local], dtype=np.int64)
445
+ for split, values in splits.items()
446
+ }
447
+
448
+
449
+ def time_to_transition_targets(rows: list[dict[str, Any]], cap_frames: int) -> np.ndarray:
450
+ labels = [norm(answer(row).get("action")) for row in rows]
451
+ targets = np.full(len(rows), float(cap_frames), dtype=np.float32)
452
+ for indices in by_episode_sorted(rows).values():
453
+ for pos, idx in enumerate(indices):
454
+ label = labels[idx]
455
+ start = row_start(rows[idx])
456
+ distance = cap_frames
457
+ for next_idx in indices[pos + 1 :]:
458
+ if labels[next_idx] != label:
459
+ distance = min(max(row_start(rows[next_idx]) - start, 0), cap_frames)
460
+ break
461
+ targets[idx] = float(distance)
462
+ return targets[:, None]
463
+
464
+
465
  def encode_labels(values: list[str]) -> tuple[np.ndarray, list[str]]:
466
  seen: OrderedDict[str, int] = OrderedDict()
467
  for value in values:
 
738
  from neural_task_models import train_classifier
739
 
740
  out_dir.mkdir(parents=True, exist_ok=True)
741
+ if train_class_count > args.max_neural_classes:
742
+ metrics = {
743
+ "status": "unsupported_large_label_space",
744
+ "task": task_id,
745
+ "task_display_name": task_display_name(task_id),
746
+ "model_family": "neural_mlp_metadata",
747
+ "source": "128_episode_qwen_jsonl_metadata",
748
+ "primary_metric": "macro_f1",
749
+ "primary_score": None,
750
+ "reason": f"train class count {train_class_count} exceeds --max-neural-classes {args.max_neural_classes}",
751
+ }
752
+ write_json(out_dir / "metrics.json", metrics)
753
+ return metrics
754
  result = train_classifier(
755
  X.astype(np.float32),
756
  y,
 
842
 
843
 
844
  def simple_multilabel(
845
+ task_id: str,
846
  rows: list[dict[str, Any]],
847
+ target_rows: list[dict[str, Any]],
848
  feature_rows: list[dict[str, Any]],
849
  X: np.ndarray,
850
  splits: dict[str, np.ndarray],
851
  out_root: Path,
852
  args: argparse.Namespace,
853
+ input_features: str,
854
  ) -> dict[str, Any]:
855
  train_objects = Counter()
856
  for idx in splits["train"]:
857
+ for obj in answer(target_rows[int(idx)]).get("objects", []) or []:
858
  key = norm(obj).lower()
859
  if key:
860
  train_objects[key] += 1
861
  vocab = [name for name, _count in train_objects.most_common(args.max_object_vocab)]
862
+ Y = object_matrix(target_rows, vocab)
863
  train_idx, val_idx, test_idx = splits["train"], splits["val"], splits["test"]
864
  freq = Y[train_idx].mean(axis=0)
865
  # Public-safe simple baseline: predict object labels seen in at least 10% of
 
868
  pred_all = np.tile((freq >= 0.10).astype(np.float32), (len(rows), 1))
869
  test_metrics = multilabel_metrics(Y[test_idx], pred_all[test_idx])
870
  val_metrics = multilabel_metrics(Y[val_idx], pred_all[val_idx]) if len(val_idx) else {}
871
+ out_dir = out_root / task_id
872
  out_dir.mkdir(parents=True, exist_ok=True)
873
  pred_rows = []
874
  for idx in test_idx:
 
878
  pred_rows.append({**feature_rows[idx], "true_objects": ";".join(true_objs), "predicted_objects": ";".join(pred_objs)})
879
  metrics = {
880
  "status": "pass",
881
+ "task": task_id,
882
+ "task_display_name": task_display_name(task_id),
883
  "model_family": "simple_train_object_frequency",
884
  "source": "128_episode_qwen_jsonl_metadata",
885
+ "input_features": input_features,
886
  "split_policy": "object vocabulary and frequencies are learned from train split only",
887
  "num_train_windows": int(len(train_idx)),
888
  "num_val_windows": int(len(val_idx)),
 
898
 
899
  neural_result = None
900
  if args.include_neural:
901
+ neural_dir = out_root / "neural_mlp" / task_id
902
  try:
903
+ neural_result = neural_multilabel(task_id, rows, feature_rows, X, Y, splits, neural_dir, args, vocab, input_features)
904
  except Exception as exc: # pragma: no cover - protects long batch runs from optional NN environment failures.
905
  neural_result = {
906
  "status": "failed",
907
+ "task": task_id,
908
+ "task_display_name": task_display_name(task_id),
909
  "model_family": "neural_mlp_metadata_multilabel",
910
  "source": "128_episode_qwen_jsonl_metadata",
911
  "primary_metric": "micro_f1",
 
917
 
918
 
919
  def neural_multilabel(
920
+ task_id: str,
921
  rows: list[dict[str, Any]],
922
  feature_rows: list[dict[str, Any]],
923
  X: np.ndarray,
 
926
  out_dir: Path,
927
  args: argparse.Namespace,
928
  vocab: list[str],
929
+ input_features: str,
930
  ) -> dict[str, Any]:
931
  from neural_task_models import train_multilabel
932
 
 
941
  pred_rows.append({**feature_rows[idx], "true_objects": ";".join(true_objs), "predicted_objects": ";".join(pred_objs)})
942
  metrics = {
943
  "status": "pass",
944
+ "task": task_id,
945
+ "task_display_name": task_display_name(task_id),
946
  "model_family": "neural_mlp_metadata_multilabel",
947
  "source": "128_episode_qwen_jsonl_metadata",
948
+ "input_features": input_features,
949
  "num_train_windows": int(len(splits["train"])),
950
  "num_val_windows": int(len(splits["val"])),
951
  "num_test_windows": int(len(splits["test"])),
 
972
  )
973
 
974
 
975
+ def action_object_relation_label(row: dict[str, Any]) -> str:
976
+ ans = answer(row)
977
+ action_label = norm(ans.get("action"))
978
+ objects = sorted({norm(obj).lower() for obj in ans.get("objects", []) or [] if norm(obj)})
979
+ if not action_label or not objects:
980
+ return ""
981
+ return f"{action_label}|{'+'.join(objects)}"
982
+
983
+
984
  def retrieval_metrics_from_scores(scores: np.ndarray, rows: list[dict[str, Any]], test_idx: np.ndarray) -> tuple[dict[str, Any], list[dict[str, Any]]]:
985
  ranks = []
986
  rank_rows = []
 
1280
  return {"simple": payload, "neural": neural_result}
1281
 
1282
 
1283
+ def regression_metrics(y_true: np.ndarray, y_pred: np.ndarray) -> dict[str, float]:
1284
+ err = y_pred - y_true
1285
+ mae = float(np.mean(np.abs(err)))
1286
+ rmse = float(np.sqrt(np.mean(err**2)))
1287
+ denom = float(np.sum((y_true - y_true.mean(axis=0, keepdims=True)) ** 2))
1288
+ r2 = 1.0 - float(np.sum(err**2)) / max(denom, 1e-12)
1289
+ return {"mae": mae, "rmse": rmse, "r2": r2}
1290
+
1291
+
1292
+ def ridge_regression_predict(
1293
+ X_train: np.ndarray,
1294
+ y_train: np.ndarray,
1295
+ X_test: np.ndarray,
1296
+ l2: float,
1297
+ ) -> tuple[np.ndarray, dict[str, np.ndarray]]:
1298
+ mean, std = fit_scaler(X_train)
1299
+ Xtr = (X_train - mean) / std
1300
+ Xte = (X_test - mean) / std
1301
+ Y = np.asarray(y_train, dtype=np.float32)
1302
+ if Y.ndim == 1:
1303
+ Y = Y[:, None]
1304
+ y_mean = Y.mean(axis=0, dtype=np.float64).astype(np.float32)
1305
+ y_std = Y.std(axis=0, dtype=np.float64).astype(np.float32)
1306
+ y_std = np.where(y_std < 1e-6, 1.0, y_std).astype(np.float32)
1307
+ Yz = ((Y - y_mean) / y_std).astype(np.float32)
1308
+ eye = np.eye(Xtr.shape[1], dtype=np.float32) * float(l2)
1309
+ W = np.linalg.solve(Xtr.T @ Xtr + eye, Xtr.T @ Yz).astype(np.float32)
1310
+ pred = Xte @ W
1311
+ pred = pred * y_std + y_mean
1312
+ return pred.astype(np.float32), {"mean": mean, "std": std, "y_mean": y_mean, "y_std": y_std, "W": W}
1313
+
1314
+
1315
+ def regression_task(
1316
+ task_id: str,
1317
+ rows: list[dict[str, Any]],
1318
+ feature_rows: list[dict[str, Any]],
1319
+ X: np.ndarray,
1320
+ y: np.ndarray,
1321
+ splits: dict[str, np.ndarray],
1322
+ out_root: Path,
1323
+ args: argparse.Namespace,
1324
+ input_features: str,
1325
+ ) -> dict[str, Any]:
1326
+ train_idx = splits["train"]
1327
+ val_idx = splits["val"]
1328
+ test_idx = splits["test"]
1329
+ out_dir = out_root / task_id
1330
+ out_dir.mkdir(parents=True, exist_ok=True)
1331
+ pred, model = ridge_regression_predict(X[train_idx], y[train_idx], X[test_idx], args.l2)
1332
+ test_metrics = regression_metrics(y[test_idx], pred)
1333
+ val_metrics = {}
1334
+ if len(val_idx):
1335
+ val_pred, _ = ridge_regression_predict(X[train_idx], y[train_idx], X[val_idx], args.l2)
1336
+ val_metrics = regression_metrics(y[val_idx], val_pred)
1337
+ pred_rows = []
1338
+ for local_k, idx in enumerate(test_idx):
1339
+ idx = int(idx)
1340
+ pred_rows.append(
1341
+ {
1342
+ **feature_rows[idx],
1343
+ "true_value": float(y[idx, 0]),
1344
+ "predicted_value": float(pred[local_k, 0]),
1345
+ "absolute_error": float(abs(pred[local_k, 0] - y[idx, 0])),
1346
+ }
1347
+ )
1348
+ metrics = {
1349
+ "status": "pass",
1350
+ "task": task_id,
1351
+ "task_display_name": task_display_name(task_id),
1352
+ "model_family": "simple_ridge_metadata",
1353
+ "source": "128_episode_qwen_jsonl_metadata",
1354
+ "input_features": input_features,
1355
+ "split_policy": "train ridge regressor on train split, report held-out test timing error",
1356
+ "num_train_windows": int(len(train_idx)),
1357
+ "num_val_windows": int(len(val_idx)),
1358
+ "num_test_windows": int(len(test_idx)),
1359
+ "splits": {"val": val_metrics, "test": test_metrics},
1360
+ "primary_metric": "mae",
1361
+ "metric_direction": "lower",
1362
+ "primary_score": test_metrics["mae"],
1363
+ }
1364
+ write_json(out_dir / "metrics.json", metrics)
1365
+ write_csv(out_dir / "predictions.csv", pred_rows)
1366
+ np.savez_compressed(out_dir / "model.npz", **model)
1367
+
1368
+ neural_result = None
1369
+ if args.include_neural:
1370
+ neural_dir = out_root / "neural_mlp" / task_id
1371
+ try:
1372
+ neural_result = neural_regression_task(task_id, rows, feature_rows, X, y, splits, neural_dir, args, input_features)
1373
+ except Exception as exc: # pragma: no cover - protects long batch runs from optional NN environment failures.
1374
+ neural_result = {
1375
+ "status": "failed",
1376
+ "task": task_id,
1377
+ "task_display_name": task_display_name(task_id),
1378
+ "model_family": "neural_mlp_metadata_regressor",
1379
+ "source": "128_episode_qwen_jsonl_metadata",
1380
+ "primary_metric": "mae",
1381
+ "metric_direction": "lower",
1382
+ "primary_score": None,
1383
+ "error": str(exc),
1384
+ }
1385
+ write_json(neural_dir / "metrics.json", neural_result)
1386
+ return {"simple": metrics, "neural": neural_result}
1387
+
1388
+
1389
+ def neural_regression_task(
1390
+ task_id: str,
1391
+ rows: list[dict[str, Any]],
1392
+ feature_rows: list[dict[str, Any]],
1393
+ X: np.ndarray,
1394
+ y: np.ndarray,
1395
+ splits: dict[str, np.ndarray],
1396
+ out_dir: Path,
1397
+ args: argparse.Namespace,
1398
+ input_features: str,
1399
+ ) -> dict[str, Any]:
1400
+ from neural_task_models import save_torch_model, train_regressor
1401
+
1402
+ out_dir.mkdir(parents=True, exist_ok=True)
1403
+ train_idx = splits["train"]
1404
+ test_idx = splits["test"]
1405
+ result = train_regressor(X.astype(np.float32), y.astype(np.float32), train_idx, test_idx, neural_config(args))
1406
+ test_metrics = regression_metrics(y[test_idx], result["pred"])
1407
+ pred_rows = []
1408
+ for local_k, idx in enumerate(test_idx):
1409
+ idx = int(idx)
1410
+ pred_rows.append(
1411
+ {
1412
+ **feature_rows[idx],
1413
+ "true_value": float(y[idx, 0]),
1414
+ "predicted_value": float(result["pred"][local_k, 0]),
1415
+ "absolute_error": float(abs(result["pred"][local_k, 0] - y[idx, 0])),
1416
+ }
1417
+ )
1418
+ metrics = {
1419
+ "status": "pass",
1420
+ "task": task_id,
1421
+ "task_display_name": task_display_name(task_id),
1422
+ "model_family": "neural_mlp_metadata_regressor",
1423
+ "source": "128_episode_qwen_jsonl_metadata",
1424
+ "input_features": input_features,
1425
+ "split_policy": "train neural regressor on train split, report held-out test timing error",
1426
+ "num_train_windows": int(len(train_idx)),
1427
+ "num_test_windows": int(len(test_idx)),
1428
+ "history": result["history"],
1429
+ "device": result["device"],
1430
+ "splits": {"test": test_metrics},
1431
+ "primary_metric": "mae",
1432
+ "metric_direction": "lower",
1433
+ "primary_score": test_metrics["mae"],
1434
+ }
1435
+ write_json(out_dir / "metrics.json", metrics)
1436
+ write_csv(out_dir / "predictions.csv", pred_rows)
1437
+ save_torch_model(
1438
+ out_dir / "model.pt",
1439
+ {
1440
+ "state_dict": result["state_dict"],
1441
+ "x_mean": result["x_mean"],
1442
+ "x_std": result["x_std"],
1443
+ "y_mean": result["y_mean"],
1444
+ "y_std": result["y_std"],
1445
+ "metrics": metrics,
1446
+ },
1447
+ )
1448
+ return metrics
1449
+
1450
+
1451
  def unsupported_record(task_id: str, out_root: Path, reason: str, primary_metric: str) -> dict[str, Any]:
1452
  payload = {
1453
  "status": "unsupported_without_raw_128_feature_blocks",
 
1548
  result = classification_baseline(task_id, rows, feature_rows, X, splits, lambda row, k=key: norm(answer(row).get(k)), out, args)
1549
  task_results.append({"task": task_id, "task_display_name": task_display_name(task_id), **result})
1550
  log("object_relevance: start")
1551
+ task_results.append(
1552
+ {
1553
+ "task": "object_relevance",
1554
+ "task_display_name": task_display_name("object_relevance"),
1555
+ **simple_multilabel(
1556
+ "object_relevance",
1557
+ rows,
1558
+ rows,
1559
+ feature_rows,
1560
+ X,
1561
+ splits,
1562
+ out,
1563
+ args,
1564
+ "frame/context metadata plus hashed prompt/options/main_task text; answer_json fields are excluded from inputs",
1565
+ ),
1566
+ }
1567
+ )
1568
  log("caption_grounding: start")
1569
  task_results.append({"task": "caption_grounding", "task_display_name": task_display_name("caption_grounding"), **caption_grounding(rows, episodes, splits, out, args)})
1570
  log("temporal_order: start")
1571
  task_results.append({"task": "temporal_order", "task_display_name": task_display_name("temporal_order"), **temporal_order(rows, X, splits, out, args)})
1572
+ log("long_horizon_next_action / next_subtask_forecast / object_set_forecast: resolving future windows")
1573
+ current_idx, future_idx = make_future_subset(rows, args.future_frames)
1574
+ current_rows = [rows[int(idx)] for idx in current_idx]
1575
+ future_rows = [rows[int(idx)] for idx in future_idx]
1576
+ current_feature_rows = [feature_rows[int(idx)] for idx in current_idx]
1577
+ current_splits = subset_splits(splits, current_idx)
1578
+ current_X = X[current_idx]
1579
+ future_action_by_id = {
1580
+ str(current_rows[pos].get("id")): norm(answer(future_rows[pos]).get("action"))
1581
+ for pos in range(len(current_rows))
1582
+ }
1583
+ future_subtask_by_id = {
1584
+ str(current_rows[pos].get("id")): norm(answer(future_rows[pos]).get("subtask"))
1585
+ for pos in range(len(current_rows))
1586
+ }
1587
+ log("long_horizon_next_action: start")
1588
+ task_results.append(
1589
+ {
1590
+ "task": "long_horizon_next_action",
1591
+ "task_display_name": task_display_name("long_horizon_next_action"),
1592
+ **classification_baseline(
1593
+ "long_horizon_next_action",
1594
+ current_rows,
1595
+ current_feature_rows,
1596
+ current_X,
1597
+ current_splits,
1598
+ lambda row: future_action_by_id.get(str(row.get("id")), ""),
1599
+ out,
1600
+ args,
1601
+ ),
1602
+ }
1603
+ )
1604
+ log("next_subtask_forecast: start")
1605
+ task_results.append(
1606
+ {
1607
+ "task": "next_subtask_forecast",
1608
+ "task_display_name": task_display_name("next_subtask_forecast"),
1609
+ **classification_baseline(
1610
+ "next_subtask_forecast",
1611
+ current_rows,
1612
+ current_feature_rows,
1613
+ current_X,
1614
+ current_splits,
1615
+ lambda row: future_subtask_by_id.get(str(row.get("id")), ""),
1616
+ out,
1617
+ args,
1618
+ ),
1619
+ }
1620
+ )
1621
+ log("action_object_relation: start")
1622
+ task_results.append(
1623
+ {
1624
+ "task": "action_object_relation",
1625
+ "task_display_name": task_display_name("action_object_relation"),
1626
+ **classification_baseline(
1627
+ "action_object_relation",
1628
+ rows,
1629
+ feature_rows,
1630
+ X,
1631
+ splits,
1632
+ action_object_relation_label,
1633
+ out,
1634
+ args,
1635
+ ),
1636
+ }
1637
+ )
1638
+ log("object_set_forecast: start")
1639
+ task_results.append(
1640
+ {
1641
+ "task": "object_set_forecast",
1642
+ "task_display_name": task_display_name("object_set_forecast"),
1643
+ **simple_multilabel(
1644
+ "object_set_forecast",
1645
+ current_rows,
1646
+ future_rows,
1647
+ current_feature_rows,
1648
+ current_X,
1649
+ current_splits,
1650
+ out,
1651
+ args,
1652
+ f"current-window frame/context metadata plus hashed prompt/options/main_task text; target object set +{args.future_frames} frames",
1653
+ ),
1654
+ }
1655
+ )
1656
+ log("time_to_transition: start")
1657
+ task_results.append(
1658
+ {
1659
+ "task": "time_to_transition",
1660
+ "task_display_name": task_display_name("time_to_transition"),
1661
+ **regression_task(
1662
+ "time_to_transition",
1663
+ rows,
1664
+ feature_rows,
1665
+ X,
1666
+ time_to_transition_targets(rows, args.transition_cap_frames),
1667
+ splits,
1668
+ out,
1669
+ args,
1670
+ f"frame/context metadata plus hashed prompt/options/main_task text; target is frames to next action boundary capped at {args.transition_cap_frames}",
1671
+ ),
1672
+ }
1673
+ )
1674
  for task_id, spec in UNSUPPORTED_TASKS.items():
1675
  log(f"{task_id}: recording unsupported status")
1676
  task_results.append({"task": task_id, "task_display_name": task_display_name(task_id), **unsupported_record(task_id, out, spec["reason"], spec["primary_metric"])})