Dataset Viewer
Auto-converted to Parquet Duplicate
Search is not available for this dataset
video
video
label
class label
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
0file-000
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
1file-001
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
2file-002
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
3file-003
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
4file-004
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
5file-005
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
6file-006
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
7file-007
8file-008
8file-008
8file-008
8file-008
End of preview. Expand in Data Studio

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Egocentric VQA Experiment Paths

本文件用于整理 flower、plate、shoe、Pack_a_first_aid_kit 四组任务的实验设置、VQA 输出、Qwen/GLM 测试结果和 JSON 文件路径。

注意:下面是你本地 Windows 路径。上传 Hugging Face 时,可以把这些文件放到对应任务文件夹下,并保留同名文件,或按实际上传路径修改。

汇总表

Task View VQA 数量 实验设置 VQA 输出 Qwen 结果 GLM 结果
flower / pick_flower cam_color 88 D:\egocentric_label\vqa_flower_config.json D:\egocentric_label\workflow_outputs_flower_cam\understanding_vqa.json D:\egocentric_label\results\understanding_qwen_flower_cam_88.json D:\egocentric_label\results\understanding_glm_flower_cam_88.json
plate left_eye 120 D:\egocentric_label\vqa_plate_config.json D:\egocentric_label\workflow_outputs_plate_left_eye\understanding_vqa.json D:\egocentric_label\results\understanding_qwen_plate_left_eye_120.json D:\egocentric_label\results\understanding_glm_plate_left_eye_120.json
shoe left_eye 119 D:\egocentric_label\vqa_shoe_config.json D:\egocentric_label\workflow_outputs_shoe_left_eye\understanding_vqa.json D:\egocentric_label\results\understanding_qwen_shoe_left_eye_119.json D:\egocentric_label\results\understanding_glm_shoe_left_eye_119.json
Pack_a_first_aid_kit / first_aid left_eye 479 D:\egocentric_label\vqa_first_aid_config.json D:\egocentric_label\workflow_outputs_first_aid_left_eye\understanding_vqa.json D:\egocentric_label\results\understanding_qwen_first_aid_left_eye_479.json D:\egocentric_label\results\understanding_glm_first_aid_left_eye_479.json

详细路径

flower / pick_flower

  • 任务名:flower
  • 视角:cam_color
  • VQA 数量:88
  • 实验设置:D:\egocentric_label\vqa_flower_config.json
  • 标注 JSON 目录:D:\egocentric_label\workflow_input\json
  • 视频目录:D:\egocentric_label\hand\pick_flower\videos\observation.images.cam_color\chunk-000
  • 类别文件:D:\egocentric_label\classes_pick_flower.txt
  • VQA 题目:D:\egocentric_label\workflow_outputs_flower_cam\understanding_vqa.json
  • VQA 切片目录:D:\egocentric_label\workflow_outputs_flower_cam\understanding_clips
  • 干扰选项:D:\egocentric_label\workflow_outputs_flower_cam\llm_distractors.json
  • Qwen 测试结果:D:\egocentric_label\results\understanding_qwen_flower_cam_88.json
  • GLM 测试结果:D:\egocentric_label\results\understanding_glm_flower_cam_88.json

plate

  • 任务名:plate
  • 视角:left_eye
  • VQA 数量:120
  • 实验设置:D:\egocentric_label\vqa_plate_config.json
  • 标注 JSON 目录:D:\egocentric_label\gripper\plate\labels
  • 视频目录:D:\egocentric_label\gripper\plate\left_eye
  • 类别文件:D:\egocentric_label\classes_plate.txt
  • VQA 题目:D:\egocentric_label\workflow_outputs_plate_left_eye\understanding_vqa.json
  • VQA 切片目录:D:\egocentric_label\workflow_outputs_plate_left_eye\understanding_clips
  • 干扰选项:D:\egocentric_label\workflow_outputs_plate_left_eye\llm_distractors.json
  • Qwen 测试结果:D:\egocentric_label\results\understanding_qwen_plate_left_eye_120.json
  • GLM 测试结果:D:\egocentric_label\results\understanding_glm_plate_left_eye_120.json

shoe

  • 任务名:shoe
  • 视角:left_eye
  • VQA 数量:119
  • 实验设置:D:\egocentric_label\vqa_shoe_config.json
  • 标注 JSON 目录:D:\egocentric_label\gripper\shoe\labels
  • 视频目录:D:\egocentric_label\gripper\shoe\videos\observation.images.left_eye\chunk-000
  • 类别文件:D:\egocentric_label\classes_shoe.txt
  • VQA 题目:D:\egocentric_label\workflow_outputs_shoe_left_eye\understanding_vqa.json
  • VQA 切片目录:D:\egocentric_label\workflow_outputs_shoe_left_eye\understanding_clips
  • 干扰选项:D:\egocentric_label\workflow_outputs_shoe_left_eye\llm_distractors.json
  • Qwen 测试结果:D:\egocentric_label\results\understanding_qwen_shoe_left_eye_119.json
  • GLM 测试结果:D:\egocentric_label\results\understanding_glm_shoe_left_eye_119.json

Pack_a_first_aid_kit / first_aid

  • 任务名:Pack_a_first_aid_kit
  • 视角:left_eye
  • VQA 数量:479
  • 实验设置:D:\egocentric_label\vqa_first_aid_config.json
  • 标注 JSON 目录:D:\egocentric_label\workflow_input\first_aid_left_eye_json
  • 视频目录:D:\egocentric_label\gripper\Pack_a_first_aid_kit\videos\observation.images.left_eye\chunk-000
  • 类别文件:D:\egocentric_label\classes_first_aid_kit.txt
  • VQA 题目:D:\egocentric_label\workflow_outputs_first_aid_left_eye\understanding_vqa.json
  • VQA 切片目录:D:\egocentric_label\workflow_outputs_first_aid_left_eye\understanding_clips
  • 干扰选项:D:\egocentric_label\workflow_outputs_first_aid_left_eye\llm_distractors.json
  • Qwen 测试结果:D:\egocentric_label\results\understanding_qwen_first_aid_left_eye_479.json
  • GLM 测试结果:D:\egocentric_label\results\understanding_glm_first_aid_left_eye_479.json

建议的 Hugging Face 目录结构

egocentric/
├─ README.md
├─ hf_experiment_manifest.json
└─ experiments/
   ├─ flower/
   │  ├─ config/
   │  │  └─ <vqa_config>.json
   │  ├─ vqa/
   │  │  ├─ understanding_vqa.json
   │  │  ├─ llm_distractors.json
   │  │  └─ understanding_clips/
   │  └─ results/
   │     ├─ <qwen_result>.json
   │     └─ <glm_result>.json
   ├─ plate/
   │  ├─ config/
   │  │  └─ <vqa_config>.json
   │  ├─ vqa/
   │  │  ├─ understanding_vqa.json
   │  │  ├─ llm_distractors.json
   │  │  └─ understanding_clips/
   │  └─ results/
   │     ├─ <qwen_result>.json
   │     └─ <glm_result>.json
   ├─ shoe/
   │  ├─ config/
   │  │  └─ <vqa_config>.json
   │  ├─ vqa/
   │  │  ├─ understanding_vqa.json
   │  │  ├─ llm_distractors.json
   │  │  └─ understanding_clips/
   │  └─ results/
   │     ├─ <qwen_result>.json
   │     └─ <glm_result>.json
   ├─ Pack_a_first_aid_kit/
   │  ├─ config/
   │  │  └─ <vqa_config>.json
   │  ├─ vqa/
   │  │  ├─ understanding_vqa.json
   │  │  ├─ llm_distractors.json
   │  │  └─ understanding_clips/
   │  └─ results/
   │     ├─ <qwen_result>.json
   │     └─ <glm_result>.json

给学长的简短说明

我已将 flower、plate、shoe、Pack_a_first_aid_kit 四组任务整理为统一的 understanding VQA 实验文件。
每组包括:实验配置 json、VQA 输出 understanding_vqa.json、切片目录 understanding_clips、干扰选项 llm_distractors.json,以及 Qwen/GLM 测试结果 json。
具体路径见 hf_experiment_manifest.json 或本 README 的详细路径部分。
Downloads last month
494