cmu_stretch

  • คำอธิบาย :

สวัสดีปฏิสัมพันธ์ครัวหุ่นยนต์ยืด

  • หน้าแรก : https://robo-affordances.github.io/

  • ซอร์สโค้ด : tfds.robotics.rtx.CmuStretch

  • รุ่น :

    • 0.1.0 (ค่าเริ่มต้น): การเปิดตัวครั้งแรก
  • ขนาดการดาวน์โหลด : Unknown size

  • ขนาดชุดข้อมูล : 728.06 MiB

  • แคชอัตโนมัติ ( เอกสาร ): No

  • แยก :

แยก ตัวอย่าง
'train' 135
  • โครงสร้างคุณสมบัติ :
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'file_path': Text(shape=(), dtype=string),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(8,), dtype=float32),
        'discount': Scalar(shape=(), dtype=float32),
        'is_first': bool,
        'is_last': bool,
        'is_terminal': bool,
        'language_embedding': Tensor(shape=(512,), dtype=float32),
        'language_instruction': Text(shape=(), dtype=string),
        'observation': FeaturesDict({
            'image': Image(shape=(128, 128, 3), dtype=uint8),
            'state': Tensor(shape=(4,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • เอกสารคุณสมบัติ :
คุณสมบัติ ระดับ รูปร่าง ประเภทD คำอธิบาย
คุณสมบัติDict
ตอนที่_ข้อมูลเมตา คุณสมบัติDict
ตอนที่_metadata/file_path ข้อความ เชือก เส้นทางไปยังไฟล์ข้อมูลต้นฉบับ
ขั้นตอน ชุดข้อมูล
ขั้นตอน/การดำเนินการ เทนเซอร์ (8,) ลอย32 การกระทำของหุ่นยนต์ประกอบด้วย [3x ee pos, 3x ee rot, 1x กริปเปอร์ binary action, 1x ยุติตอน]
ขั้นตอน/ส่วนลด สเกลาร์ ลอย32 ส่วนลดหากมีให้ ค่าเริ่มต้นคือ 1
ขั้นตอน/is_first เทนเซอร์ บูล
ขั้นตอน/is_last เทนเซอร์ บูล
ขั้นตอน/is_terminal เทนเซอร์ บูล
ขั้นตอน/ภาษา_embedding เทนเซอร์ (512,) ลอย32 การฝังภาษาโคน่า ดู https://tfhub.dev/google/universal-sentence-encoder-large/5
ขั้นตอน/Language_instruction ข้อความ เชือก การสอนภาษา.
ขั้นตอน/การสังเกต คุณสมบัติDict
ขั้นตอน/การสังเกต/ภาพ ภาพ (128, 128, 3) uint8 การสังเกต RGB ของกล้องหลัก
ขั้นตอน/การสังเกต/สถานะ เทนเซอร์ (4,) ลอย32 สถานะของหุ่นยนต์ ประกอบด้วย [3x มุมข้อต่อของหุ่นยนต์/ee pos, 1x ตำแหน่งกริปเปอร์]
ขั้นตอน/รางวัล สเกลาร์ ลอย32 รางวัลหากมีให้ 1 ในขั้นตอนสุดท้ายสำหรับการสาธิต
  • การอ้างอิง :
@inproceedings{bahl2023affordances,
  title={Affordances from Human Videos as a Versatile Representation for Robotics},
  author={Bahl, Shikhar and Mendonca, Russell and Chen, Lili and Jain, Unnat and Pathak, Deepak},
  booktitle={CVPR},
  year={2023}
}
@article{mendonca2023structured,
  title={Structured World Models from Human Videos},
  author={Mendonca, Russell and Bahl, Shikhar and Pathak, Deepak},
  journal={CoRL},
  year={2023}
}