stanford_robocook_converted_externally_to_rlds

  • คำอธิบาย :

แฟรงก้าเตรียมเกี๊ยวด้วยเครื่องมือต่างๆ

แยก ตัวอย่าง
'train' 2,460
  • โครงสร้างคุณสมบัติ :
FeaturesDict({
    'episode_metadata': FeaturesDict({
        'extrinsics_1': Tensor(shape=(4, 4), dtype=float32),
        'extrinsics_2': Tensor(shape=(4, 4), dtype=float32),
        'extrinsics_3': Tensor(shape=(4, 4), dtype=float32),
        'extrinsics_4': Tensor(shape=(4, 4), dtype=float32),
        'file_path': Text(shape=(), dtype=string),
    }),
    'steps': Dataset({
        'action': Tensor(shape=(7,), 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({
            'depth_1': Tensor(shape=(256, 256), dtype=float32),
            'depth_2': Tensor(shape=(256, 256), dtype=float32),
            'depth_3': Tensor(shape=(256, 256), dtype=float32),
            'depth_4': Tensor(shape=(256, 256), dtype=float32),
            'image_1': Image(shape=(256, 256, 3), dtype=uint8),
            'image_2': Image(shape=(256, 256, 3), dtype=uint8),
            'image_3': Image(shape=(256, 256, 3), dtype=uint8),
            'image_4': Image(shape=(256, 256, 3), dtype=uint8),
            'state': Tensor(shape=(7,), dtype=float32),
        }),
        'reward': Scalar(shape=(), dtype=float32),
    }),
})
  • เอกสารคุณสมบัติ :
คุณสมบัติ ระดับ รูปร่าง ประเภทD คำอธิบาย
คุณสมบัติDict
ตอนที่_ข้อมูลเมตา คุณสมบัติDict
ตอนที่_metadata/extrinsics_1 เทนเซอร์ (4, 4) ลอย32 กล้อง 1 เมทริกซ์ภายนอก
ตอนที่_ข้อมูลเมตา/ภายนอก_2 เทนเซอร์ (4, 4) ลอย32 กล้อง 2 เมทริกซ์ภายนอก
ตอนที่_ข้อมูลเมตา/ภายนอก_3 เทนเซอร์ (4, 4) ลอย32 กล้อง 3 เมทริกซ์ภายนอก
ตอนที่_ข้อมูลเมตา/ภายนอก_4 เทนเซอร์ (4, 4) ลอย32 กล้อง 4 เมทริกซ์ภายนอก
ตอนที่_metadata/file_path ข้อความ เชือก เส้นทางไปยังไฟล์ข้อมูลต้นฉบับ
ขั้นตอน ชุดข้อมูล
ขั้นตอน/การดำเนินการ เทนเซอร์ (7,) ลอย32 การกระทำของหุ่นยนต์ประกอบด้วย [ความเร็วเชิงมุมของหุ่นยนต์ที่ปลายเอฟเฟกต์ 3x ความเร็วเชิงมุมของหุ่นยนต์ที่ปลายเอฟเฟกต์ 3x ความเร็วของกริปเปอร์ 1x]
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/ความลึก_1 เทนเซอร์ (256, 256) ลอย32 กล้องที่ 1 การสังเกตเชิงลึก
ขั้นตอน/การสังเกต/ความลึก_2 เทนเซอร์ (256, 256) ลอย32 กล้องที่ 2 การสังเกตเชิงลึก
ขั้นตอน/การสังเกต/ความลึก_3 เทนเซอร์ (256, 256) ลอย32 กล้องที่ 3 การสังเกตเชิงลึก
ขั้นตอน/การสังเกต/ความลึก_4 เทนเซอร์ (256, 256) ลอย32 กล้อง 4 การสังเกตเชิงลึก
ขั้นตอน/การสังเกต/image_1 ภาพ (256, 256, 3) uint8 กล้อง 1 การสังเกต RGB
ขั้นตอน/การสังเกต/image_2 ภาพ (256, 256, 3) uint8 การสังเกต RGB ของกล้อง 2
ขั้นตอน/ข้อสังเกต/image_3 ภาพ (256, 256, 3) uint8 การสังเกต RGB ของกล้อง 3
ขั้นตอน/การสังเกต/image_4 ภาพ (256, 256, 3) uint8 การสังเกต RGB ของกล้อง 4
ขั้นตอน/การสังเกต/สถานะ เทนเซอร์ (7,) ลอย32 สถานะของหุ่นยนต์ประกอบด้วย [ตำแหน่งปลายเอฟเฟกต์ของหุ่นยนต์ 3x มุมออยเลอร์ปลายเอฟเฟกต์ของหุ่นยนต์ 3x ตำแหน่งมือจับ 1x]
ขั้นตอน/รางวัล สเกลาร์ ลอย32 รางวัลหากมีให้ 1 ในขั้นตอนสุดท้ายสำหรับการสาธิต
@article{shi2023robocook,
  title={RoboCook: Long-Horizon Elasto-Plastic Object Manipulation with Diverse Tools},
  author={Shi, Haochen and Xu, Huazhe and Clarke, Samuel and Li, Yunzhu and Wu, Jiajun},
  journal={arXiv preprint arXiv:2306.14447},
  year={2023}
}