avatar
Articles
165
Tags
34
Categories
0
Home
Archives
Tags
Categories
List
  • Music
  • Movie
Link
About
ALTNT's Hexo Blog
Home
Archives
Tags
Categories
List
  • Music
  • Movie
Link
About

Diffusion Policy

Created2024-08-23|Updated2024-08-23
|Post Views:
Author: ALTNT
Link: http://blog.705553939.xyz/2024/08/23/Ros%E6%9C%BA%E5%99%A8%E4%BA%BA%E7%9B%B8%E5%85%B3/%E6%9C%BA%E5%99%A8%E4%BA%BA%E7%9B%B8%E5%85%B3%E8%AE%BA%E6%96%87/diffusionpolicy/
Copyright Notice: All articles in this blog are licensed under CC BY-NC-SA 4.0 unless stating additionally.
Ros机器人
Previous
Markdown 特殊符号及数学公式整理
Next
Generative Skill Chaining: Long-Horizon Skill Planning with Diffusion Models
Related Articles
2024-06-05
模仿学习
2024-06-05
ROS通信机制
2024-08-23
STAP: Sequencing Task-Agnostic Policies
2024-06-05
接下来可能看的论文
2024-08-23
Generative Skill Chaining: Long-Horizon Skill Planning with Diffusion Models
2024-08-23
Learning Symbolic Representations for Planning with Parameterized Skills
avatar
ALTNT
Articles
165
Tags
34
Categories
0
Follow Me
Announcement
This is my Blog
Contents
  1. 1. Diffusion Policy: Visuomotor Policy Learning via Action Diffusion
    1. 1.1. 已知的创新点:
    2. 1.2. 摘要
    3. 1.3. 1介绍
    4. 1.4. difussion policy formulation
      1. 1.4.1. 2.1扩散概率模型去噪
      2. 1.4.2. 2.2DDPM Training
      3. 1.4.3. 2.3 用于视觉运动策略学习的扩散(Diffusion for Visuomotor Policy Learning)
    5. 1.5. 具体实现
      1. 1.5.1. 3.1 Network Architecture Options
        1. 1.5.1.1. CNN-based Diffusion Policy
        2. 1.5.1.2. Time-series diffusion transformer
        3. 1.5.1.3. 建议
      2. 1.5.2. 3.2 Visual Encoder
      3. 1.5.3. 3.3 Noise Schedule(噪声计划? ——就是随训练调整预测噪声模型中的参数)
      4. 1.5.4. 3.4 Accelerating Inference for Real-time Control 实时控制加速推理
    6. 1.6. 4 扩散策略的4个有趣性质
      1. 1.6.1. 4.1模型多模态(即多种模式)动作分布
      2. 1.6.2. 4.2 Synergy with Position Control与位置控制协同
      3. 1.6.3. 4.3动作序列预测的好处
      4. 1.6.4. 4.4 Training Stability
        1. 1.6.4.1. 4.4.1为什么用于隐式策略的基于能量的模型(EBM)难以训练:
        2. 1.6.4.2. 4.4.2扩散策略和 DDPM为什么好训练:
      5. 1.6.5. 4.5与控制理论的联系
    7. 1.7. 5 Evaluation
      1. 1.7.1. 5.1仿真环境和数据集
Recent Post
周报2025年6月8日2025-06-09
The 20 m Africa rice distribution map of 20232025-06-04
DCM————DeepCropMapping: A multi-temporal deep learning approach with improved spatial generalizability for dynamic corn and soybean mapping2025-06-04
重新复现CACM记录2025-06-04
机器学习相关概念2025-05-26
©2020 - 2025 By ALTNT
Framework Hexo|Theme Butterfly