duanyiqun / DiffusionDepth

PyTorch Implementation of introducing diffusion approach to 3D depth perception ECCV 2024
https://arxiv.org/abs/2303.05021
Apache License 2.0
306 stars 17 forks source link

noise prediction #18

Closed xiangli93 closed 1 year ago

xiangli93 commented 1 year ago

论文中是预测深度和加噪深度的差异。代码是预测噪声和设定噪声的差异。是一样的吗

duanyiqun commented 1 year ago

可以参考一下 https://github.com/huggingface/diffusers 主流diffusion model的实现方式大部分都是采样n 个t时刻算预测噪声和diffuse 结果的差异,这里是一种近似,可以提升训练效率。

xiangli93 commented 1 year ago

预测目标 训练效率会提高?