代码地址:isl-org/DPT: Dense Prediction Transformers (github.com)
这篇文章的环境非常好配,几乎没有报什么error,按照readme里给的步骤一两步就搞定了。
Setup
- Download the model weights and place them in the
weights
folderSet up dependencies:
pip install -r requirements.txtUsage
Place one or more input images in the folder
input
.Run a monocular depth estimation model:
python run_monodepth.pyOr run a semantic segmentation model:
python run_segmentation.pyThe results are written to the folder
output_monodepth
andoutput_semseg
, respectively.
深度估计的weights选择的是:
语义分割的weights选择的是:
下面来看看稿主跑出来的结果和稿主对此的评价吧:
1.视野较为开阔的风景照:(守序善良)
2.桌面物体近照:(中立善良)
3.小动物:(混乱邪恶)
4.特殊场景:(守序邪恶)
素描画
电影院拍摄幕布
贴了贴纸的墙
P过的图
5.二次元:(中立邪恶)