IMAGE COMPLEMENTATION STACK SOLUTION METHOD IN SIGGRAPH 2017_impainting

I'm a beginner at programming and I've only touched the code for a few days, but I really wanted to supplement the image with AI, so I started programming that I'm not used to.

What you tried
SIGGRAPH thesis

I spent a few days on my own building the environment using Docker and Ubuntu's LTS version 20.4 to install and use the package.However, this error occurred in the process of complementing the image and it continues to stack.

root@82a52e3b8c9f:/mount#thimpaint.lua --input example.png --mask example_mask.png
{
gpu —false
mask: "example_mask.png"
postproc:false
model: "completionnet_places2_freeform.t7"
maxdim —600
input: "example.png"
}
Loading model...
/root/torch/install/bin/luajit:impaint.lua:78:inconsistent tensor size, expected
tensor [4 x 336 x 600] and mask [3 x 336 x 600] to have the same number of elements,
but got 806400 and 604800 elements respectively at
/root/torch/pkg/torch/lib/TH/generic/THTensorMath.c:138
stack traceback:
[C]: in function 'maskedFill'
impaint.lua:78:in main chunk
[C]: in function 'dofile'
/root/torch/install/lib/luarocks/rocks/trepl/scm-1/bin/th:150:in main chunk
[C]—at 0x00405d50

I looked up the error code on the Internet, but I'm at a loss because nothing fits me well.As an amateur, the size of the image was unified, and I made it with example.png and example_mask.png with the first name.However, it has not been resolved at all.

And there was no article in the SIGGRAPH paper that said such an error occurred, and I couldn't find a way to correct it, and I was so angry that I couldn't help but wonder why I was the only one who had to go through this.Can someone solve this problem?