MotionCLIP icon indicating copy to clipboard operation
MotionCLIP copied to clipboard

the body of Text Motion render result is opposite

Open Zessay opened this issue 2 years ago • 5 comments

I test the text_motion script according to the readme, but the body in the render result is opposite. I don't modify any config and have no idea whether it is normal.

image

Zessay avatar Sep 09 '22 02:09 Zessay

Hi @Zessay ! What dataset did you use? please share the cmd you run to produce this result.

GuyTevet avatar Sep 10 '22 15:09 GuyTevet

Hi @Zessay ! What dataset did you use? please share the cmd you run to produce this result.

Hi, I use the pretrained model in the readme and just test with the given command line

python -m src.visualize.text2motion ./exps/paper-model/checkpoint_0100.pth.tar --input_file assets/paper_texts.txt

Zessay avatar Sep 13 '22 02:09 Zessay

I had the same issue, and I suppose its due to some miss match between data formats. There is a hack in anim.py that solves this issue, it did for me anyway, just uncomment line 105 and 106 # # invert axis motion[:, 1, :] = -motion[:, 1, :] motion[:, 2, :] = -motion[:, 2, :] # this hack is not needed for amass

larsh0103 avatar Sep 19 '22 13:09 larsh0103

Thanks! This is what I did 😄

Zessay avatar Sep 22 '22 01:09 Zessay

@larsh0103 Thx your tips! Do u know how to render the image from the text-to-emotion script?

ywdong avatar Mar 21 '23 07:03 ywdong