tianxintao / online-motion-style-transfer Goto Github PK

Hi, thanks for your great work! I used deep-motion-editing before for bvh rendering, but I find their results too slender and not as pretty as yours. Could you please share your code about bvh rendering or give me some tips.

Thanks for your amazing code! I am wondering if it could be possible that you provide pretrained model for FID, and elaborate on the detailed training procedures of FID

Hi Tianxin,

As far as I know, the evaluation metric in the paper Style-ERD is FMD, not FID. Is the fid.py the feature extractor mentioned in the paper?

If yes, I have some questions about this metric.
What's the motivation why you didn't use the FMD in the paper [38]? Do you have the results of FMD in the paper [38]? What are the differences/advantages of your "Denoise Autoencoder"? Why it's called the denoise autoencoder？ I feel a bit confused about the metric in your paper.

Thanks a lot for your time!

Is it a bug in the dataloader?

When you want to generate the output style, the if statement is checking the input_style, not 'output_style'.

if input_style.sum() == 0.0:
    output_style[np.random.randint(0, len(style_labels))] = 1

Refer to:

First of all, thank you very much for taking the time to see my question.
Secondly, I want to get the dataset of Xia used in the paper, but I can't find the dataset.I find you only provide part of test data. If it's convenient, can you provide the method or link to find the dataset? Thank you very much again.