Yifeng

I did the following things and mxnet is installed successfully, but **I haven't checked the performance(mAP) of FGFA with this mxnet**. 1. After finishing these commands: ``` git clone --recursive...

I encounter the same problem, so is there any solution?

Thank you all the same > On Sep 6, 2019, at 7:47 PM, JBramauer wrote: > > I'm sorry, but I don't have access to a pepper robot any more....

Thank you very much for your help! 😊

The official VoxCeleb2 website has already removed the file containing youtube URLs. Currently, the publicly available VoxCeleb2 URL file can be found here: https://mmai.io/datasets/voxceleb/ (which includes both VoxCeleb1 and 2).

> Sorry. I am clearly missed something. Can you clarify please. > > What I am must do for getting result video where man from "cnn.mp4" video speak with voice...

Thanks for your attention. The inference code and checkpoint for DreamTalk are now available at https://github.com/ali-vilab/dreamtalk . As of now, we do not have plans to release the training code....

The code has been tested without issues on Ubuntu. Based on the error message you received, you might consider installing an audio I/O backend.

您好. 您的错误应该是python版本的问题. 可以尝试更高版本 比如3.8 或 3.9 等等( 根据https://stackoverflow.com/questions/75529492/importerror-cannot-import-name-ordereddict-from-typing 只要3.7.2以上即可) 您这个错误提示是python type hint的问题, 和cuda版本等无关, 和python版本有关.

您可以尝试 windows系统: pip install soundfile linux系统: pip install sox 参见 #2