yoctta

the batch size is set to 48 in the paper but 16 in the config code, because we use distributed data parallel. Learning rate set as 1e-4 is OK. And...

Hello , We use the face detector on c0 videos for getting the face landmarks and use the same landmarks to preprocess c23 ,c40 videos for convenient. From: SamHSlva Sent:...

• Retinaface is not really needed because I’ve released the pre-computed landmarks, you can edit the script to comment it. The server will dispatch jobs to one or multiple clients...

Please check the update, you should pretrain the backbone and init the multi-attention model with a pretrained checkpoint From: IItaly Sent: 2021年7月6日 16:05 To: yoctta/multiple-attention Cc: yoctta; Mention Subject: [yoctta/multiple-attention]...

I've updated main.py to clearify it. ckpt_35.pth should be a pretrained backbone model. From: ***@***.*** ***@***.***> on behalf of IItaly ***@***.***>Sent: Tuesday, July 6, 2021, 16:05To: yoctta/multiple-attentionCc: yoctta; MentionSubject: [yoctta/multiple-attention]...

You can try to run it, I think the code can also work with 1 GPU, but limited batch size may be an issue.

> I'm knew to this field and I'm also not able to run this code please guide!! Please show your dataset directories and error message.

The entrance is main.py, and you should run pretrain first. From: Dave Alexis Sent: Monday, May 23, 2022 10:38 AM To: yoctta/multiple-attention Cc: Subscribed Subject: [yoctta/multiple-attention] Anyone successfully run this...

是计算所的吗，应该给你们发过了 From: Daichi Zhang Sent: Monday, October 25, 2021 8:14 PM To: yoctta/multiple-attention Cc: Subscribed Subject: Re: [yoctta/multiple-attention] How to load the pretrained weights? (#11) hello, have u solved the...

Init Distributed Data Parallel for training with more than 1 computers. From: Xin-hhh Sent: Wednesday, November 3, 2021 3:30 PM To: yoctta/multiple-attention Cc: Subscribed Subject: [yoctta/multiple-attention] What address does the...