deropty

significant GPU memory cost on PRID

1

Here is my configuration. ```python model: name: 'resnet50' pretrained: True data: type: 'video' sources: ['prid2011'] targets: ['prid2011'] height: 224 width: 112 combineall: False transforms: ['random_flip'] save_dir: 'log/osnet_x1_0_prid2011_softmax_cosinelr' loss: name: 'softmax'...

Is there a DDP training code?

7

Thank you for sharing the fantastic work. As I do not have the SLURM cluster, Is there the DDP training code? Or anyone can help?