GCN-D MMdistributedDataparallel train

yl-1993 / learn-to-cluster

Learning to Cluster Faces (CVPR 2019, CVPR 2020)

MIT License

705 stars 143 forks source link

GCN-D MMdistributedDataparallel train #95

Closed joelulu closed 1 year ago

joelulu commented 2 years ago

你好，我想用MMdistributedDataparallel训练GCN-D 更改了--distributed 为True 报错： Default process group has not been initialized, please make sure to call init_process_group

ThomaswellY commented 2 years ago

the same question, the codes seems not supportable for distributed gpus

yl-1993 commented 2 years ago

Yes. It currently does not support distributed training. We welcome any related PR!