更新 'train_for_c2net.py'

This commit is contained in:
liuzx 2023-06-12 10:55:50 +08:00
parent 81273b6c2c
commit 7aab8b05c6
1 changed files with 8 additions and 5 deletions

View File

@ -112,10 +112,13 @@ if __name__ == "__main__":
if (args.epoch_size):
epoch_size = args.epoch_size
print('epoch_size is: ', epoch_size)
model.train(epoch_size, ds_train,callbacks=[time_cb, ckpoint_cb,LossMonitor()])
# set callback functions
callback =[time_cb,LossMonitor()]
local_rank=int(os.getenv('RANK_ID'))
# callback =[time_cb,LossMonitor()]
# local_rank=int(os.getenv('RANK_ID'))
# for data parallel, only save checkpoint on rank 0
if local_rank==0 :
callback.append(ckpoint_cb)
model.train(epoch_size,ds_train,callbacks=callback)
# if local_rank==0 :
# callback.append(ckpoint_cb)
# model.train(epoch_size,
# ds_train,
# callbacks=callback)