forked from mindspore-Ecosystem/mindspore
dataset sink is false when run in CPU
This commit is contained in:
parent
067a838e87
commit
fafa0d73fc
|
@ -40,6 +40,9 @@ if __name__ == "__main__":
|
|||
|
||||
args = parser.parse_args()
|
||||
|
||||
if args.device_target == "CPU":
|
||||
args.dataset_sink_mode = False
|
||||
|
||||
context.set_context(mode=context.GRAPH_MODE, device_target=args.device_target)
|
||||
ds_train = create_dataset(os.path.join(args.data_path, "train"),
|
||||
cfg.batch_size,
|
||||
|
|
Loading…
Reference in New Issue