RuntimeError: DataLoader worker is killed by signal: Killed.

[jjhhan]

When I update to 0.9， The training process that was normally running under 0.8 collapsed.
Originally only 10 g was needed, but it suddenly soared to 60 g after reading the data.This is a screenshot of my Linux - htop command. Not long after I started training

I have tried to modify num_workers and cases_num, also tried the original monai.data.dataset.It doesn't work.
Mycode is : I've deleted most of transforms
train_transforms = [
LoadImaged(keys=['image', 'label']),
EnsureChannelFirstd(keys=['image', 'label']),
CropForegroundd(keys=['image', 'label'], source_key='image'), ]
train_transforms = Compose(train_transforms)
train_ds = CacheDataset(
data=train_dicts, transform=train_transforms, cache_num=6, copy_cache=True,
cache_rate=1.0, num_workers=4)
train_loader = DataLoader(train_ds, batch_size=2, shuffle=False, collate_fn=list_data_collate, num_workers=4, pin_memory=True)