when I set use_ema=True in run_cli(), the error is shown as below.So the ema is not supported in the version of pytorch lightning?
File "/data/BEVDepth/callbacks/ema.py", line 82, in on_fit_start
callback.on_fit_start(self, self.lightning_module)
File "/data/BEVDepth/callbacks/ema.py", line 82, in on_fit_start
trainer.ema_model = ModelEMA(trainer.model.module.module.model.cuda(),
File "/home/gongzheng/anaconda3/envs/bev/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1
131, in __getattr__
trainer.ema_model = ModelEMA(trainer.model.module.module.model.cuda(),
File "/home/gongzheng/anaconda3/envs/bev/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1
131, in __getattr__
type(self).__name__, name))
AttributeError: 'BEVDepthLightningModel' object has no attribute 'module'
type(self).__name__, name))
when I set use_ema=True in run_cli(), the error is shown as below.So the ema is not supported in the version of pytorch lightning?