RuntimeError: grad can be implicitly created only for scalar outputs

[realmadridchenwentao]

When I fine-tune the pre-trained model weights in SIIM segmentation tasks, the following error is reported:

  | Name  | Type      | Params
------------------------------------
0 | model | Unet      | 32.5 M
1 | loss  | MixedLoss | 0     
------------------------------------
32.5 M    Trainable params
0         Non-trainable params
32.5 M    Total params
Epoch 0:   0%|          | 0/669 [00:01<?, ?it/s]                      
Traceback (most recent call last):
  File "run.py", line 167, in <module>
    main(cfg, args)
  File "run.py", line 106, in main
    trainer.fit(model, dm)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 473, in fit
    results = self.accelerator_backend.train()
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/accelerators/dp_accelerator.py", line 110, in train
    results = self.train_or_test()
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/accelerators/accelerator.py", line 69, in train_or_test
    results = self.trainer.train()
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 524, in train
    self.train_loop.run_training_epoch()
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 572, in run_training_epoch
    batch_output = self.run_training_batch(batch, batch_idx, dataloader_idx)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 730, in run_training_batch
    self.optimizer_step(optimizer, opt_idx, batch_idx, train_step_and_backward_closure)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 513, in optimizer_step
    using_lbfgs=is_lbfgs,
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/core/lightning.py", line 1261, in optimizer_step
    optimizer.step(closure=optimizer_closure)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/core/optimizer.py", line 286, in step
    self.__optimizer_step(*args, closure=closure, profiler_name=profiler_name, **kwargs)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/core/optimizer.py", line 140, in __optimizer_step
    trainer.precision_connector.backend.optimizer_step(trainer, optimizer, closure)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/plugins/native_amp.py", line 75, in optimizer_step
    closure()
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 725, in train_step_and_backward_closure
    self.trainer.hiddens
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 828, in training_step_and_backward
    self.backward(result, optimizer, opt_idx)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/trainer/training_loop.py", line 851, in backward
    result.closure_loss, optimizer, opt_idx, *args, **kwargs
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/accelerators/accelerator.py", line 99, in backward
    closure_loss, optimizer, opt_idx, *args, **kwargs
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/plugins/native_amp.py", line 47, in backward
    model.backward(closure_loss, optimizer, opt_idx)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/pytorch_lightning/core/lightning.py", line 1158, in backward
    loss.backward(*args, **kwargs)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/torch/tensor.py", line 221, in backward
    torch.autograd.backward(self, gradient, retain_graph, create_graph)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/torch/autograd/__init__.py", line 126, in backward
    grad_tensors_ = _make_grads(tensors, grad_tensors_)
  File "/home/wentaochen/anaconda3/envs/gloria/lib/python3.7/site-packages/torch/autograd/__init__.py", line 50, in _make_grads
    raise RuntimeError("grad can be implicitly created only for scalar outputs")
RuntimeError: grad can be implicitly created only for scalar outputs

Hope you can give me some advice! Thank you very much!