AttributeError: 'MMDistributedDataParallel' object has no attribute '_sync_params'

I am trying to train a Detector using the latest mmcv version and torch 1.11. I installed mmcv simply by:

git clone https://github.com/open-mmlab/mmcv.git
pip install -r requirements/optional.txt
MMCV_WITH_OPS=1 pip install -e .

After launching the training script, I get an error:

Traceback (most recent call last):
  File "/workspace/Project/launch_training.py", line 190, in <module>
    main()
  File "/workspace/Project/launch_training.py", line 179, in main
    train_detector(
  File "/opt/conda/lib/python3.8/site-packages/mmdet/apis/train.py", line 208, in train_detector
    runner.run(data_loaders, cfg.workflow)
  File "/opt/conda/lib/python3.8/site-packages/mmcv/runner/epoch_based_runner.py", line 127, in run
    epoch_runner(data_loaders[i], **kwargs)
  File "/opt/conda/lib/python3.8/site-packages/mmcv/runner/epoch_based_runner.py", line 50, in train
    self.run_iter(data_batch, train_mode=True, **kwargs)
  File "/opt/conda/lib/python3.8/site-packages/mmcv/runner/epoch_based_runner.py", line 29, in run_iter
    outputs = self.model.train_step(data_batch, self.optimizer,
  File "/opt/conda/lib/python3.8/site-packages/mmcv/parallel/distributed.py", line 48, in train_step
    self._sync_params()
  File "/opt/conda/lib/python3.8/site-packages/torch/nn/modules/module.py", line 1185, in __getattr__
    raise AttributeError("'{}' object has no attribute '{}'".format(
AttributeError: 'MMDistributedDataParallel' object has no attribute '_sync_params'

I did some digging into the issue and I found this PR here: https://github.com/pytorch/pytorch/pull/64514 which indicates that _sync_params is deprecated from PyTorch. Is there any plan to fix this issue soon or is there a workaround this?

[teamwong111]

Any updates on this? I met the same problem, and currently, it seems not straightforward to install PyTorch 1.10 other than building from the source. Thanks!

We will fix it ASAP. I have submitted a pr.

[teamwong111]

Which version of PyTorch do you use? The latest stable PyTorch(1.10.2) still uses this function, as far as I know. We will handle this problem when PyTorch1.11.0 is released.