6766 data analyser label argmax (#6852)

Fixes #6766

### Description
the label might have been processed during the GPU transform,
in the retry on CPU, the argmax should be skipped in this case.

### Types of changes
<!--- Put an `x` in all the boxes that apply, and remove the not
applicable items -->
- [x] Non-breaking change (fix or new feature that would not break
existing functionality).
- [ ] Breaking change (fix or new feature that would cause existing
functionality to change).
- [ ] New tests added to cover the changes.
- [ ] Integration tests passed locally by running `./runtests.sh -f -u
--net --coverage`.
- [ ] Quick tests passed locally by running `./runtests.sh --quick
--unittests --disttests`.
- [ ] In-line docstrings updated.
- [ ] Documentation updated, tested `make html` command in the `docs/`
folder.

Signed-off-by: Wenqi Li <[email protected]>

monai/apps/auto3dseg/data_analyzer.py

@@ -332,9 +332,11 @@ def _get_all_case_stats(
             batch_data = batch_data[0]
             try:
                 batch_data[self.image_key] = batch_data[self.image_key].to(device)
+                _label_argmax = False
                 if self.label_key is not None:
                     label = batch_data[self.label_key]
                     label = torch.argmax(label, dim=0) if label.shape[0] > 1 else label[0]
+                    _label_argmax = True  # track if label is argmaxed
                     batch_data[self.label_key] = label.to(device)
                 d = summarizer(batch_data)
             except BaseException as err:
@@ -348,7 +350,8 @@ def _get_all_case_stats(
                     batch_data[self.image_key] = batch_data[self.image_key].to("cpu")
                     if self.label_key is not None:
                         label = batch_data[self.label_key]
-                        label = torch.argmax(label, dim=0) if label.shape[0] > 1 else label[0]
+                        if not _label_argmax:
+                            label = torch.argmax(label, dim=0) if label.shape[0] > 1 else label[0]
                         batch_data[self.label_key] = label.to("cpu")
                     d = summarizer(batch_data)