compat. torch 1.7/1.8 sw integration

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

monai/data/meta_tensor.py

@@ -66,6 +66,8 @@ class MetaTensor(MetaObj, torch.Tensor):
           not work if `im` is of type `MetaTensor`. This can be resolved with
           `torch.jit.trace(net, im.as_tensor())`.
         - For pytorch < 1.8, sharing `MetaTensor` instances across processes may not be supported.
+        - For pytorch < 1.9, next(iter(meta_tensor)) returns a torch.Tensor.
+          see: https://github.com/pytorch/pytorch/issues/54457
         - A warning will be raised if in the constructor `affine` is not `None` and
           `meta` already contains the key `affine`.
         - You can query whether the `MetaTensor` is a batch with the `is_batch` attribute.

monai/inferers/utils.py

@@ -137,7 +137,7 @@ def sliding_window_inference(
         diff = max(roi_size[k - 2] - inputs.shape[k], 0)
         half = diff // 2
         pad_size.extend([half, diff - half])
-    inputs = F.pad(inputs, pad=pad_size, mode=look_up_option(padding_mode, PytorchPadMode).value, value=cval)
+    inputs = F.pad(inputs, pad=pad_size, mode=look_up_option(padding_mode, PytorchPadMode), value=cval)
 
     scan_interval = _get_scan_interval(image_size, roi_size, num_spatial_dims, overlap)
 

tests/test_integration_sliding_window.py

@@ -24,7 +24,7 @@
 from monai.networks import eval_mode, predict_segmentation
 from monai.networks.nets import UNet
 from monai.transforms import AddChannel, SaveImage
-from monai.utils import set_determinism
+from monai.utils import pytorch_after, set_determinism
 from tests.utils import DistTestCase, TimedCall, make_nifti_image, skip_if_quick
 
 
@@ -47,8 +47,11 @@ def _sliding_window_processor(_engine, batch):
             return predict_segmentation(seg_probs)
 
     def save_func(engine):
-        for m in engine.state.output:
-            saver(m)
+        if pytorch_after(1, 9, 1):
+            for m in engine.state.output:
+                saver(m)
+        else:
+            saver(engine.state.output[0])
 
     infer_engine = Engine(_sliding_window_processor)
     infer_engine.add_event_handler(Events.ITERATION_COMPLETED, save_func)