TOSA: Elementwise Rank > 4 and Batch > 1
Added support for elementwise operations:
-Support for up to Rank == 6
-Support for Batch > 1 for Rank == 4
-For binary elementwise ops this includes handling
of broadcasting in dimensions above H-dimension
Signed-off-by: Patrik Gustavsson <patrik.gustavsson@arm.com>
Change-Id: I73850bbfb288077a99bd2ceecbf989172016da24
diff --git a/ethosu/vela/tensor.py b/ethosu/vela/tensor.py
index 37fd06e..2e70d72 100644
--- a/ethosu/vela/tensor.py
+++ b/ethosu/vela/tensor.py
@@ -632,7 +632,7 @@
self, coord: Optional[Shape] = None, shape4D: Optional[Shape4D] = None
) -> Tuple[Optional[Shape], Optional[Shape]]:
if coord is None:
- coord = [0] * len(self.storage_shape)
+ coord = [0] * min(len(self.storage_shape), 4)
if shape4D and self.is_standard_fm:
augmented_shape = self.get_4D_storage_shape_for_shape(shape4D).as_list()