MLBEDSW-2435: Fix for cascading upscaling operators
Fixed a coordinate issue which caused the compiler to crash when
cascading upscaling operators such as ResizeBilinear.
Signed-off-by: Jacob Bohlin <jacob.bohlin@arm.com>
Change-Id: I982863573b0e5829e6d0c255dbbc308cb332a37a
diff --git a/ethosu/vela/high_level_command_stream_generator.py b/ethosu/vela/high_level_command_stream_generator.py
index 0cd3ad2..ab72fbc 100644
--- a/ethosu/vela/high_level_command_stream_generator.py
+++ b/ethosu/vela/high_level_command_stream_generator.py
@@ -75,9 +75,13 @@
strides = None
skirt = None
+ upscaling = 1
if ps.primary_op is not None:
strides = ps.primary_op.attrs.get("strides", None)
skirt = ps.primary_op.attrs.get("skirt", None)
+ if ps.primary_op.type in set(("Conv2DBackpropInputSwitchedBias", "ResizeBilinear")):
+ upscaling = ofm_tensor.shape[-3] // ifm_tensor.shape[-3]
+ assert ofm_tensor.shape[-2] == (ifm_tensor.shape[-2] * upscaling)
concat_axis = 0
concat_offset = 0
@@ -113,13 +117,13 @@
if ifm_tensor.shape != []:
ifm_box, _, _ = ofm_box.transform_with_strides_and_skirt(
- strides, skirt, ifm_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0]
+ strides, skirt, ifm_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0], upscaling
)
else:
ifm_box = Box([], [])
if ifm2_tensor is not None and ifm2_tensor.shape != []:
ifm2_box, _, _ = ofm_box.transform_with_strides_and_skirt(
- strides, skirt, ifm2_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[1]
+ strides, skirt, ifm2_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[1], upscaling
)
else:
ifm2_box = Box([], [])
@@ -127,7 +131,7 @@
for intermediate in ps.intermediates:
if intermediate != None and intermediate.shape != [] and intermediate.purpose == TensorPurpose.FeatureMap:
intermediate_box, _, _ = ofm_box.transform_with_strides_and_skirt(
- strides, skirt, intermediate.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0]
+ strides, skirt, intermediate.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0], upscaling
)
yield from dma_if_necessary(ps, intermediate_box, intermediate)
@@ -214,13 +218,13 @@
k_height = weight_tensor.shape[0]
ifm_box, pad_top, pad_bottom = ofm_box.transform_with_strides_and_skirt(
- strides, skirt, ifm_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0], k_height
+ strides, skirt, ifm_tensor.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0], k_height, upscaling
)
for intermediate in ps.intermediates:
if intermediate != None and intermediate.shape != [] and intermediate.purpose == TensorPurpose.FeatureMap:
intermediate_box, _, _ = ofm_box.transform_with_strides_and_skirt(
- strides, skirt, intermediate.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0]
+ strides, skirt, intermediate.shape, npu_block_type, concat_axis, concat_offset, split_offsets[0], upscaling
)
yield from dma_if_necessary(ps, intermediate_box, intermediate)