Provide GatherND with original layout for inputs and output (#3002)

* Provide GatherND with original layout for inputs and output

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>

* Fix code review #1

Signed-off-by: Roman Kazantsev <roman.kazantsev@intel.com>
This commit is contained in:
Roman Kazantsev 2020-11-10 17:24:04 +03:00 committed by GitHub
parent 4a0170dcb0
commit e3b879ad3b
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
3 changed files with 201 additions and 0 deletions

View File

@ -28,6 +28,7 @@ extensions/back/GroupedConvWeightsNormalize.py
extensions/back/insert_compatibility_l2normalization.py extensions/back/insert_compatibility_l2normalization.py
extensions/back/InterpolateReshape.py extensions/back/InterpolateReshape.py
extensions/back/kaldi_remove_memory_output.py extensions/back/kaldi_remove_memory_output.py
extensions/back/LayoutChangeForGatherND.py
extensions/back/LeakyReLUMutation.py extensions/back/LeakyReLUMutation.py
extensions/back/LRNToNorm.py extensions/back/LRNToNorm.py
extensions/back/MatMulNormalizer.py extensions/back/MatMulNormalizer.py

View File

@ -0,0 +1,61 @@
"""
Copyright (C) 2018-2020 Intel Corporation
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
"""
import numpy as np
from extensions.ops.transpose import Transpose
from mo.front.common.partial_infer.utils import int64_array
from mo.front.tf.graph_utils import create_op_with_const_inputs
from mo.graph.graph import Graph, Port
from mo.back.replacement import BackReplacementPattern
class LayoutChangeForGatherND(BackReplacementPattern):
"""
Return original layout for inputs and output of GatherND operation
since the operation is designed for NHWC layout.
"""
enabled = True
force_shape_inference = True
graph_condition = [lambda graph: graph.graph['fw'] == 'tf']
@staticmethod
def insert_transpose(graph: Graph, input_port: Port, before_input=True):
input_rank = len(input_port.data.get_shape())
if input_rank > 3:
if before_input:
axis_order = np.concatenate((int64_array([0]),
int64_array(list(range(2, input_rank))),
int64_array([1])))
source_node = input_port.get_source().node
transpose_name = source_node.soft_get('name', source_node.id) + '/TransposeToNHWC'
else:
axis_order = np.concatenate(
(int64_array([0]),
int64_array([input_rank - 1]),
int64_array(list(range(1, input_rank - 1)))))
transpose_name = input_port.node.soft_get('name', input_port.node.id) + '/TransposeToNCHW'
input_port.node['need_shape_inference'] = True
input_port.node['override_output_shape'] = True
transpose = create_op_with_const_inputs(graph, Transpose, {1: axis_order}, {'name': transpose_name})
input_port.get_connection().insert_node(transpose)
transpose['need_shape_inference'] = True
transpose['override_output_shape'] = True
def find_and_replace_pattern(self, graph: Graph):
for gathernd in graph.get_op_nodes(type='GatherND'):
self.insert_transpose(graph, gathernd.in_port(0), before_input=True)
self.insert_transpose(graph, gathernd.in_port(1), before_input=True)
self.insert_transpose(graph, gathernd.out_port(0), before_input=False)

View File

@ -0,0 +1,139 @@
"""
Copyright (C) 2020 Intel Corporation
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
"""
import unittest
import numpy as np
from extensions.back.LayoutChangeForGatherND import LayoutChangeForGatherND
from mo.front.common.partial_infer.utils import int64_array
from mo.utils.ir_engine.compare_graphs import compare_graphs
from mo.utils.unittest.graph import build_graph
nodes_attributes = {
'placeholder_1': {'shape': None, 'type': 'Parameter', 'kind': 'op', 'op': 'Parameter'},
'placeholder_1_data': {'value': None, 'shape': None, 'kind': 'data', 'data_type': None},
'placeholder_2': {'shape': None, 'type': 'Parameter', 'kind': 'op', 'op': 'Parameter'},
'placeholder_2_data': {'value': None, 'shape': None, 'kind': 'data', 'data_type': None},
# GatherND
'gathernd': {'type': 'GatherND', 'kind': 'op', 'op': 'GatherND'},
'gathernd_data': {'value': None, 'shape': None, 'kind': 'data'},
# Result layer
'result': {'type': 'Result', 'kind': 'op', 'op': 'Result'},
# Transpose layers
'transpose_1': {'type': 'Transpose', 'kind': 'op', 'op': 'Transpose', 'need_shape_inference': True},
'transpose_1_data': {'value': None, 'shape': None, 'kind': 'data'},
'axis_1_const': {'type': 'Const', 'kind': 'op', 'op': 'Const', 'value': None},
'axis_1_const_data': {'kind': 'data', 'value': None, 'shape': None},
'transpose_2': {'type': 'Transpose', 'kind': 'op', 'op': 'Transpose', 'need_shape_inference': True},
'transpose_2_data': {'value': None, 'shape': None, 'kind': 'data'},
'axis_2_const': {'type': 'Const', 'kind': 'op', 'op': 'Const', 'value': None},
'axis_2_const_data': {'kind': 'data', 'value': None, 'shape': None},
'transpose_3': {'type': 'Transpose', 'kind': 'op', 'op': 'Transpose', 'need_shape_inference': True},
'transpose_3_data': {'value': None, 'shape': None, 'kind': 'data'},
'axis_3_const': {'type': 'Const', 'kind': 'op', 'op': 'Const', 'value': None},
'axis_3_const_data': {'kind': 'data', 'value': None, 'shape': None},
}
class LayoutChangeForGatherNDTests(unittest.TestCase):
def test_tf_all_ports(self):
graph = build_graph(nodes_attributes,
[('placeholder_1', 'placeholder_1_data'),
('placeholder_2', 'placeholder_2_data'),
('placeholder_1_data', 'gathernd'),
('placeholder_2_data', 'gathernd'),
('gathernd', 'gathernd_data'),
('gathernd_data', 'result'),
],
{'placeholder_1_data': {'shape': np.array([1, 3, 224, 224])},
'placeholder_2_data': {'shape': np.array([1, 3, 224, 224])},
'gathernd_data': {'shape': np.array([1, 3, 224, 224])},
})
graph.graph['fw'] = 'tf'
graph_ref = build_graph(nodes_attributes,
[('placeholder_1', 'placeholder_1_data'),
('placeholder_2', 'placeholder_2_data'),
('placeholder_1_data', 'transpose_1'),
('axis_1_const', 'axis_1_const_data'),
('axis_1_const_data', 'transpose_1'),
('transpose_1', 'transpose_1_data'),
('placeholder_2_data', 'transpose_2'),
('axis_2_const', 'axis_2_const_data'),
('axis_2_const_data', 'transpose_2'),
('transpose_2', 'transpose_2_data'),
('transpose_1_data', 'gathernd'),
('transpose_2_data', 'gathernd'),
('gathernd', 'gathernd_data'),
('gathernd_data', 'transpose_3'),
('axis_3_const', 'axis_3_const_data'),
('axis_3_const_data', 'transpose_3'),
('transpose_3', 'transpose_3_data'),
('transpose_3_data', 'result'),
],
{'placeholder_1_data': {'shape': np.array([1, 3, 224, 224])},
'placeholder_2_data': {'shape': np.array([1, 3, 224, 224])},
'axis_1_const_data': {'value': int64_array([0, 2, 3, 1])},
'axis_2_const_data': {'value': int64_array([0, 2, 3, 1])},
'gathernd_data': {'shape': np.array([1, 3, 224, 224])},
'axis_3_const_data': {'value': int64_array([0, 3, 1, 2])},
})
pattern = LayoutChangeForGatherND()
pattern.find_and_replace_pattern(graph)
(flag, resp) = compare_graphs(graph, graph_ref, 'result', check_op_attrs=True)
self.assertTrue(flag, resp)
def test_tf_one_ports(self):
graph = build_graph(nodes_attributes,
[('placeholder_1', 'placeholder_1_data'),
('placeholder_2', 'placeholder_2_data'),
('placeholder_1_data', 'gathernd'),
('placeholder_2_data', 'gathernd'),
('gathernd', 'gathernd_data'),
('gathernd_data', 'result'),
],
{'placeholder_1_data': {'shape': np.array([1, 3, 224, 224])},
'placeholder_2_data': {'shape': np.array([1, 3])},
'gathernd_data': {'shape': np.array([1, 3])},
})
graph.graph['fw'] = 'tf'
graph_ref = build_graph(nodes_attributes,
[('placeholder_1', 'placeholder_1_data'),
('placeholder_2', 'placeholder_2_data'),
('placeholder_1_data', 'transpose_1'),
('axis_1_const', 'axis_1_const_data'),
('axis_1_const_data', 'transpose_1'),
('transpose_1', 'transpose_1_data'),
('transpose_1_data', 'gathernd'),
('placeholder_2_data', 'gathernd'),
('gathernd', 'gathernd_data'),
('gathernd_data', 'result'),
],
{'placeholder_1_data': {'shape': np.array([1, 3, 224, 224])},
'placeholder_2_data': {'shape': np.array([1, 3])},
'axis_1_const_data': {'value': int64_array([0, 2, 3, 1])},
'gathernd_data': {'shape': np.array([1, 3])}
})
pattern = LayoutChangeForGatherND()
pattern.find_and_replace_pattern(graph)
(flag, resp) = compare_graphs(graph, graph_ref, 'result', check_op_attrs=True)
self.assertTrue(flag, resp)