[ MO ] Random Uniform Replacer (#1814)
This commit is contained in:
@@ -540,6 +540,7 @@ extensions/middle/pass_separator.py
|
||||
extensions/middle/permute_tensor_iterator.py
|
||||
extensions/middle/preprocessing.py
|
||||
extensions/middle/quantize_fuses.py
|
||||
extensions/middle/RandomUniformReplacer.py
|
||||
extensions/middle/ReluQuantizeFuse.py
|
||||
extensions/middle/RemoveDuplicationMemory.py
|
||||
extensions/middle/RemoveIdentity.py
|
||||
|
||||
66
model-optimizer/extensions/middle/RandomUniformReplacer.py
Normal file
66
model-optimizer/extensions/middle/RandomUniformReplacer.py
Normal file
@@ -0,0 +1,66 @@
|
||||
"""
|
||||
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 typing import Dict
|
||||
|
||||
from mo.front.tf.graph_utils import create_op_with_const_inputs
|
||||
from mo.graph.graph import Graph, Node
|
||||
from mo.middle.replacement import MiddleReplacementPattern
|
||||
from mo.ops.broadcast import Broadcast
|
||||
|
||||
|
||||
class RandomUniformReplacer(MiddleReplacementPattern):
|
||||
"""
|
||||
Replaces RandomUniform operation with Broadcast of ones in sub-graph:
|
||||
|
||||
ShapeOf ---> RandomUniform ---> Mul
|
||||
|
||||
"""
|
||||
|
||||
enabled = True
|
||||
|
||||
@staticmethod
|
||||
def pattern():
|
||||
return dict(
|
||||
nodes=[
|
||||
('shape', dict(op='ShapeOf')),
|
||||
('shape_data', dict()),
|
||||
('random_uniform', dict(op='RandomUniform')),
|
||||
('random_uniform_data', dict()),
|
||||
('mul', dict(op='Mul')),
|
||||
('mul_const', dict(op='Const')),
|
||||
('mul_const_data', dict())
|
||||
],
|
||||
edges=[
|
||||
('shape', 'shape_data'),
|
||||
('shape_data', 'random_uniform'),
|
||||
('random_uniform', 'random_uniform_data'),
|
||||
('random_uniform_data', 'mul'),
|
||||
('mul_const', 'mul_const_data'),
|
||||
('mul_const_data', 'mul')
|
||||
]
|
||||
)
|
||||
|
||||
@staticmethod
|
||||
def replace_pattern(graph: Graph, match: Dict[str, Node]):
|
||||
node = match['random_uniform']
|
||||
node_name = node.soft_get('name', node.id)
|
||||
data_type = match['mul_const'].out_port(0).get_data_type()
|
||||
broadcast_node = create_op_with_const_inputs(graph, Broadcast, port_value_dict={0: np.array([1], dtype=data_type)},
|
||||
op_attrs={'name': node_name + '/Broadcast', 'mode': 'numpy'})
|
||||
node.in_port(0).get_connection().set_destination(broadcast_node.in_port(1))
|
||||
node.out_port(0).get_connection().set_source(broadcast_node.out_port(0))
|
||||
@@ -0,0 +1,98 @@
|
||||
"""
|
||||
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 unittest
|
||||
|
||||
import numpy as np
|
||||
|
||||
from extensions.middle.RandomUniformReplacer import RandomUniformReplacer
|
||||
from mo.utils.ir_engine.compare_graphs import compare_graphs
|
||||
from mo.utils.unittest.graph import build_graph
|
||||
|
||||
nodes_attributes = {
|
||||
'shape': {'kind': 'op', 'op': 'ShapeOf'},
|
||||
'shape_data': {'kind': 'data'},
|
||||
'random_uniform': {'kind': 'op', 'op': 'RandomUniform'},
|
||||
'random_uniform_data': {'kind': 'data'},
|
||||
'mul': {'kind': 'op', 'op': 'Mul'},
|
||||
'mul_const': {'kind': 'op', 'op': 'Const'},
|
||||
'mul_const_data': {'kind': 'data', 'value': np.array([1], dtype=np.int32)},
|
||||
|
||||
'broadcast': {'kind': 'op', 'op': 'Broadcast'},
|
||||
'broadcast_const': {'kind': 'op', 'op': 'Const'},
|
||||
'broadcast_const_data': {'kind': 'data', 'value': np.array([1], dtype=np.int32)},
|
||||
}
|
||||
|
||||
|
||||
class RandomUniformReplacerTest(unittest.TestCase):
|
||||
def test_1(self):
|
||||
graph = build_graph(nodes_attributes,
|
||||
edges=[
|
||||
('shape', 'shape_data'),
|
||||
('shape_data', 'random_uniform'),
|
||||
('random_uniform', 'random_uniform_data'),
|
||||
('random_uniform_data', 'mul', {'in': 0}),
|
||||
('mul_const', 'mul_const_data'),
|
||||
('mul_const_data', 'mul', {'in': 1})
|
||||
],
|
||||
nodes_with_edges_only=True)
|
||||
|
||||
ref_graph = build_graph(nodes_attributes,
|
||||
edges=[
|
||||
('shape', 'shape_data'),
|
||||
('shape_data', 'broadcast', {'in': 1}),
|
||||
('broadcast_const', 'broadcast_const_data'),
|
||||
('broadcast_const_data', 'broadcast', {'in': 0}),
|
||||
('broadcast', 'random_uniform_data'),
|
||||
('random_uniform_data', 'mul', {'in': 0}),
|
||||
('mul_const', 'mul_const_data'),
|
||||
('mul_const_data', 'mul', {'in': 1})
|
||||
],
|
||||
nodes_with_edges_only=True)
|
||||
|
||||
RandomUniformReplacer().find_and_replace_pattern(graph)
|
||||
|
||||
flag, resp = compare_graphs(graph, ref_graph, 'mul', check_op_attrs=True)
|
||||
self.assertTrue(flag, resp)
|
||||
|
||||
def test_2(self):
|
||||
graph = build_graph(nodes_attributes,
|
||||
edges=[
|
||||
('shape', 'shape_data'),
|
||||
('shape_data', 'random_uniform'),
|
||||
('random_uniform', 'random_uniform_data'),
|
||||
('random_uniform_data', 'mul', {'in': 1}),
|
||||
('mul_const', 'mul_const_data'),
|
||||
('mul_const_data', 'mul', {'in': 0})
|
||||
],
|
||||
nodes_with_edges_only=True)
|
||||
|
||||
ref_graph = build_graph(nodes_attributes,
|
||||
edges=[
|
||||
('shape', 'shape_data'),
|
||||
('shape_data', 'broadcast', {'in': 1}),
|
||||
('broadcast_const', 'broadcast_const_data'),
|
||||
('broadcast_const_data', 'broadcast', {'in': 0}),
|
||||
('broadcast', 'random_uniform_data'),
|
||||
('random_uniform_data', 'mul', {'in': 1}),
|
||||
('mul_const', 'mul_const_data'),
|
||||
('mul_const_data', 'mul', {'in': 0})
|
||||
],
|
||||
nodes_with_edges_only=True)
|
||||
|
||||
RandomUniformReplacer().find_and_replace_pattern(graph)
|
||||
|
||||
flag, resp = compare_graphs(graph, ref_graph, 'mul', check_op_attrs=True)
|
||||
self.assertTrue(flag, resp)
|
||||
Reference in New Issue
Block a user