def test_tf_resample_infer(self):
graph = build_graph(
nodes_attributes, [('node_1', 'resample'), ('resample', 'node_3'),
('node_3', 'op_output')], {
'node_3': {
'shape': None
},
'node_1': {
'shape': np.array([1, 3, 227, 227])
},
'resample': {
'antialias': 1,
'height': 384,
'width': 512,
'resample_type': 'LINEAR',
'factor': 1.0
}
})
graph.graph['layout'] = 'NCHW'
resample_node = Node(graph, 'resample')
ResampleOp.resample_infer(resample_node)
exp_shape = np.array([1, 3, 384, 512])
res_shape = graph.node['node_3']['shape']
for i in range(0, len(exp_shape)):
self.assertEqual(exp_shape[i], res_shape[i])
def test_tf_infer(self):
new_width = 100
new_height = 125
new_attrs = nodes_attributes.copy()
new_attrs.update({
'new_shape': {
'value': np.array([new_height, new_width]),
'type': 'Const',
'kind': 'op'
}
})
graph = build_graph(
new_attrs, [('node_1', 'resample'), ('new_shape', 'resample'),
('resample', 'node_3'), ('node_3', 'op_output')], {
'node_3': {
'shape': None
},
'node_1': {
'shape': np.array([1, 224, 227, 3])
},
'resample': {
'antialias': 1,
'resample_type': 'LINEAR',
'factor': 1.0,
'fw': 'tf'
}
})
graph.graph['layout'] = 'NHWC'
resample_node = Node(graph, 'resample')
ResampleOp.resample_infer(resample_node)
exp_shape = np.array([1, new_height, new_width, 3])
res_shape = graph.node['node_3']['shape']
for i in range(0, len(exp_shape)):
self.assertEqual(exp_shape[i], res_shape[i])
def replace_sub_graph(self, graph: nx.MultiDiGraph, match: dict):
log.debug('Matched NearestNeighborUpsampling pattern: {}'.format(
[node.id for node in match.values()]))
try:
input_height = match['pack_1'].in_node(1).value.item()
input_width = match['pack_1'].in_node(3).value.item()
height_scale = match['mul_const'].shape[-4]
width_scale = match['mul_const'].shape[-2]
except Exception as ex:
log.warning(
'Failed to determine scaling parameters from the topology. Do not apply pattern.'
)
return
resample_op = ResampleOp(
graph, {
'width': input_width * width_scale,
'height': input_height * height_scale,
'name': 'Resample_',
'antialias': 0,
'resample_type': 'caffe.ResampleParameter.NEAREST'
})
resample_node = resample_op.create_node([match['op']])
replace_node(match['reshape_2'], resample_node)
graph.remove_nodes_from(
[node.id for node in match.values() if node.id != match['op'].id])
def extract(node):
mapping_rule = {
'resample_type': 'caffe.ResampleParameter.NEAREST',
'fw': 'tf',
'antialias': 0
}
ResampleOp.update_node_stat(node, mapping_rule)
return __class__.enabled
def replace_pattern(graph: Graph, match: dict):
node = match['interpolate']
# common
mode = node.mode
assert mode in ['linear', 'nearest', 'cubic', 'area']
in_shape = node.in_port(0).data.get_shape()
assert in_shape is not None and len(in_shape) == 4
out_shape = node.out_port(0).data.get_shape()
assert out_shape is not None and len(out_shape) == 4
in_height, in_width = in_shape[2], in_shape[3]
out_height, out_width = out_shape[2], out_shape[3]
factor = factor_update(
None if not node.has_valid('factor') else node.factor,
[float(out_height) / in_height,
float(out_width) / in_width], [in_height, in_width],
[out_height, out_width], node.soft_get('name'))
update_attrs = {
'width': out_width,
'height': out_height,
'factor': factor,
}
if (node.has_valid('shrink_factor')
and node.has_valid('zoom_factor')) or factor is None:
del update_attrs['factor']
if node.has('factor'):
del node['factor']
if ((node.has_valid('shrink_factor') and node.shrink_factor != 1) or
(node.has_valid('zoom_factor') and node.zoom_factor != 1) or 'factor' in update_attrs) \
and ((not node.has_valid('width') or node.width == 0) and
(not node.has_valid('height') or node.height == 0)):
update_attrs['width'] = 0
update_attrs['height'] = 0
# specific
if mode in ['nearest', 'cubic', 'area'
] or node.has_and_set('convert_to_resample'):
assert not node.align_corners
assert node.pads_begin == 0 and node.pads_end == 0
update_attrs[
'resample_type'] = InterpolateToInterpOrResample.type_map[mode]
ResampleOp.update_node_stat(node, update_attrs)
node.in_port(1).disconnect()
elif mode == 'linear':
update_attrs.update({
'pad_beg': node.pads_begin,
'pad_end': node.pads_end,
'align_corners': node.align_corners,
})
InterpOp.update_node_stat(node, update_attrs)
node.in_port(1).disconnect()
node['force_precision_in_ports'] = None
def extract(node):
attrs = get_mxnet_layer_attrs(node.symbol_dict)
node_attrs = {
'type': 'Resample',
'factor': attrs.int("scale", 1),
'resample_type': 'caffe.ResampleParameter.NEAREST',
'antialias': 0
}
# update the attributes of the node
ResampleOp.update_node_stat(node, node_attrs)
return __class__.enabled
def extract(node):
mode = onnx_attr(node, 'mode', 's', default='nearest', dst_type=lambda x: x.decode())
scales = onnx_attr(node, 'scales', 'floats', dst_type=lambda x: np.array(x, dtype=np.float32))
width_scale = onnx_attr(node, 'width_scale', 'f')
height_scale = onnx_attr(node, 'height_scale', 'f')
supported_modes = ['nearest', 'linear']
if mode not in supported_modes:
raise Error(
'Error decoding Upsample node {}, mode = {} is not in the list of supported modes {}.',
node.name,
mode,
supported_modes
)
# TODO: this is a temporary limitation
if mode != 'nearest':
raise Error(
'Upsample mode {} for node {} is not supported. Only nearest is supported.',
mode,
node.name
)
# TODO: this is a temporary limitation
if scales is not None:
raise Error(
'Upsample scales attribute is defined for node {}. Only scale_width and scale_height are supported.',
node.name
)
if width_scale is None or height_scale is None:
raise Error(
'One/both of widths_scale = {} and height_scale = {} is not defined for Upsampe node {}.',
width_scale,
height_scale,
node.name
)
if width_scale != height_scale:
raise Error(
'Upsample node {} have different widths_scale = {} and height_scale = {}. It is not supported; they should match.',
node.name,
width_scale,
height_scale
)
mode_to_resample_type = {'nearest': 'caffe.ResampleParameter.NEAREST'}
assert mode in mode_to_resample_type
assert width_scale == height_scale
assert width_scale is not None
ResampleOp.update_node_stat(node, {'resample_type': mode_to_resample_type[mode], 'factor': width_scale, 'antialias': 0})
return __class__.enabled
def test_caffe_factor_infer(self):
factor = 3.0
graph = build_graph(nodes_attributes,
[('node_1', 'resample'),
('resample', 'node_3')],
{'node_3': {'is_output': True, 'shape': None},
'node_1': {'shape': np.array([1, 3, 224, 227])},
'resample': {'antialias': 1,
'resample_type': 'LINEAR',
'factor': factor}
})
graph.graph['layout'] = 'NCHW'
resample_node = Node(graph, 'resample')
ResampleOp.resample_infer(resample_node)
exp_shape = np.array([1, 3, 224 * factor, 227 * factor])
res_shape = graph.node['node_3']['shape']
for i in range(0, len(exp_shape)):
self.assertEqual(exp_shape[i], res_shape[i])
def replace_pattern(graph: Graph, match: dict):
node = match['interpolate']
assert 1 in node.in_ports() and not node.in_port(1).disconnected() and \
node.in_port(1).data.get_value() is not None, 'Interpolate node {} is corrupted: no 1-port input found'
# common
mode = node.mode
assert mode in ['linear', 'nearest', 'cubic', 'area']
in_shape = node.in_port(0).data.get_shape()
assert in_shape is not None and len(in_shape) in [4, 5]
out_shape = node.out_port(0).data.get_shape()
assert out_shape is not None and len(out_shape) in [4, 5]
in_height, in_width = in_shape[2], in_shape[3]
out_height, out_width = out_shape[2], out_shape[3]
factor = factor_update(
None if not node.has_valid('factor') else node.factor,
[float(out_height) / in_height,
float(out_width) / in_width], [in_height, in_width],
[out_height, out_width], node.soft_get('name'))
update_attrs = {
'width': out_width,
'height': out_height,
'factor': factor,
}
if (node.has_valid('shrink_factor')
and node.has_valid('zoom_factor')) or factor is None:
del update_attrs['factor']
if node.has('factor'):
del node['factor']
if ((node.has_valid('shrink_factor') and node.shrink_factor != 1) or
(node.has_valid('zoom_factor') and node.zoom_factor != 1) or 'factor' in update_attrs) \
and ((not node.has_valid('width') or node.width == 0) and
(not node.has_valid('height') or node.height == 0)):
update_attrs['width'] = 0
update_attrs['height'] = 0
# specific
if mode in ['nearest', 'cubic', 'area'
] or node.has_and_set('convert_to_resample'):
assert not node.align_corners
assert node.pads_begin == 0 and node.pads_end == 0
update_attrs[
'resample_type'] = InterpolateToInterpOrResample.type_map[mode]
ResampleOp.update_node_stat(node, update_attrs)
if not graph.graph[
'cmd_params'].keep_shape_ops or graph.graph['fw'] != 'tf':
node.in_port(1).disconnect()
else:
# we avoid making resample non-reshapable for tf version
shape = Shape(graph, {}).create_node()
node.in_port(0).get_source().connect(shape.in_port(0))
batch = node_to_get_batch_value(shape)
features = node_to_get_features_dimension_value(shape)
full_shape = new_shape_node_from_shape_nodes(
[batch, features,
node.in_port(1).get_source().node])
node.in_port(1).get_connection().set_source(
full_shape.out_port(0))
full_shape['override_output_shape'] = True
elif mode == 'linear':
assert len(in_shape) == 4, 'Interp does not support 5D input'
update_attrs.update({
'pad_beg': node.pads_begin,
'pad_end': node.pads_end,
'align_corners': node.align_corners,
})
InterpOp.update_node_stat(node, update_attrs)
node.in_port(1).disconnect()