tf_optimizer.py

#   Copyright (c) 2019  PaddlePaddle Authors. All Rights Reserved.
#
# 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.

# TODO useless node remove
from x2paddle.op_mapper.tf_op_mapper import TFOpMapper
from x2paddle.core.fluid_code import Layer
from x2paddle.core.util import *
import copy as cp


class TFOptimizer(object):
    activation_ops = {
        'Relu': 'relu',
        'Sigmoid': 'sigmoid',
        'Relu6': 'relu6',
        'swish_f32': 'swish'
    }
    layers_with_act = [
        'Conv2D', 'BiasAdd', 'DepthwiseConv2dNative', 'Conv2DBackpropInput',
        'FusedBatchNorm', 'conv2d', 'elementwise_add', 'conv2d_transpose',
        'batch_norm'
    ]
    layers_with_bias = [
        'Conv2D', 'DepthwiseConv2dNative', 'Conv2DBackpropInput', 'conv2d',
        'conv2d_transpose'
    ]

    def __init__(self, op_mapper):
        self.op_mapper = op_mapper
        self.graph = op_mapper.graph

    def delete_redundance_code(self):
        for node_name in self.graph.topo_sort:
            if node_name in self.op_mapper.omit_nodes:
                node = self.graph.get_node(node_name)
                if node is None:
                    continue
                omit_freq = self.op_mapper.omit_nodes.count(node_name)
                if len(node.outputs) <= omit_freq:
                    node.fluid_code.clear()

                    # remove node from graph
                    input_names = node.inputs
                    output_names = node.outputs
                    for in_name in input_names:
                        in_node = self.graph.get_node(in_name)
                        index = in_node.outputs.index(node_name)
                        del in_node.outputs[index]
                    for out_name in output_names:
                        out_node = self.graph.get_node(out_name)
                        index = out_node.inputs.index(node_name)
                        del out_node.inputs[index]
                    del self.graph.node_map[node_name]

    def strip_graph(self):
        visited_nodes = set()

        def visit(node_name):
            if node_name in visited_nodes:
                return
            visited_nodes.add(node_name)
            input_names = self.graph.get_node(node_name).inputs
            for in_name in input_names:
                visit(in_name)

        for node_name in self.graph.output_nodes:
            visit(node_name)

        for i, node_name in enumerate(self.graph.topo_sort):
            if node_name not in visited_nodes:
                node = self.graph.get_node(node_name)
                if node is None:
                    continue
                input_names = node.inputs
                output_names = node.outputs
                for in_name in input_names:
                    in_node = self.graph.get_node(in_name)
                    index = in_node.outputs.index(node_name)
                    del in_node.outputs[index]
                for out_name in output_names:
                    out_node = self.graph.get_node(out_name)
                    index = out_node.inputs.index(node_name)
                    del out_node.inputs[index]
                del self.graph.node_map[node_name]

    def merge_activation(self):
        act_nodes = list()
        for node_name in self.graph.topo_sort:
            node = self.graph.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in self.activation_ops:
                act_nodes.append(node_name)

        for act_node_name in act_nodes:
            node = self.graph.get_node(act_node_name)
            input = self.graph.get_node(node.inputs[0])
            if input.layer_type not in self.layers_with_act:
                continue
            if len(input.fluid_code.layers) == 0:
                continue
            if 'act' in input.fluid_code.layers[
                    -1].param_attr and input.fluid_code.layers[-1].param_attr[
                        'act'] is not None:
                continue
            if len(input.outputs) != 1:
                continue
            index = -1
            for i in range(len(input.fluid_code.layers)):
                if input.fluid_code.layers[i].op in self.layers_with_act:
                    index = i
                    break
            input.fluid_code.layers[index].param_attr['act'] = string(
                self.activation_ops[node.layer_type])
            input.fluid_code.layers[-1].output = node.fluid_code.layers[
                0].output
            self.graph.remove_node(act_node_name)

    def merge_bias(self):
        for node_name in self.graph.topo_sort:
            node = self.graph.get_node(node_name)
            if node is None:
                continue
            if node.layer_type == "BiasAdd":
                input = self.graph.get_node(node.inputs[0])
                if input.layer_type not in self.layers_with_bias:
                    continue
                if len(input.outputs) != 1:
                    continue
                if len(input.fluid_code.layers) == 0:
                    continue
                bias_with_act = False
                if 'act' in node.fluid_code.layers[-1].param_attr:
                    bias_with_act = True
                layer_with_act = False
                index = -1
                for i in range(len(input.fluid_code.layers)):
                    if input.fluid_code.layers[i].op in self.layers_with_bias:
                        index = i
                        break
                if 'act' in input.fluid_code.layers[
                        index].param_attr and input.fluid_code.layers[
                            index].param_attr['act'] is not None:
                    layer_with_act = True

                if bias_with_act and layer_with_act:
                    continue
                if not input.fluid_code.layers[index].param_attr['bias_attr']:
                    bias_name = node.inputs[1]
                    input.fluid_code.layers[index].param_attr[
                        'bias_attr'] = string(bias_name)
                    input.fluid_code.layers[-1].output = node.fluid_code.layers[
                        0].output
                    if bias_with_act:
                        input.fluid_code.layers[index].param_attr[
                            'act'] = node.fluid_code.layers[-1].param_attr[
                                'act']
                    node.fluid_code.clear()
                    self.graph.remove_node(node.layer_name)

    def remove_transpose(self):
        graph_copy = cp.deepcopy(self.graph)
        nhwc_insensitive_ops = [
            'Relu', 'Relu6', 'Abs', 'Sigmoid', 'Exp', 'Rsqrt', 'swish_f32',
            'LeakyRelu', 'Cast'
        ]
        elementwise_ops = [
            'Sub', 'Add', 'RealDiv', 'Maximum', 'Mul', 'FloorDiv',
            'GreaterEqual'
        ]
        for node_name in self.graph.topo_sort:
            node = graph_copy.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in nhwc_insensitive_ops:
                graph_copy.remove_node(node_name)

        optimize_ops = [
            'Conv2D', 'MaxPool', 'FusedBatchNorm', 'DepthwiseConv2dNative',
            'AvgPool', 'Pad', 'Conv2DBackpropInput', 'ResizeNearestNeighbor',
            'ResizeBilinear', "Placeholder"
        ]

        for node_name in self.graph.topo_sort:
            node = graph_copy.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in elementwise_ops:
                is_nhwc = True
                for in_name in node.inputs:
                    in_node = graph_copy.get_node(in_name)
                    if hasattr(in_node, "is_nhwc"):
                        if not in_node.is_nhwc:
                            is_nhwc = False
                    else:
                        if len(in_node.fluid_code.layers) < 2:
                            is_nhwc = False
                            continue
                        if in_node.fluid_code.layers[
                                -1].op != "transpose" or in_node.fluid_code.layers[
                                    -1].param_attr["perm"] != [0, 2, 3, 1]:
                            is_nhwc = False
                            continue
                node.is_nhwc = is_nhwc

        for i in range(len(self.graph.topo_sort)):
            node_name = self.graph.topo_sort[-1 * i - 1]
            node = graph_copy.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in elementwise_ops:
                can_be_removed = True
                if len(node.fluid_code.layers) > 1:
                    can_be_removed = False
                if not node.is_nhwc:
                    can_be_removed = False
                for out_name in node.outputs:
                    out_node = graph_copy.get_node(out_name)
                    if hasattr(out_node, "is_nhwc"):
                        if not out_node.is_nhwc:
                            can_be_removed = False
                    else:
                        if len(out_node.fluid_code.layers) < 2:
                            can_be_removed = False
                            break
                        if out_node.fluid_code.layers[
                                0].op != "transpose" or out_node.fluid_code.layers[
                                    0].param_attr["perm"] != [0, 3, 1, 2]:
                            can_be_removed = False
                            break
                node.can_be_removed = can_be_removed

        for node_name in self.graph.topo_sort:
            node = graph_copy.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in optimize_ops:
                if node.fluid_code.layers[
                        -1].op != "transpose" or node.fluid_code.layers[
                            -1].param_attr["perm"] != [0, 2, 3, 1]:
                    continue
                can_be_removed = True
                output_names = node.outputs
                for out_name in output_names:
                    out_node = graph_copy.get_node(out_name)
                    if hasattr(out_node, "can_be_removed"):
                        if not out_node.can_be_removed:
                            can_be_removed = False
                            break
                    elif out_node.fluid_code.layers[
                            0].op != "transpose" or out_node.fluid_code.layers[
                                0].param_attr["perm"] != [0, 3, 1, 2]:
                        can_be_removed = False
                        break
                if can_be_removed and len(node.fluid_code.layers) > 1:
                    true_node = self.graph.get_node(node_name)
                    if true_node.layer_type == "Placeholder":
                        index = self.graph.input_nodes.index(
                            true_node.fluid_code.layers[-2].output)
                        if isinstance(true_node.fluid_code.layers[-1].output,
                                      str):
                            self.graph.input_nodes[
                                index] = true_node.fluid_code.layers[-1].output
                        else:
                            self.graph.input_nodes[
                                index] = true_node.fluid_code.layers[
                                    -1].output.layer_name
                    true_node.fluid_code.layers[
                        -2].output = true_node.fluid_code.layers[-1].output
                    node.removed = True
                    del true_node.fluid_code.layers[-1]
                    for out_name in output_names:
                        out_node = self.graph.get_node(out_name)
                        if out_node.layer_type in elementwise_ops:
                            continue
                        out_node.fluid_code.layers[
                            1].inputs = out_node.fluid_code.layers[0].inputs
                        del out_node.fluid_code.layers[0]

        for node_name in self.graph.topo_sort:
            node = graph_copy.get_node(node_name)
            if node is None:
                continue
            if node.layer_type in elementwise_ops:
                if not node.can_be_removed:
                    true_node = self.graph.get_node(node_name)
                    for i, in_name in enumerate(node.inputs):
                        in_node = graph_copy.get_node(in_name)
                        if hasattr(in_node, "is_nhwc") and in_node.is_nhwc:
                            if i == 0:
                                l = Layer()
                                l.op = "transpose"
                                l.inputs = true_node.fluid_code.layers[
                                    0].inputs["x"]
                                l.param_attr = {"perm": [0, 2, 3, 1]}
                                l.output = "nhwc_" + l.inputs.layer_name
                                true_node.fluid_code.layers[0].inputs[
                                    "x"] = l.output
                                true_node.fluid_code.layers.insert(0, l)
                            elif i == 1:
                                l = Layer()
                                l.op = "transpose"
                                l.inputs = true_node.fluid_code.layers[
                                    0].inputs["y"]
                                l.param_attr = {"perm": [0, 2, 3, 1]}
                                l.output = "nhwc_" + l.inputs.layer_name
                                true_node.fluid_code.layers[0].inputs[
                                    "y"] = l.output
                                true_node.fluid_code.layers.insert(0, l)
                            else:
                                raise Exception("Unexpected situation happend")
                    continue
                else:
                    for out_name in node.outputs:
                        out_node = self.graph.get_node(out_name)
                        if out_node.layer_type not in elementwise_ops:
                            assert out_node.fluid_code.layers[
                                0].op == "transpose", "unexpected situation happend"
                            out_node.fluid_code.layers[
                                1].inputs = out_node.fluid_code.layers[0].inputs
                            del out_node.fluid_code.layers[0]

    def make_nchw_input_output(self):
        for i, name in enumerate(self.graph.input_nodes):
            node = self.graph.get_node(name)
            if len(node.out_shapes[0]) == 4 and node.tf_data_format == "NHWC":
                shape = node.fluid_code.layers[0].param_attr["shape"]
                shape = [shape[i] for i in [0, 3, 1, 2]]
                node.fluid_code.layers[0].param_attr["shape"] = shape
                node.fluid_code.layers[0].output = "nhwc_" + name
                attr = {"perm": [0, 2, 3, 1]}
                node.fluid_code.add_layer("transpose",
                                          inputs="nhwc_" + name,
                                          output=node,
                                          param_attr=attr)
                self.graph.input_nodes[i] = "nhwc_" + name
        for i, name in enumerate(self.graph.output_nodes):
            node = self.graph.get_node(name)
            if node.layer_type != "transpose":
                if node.fluid_code.layers[-1].op == "transpose":
                    node.fluid_code.layers[-2].output = name
                    del node.fluid_code.layers[-1]

    def merge_batch_norm(self):
        for i, name in enumerate(self.graph.topo_sort):
            node = self.graph.get_node(name)
            if node is None:
                continue
            is_batch_norm = True
            if node.layer_type == "Add":
                in_nodes0 = [
                    self.graph.get_node(in_name) for in_name in node.inputs
                ]
                if in_nodes0[0].layer_type != "Mul" or in_nodes0[
                        1].layer_type != "Sub":
                    is_batch_norm = False
                    continue

                in_nodes1 = [
                    self.graph.get_node(in_name)
                    for in_name in in_nodes0[0].inputs
                ]
                in_nodes2 = [
                    self.graph.get_node(in_name)
                    for in_name in in_nodes0[1].inputs
                ]
                if len(in_nodes1[0].out_shapes[0]) != 4:
                    is_batch_norm = False
                    continue
                if in_nodes1[1].layer_type != "Mul":
                    is_batch_norm = False
                    continue

                if in_nodes2[0].layer_type != "Const" or in_nodes2[
                        1].layer_type != "Mul":
                    is_batch_norm = False
                    continue

                in_nodes3 = [
                    self.graph.get_node(in_name)
                    for in_name in in_nodes1[1].inputs
                ]
                if in_nodes3[0].layer_type != "Rsqrt" or in_nodes3[
                        1].layer_type != "Const":
                    is_batch_norm = False
                    continue

                in_nodes4 = [
                    self.graph.get_node(in_name)
                    for in_name in in_nodes2[1].inputs
                ]
                if in_nodes4[0].layer_type != "Const" or in_nodes4[
                        1].layer_name != in_nodes1[1].layer_name:
                    is_batch_norm = False
                    continue

                in_nodes5 = self.graph.get_node(in_nodes3[0].inputs[0])
                if in_nodes5.layer_type != "Add":
                    is_batch_norm = False
                    continue

                in_nodes6 = [
                    self.graph.get_node(in_name) for in_name in in_nodes5.inputs
                ]
                if in_nodes6[0].layer_type != "Const" or in_nodes6[
                        1].layer_type != "Const":
                    is_batch_norm = False
                    continue

                conv_shape = in_nodes1[0].out_shapes[0]
                if conv_shape[3] < 0:
                    is_batch_norm = False
                    continue

                # moving_variance
                if in_nodes6[0].value.size != conv_shape[3]:
                    is_batch_norm = False
                    continue

                # epsilon
                if in_nodes6[1].value.size != 1:
                    is_batch_norm = False
                    continue

                # gamma
                if in_nodes3[1].value.size != conv_shape[3]:
                    is_batch_norm = False
                    continue

                # moving_mean
                if in_nodes4[0].value.size != conv_shape[3]:
                    is_batch_norm = False
                    continue

                # beta
                if in_nodes2[0].value.size != conv_shape[3]:
                    is_batch_norm = False
                    continue

                if is_batch_norm:
                    index = in_nodes1[0].outputs.index(in_nodes0[0].layer_name)
                    del in_nodes1[0].outputs[index]
                    node.layer_type = "FusedBatchNorm"
                    node.inputs = [in_nodes1[0].layer_name]
                    node.outputs = node.outputs
                    act = node.fluid_code.layers[-1].param_attr.get("act", None)
                    node.fluid_code.clear()
                    attr = {
                        "epsilon": in_nodes6[1].value,
                        "param_attr": string(in_nodes3[1].layer_name),
                        "bias_attr": string(in_nodes2[0].layer_name),
                        "moving_mean_name": string(in_nodes4[0].layer_name),
                        "moving_variance_name": string(in_nodes6[0].layer_name),
                        "is_test": True,
                        "act": act
                    }

                    node.fluid_code.add_layer("batch_norm",
                                              inputs=cp.copy(in_nodes1[0]),
                                              output=node,
                                              param_attr=attr)

                del self.graph.node_map[in_nodes0[0].layer_name]
                del self.graph.node_map[in_nodes0[1].layer_name]
                del self.graph.node_map[in_nodes1[1].layer_name]
                del self.graph.node_map[in_nodes2[1].layer_name]
                del self.graph.node_map[in_nodes3[0].layer_name]
                del self.graph.node_map[in_nodes4[0].layer_name]
                del self.graph.node_map[in_nodes5.layer_name]