# Copyright (c) 2020 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. import numpy as np import copy from x2paddle.optimizer.pattern_matcher import FuseBase from x2paddle.core.program import PaddleGraph, PaddleLayer from x2paddle.core.util import * class DygraphAdaptivePool2dFuser(FuseBase): def __init__(self): super(DygraphAdaptivePool2dFuser, self).__init__(graph_type="dygraph") self.patterns = list() def build_pattern(self): """ 描述需要替换的adaptive pool2d图结构。 adaptive pool2d层模式python实现代码示例: 模式一： x68 = prim.shape(input=x60) x69 = len(x68) x70 = x69 <= 2 if x70 : raise RaiseException('Exception') x73 = [] x74 = x68[-2: 2147483647: 1] x75 = len(x74) x76 = [2, x75] x77 = min(x76) for _x79 in range(x77): x80 = [6, 6][_x79] x73.append(x80) x81 = paddle.nn.functional.adaptive_avg_pool2d(input=x60, pool_size=x73, pool_type='avg') 模式二： x64 = x60.shape x65 = len(x64) x66 = x65 > 2 if x66 : pass else: raise RaiseException('AssertionError: ') x69 = self.pool2d3(x60) """ def gen_name(id): return "x" + str(id) # 模式一： pattern = PaddleGraph(graph_type="dygraph") pattern.add_layer( "prim.shape", inputs={'input': "pool-input-0"}, outputs=[gen_name(1)]) pattern.add_layer( "prim.len", inputs={"input": gen_name(1)}, outputs=[gen_name(6)]) pattern.add_layer( "prim.le", inputs={"x": gen_name(6)}, outputs=[gen_name(8)], y=2) pattern.add_layer("prim.if", {'input': gen_name(8)}, [gen_name(9)]) if_layer = pattern.layers[list(pattern.layers.keys())[-1]] pattern_block0 = PaddleGraph(parent_layer=if_layer, graph_type="dygraph") pattern_block0.add_layer( "prim.exception", inputs={}, outputs=[gen_name(9)], input="Exception") if_layer.add_block(pattern_block0) pattern_block1 = PaddleGraph(parent_layer=if_layer, graph_type="dygraph") if_layer.add_block(pattern_block1) pattern.add_layer("prim.list", inputs={}, outputs=[gen_name(10)]) pattern.add_layer( "prim.slice", inputs={"input": gen_name(1), }, outputs=[gen_name(12)]) pattern.add_layer( "prim.len", inputs={"input": gen_name(12)}, outputs=[gen_name(14)]) pattern.add_layer( "prim.list", inputs={"input1": gen_name(14)}, outputs=[gen_name(15)]) pattern.add_layer( "prim.min", inputs={"input": gen_name(15)}, outputs=[gen_name(16)]) pattern.add_layer("prim.loop", {'input': gen_name(16)}, [gen_name(17), gen_name(18)]) loop_layer = pattern.layers[list(pattern.layers.keys())[-1]] pattern_block = PaddleGraph(loop_layer, graph_type="dygraph") pattern_block.add_layer( "prim.getitem", inputs={"index": gen_name(18)}, outputs=[gen_name(19)]) pattern_block.add_layer( "prim.append", inputs={"list": gen_name(10), "index": gen_name(19)}, outputs=[gen_name(20)]) loop_layer.inputs["input-0"] = gen_name(10) loop_layer.add_block(pattern_block) pool_attrs = {'pool_type': string("avg")} pattern.add_layer( "paddle.nn.functional.adaptive_avg_pool2d", inputs={'input': "pool-input-0", "pool_size": gen_name(10)}, outputs=[gen_name(21)], **pool_attrs) pattern.build(inputs={"input-0": "pool-input-0", }) self.patterns.append(pattern) # 模式二： pattern = PaddleGraph(graph_type="dygraph") pattern.add_layer( "prim.shape", inputs={'input': "pool-input-0"}, outputs=[gen_name(0)]) pattern.add_layer( "prim.len", inputs={"input": gen_name(0)}, outputs=[gen_name(1)]) pattern.add_layer( "prim.gt", inputs={"x": gen_name(1)}, outputs=[gen_name(2)], y=2) pattern.add_layer("prim.if", {'input': gen_name(2)}, [gen_name(3)]) if_layer = pattern.layers[list(pattern.layers.keys())[-1]] pattern_block0 = PaddleGraph(parent_layer=if_layer, graph_type="dygraph") if_layer.add_block(pattern_block0) pattern_block1 = PaddleGraph(parent_layer=if_layer, graph_type="dygraph") pattern_block1.add_layer( "prim.exception", inputs={}, outputs=[gen_name(4)], input="Exception") if_layer.add_block(pattern_block1) pattern.add_layer( "paddle.nn.AdaptiveAvgPool2D", inputs={"input": "pool-input-0"}, outputs=["pool1", gen_name(5)]) pattern.build(inputs={"input-0": "pool-input-0", "input-1": "pool-input-0",}) self.patterns.append(pattern) def insert_new_layer(self, graph, parameters, matches): parameters = graph.parameters new_layer = self.gen_new_layer(parameters, matches) new_layer_id = list(matches.keys())[0] graph.layers[new_layer_id] = new_layer matches.pop(new_layer_id) def gen_new_layer(self, parameters, matches): layers_id = list(matches.keys()) if matches[layers_id[-1]].kernel == "paddle.nn.functional.adaptive_avg_pool2d": layer = matches[layers_id[11]] pool_size = layer.attrs["list"] layer = matches[layers_id[0]] input_name = layer.inputs["input"] layer = matches[layers_id[-1]] output_name = layer.outputs[0] attrs = dict() attrs["output_size"] = pool_size new_layer = PaddleLayer( layers_id[0], "paddle.nn.functional.adaptive_avg_pool2d", inputs={"input": input_name}, outputs=[output_name], **attrs) else: new_layer = copy.deepcopy(matches[layers_id[-1]]) return new_layer