diff --git a/x2paddle/core/program.py b/x2paddle/core/program.py index 88b9df330a17f98e87b33cb07204925b0df897c5..9f35a8b0a0d2d438d76ac7a3a2f9c4ed58186b2e 100644 --- a/x2paddle/core/program.py +++ b/x2paddle/core/program.py @@ -1,4 +1,4 @@ -# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved. +# 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. diff --git a/x2paddle/decoder/pytorch_decoder.py b/x2paddle/decoder/pytorch_decoder.py index 9befec388e964e8d7cd0187e2dcbea0830756a0e..c1a626d58bd352a10f020c95867fffbc1dbf6b47 100644 --- a/x2paddle/decoder/pytorch_decoder.py +++ b/x2paddle/decoder/pytorch_decoder.py @@ -1,4 +1,4 @@ -# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved. +# 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. diff --git a/x2paddle/op_mapper/pytorch2paddle/aten.py b/x2paddle/op_mapper/pytorch2paddle/aten.py index f5289fd5ac2864db8ca96a4c438737209df5ecb4..d49f2680bf83afe9861650d338ba8103ae36f752 100644 --- a/x2paddle/op_mapper/pytorch2paddle/aten.py +++ b/x2paddle/op_mapper/pytorch2paddle/aten.py @@ -28,6 +28,32 @@ dtype_dict = { } +def aten_abs(mapper, graph, node): + """ 构造获取绝对值的PaddleLayer。 + + TorchScript示例: + %n0.3 : Tensor = aten::abs(%n.3) + 参数含义: + %n0.3 (Tensor): 绝对值后的Tensor。 + %n.3 (Tensor): 绝对值前的Tensor。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%n.3 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + + graph.add_layer( + "paddle.fluid.layers.abs", inputs=layer_inputs, outputs=layer_outputs) + return current_inputs, current_outputs + + def aten_adaptive_avg_pool2d(mapper, graph, node): """ 构造average adaptive pool2d的PaddleLayer。 @@ -279,7 +305,10 @@ def aten_arange(mapper, graph, node): layer_inputs["end"] = inputs_name[0] current_inputs.append(inputs_name[0]) # 处理输入1，即%43，代表dtype - layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[1]]] + if mapper.attrs[inputs_name[1]] is None: + layer_attrs["dtype"] = None + else: + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[1]]] elif len(inputs_name) == 6: # %position_ids.1 : Tensor = aten::arange(%51, %52, %43, %45, %42, %46) # 输入的后三者分别代表layout、device、是否使用梯度 @@ -300,7 +329,10 @@ def aten_arange(mapper, graph, node): layer_inputs["end"] = inputs_name[1] current_inputs.append(inputs_name[1]) # 处理输入2，即%43，代表dtype - layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[2]]] + if mapper.attrs[inputs_name[2]] is None: + layer_attrs["dtype"] = None + else: + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[2]]] elif len(inputs_name) == 7: # %position_ids.1 : Tensor = aten::arange(%51, %52, %53, %43, %45, %42, %46) # 输入的后三者分别代表layout、device、是否使用梯度 @@ -329,7 +361,10 @@ def aten_arange(mapper, graph, node): layer_inputs["step"] = inputs_name[2] current_inputs.append(inputs_name[2]) # 处理输入3，即%43，代表dtype - layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[3]]] + if mapper.attrs[inputs_name[3]] is None: + layer_attrs["dtype"] = None + else: + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[3]]] else: raise Exception("Unknown aten::arange signature taking " + str( len(inputs_name)) + " arguments.") @@ -582,6 +617,104 @@ def aten___contains__(mapper, graph, node): return current_inputs, current_outputs +def aten_constant_pad_nd(mapper, graph, node): + """ 构造填充固定值的PaddleLayer。 + + TorchScript示例: + %58 : Tensor = aten::constant_pad_nd(%input1.24, %4876, %42) + 参数含义: + %58 (Tensor): 输出，填充后的Tensor。 + %input1.24 (Tensor): 需要填充的Tensor。 + %4876 (list): 填充大小。 + %42 (-): 填充值。 + """ + if "constant_pad" in mapper.dygraph_name_id: + mapper.dygraph_name_id["constant_pad"] += 1 + else: + mapper.dygraph_name_id["constant_pad"] = 0 + constant_pad_name = "constant_pad" + str(mapper.dygraph_name_id[ + "constant_pad"]) + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [constant_pad_name, output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%input1.24 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["input"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + # 处理输入1，即%4876 + layer_attrs["padding"] = mapper.attrs[inputs_name[1]] + # 处理输入2，即%42 + layer_attrs["value"] = mapper.attrs[inputs_name[2]] + + graph.add_layer( + "fluid.layers.shape", + inputs={"input": inputs_name[0]}, + outputs=[inputs_name[0] + "_shape"]) + graph.add_layer( + "prim.len", + inputs={"input": inputs_name[0] + "_shape"}, + outputs=[inputs_name[0] + "_len"]) + + def add_pad_layers(kernel, dim): + graph.add_layer( + "prim.ne", + inputs={"x": inputs_name[0] + "_len"}, + outputs=[inputs_name[0] + "_cond"], + y=dim) + graph.add_layer( + "prim.if", {'input': inputs_name[0] + "_cond"}, + outputs=[inputs_name[0] + "_if", output_name]) + if_layer = graph.layers[list(graph.layers.keys())[-1]] + block = PaddleGraph(if_layer, graph_type="dygraph") + block.add_layer( + "prim.sub", + inputs={"y": inputs_name[0] + "_len"}, + outputs=[inputs_name[0] + "_len0"], + x=dim) + block.add_layer( + "prim.len2list", + inputs={"len": inputs_name[0] + "_len0"}, + outputs=[inputs_name[0] + "_list"]) + block.add_layer( + "paddle.tensor.unsqueeze", + inputs={"x": inputs_name[0], + "axis": inputs_name[0] + "_list"}, + outputs=[inputs_name[0] + "_var"]) + block.add_layer( + kernel, + inputs={"input": inputs_name[0] + "_var"}, + outputs=layer_outputs, + **layer_attrs) + block.add_layer( + "paddle.tensor.squeeze", + inputs={"x": output_name, + "axis": inputs_name[0] + "_list"}, + outputs=[output_name]) + if_layer.add_block(block) + block = PaddleGraph(if_layer, graph_type="dygraph") + layer_inputs["input"] = inputs_name[0] + block.add_layer( + kernel, inputs=layer_inputs, outputs=layer_outputs, **layer_attrs) + if_layer.add_block(block) + if_layer.inputs["input-0"] = inputs_name[0] + if_layer.inputs["input-1"] = inputs_name[0] + "_len" + + if len(layer_attrs["padding"]) == 2: + add_pad_layers("paddle.nn.ConstantPad1d", 3) + elif len(layer_attrs["padding"]) == 4: + add_pad_layers("paddle.nn.ConstantPad2d", 4) + elif len(layer_attrs["padding"]) == 6: + add_pad_layers("paddle.nn.ConstantPad3d", 5) + else: + raise Exception("The lenght of padding list must be 2, 4 or 6!") + return current_inputs, current_outputs + + def aten_contiguous(mapper, graph, node): """ 构造在内存中连续存储的PaddleLayer。 @@ -1044,7 +1177,10 @@ def aten_embedding(mapper, graph, node): # 获取当前节点输入的list current_inputs = list(layer_inputs.values()) # 处理输入2，即%45 - layer_attrs["padding_idx"] = mapper.attrs[inputs_name[2]] + if mapper.attrs[inputs_name[2]] == -1: + layer_attrs["padding_idx"] = None + else: + layer_attrs["padding_idx"] = mapper.attrs[inputs_name[2]] # 处理输入4，即%46 # layer_attrs["sparse"] = mapper.attrs[inputs_name[4]] layer_attrs["is_sparse"] = mapper.attrs[inputs_name[4]] @@ -1153,7 +1289,7 @@ def aten_expand(mapper, graph, node): y=string("VarType.BOOL")) graph.add_layer( "prim.if", {'input': inputs_name[0] + "_cond"}, - outputs=[inputs_name[0] + "_if1"]) + outputs=[inputs_name[0] + "_if1", inputs_name[1] + "_var"]) if_layer = graph.layers[list(graph.layers.keys())[-1]] block = PaddleGraph(if_layer, graph_type="dygraph") block.add_layer( @@ -1471,6 +1607,52 @@ def aten_floor_divide(mapper, graph, node): return current_inputs, current_outputs +def aten_full_like(mapper, graph, node): + """ 构造创建一个与输入具有相同的形状并且数据类型固定的Tensor的PaddleLayer。 + + TorchScript示例: + %159 : Tensor = aten::full_like(%val_if_large.3, %51, %50, %62, %53, %65, %66) + 参数含义: + %159 (Tensor): 输出，全为固定值的Tensor。 + %val_if_large.3 (Tensor): 类似形状的Tensor。 + %51 (int/float/bool): 填充值。 + %50 (int): dtype。 + %62 (int): layout。 + %53 (int): device。 + %65 (bool): 是否计算梯度。 + %66 (int): 内存形式。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%val_if_large.3 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + # 处理输入1，即%51 + if inputs_name[1] in mapper.attrs: + layer_attrs["fill_value"] = mapper.attrs[inputs_name[1]] + else: + mapper._check_input(graph, inputs_node[1], inputs_name[1], + current_outputs) + layer_inputs["fill_value"] = inputs_name[1] + current_inputs.append(inputs_name[1]) + # 处理输入2，即%50，代表dtype + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[2]]] + + graph.add_layer( + "paddle.full_like", + inputs=layer_inputs, + outputs=layer_outputs, + **layer_attrs) + return current_inputs, current_outputs + + def aten_gelu(mapper, graph, node): """ 构造GeLU激活的PaddleLayer。 @@ -1875,6 +2057,32 @@ def aten_len(mapper, graph, node): return current_inputs, current_outputs +def aten_log(mapper, graph, node): + """ 构构造log的PaddleLayer。 + + TorchScript示例: + %787 : Tensor = aten::log(%786) + 参数含义: + %787 (Tensor): 输出，取log的Tensor。 + %786 (Tensor): 需要获取log的Tensor。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%786 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + + graph.add_layer( + "fluid.layers.log", inputs=layer_inputs, outputs=layer_outputs) + return current_inputs, current_outputs + + def aten_lt(mapper, graph, node): """ 构造对比大小的PaddleLayer。 @@ -1999,6 +2207,136 @@ def aten_masked_fill_(mapper, graph, node): return current_inputs, current_outputs +def aten_masked_fill(mapper, graph, node): + """ 构造填充mask的PaddleLayer。 + + TorchScript示例: + %input.4 : Tensor = aten::masked_fill(%scores.2, %mask.2, %46) + 参数含义: + %input.4 (Tensor): 输出，填充后的结果。 + %scores.2 (Tensor): 需要填充的Tensor。 + %mask.2 (Tensor): bool型的Tensor，哪些位置需要填充。 + %46 (-): 填充的值。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输入的list + current_inputs = [] + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%input.4 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + current_inputs.append(inputs_name[0]) + graph.add_layer( + "prim.type", + inputs={"input": inputs_name[0]}, + outputs=[inputs_name[0] + "_type"]) + # 处理输入1，即%scores.2 + mapper._check_input(graph, inputs_node[1], inputs_name[1], current_outputs) + current_inputs.append(inputs_name[1]) + graph.add_layer( + "paddle.logical_not", + inputs={"x": inputs_name[1]}, + outputs=[inputs_name[1] + "_not"]) + graph.add_layer( + "fluid.layers.cast", + inputs={"x": inputs_name[1]}, + outputs=[inputs_name[1] + "_mask"], + dtype=inputs_name[0] + "_type") + graph.add_layer( + "fluid.layers.cast", + inputs={"x": inputs_name[1] + "_not"}, + outputs=[inputs_name[1] + "_not_mask"], + dtype=inputs_name[0] + "_type") + graph.add_layer( + "paddle.multiply", + inputs={"x": inputs_name[0], + "y": inputs_name[1] + "_not_mask"}, + outputs=[inputs_name[0] + "_not_mask"]) + # 处理输入2，即%46 + mapper._check_input(graph, inputs_node[2], inputs_name[2], current_outputs) + graph.add_layer( + "prim.eq", + inputs={"x": inputs_name[2]}, + outputs=[inputs_name[2] + "_cond1"], + y="-float('inf')") + graph.add_layer( + "prim.eq", + inputs={"x": inputs_name[2]}, + outputs=[inputs_name[2] + "_cond2"], + y="float('inf')") + graph.add_layer( + "prim.or", + inputs={ + "x": inputs_name[2] + "_cond1", + "y": inputs_name[2] + "_cond2" + }, + outputs=[inputs_name[2] + "_cond"]) + graph.add_layer( + "prim.if", {'input': inputs_name[2] + "_cond"}, + outputs=[inputs_name[2] + "_if"]) + if_layer = graph.layers[list(graph.layers.keys())[-1]] + block = PaddleGraph(if_layer, graph_type="dygraph") + block.add_layer( + "prim.equal", + inputs={"input": inputs_name[1] + "_mask"}, + outputs=[inputs_name[2] + "_1"]) + if_layer.add_block(block) + block = PaddleGraph(if_layer, graph_type="dygraph") + block.add_layer( + "prim.mul", + inputs={"x": inputs_name[1] + "_mask", + "y": inputs_name[2]}, + outputs=[inputs_name[2] + "_1"]) + if_layer.add_block(block) + if_layer.inputs["input-0"] = inputs_name[1] + "_mask" + if_layer.inputs["input-1"] = inputs_name[2] + if_layer.outputs.append(inputs_name[2] + "_1") + graph.add_layer( + "fluid.layers.elementwise_add", + inputs={"x": inputs_name[2] + "_1", + "y": inputs_name[0] + "_not_mask"}, + outputs=layer_outputs) + return current_inputs, current_outputs + + +def aten_max(mapper, graph, node): + """ 构造获取最大值的PaddleLayer。 + + TorchScript示例: + %val_if_large0.3 : Tensor = aten::max(%val_if_large.3, %159) + 参数含义: + %val_if_large0.3 (Tensor): 输出，对比后的结果。 + %val_if_large.3 (Tensor): 输入，需要对比的Tensor1。 + %159 (Tensor): 输入，需要对比的Tensor2。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + input_type = list(node.inputs())[1].type() + if str(input_type) == "Tensor": + # 处理输入0，即%val_if_large.3 + mapper._check_input(graph, inputs_node[0], inputs_name[0], + current_outputs) + layer_inputs["x"] = inputs_name[0] + # 处理输入1，即%159 + mapper._check_input(graph, inputs_node[1], inputs_name[1], + current_outputs) + layer_inputs["y"] = inputs_name[1] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + graph.add_layer( + "paddle.maximum", inputs=layer_inputs, outputs=layer_outputs) + else: + pass + return current_inputs, current_outputs + + def aten_max_pool2d(mapper, graph, node): """ 构造最大池化的PaddleLayer。 @@ -2085,6 +2423,41 @@ def aten_matmul(mapper, graph, node): return current_inputs, current_outputs +def aten_min(mapper, graph, node): + """ 构造获取最小值的PaddleLayer。 + + TorchScript示例: + %val_if_large0.3 : Tensor = aten::min(%val_if_large.3, %159) + 参数含义: + %val_if_large0.3 (Tensor): 输出，对比后的结果。 + %val_if_large.3 (Tensor): 输入，需要对比的Tensor1。 + %159 (Tensor): 输入，需要对比的Tensor2。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + input_type = list(node.inputs())[1].type() + if str(input_type) == "Tensor": + # 处理输入0，即%val_if_large.3 + mapper._check_input(graph, inputs_node[0], inputs_name[0], + current_outputs) + layer_inputs["x"] = inputs_name[0] + # 处理输入1，即%159 + mapper._check_input(graph, inputs_node[1], inputs_name[1], + current_outputs) + layer_inputs["y"] = inputs_name[1] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + graph.add_layer( + "paddle.minimum", inputs=layer_inputs, outputs=layer_outputs) + else: + pass + return current_inputs, current_outputs + + def aten_mean(mapper, graph, node): """ 构造求均值的PaddleLayer。 @@ -2163,6 +2536,36 @@ def aten_mul(mapper, graph, node): return current_inputs, current_outputs +def aten_mul_(mapper, graph, node): + """ 构造数值相乘的PaddleLayer。 + + TorchScript示例: + %size_prods.39 : int = aten::mul_(%size_prods.38, %114) + 参数含义: + %size_prods.39 (Tensor): 输出，相乘后的结果。 + %size_prods.38 (-): 数值1。 + %114 (-): 数值2。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%size_prods.38 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 处理输入1，即%114 + mapper._check_input(graph, inputs_node[1], inputs_name[1], current_outputs) + layer_inputs["y"] = inputs_name[1] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + current_outputs = layer_outputs + + graph.add_layer("prim.mul", inputs=layer_inputs, outputs=layer_outputs) + return current_inputs, current_outputs + + def aten_ne(mapper, graph, node): """ 构造判断数值是否不相等的PaddleLayer。 @@ -2242,6 +2645,46 @@ def aten___not__(mapper, graph, node): return current_inputs, current_outputs +def aten_ones(mapper, graph, node): + """ 构造创建固定形状、数据类型且值全为0的Tensor的PaddleLayer。 + + TorchScript示例: + %input.49 : Tensor = aten::ones(%23, %8, %6, %24, %5) + 参数含义: + %input.49 (Tensor): 输出，全0的Tensor。 + %23 (list): 形状。 + %8 (int): 类型dtype。 + %6 (int): layout。 + %4995 (Device): 设备。 + %4995 (bool): 是否计算梯度。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + current_inputs = [] + # 处理输入0，即%23，代表end + if inputs_name[0] in mapper.attrs: + layer_attrs["shape"] = mapper.attrs[inputs_name[0]] + else: + mapper._check_input(graph, inputs_node[0], inputs_name[0], + current_outputs) + layer_inputs["shape"] = inputs_name[0] + current_inputs.append(inputs_name[0]) + # 处理输入1，即%8，代表dtype + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[1]]] + + graph.add_layer( + "paddle.ones", + inputs=layer_inputs, + outputs=layer_outputs, + **layer_attrs) + return current_inputs, current_outputs + + def aten_permute(mapper, graph, node): """ 构造对bool型取负的PaddleLayer。 @@ -2418,6 +2861,45 @@ def aten_relu6(mapper, graph, node): return current_inputs, current_outputs +def aten_repeat(mapper, graph, node): + """ 构造根据参数对输入各维度进行复制的PaddleLayer。 + + TorchScript示例: + 701 : Tensor = aten::repeat(%699, %700) + 参数含义: + %701 (Tensor): 输出，复制后的Tensor。 + %699 (Tensor): 需要复制的Tensor。 + %700 (list): 指定每个维度复制的次数。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%699 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入、输出的list + current_inputs = list(layer_inputs.values()) + # 处理输入1，即%700 + if inputs_name[1] in mapper.attrs: + layer_attrs["repeat_times"] = mapper.attrs[inputs_name[1]] + else: + mapper._check_input(graph, inputs_node[1], inputs_name[1], + current_outputs) + layer_inputs["repeat_times"] = inputs_name[1] + current_inputs.append(inputs_name[1]) + + graph.add_layer( + "paddle.tile", + inputs=layer_inputs, + outputs=layer_outputs, + **layer_attrs) + return current_inputs, current_outputs + + def aten_reshape(mapper, graph, node): """ 构造调整大小的PaddleLayer。 @@ -2933,6 +3415,70 @@ def aten_softplus(mapper, graph, node): return current_inputs, current_outputs +def aten_sqrt(mapper, graph, node): + """ 构构造sqrt的PaddleLayer。 + + TorchScript示例: + %787 : Tensor = aten::sqrt(%786) + 参数含义: + %787 (Tensor): 输出，取sqrt的Tensor。 + %786 (Tensor): 需要获取sqrt的Tensor。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%786 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + + graph.add_layer( + "fluid.layers.sqrt", inputs=layer_inputs, outputs=layer_outputs) + return current_inputs, current_outputs + + +def aten_squeeze(mapper, graph, node): + """ 构造删除位数为1的维度的PaddleLayer。 + + TorchScript示例: + %12 : Tensor = aten::squeeze(%start_logits.1, %4) + 参数含义: + %12 (Tensor): 输出，删除维度后的Tensor。 + %start_logits.1 (Tensor): 需要删除维度的Tensor。 + %4 (int): 维度。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%start_logits.1 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + # 处理输入1，即%4 + if inputs_name[1] in mapper.attrs: + layer_attrs["axis"] = mapper.attrs[inputs_name[1]] + else: + mapper._check_input(graph, inputs_node[1], inputs_name[1], + current_outputs) + layer_inputs["axis"] = inputs_name[1] + current_inputs.append(inputs_name[1]) + graph.add_layer( + "paddle.tensor.squeeze", + inputs=layer_inputs, + outputs=layer_outputs, + **layer_attrs) + return current_inputs, current_outputs + + def aten_stack(mapper, graph, node): """ 构造堆叠Tensor的PaddleLayer。 @@ -3568,3 +4114,40 @@ def aten_zeros(mapper, graph, node): outputs=layer_outputs, **layer_attrs) return current_inputs, current_outputs + + +def aten_zeros_like(mapper, graph, node): + """ 构造创建与输入Tensor形状一致的、数据类型且值全为0的Tensor的PaddleLayer。 + + TorchScript示例: + %782 : Tensor = aten::zeros_like(%n.2, %655, %670, %662, %671, %672) + 参数含义: + %782 (Tensor): 输出，全0的Tensor。 + %n.2 (Tensor): 标准Tensor。 + %655 (int): 类型dtype。 + %670 (int): layout。 + %662 (Device): 设备。 + %671 (bool): 是否计算梯度。 + %672 (memory_format): 存储类型。 + """ + output_name = mapper._get_outputs_name(node)[0] + layer_outputs = [output_name] + layer_inputs = {} + layer_attrs = {} + inputs_name, inputs_node = mapper._get_inputs_name(node) + # 获取当前节点输出的list + current_outputs = [output_name] + # 处理输入0，即%n.2 + mapper._check_input(graph, inputs_node[0], inputs_name[0], current_outputs) + layer_inputs["x"] = inputs_name[0] + # 获取当前节点输入的list + current_inputs = list(layer_inputs.values()) + # 处理输入1，即%655，代表dtype + layer_attrs["dtype"] = dtype_dict[mapper.attrs[inputs_name[1]]] + + graph.add_layer( + "paddle.zeros_like", + inputs=layer_inputs, + outputs=layer_outputs, + **layer_attrs) + return current_inputs, current_outputs diff --git a/x2paddle/op_mapper/pytorch2paddle/prim.py b/x2paddle/op_mapper/pytorch2paddle/prim.py index 801a6f923d7449f50ce8a7d63361f676c457ef73..834a409f1ce9b373541a7253c6e9ca73e17e92de 100644 --- a/x2paddle/op_mapper/pytorch2paddle/prim.py +++ b/x2paddle/op_mapper/pytorch2paddle/prim.py @@ -32,7 +32,6 @@ def prim_Constant(mapper, graph, node): if isinstance(value, str): value = string(value) if str(output_type) == "Tensor": - # value = "paddle.to_tensor({})".format(value) value = "{}".format(value) if "inf" in str(value): diff --git a/x2paddle/op_mapper/pytorch2paddle/prim2code.py b/x2paddle/op_mapper/pytorch2paddle/prim2code.py index 04bf22daab1a346befc561ac440d63928cf21d27..d16197f6793fccc7e236edbc21e5abb01ce98c37 100644 --- a/x2paddle/op_mapper/pytorch2paddle/prim2code.py +++ b/x2paddle/op_mapper/pytorch2paddle/prim2code.py @@ -1,4 +1,4 @@ -# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved. +# 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. diff --git a/x2paddle/op_mapper/pytorch2paddle/pytorch_op_mapper.py b/x2paddle/op_mapper/pytorch2paddle/pytorch_op_mapper.py index 26564e857aabfa136f943637ffd42a51b64e5612..2d63411879009003d80485602ae28e2a498b0035 100644 --- a/x2paddle/op_mapper/pytorch2paddle/pytorch_op_mapper.py +++ b/x2paddle/op_mapper/pytorch2paddle/pytorch_op_mapper.py @@ -1,4 +1,4 @@ -# Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved. +# 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. diff --git a/x2paddle/optimizer/fusion/constant_fuse_pass.py b/x2paddle/optimizer/fusion/constant_fuse_pass.py index 26b24c4c5ffa3bca1b220feeb37e88b7176de429..62d3552bcd4ff041dca81fd03ee6ef829ebe64ba 100644 --- a/x2paddle/optimizer/fusion/constant_fuse_pass.py +++ b/x2paddle/optimizer/fusion/constant_fuse_pass.py @@ -1,3 +1,5 @@ +# 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 diff --git a/x2paddle/optimizer/fusion/dropout_fuse_pass.py b/x2paddle/optimizer/fusion/dropout_fuse_pass.py index 52104ac696cef1e15eb65073fac22c9ad0381b18..e004b1de88c1677e81007ac696adf6ae4713e8ff 100644 --- a/x2paddle/optimizer/fusion/dropout_fuse_pass.py +++ b/x2paddle/optimizer/fusion/dropout_fuse_pass.py @@ -1,3 +1,5 @@ +# 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