注意
转到末尾以下载完整的示例代码。
（测试版）使用 FX 构建简单的 CPU 性能分析器#
创建于：2021年3月4日 | 最后更新：2025年7月14日 | 最后验证：未验证
作者: James Reed
在本教程中，我们将使用 FX 来完成以下操作
以一种我们可以检查和收集关于代码结构和执行的统计数据的方式，来捕获 PyTorch Python 代码
构建一个小型类，用作一个简单的性能“分析器”，从实际运行中收集模型每个部分的运行时统计数据。
在本教程中，我们将使用 torchvision ResNet18 模型进行演示。
import torch
import torch.fx
import torchvision.models as models
rn18 = models.resnet18()
rn18.eval()
ResNet(
(conv1): Conv2d(3, 64, kernel_size=(7, 7), stride=(2, 2), padding=(3, 3), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(maxpool): MaxPool2d(kernel_size=3, stride=2, padding=1, dilation=1, ceil_mode=False)
(layer1): Sequential(
(0): BasicBlock(
(conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
(1): BasicBlock(
(conv1): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(layer2): Sequential(
(0): BasicBlock(
(conv1): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(64, 128, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(128, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(layer3): Sequential(
(0): BasicBlock(
(conv1): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(128, 256, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(256, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(layer4): Sequential(
(0): BasicBlock(
(conv1): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(downsample): Sequential(
(0): Conv2d(256, 512, kernel_size=(1, 1), stride=(2, 2), bias=False)
(1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(1): BasicBlock(
(conv1): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn1): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
(relu): ReLU(inplace=True)
(conv2): Conv2d(512, 512, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False)
(bn2): BatchNorm2d(512, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)
)
)
(avgpool): AdaptiveAvgPool2d(output_size=(1, 1))
(fc): Linear(in_features=512, out_features=1000, bias=True)
)
现在我们有了模型，我们想更深入地了解其性能。也就是说，对于以下调用，模型的哪些部分耗时最长？
input = torch.randn(5, 3, 224, 224)
output = rn18(input)
回答这个问题的一个常用方法是遍历程序源代码，在程序的各个点添加收集时间戳的代码，并比较这些时间戳之间的差异，以查看时间戳之间的区域耗时多久。
这种技术当然适用于 PyTorch 代码，但如果我们不必复制和编辑模型代码，尤其是那些我们没有编写的代码（比如这个 torchvision 模型），那就更好了。因此，我们将使用 FX 来自动化这个“插桩”过程，而无需修改任何源代码。
首先，让我们完成一些导入操作（我们将在后面的代码中使用所有这些导入）。
import statistics, tabulate, time
from typing import Any, Dict, List
from torch.fx import Interpreter
注意
tabulate 是一个外部库，并非 PyTorch 的依赖项。我们将用它来更轻松地可视化性能数据。请确保您已从您喜欢的 Python 包源安装了它。
使用符号跟踪捕获模型#
接下来，我们将使用 FX 的符号跟踪机制，将我们模型的定义捕获到一个我们可以操作和检查的数据结构中。
traced_rn18 = torch.fx.symbolic_trace(rn18)
print(traced_rn18.graph)
graph():
%x : torch.Tensor [num_users=1] = placeholder[target=x]
%conv1 : [num_users=1] = call_module[target=conv1](args = (%x,), kwargs = {})
%bn1 : [num_users=1] = call_module[target=bn1](args = (%conv1,), kwargs = {})
%relu : [num_users=1] = call_module[target=relu](args = (%bn1,), kwargs = {})
%maxpool : [num_users=2] = call_module[target=maxpool](args = (%relu,), kwargs = {})
%layer1_0_conv1 : [num_users=1] = call_module[target=layer1.0.conv1](args = (%maxpool,), kwargs = {})
%layer1_0_bn1 : [num_users=1] = call_module[target=layer1.0.bn1](args = (%layer1_0_conv1,), kwargs = {})
%layer1_0_relu : [num_users=1] = call_module[target=layer1.0.relu](args = (%layer1_0_bn1,), kwargs = {})
%layer1_0_conv2 : [num_users=1] = call_module[target=layer1.0.conv2](args = (%layer1_0_relu,), kwargs = {})
%layer1_0_bn2 : [num_users=1] = call_module[target=layer1.0.bn2](args = (%layer1_0_conv2,), kwargs = {})
%add : [num_users=1] = call_function[target=operator.add](args = (%layer1_0_bn2, %maxpool), kwargs = {})
%layer1_0_relu_1 : [num_users=2] = call_module[target=layer1.0.relu](args = (%add,), kwargs = {})
%layer1_1_conv1 : [num_users=1] = call_module[target=layer1.1.conv1](args = (%layer1_0_relu_1,), kwargs = {})
%layer1_1_bn1 : [num_users=1] = call_module[target=layer1.1.bn1](args = (%layer1_1_conv1,), kwargs = {})
%layer1_1_relu : [num_users=1] = call_module[target=layer1.1.relu](args = (%layer1_1_bn1,), kwargs = {})
%layer1_1_conv2 : [num_users=1] = call_module[target=layer1.1.conv2](args = (%layer1_1_relu,), kwargs = {})
%layer1_1_bn2 : [num_users=1] = call_module[target=layer1.1.bn2](args = (%layer1_1_conv2,), kwargs = {})
%add_1 : [num_users=1] = call_function[target=operator.add](args = (%layer1_1_bn2, %layer1_0_relu_1), kwargs = {})
%layer1_1_relu_1 : [num_users=2] = call_module[target=layer1.1.relu](args = (%add_1,), kwargs = {})
%layer2_0_conv1 : [num_users=1] = call_module[target=layer2.0.conv1](args = (%layer1_1_relu_1,), kwargs = {})
%layer2_0_bn1 : [num_users=1] = call_module[target=layer2.0.bn1](args = (%layer2_0_conv1,), kwargs = {})
%layer2_0_relu : [num_users=1] = call_module[target=layer2.0.relu](args = (%layer2_0_bn1,), kwargs = {})
%layer2_0_conv2 : [num_users=1] = call_module[target=layer2.0.conv2](args = (%layer2_0_relu,), kwargs = {})
%layer2_0_bn2 : [num_users=1] = call_module[target=layer2.0.bn2](args = (%layer2_0_conv2,), kwargs = {})
%layer2_0_downsample_0 : [num_users=1] = call_module[target=layer2.0.downsample.0](args = (%layer1_1_relu_1,), kwargs = {})
%layer2_0_downsample_1 : [num_users=1] = call_module[target=layer2.0.downsample.1](args = (%layer2_0_downsample_0,), kwargs = {})
%add_2 : [num_users=1] = call_function[target=operator.add](args = (%layer2_0_bn2, %layer2_0_downsample_1), kwargs = {})
%layer2_0_relu_1 : [num_users=2] = call_module[target=layer2.0.relu](args = (%add_2,), kwargs = {})
%layer2_1_conv1 : [num_users=1] = call_module[target=layer2.1.conv1](args = (%layer2_0_relu_1,), kwargs = {})
%layer2_1_bn1 : [num_users=1] = call_module[target=layer2.1.bn1](args = (%layer2_1_conv1,), kwargs = {})
%layer2_1_relu : [num_users=1] = call_module[target=layer2.1.relu](args = (%layer2_1_bn1,), kwargs = {})
%layer2_1_conv2 : [num_users=1] = call_module[target=layer2.1.conv2](args = (%layer2_1_relu,), kwargs = {})
%layer2_1_bn2 : [num_users=1] = call_module[target=layer2.1.bn2](args = (%layer2_1_conv2,), kwargs = {})
%add_3 : [num_users=1] = call_function[target=operator.add](args = (%layer2_1_bn2, %layer2_0_relu_1), kwargs = {})
%layer2_1_relu_1 : [num_users=2] = call_module[target=layer2.1.relu](args = (%add_3,), kwargs = {})
%layer3_0_conv1 : [num_users=1] = call_module[target=layer3.0.conv1](args = (%layer2_1_relu_1,), kwargs = {})
%layer3_0_bn1 : [num_users=1] = call_module[target=layer3.0.bn1](args = (%layer3_0_conv1,), kwargs = {})
%layer3_0_relu : [num_users=1] = call_module[target=layer3.0.relu](args = (%layer3_0_bn1,), kwargs = {})
%layer3_0_conv2 : [num_users=1] = call_module[target=layer3.0.conv2](args = (%layer3_0_relu,), kwargs = {})
%layer3_0_bn2 : [num_users=1] = call_module[target=layer3.0.bn2](args = (%layer3_0_conv2,), kwargs = {})
%layer3_0_downsample_0 : [num_users=1] = call_module[target=layer3.0.downsample.0](args = (%layer2_1_relu_1,), kwargs = {})
%layer3_0_downsample_1 : [num_users=1] = call_module[target=layer3.0.downsample.1](args = (%layer3_0_downsample_0,), kwargs = {})
%add_4 : [num_users=1] = call_function[target=operator.add](args = (%layer3_0_bn2, %layer3_0_downsample_1), kwargs = {})
%layer3_0_relu_1 : [num_users=2] = call_module[target=layer3.0.relu](args = (%add_4,), kwargs = {})
%layer3_1_conv1 : [num_users=1] = call_module[target=layer3.1.conv1](args = (%layer3_0_relu_1,), kwargs = {})
%layer3_1_bn1 : [num_users=1] = call_module[target=layer3.1.bn1](args = (%layer3_1_conv1,), kwargs = {})
%layer3_1_relu : [num_users=1] = call_module[target=layer3.1.relu](args = (%layer3_1_bn1,), kwargs = {})
%layer3_1_conv2 : [num_users=1] = call_module[target=layer3.1.conv2](args = (%layer3_1_relu,), kwargs = {})
%layer3_1_bn2 : [num_users=1] = call_module[target=layer3.1.bn2](args = (%layer3_1_conv2,), kwargs = {})
%add_5 : [num_users=1] = call_function[target=operator.add](args = (%layer3_1_bn2, %layer3_0_relu_1), kwargs = {})
%layer3_1_relu_1 : [num_users=2] = call_module[target=layer3.1.relu](args = (%add_5,), kwargs = {})
%layer4_0_conv1 : [num_users=1] = call_module[target=layer4.0.conv1](args = (%layer3_1_relu_1,), kwargs = {})
%layer4_0_bn1 : [num_users=1] = call_module[target=layer4.0.bn1](args = (%layer4_0_conv1,), kwargs = {})
%layer4_0_relu : [num_users=1] = call_module[target=layer4.0.relu](args = (%layer4_0_bn1,), kwargs = {})
%layer4_0_conv2 : [num_users=1] = call_module[target=layer4.0.conv2](args = (%layer4_0_relu,), kwargs = {})
%layer4_0_bn2 : [num_users=1] = call_module[target=layer4.0.bn2](args = (%layer4_0_conv2,), kwargs = {})
%layer4_0_downsample_0 : [num_users=1] = call_module[target=layer4.0.downsample.0](args = (%layer3_1_relu_1,), kwargs = {})
%layer4_0_downsample_1 : [num_users=1] = call_module[target=layer4.0.downsample.1](args = (%layer4_0_downsample_0,), kwargs = {})
%add_6 : [num_users=1] = call_function[target=operator.add](args = (%layer4_0_bn2, %layer4_0_downsample_1), kwargs = {})
%layer4_0_relu_1 : [num_users=2] = call_module[target=layer4.0.relu](args = (%add_6,), kwargs = {})
%layer4_1_conv1 : [num_users=1] = call_module[target=layer4.1.conv1](args = (%layer4_0_relu_1,), kwargs = {})
%layer4_1_bn1 : [num_users=1] = call_module[target=layer4.1.bn1](args = (%layer4_1_conv1,), kwargs = {})
%layer4_1_relu : [num_users=1] = call_module[target=layer4.1.relu](args = (%layer4_1_bn1,), kwargs = {})
%layer4_1_conv2 : [num_users=1] = call_module[target=layer4.1.conv2](args = (%layer4_1_relu,), kwargs = {})
%layer4_1_bn2 : [num_users=1] = call_module[target=layer4.1.bn2](args = (%layer4_1_conv2,), kwargs = {})
%add_7 : [num_users=1] = call_function[target=operator.add](args = (%layer4_1_bn2, %layer4_0_relu_1), kwargs = {})
%layer4_1_relu_1 : [num_users=1] = call_module[target=layer4.1.relu](args = (%add_7,), kwargs = {})
%avgpool : [num_users=1] = call_module[target=avgpool](args = (%layer4_1_relu_1,), kwargs = {})
%flatten : [num_users=1] = call_function[target=torch.flatten](args = (%avgpool, 1), kwargs = {})
%fc : [num_users=1] = call_module[target=fc](args = (%flatten,), kwargs = {})
return fc
这为我们提供了 ResNet18 模型的图表示。一个图由一系列相互连接的节点组成。每个节点代表 Python 代码中的一个调用点（无论是对函数、模块还是方法的调用），而边（在每个节点上表示为 args 和 kwargs）代表在这些调用点之间传递的值。有关图表示和 FX 其他 API 的更多信息，请参阅 FX 文档 https://pytorch.ac.cn/docs/stable/fx.html。
创建分析解释器#
接下来，我们将创建一个继承自 torch.fx.Interpreter 的类。虽然 symbolic_trace 生成的 GraphModule 会编译 Python 代码，在您调用 GraphModule 时运行，但运行 GraphModule 的另一种方法是逐个执行 Graph 中的每个 Node。这正是 Interpreter 提供的功能：它逐个节点地解释图。
通过继承 Interpreter，我们可以重写各种功能并安装我们想要的分析行为。我们的目标是拥有一个对象，我们可以向其传递一个模型，调用该模型1次或多次，然后获取关于这些运行期间模型及其各部分耗时的统计数据。
让我们定义我们的 ProfilingInterpreter 类
class ProfilingInterpreter(Interpreter):
def __init__(self, mod : torch.nn.Module):
# Rather than have the user symbolically trace their model,
# we're going to do it in the constructor. As a result, the
# user can pass in any ``Module`` without having to worry about
# symbolic tracing APIs
gm = torch.fx.symbolic_trace(mod)
super().__init__(gm)
# We are going to store away two things here:
#
# 1. A list of total runtimes for ``mod``. In other words, we are
# storing away the time ``mod(...)`` took each time this
# interpreter is called.
self.total_runtime_sec : List[float] = []
# 2. A map from ``Node`` to a list of times (in seconds) that
# node took to run. This can be seen as similar to (1) but
# for specific sub-parts of the model.
self.runtimes_sec : Dict[torch.fx.Node, List[float]] = {}
######################################################################
# Next, let's override our first method: ``run()``. ``Interpreter``'s ``run``
# method is the top-level entry point for execution of the model. We will
# want to intercept this so that we can record the total runtime of the
# model.
def run(self, *args) -> Any:
# Record the time we started running the model
t_start = time.time()
# Run the model by delegating back into Interpreter.run()
return_val = super().run(*args)
# Record the time we finished running the model
t_end = time.time()
# Store the total elapsed time this model execution took in the
# ``ProfilingInterpreter``
self.total_runtime_sec.append(t_end - t_start)
return return_val
######################################################################
# Now, let's override ``run_node``. ``Interpreter`` calls ``run_node`` each
# time it executes a single node. We will intercept this so that we
# can measure and record the time taken for each individual call in
# the model.
def run_node(self, n : torch.fx.Node) -> Any:
# Record the time we started running the op
t_start = time.time()
# Run the op by delegating back into Interpreter.run_node()
return_val = super().run_node(n)
# Record the time we finished running the op
t_end = time.time()
# If we don't have an entry for this node in our runtimes_sec
# data structure, add one with an empty list value.
self.runtimes_sec.setdefault(n, [])
# Record the total elapsed time for this single invocation
# in the runtimes_sec data structure
self.runtimes_sec[n].append(t_end - t_start)
return return_val
######################################################################
# Finally, we are going to define a method (one which doesn't override
# any ``Interpreter`` method) that provides us a nice, organized view of
# the data we have collected.
def summary(self, should_sort : bool = False) -> str:
# Build up a list of summary information for each node
node_summaries : List[List[Any]] = []
# Calculate the mean runtime for the whole network. Because the
# network may have been called multiple times during profiling,
# we need to summarize the runtimes. We choose to use the
# arithmetic mean for this.
mean_total_runtime = statistics.mean(self.total_runtime_sec)
# For each node, record summary statistics
for node, runtimes in self.runtimes_sec.items():
# Similarly, compute the mean runtime for ``node``
mean_runtime = statistics.mean(runtimes)
# For easier understanding, we also compute the percentage
# time each node took with respect to the whole network.
pct_total = mean_runtime / mean_total_runtime * 100
# Record the node's type, name of the node, mean runtime, and
# percent runtime.
node_summaries.append(
[node.op, str(node), mean_runtime, pct_total])
# One of the most important questions to answer when doing performance
# profiling is "Which op(s) took the longest?". We can make this easy
# to see by providing sorting functionality in our summary view
if should_sort:
node_summaries.sort(key=lambda s: s[2], reverse=True)
# Use the ``tabulate`` library to create a well-formatted table
# presenting our summary information
headers : List[str] = [
'Op type', 'Op', 'Average runtime (s)', 'Pct total runtime'
]
return tabulate.tabulate(node_summaries, headers=headers)
注意
我们使用 Python 的 time.time 函数来获取挂钟时间戳并进行比较。这不是测量性能最精确的方法，只能给我们一个初步的近似值。我们仅为了本教程的演示目的而使用这种简单技术。
研究 ResNet18 的性能#
我们现在可以使用 ProfilingInterpreter 来检查我们的 ResNet18 模型的性能特征；
interp = ProfilingInterpreter(rn18)
interp.run(input)
print(interp.summary(True))
Op type Op Average runtime (s) Pct total runtime
------------- --------------------- --------------------- -------------------
call_module maxpool 0.00455713 8.19317
call_module conv1 0.00452542 8.13616
call_module layer4_0_conv2 0.00322938 5.80603
call_module layer4_1_conv1 0.00296068 5.32294
call_module layer4_1_conv2 0.0028646 5.1502
call_module layer1_0_conv1 0.00285673 5.13605
call_module layer1_0_conv2 0.00283647 5.09962
call_module layer1_1_conv2 0.00257945 4.63754
call_module layer2_1_conv2 0.00231075 4.15445
call_module layer1_1_conv1 0.00228453 4.1073
call_module layer3_1_conv1 0.00217485 3.91012
call_module layer3_0_conv2 0.0021069 3.78796
call_module layer2_1_conv1 0.00205278 3.69065
call_module layer3_1_conv2 0.00205112 3.68765
call_module layer2_0_conv2 0.00203347 3.65593
call_module layer4_0_conv1 0.00182629 3.28344
call_module layer3_0_conv1 0.00165868 2.9821
call_module bn1 0.00134087 2.41071
call_module layer2_0_conv1 0.00128579 2.3117
call_module layer2_0_downsample_0 0.000769615 1.38367
call_module layer3_0_downsample_0 0.000460863 0.828575
call_module layer4_0_downsample_0 0.000448465 0.806286
call_function add 0.000412226 0.741131
call_function add_1 0.000394344 0.708983
call_module layer1_0_bn1 0.000311375 0.559813
call_module layer1_0_bn2 0.000282764 0.508376
call_module relu 0.000280619 0.504518
call_module layer1_1_bn2 0.000271559 0.488229
call_function add_3 0.000256062 0.460367
call_module layer2_0_bn1 0.000194073 0.348919
call_module fc 0.000189066 0.339917
call_module layer2_1_bn2 0.000165462 0.297481
call_module layer4_1_bn2 0.000159979 0.287622
call_module layer1_1_bn1 0.000138998 0.249901
call_module layer2_0_downsample_1 0.000130892 0.235327
call_module avgpool 0.000122786 0.220753
call_module layer3_1_bn2 0.000111818 0.201036
call_module layer4_0_bn2 0.000100136 0.180032
call_module layer1_0_relu 9.2268e-05 0.165887
call_module layer1_0_relu_1 9.08375e-05 0.163315
call_module layer3_0_bn2 9.01222e-05 0.162029
call_module layer4_1_bn1 8.84533e-05 0.159028
call_module layer1_1_relu_1 8.32081e-05 0.149598
call_module layer2_0_bn2 8.2016e-05 0.147455
call_module layer2_1_bn1 8.17776e-05 0.147026
call_function add_2 7.84397e-05 0.141025
call_function add_5 7.58171e-05 0.13631
call_module layer4_0_downsample_1 7.39098e-05 0.132881
call_module layer4_0_bn1 7.15256e-05 0.128594
call_module layer3_0_bn1 7.03335e-05 0.126451
call_module layer3_0_downsample_1 7.03335e-05 0.126451
call_module layer1_1_relu 6.81877e-05 0.122593
call_function add_7 6.65188e-05 0.119593
call_module layer3_1_bn1 6.62804e-05 0.119164
call_function add_6 6.10352e-05 0.109734
call_function add_4 5.48363e-05 0.0985889
call_module layer4_1_relu 5.26905e-05 0.0947311
call_module layer2_0_relu 4.98295e-05 0.0895873
call_module layer2_1_relu_1 4.79221e-05 0.0861581
call_module layer4_0_relu 4.72069e-05 0.0848722
call_module layer4_0_relu_1 4.52995e-05 0.081443
call_module layer2_0_relu_1 4.45843e-05 0.0801571
call_module layer4_1_relu_1 4.45843e-05 0.0801571
call_module layer2_1_relu 4.17233e-05 0.0750133
call_module layer3_1_relu 3.69549e-05 0.0664403
call_module layer3_0_relu 3.62396e-05 0.0651544
call_module layer3_0_relu_1 3.55244e-05 0.0638685
call_module layer3_1_relu_1 3.40939e-05 0.0612966
call_function flatten 2.57492e-05 0.0462939
placeholder x 2.24113e-05 0.0402929
output output 9.53674e-06 0.0171459
这里有两点需要指出：
MaxPool2d耗时最长。这是一个已知问题：pytorch/pytorch#51393
结论#
正如我们所见，使用 FX 我们可以轻松地将 PyTorch 程序（甚至是我们没有源代码的程序！）捕获为机器可解释的格式，并用于分析，比如我们在这里所做的性能分析。FX 为处理 PyTorch 程序开辟了一个激动人心的可能性世界。
最后，由于 FX 仍处于测试阶段，我们非常乐意听到您关于使用它的任何反馈。请随时使用 PyTorch 论坛（https://discuss.pytorch.org/）和问题跟踪器（pytorch/pytorch#issues）提供您可能有的任何反馈。
脚本总运行时间： (0 分 0.303 秒)
