%matplotlib inline

数据并行(选读)

Authors: Sung Kim and Jenny Kang

在这个教程里,我们将学习如何使用 DataParallel 来使用多GPU。

PyTorch非常容易就可以使用多GPU,用如下方式把一个模型放到GPU上:


    device = torch.device("cuda:0")
    model.to(device)

GPU: 然后复制所有的张量到GPU上:


    mytensor = my_tensor.to(device)

请注意,只调用my_tensor.to(device)并没有复制张量到GPU上,而是返回了一个copy。所以你需要把它赋值给一个新的张量并在GPU上使用这个张量。

在多GPU上执行前向和反向传播是自然而然的事。 但是PyTorch默认将只使用一个GPU。

使用DataParallel可以轻易的让模型并行运行在多个GPU上。


    model = nn.DataParallel(model)

这才是这篇教程的核心,接下来我们将更详细的介绍它。

导入和参数

导入PyTorch模块和定义参数。

import torch
import torch.nn as nn
from torch.utils.data import Dataset, DataLoader

# Parameters and DataLoaders
input_size = 5
output_size = 2

batch_size = 30
data_size = 100

Device

device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")

虚拟数据集

制作一个虚拟(随机)数据集, 你只需实现 __getitem__

class RandomDataset(Dataset):

    def __init__(self, size, length):
        self.len = length
        self.data = torch.randn(length, size)

    def __getitem__(self, index):
        return self.data[index]

    def __len__(self):
        return self.len

rand_loader = DataLoader(dataset=RandomDataset(input_size, data_size),
                         batch_size=batch_size, shuffle=True)

简单模型

作为演示,我们的模型只接受一个输入,执行一个线性操作,然后得到结果。 说明:DataParallel能在任何模型(CNN,RNN,Capsule Net等)上使用。

我们在模型内部放置了一条打印语句来打印输入和输出向量的大小。

请注意批次的秩为0时打印的内容。

class Model(nn.Module):
    # Our model

    def __init__(self, input_size, output_size):
        super(Model, self).__init__()
        self.fc = nn.Linear(input_size, output_size)

    def forward(self, input):
        output = self.fc(input)
        print("\tIn Model: input size", input.size(),
              "output size", output.size())

        return output

创建一个模型和数据并行

这是本教程的核心部分。

首先,我们需要创建一个模型实例和检测我们是否有多个GPU。 如果有多个GPU,使用nn.DataParallel来包装我们的模型。 然后通过model.to(device)把模型放到GPU上。

model = Model(input_size, output_size)
if torch.cuda.device_count() > 1:
    print("Let's use", torch.cuda.device_count(), "GPUs!")
    # dim = 0 [30, xxx] -> [10, ...], [10, ...], [10, ...] on 3 GPUs
    model = nn.DataParallel(model)

model.to(device)
Model(
  (fc): Linear(in_features=5, out_features=2, bias=True)
)

运行模型

现在可以看到输入和输出张量的大小。

for data in rand_loader:
    input = data.to(device)
    output = model(input)
    print("Outside: input size", input.size(),
          "output_size", output.size())
    In Model: input size torch.Size([30, 5]) output size torch.Size([30, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([30, 5]) output size torch.Size([30, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([30, 5]) output size torch.Size([30, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
Outside: input size torch.Size([10, 5]) output_size torch.Size([10, 2])

结果

当没有或者只有一个GPU时,对30个输入和输出进行批处理,得到了期望的一样得到30个输入和输出,但是如果你有多个GPU,你得到如下的结果。

2 GPUs ~

If you have 2, you will see:

.. code:: bash

# on 2 GPUs
Let's use 2 GPUs!
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
    In Model: input size torch.Size([15, 5]) output size torch.Size([15, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([5, 5]) output size torch.Size([5, 2])
    In Model: input size torch.Size([5, 5]) output size torch.Size([5, 2])
Outside: input size torch.Size([10, 5]) output_size torch.Size([10, 2])

3 GPUs ~

If you have 3 GPUs, you will see:

.. code:: bash

Let's use 3 GPUs!
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
    In Model: input size torch.Size([10, 5]) output size torch.Size([10, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
Outside: input size torch.Size([10, 5]) output_size torch.Size([10, 2])

8 GPUs ~~

If you have 8, you will see:

.. code:: bash

Let's use 8 GPUs!
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([4, 5]) output size torch.Size([4, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
Outside: input size torch.Size([30, 5]) output_size torch.Size([30, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
    In Model: input size torch.Size([2, 5]) output size torch.Size([2, 2])
Outside: input size torch.Size([10, 5]) output_size torch.Size([10, 2])

总结

DataParallel会自动的划分数据,并将作业发送到多个GPU上的多个模型。 并在每个模型完成作业后,收集合并结果并返回。

更多信息请看这里: https://pytorch.org/tutorials/beginner/former_torchies/parallelism_tutorial.html.