PyTorch学习笔记（二）——TensorBoard

1 用途

1、训练过程中loss是如何变化的，是否正常或是否按预想的变化，选择什么样的模型

2、模型在不同阶段的输出

2 需要导入的类和常用的方法

from torch.utils.tensorboard import SummaryWriter
writer.add_image()
writer.add_scalar()

查看SummaryWriter的官方文档

直接向log_dir文件夹写入事件文件，可以被TensorBoard进行解析

class SummaryWriter:
    """Writes entries directly to event files in the log_dir to be
    consumed by TensorBoard.

    The `SummaryWriter` class provides a high-level API to create an event file
    in a given directory and add summaries and events to it. The class updates the
    file contents asynchronously. This allows a training program to call methods
    to add data to the file directly from the training loop, without slowing down
    training.
    """

实例化SummaryWriter类

writer = SummaryWriter("logs")   # 把对应的事件文件存储到logs文件夹下

 2.1 add_scalar()

def add_scalar(
        self,
        tag,
        scalar_value,
        global_step=None,
        walltime=None,
        new_style=False,
        double_precision=False,
    ):
        """Add scalar data to summary.

        Args:
            tag (str): Data identifier
            scalar_value (float or string/blobname): Value to save
            global_step (int): Global step value to record
            walltime (float): Optional override default walltime (time.time())
              with seconds after epoch of event
            new_style (boolean): Whether to use new style (tensor field) or old
              style (simple_value field). New style could lead to faster data loading.
        Examples::

            from torch.utils.tensorboard import SummaryWriter
            writer = SummaryWriter()
            x = range(100)
            for i in x:
                writer.add_scalar('y=2x', i * 2, i)
            writer.close()

        Expected result:

        .. image:: _static/img/tensorboard/add_scalar.png
           :scale: 50 %

        """
        torch._C._log_api_usage_once("tensorboard.logging.add_scalar")
        if self._check_caffe2_blob(scalar_value):
            from caffe2.python import workspace

            scalar_value = workspace.FetchBlob(scalar_value)

        summary = scalar(
            tag, scalar_value, new_style=new_style, double_precision=double_precision
        )
        self._get_file_writer().add_summary(summary, global_step, walltime)

 参数：

• tag：类似于图标的标题title
• scalar_value：（y轴）对应的需要保存的数值
• global_step：（x轴）步数，训练到多少步

举例：

from torch.utils.tensorboard import SummaryWriter   #导入SummaryWriter类

#创建实例
writer=SummaryWriter("logs")   #把对应的事件文件存储到logs文件夹下
 
for i in range(100):
    writer.add_scalar("y=2x",2*i,i)
 
writer.close()

运行后多了一个logs文件夹，里面是TensorBoard的一些事件文件。

打开事件文件的方法：在Terminal里输入：tensorboard --logdir=logs

默认打开的都是6006端口，也可以指定端口：​tensorboard --logdir=logs --port=6007

 

每向writer中写入新的事件，也记录了上一个事件，会出现过拟合现象，例如

解决方法：删除logs下的文件，重新运行代码

  2.2 add_image()

    def add_image(
        self, tag, img_tensor, global_step=None, walltime=None, dataformats="CHW"
    ):
        """Add image data to summary.

        Note that this requires the ``pillow`` package.

        Args:
            tag (str): Data identifier
            img_tensor (torch.Tensor, numpy.ndarray, or string/blobname): Image data
            global_step (int): Global step value to record
            walltime (float): Optional override default walltime (time.time())
              seconds after epoch of event
            dataformats (str): Image data format specification of the form
              CHW, HWC, HW, WH, etc.
        Shape:
            img_tensor: Default is :math:`(3, H, W)`. You can use ``torchvision.utils.make_grid()`` to
            convert a batch of tensor into 3xHxW format or call ``add_images`` and let us do the job.
            Tensor with :math:`(1, H, W)`, :math:`(H, W)`, :math:`(H, W, 3)` is also suitable as long as
            corresponding ``dataformats`` argument is passed, e.g. ``CHW``, ``HWC``, ``HW``.

        Examples::

            from torch.utils.tensorboard import SummaryWriter
            import numpy as np
            img = np.zeros((3, 100, 100))
            img[0] = np.arange(0, 10000).reshape(100, 100) / 10000
            img[1] = 1 - np.arange(0, 10000).reshape(100, 100) / 10000

            img_HWC = np.zeros((100, 100, 3))
            img_HWC[:, :, 0] = np.arange(0, 10000).reshape(100, 100) / 10000
            img_HWC[:, :, 1] = 1 - np.arange(0, 10000).reshape(100, 100) / 10000

            writer = SummaryWriter()
            writer.add_image('my_image', img, 0)

            # If you have non-default dimension setting, set the dataformats argument.
            writer.add_image('my_image_HWC', img_HWC, 0, dataformats='HWC')
            writer.close()

        Expected result:

        .. image:: _static/img/tensorboard/add_image.png
           :scale: 50 %

        """
        torch._C._log_api_usage_once("tensorboard.logging.add_image")
        if self._check_caffe2_blob(img_tensor):
            from caffe2.python import workspace

            img_tensor = workspace.FetchBlob(img_tensor)
        self._get_file_writer().add_summary(
            image(tag, img_tensor, dataformats=dataformats), global_step, walltime
        )

参数：

• tag：对应图像的title
• img_tensor：图像的数据类型，只能是torch.Tensor、numpy.array、string/blobname
• global_step：训练步骤，int 类型

举例：

from torch.utils.tensorboard import SummaryWriter
from PIL import Image
import numpy as np

writer = SummaryWriter("logs")
img_path = "dataset/train/ants/5650366_e22b7e1065.jpg"
#<class 'PIL.JpegImagePlugin.JpegImageFile'> 类型不符合要求
img_PIL = Image.open(img_path)
#把PIL类型的img变量转换为numpy类型（add_image() 函数所需要的图像的数据类型），重新赋值给img_array
img_array = np.array(img_PIL)
print(type(img_array))  #<class 'numpy.ndarray'>
print(img_array.shape)  #(375, 500, 3) 即(H,W,C)(高度，宽度，通道)

writer.add_image("test",img_array,2,dataformats='HWC')

 在一个title下，通过滑块显示每一步的图形，可以直观地观察训练中给model提供了哪些数据，或者想对model进行测试时，可以看到每个阶段的输出结果。如果想要单独显示，重命名一下title即可