[pytorch入门] 2. tensorboard

tensorboard简介

TensorBoard 是一组用于数据可视化的工具。它包含在流行的开源机器学习库 Tensorflow 中.但是也可以独立安装，服务Pytorch等其他的框架
可以常常用来观察训练过程中每一阶段如何输出的

• 安装

  pip install tensorboard
  

• 启动

  tensorboard --logdir=<directory_name>
  

  会默认在6006端口打开，也可以自行制定窗口，如:

  tensorboard --logdir=logs --port=6007
  

用法

1. 所在类：

   from torch.utils.tensorboard import SummaryWriter
   

   介绍：

   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.
       """
   

2. 创建对象

   writer = SummaryWriter('logs') # 说明写入哪个文件夹
   

3. 常用方法

   writer.add_image()   # 图像方式
   writer.add_scalar()  # 坐标方式
   
   writer.close()  # 使用完之后需要close
   

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.
        添加标量数据到summary中

        Args:
            tag (str): Data identifier 图表标题
            scalar_value (float or string/blobname): Value to save 数值（y轴）
            global_step (int): Global step value to record 训练到多少步（x轴）
            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 %

        """

注意：向writer中写入新事件的同时她也会保留上一个事件，这就会导致一些拟合出现问题
解决：删除之前的log文件，重新生成

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 注意数据的类型
            global_step (int): Global step value to record
            后面不用管
            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``.
            """

实践

如在tensorboard中展示图片：

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

writer = SummaryWriter('logs')
image_path = './dataset2/train/ants_image/0013035.jpg'
img_PIL = Image.open(image_path)
img_array = np.array(img_PIL)
print(type(img_array))
print(img_array.shape)

writer.add_image("test",img_array,1,dataformats='HWC') # 展示读取的图片


for i in range(100):
    writer.add_scalar('y=2x', 3*i, i)     # 绘图

writer.close()

• writer.add_image中的参数

  def add_image(
          self, tag, img_tensor, global_step=None, walltime=None, dataformats="CHW"
      ):
  

  名称、图形向量（ndarray类型），第几步（是滑动翻页那种的，这里相当于设定是第几页，每次向后设定时不会清除原来的数据）

当前代码效果如图：

修改图片后：

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

writer = SummaryWriter('logs')
image_path = './dataset2/train/ants_image/5650366_e22b7e1065.jpg'
img_PIL = Image.open(image_path)
img_array = np.array(img_PIL)
print(type(img_array))
print(img_array.shape)

# 这里更新，说明为第二步
writer.add_image("test",img_array,2,dataformats='HWC')


for i in range(100):
    writer.add_scalar('y=2x', 3*i, i)

writer.close()

拖拉就会发现有两张图