【解析】TensorBoard 可视化

在深度学习之中，可视化是必不可少的关键一步，虽然很多时候可以将需要的信息在输出窗口打印出来，但是能表达的信息是在太少，比如 loss趋势图、acc趋势图、权重分布图等等，都难以描述出来；Tensorboard 本是为了 Tensorflow而设计的一款可视化工具，后期 Pytorch 也引入支持，极大的小伙伴们的学习难度，虽然也有其他可视化工具支持，但是 Tensorboard可以直接迁移过来使用，上手还是更快。

官方链接：https://pytorch.org/docs/stable/tensorboard.html

毫无疑问，在使用 tensorboard 之前，需要先安装

pip install tensorboard

Tensorboard 主要的类是 SummaryWriter，所以在使用 Tensorboard 的时候，我们需要先创建一个 SummaryWriter 实例

from torch.utils.tensorboard import SummaryWriter

writer = SummaryWriter()

在 PyCharm 的 Terminal 中输入 以下命令即可
注意⚠️：Terminal 的路径必须是在本工程下，不然可能会报错 或 得到意想不到的事！

# 在PyCharm 的 Terminal 中输入
tensorboard --logdir=runs

# 退出 Tensorboard
CTRL + C 键

1、SummaryWriter主要方法

1.1 add_scalar

add_scalar(tag, scalar_value, global_step=None, walltime=None)

Add scalar data to summary.

• tag (string) – 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

1.2 add_scalars

add_scalars(main_tag, tag_scalar_dict, global_step=None, walltime=None)

Adds many scalar data to summary.

• main_tag (string) – The parent name for the tags
• tag_scalar_dict (dict) – Key-value pair storing the tag and corresponding values
• global_step (int) – Global step value to record
• walltime (float) – Optional override default walltime (time.time()) seconds after epoch of event

1.3 add_graph

add_graph(model, input_to_model=None, verbose=False)

Add graph data to summary.

Parameters

• model (torch.nn.Module) – Model to draw.
• input_to_model (torch.Tensor or list of torch.Tensor) – A variable or a tuple of variables to be fed.
• verbose (bool) – Whether to print graph structure in console.

1.4 add_histogram

add_histogram(tag, values, global_step=None, bins='tensorflow', walltime=None, max_bins=None)

Add histogram to summary.

• tag (string) – Data identifier
• values (torch.Tensor, numpy.array, or string/blobname) – Values to build histogram
• global_step (int) – Global step value to record
• bins (string) – One of {‘tensorflow’,’auto’, ‘fd’, …}. This determines how the bins are made.
• walltime (float) – Optional override default walltime (time.time()) seconds after epoch of event

1.5 add_image

add_image(tag, img_tensor, global_step=None, walltime=None, dataformats='CHW')

Add image data to summary.

Note that this requires the pillow package.

• tag (string) – Data identifier
• img_tensor (torch.Tensor, numpy.array, 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

Shape:
img_tensor: Default is $(3,H,W)$. You can use torchvision.utils.make_grid() to convert a batch of tensor into $3\times H\times W$ format or call add_images and let us do the job. Tensor with $(1,H,W),(H,W),(H,W,3)$ is also suitable as long as corresponding dataformats argument is passed, e.g. CHW, HWC, HW.

1.6 add_images

add_images(tag, img_tensor, global_step=None, walltime=None, dataformats='NCHW')

Add batched image data to summary.

Note that this requires the pillow package.

• tag (string) – Data identifier
• img_tensor (torch.Tensor, numpy.array, 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 (string) – Image data format specification of the form NCHW, NHWC, CHW, HWC, HW, WH, etc.

Shape:
img_tensor: Default is (N,3,H,W). If dataformats is specified, other shape will be accepted. e.g. NCHW or NHWC.

2、测试

测试代码

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

writer = SummaryWriter()

for n_iter in range(100):
    writer.add_scalar('Loss/train', np.random.random(), n_iter)
    writer.add_scalar('Loss/test', np.random.random(), n_iter)
    writer.add_scalar('Accuracy/train', np.random.random(), n_iter)
    writer.add_scalar('Accuracy/test', np.random.random(), n_iter)

writer.close()
    

当运行完 测试代码/项目工程之后，默认情况下，会在项目下生成一个 rans 的文件夹，文件夹内就是运行的结果，接下来，我们需要在 PyCharm 的 Terminal 中输入 以下命令即可
注意⚠️：Terminal 的路径必须是在本工程下，不然可能会报错 或 得到意想不到的事！

在Terminal 运行完之后，会看到 一个蓝色的 http://localhost:6006/，点击就会自动进入 Tensorboard 的网页界面，见效果图

# 在PyCharm 的 Terminal 中输入
tensorboard --logdir=runs

# 退出 Tensorboard
CTRL + C 键

Win系统

Mac

效果

问题

对于win系统用户而言，可能出现没有相应的趋势图，原因又可能是 浏览器的问题，在此推荐使用火狐浏览器，即可轻松解决