blocks|key|2870499|text|如果你在内存缓冲区中有图像，这是很容易做到的。下面，我展示了一个示例，其中pyplot被保存到缓冲区，然后转换为TF图像表示，然后将其发送到图像摘要。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|2870500|import+io
import+matplotlib.pyplot+as+plt
import+tensorflow+as+tf


def+gen_plot():
++++"""Create+a+pyplot+plot+and+save+to+buffer."""
++++plt.figure()
++++plt.plot([1,+2])
++++plt.title("test")
++++buf+=+io.BytesIO()
++++plt.savefig(buf,+format='png')
++++buf.seek(0)
++++return+buf


#+Prepare+the+plot
plot_buf+=+gen_plot()

#+Convert+PNG+buffer+to+TF+image
image+=+tf.image.decode_png(plot_buf.getvalue(),+channels=4)

#+Add+the+batch+dimension
image+=+tf.expand_dims(image,+0)

#+Add+image+summary
summary_op+=+tf.summary.image("plot",+image)

#+Session
with+tf.Session()+as+sess:
++++#+Run
++++summary+=+sess.run(summary_op)
++++#+Write+summary
++++writer+=+tf.train.SummaryWriter('./logs')
++++writer.add_summary(summary)
++++writer.close()|code-block|syntax|javascript|2870501|这提供了以下TensorBoard可视化：|2870502|​|2870503|📷|atomic|offset|length|2870504|2870505|entityMap|0|IMAGE|mutability|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/e64b47a99783e75f2e7b2d8ef9b289fa.png|imageAlt^0|0|0|0|0|0|1|0|0|0^^$0|@$1|2|3|4|5|6|7|Z|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|10|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|11|8|@]|9|@]|A|$]]|$1|I|3|J|5|6|7|12|8|@]|9|@]|A|$]]|$1|K|3|L|5|M|7|13|8|@]|9|@$N|14|O|15|1|16]]|A|$]]|$1|P|3|J|5|6|7|17|8|@]|9|@]|A|$]]|$1|Q|3|-4|5|6|7|18|8|@]|9|@]|A|$]]]|R|$S|$5|T|U|V|A|$W|X|Y|-4]]]]

It is quite easy to do if you have the image in a memory buffer. Below, I show an example, where a pyplot is saved to a buffer and then converted to a TF image representation which is then sent to an image summary.

<pre><code>import io
import matplotlib.pyplot as plt
import tensorflow as tf


def gen_plot():
 """Create a pyplot plot and save to buffer."""
 plt.figure()
 plt.plot([1, 2])
 plt.title("test")
 buf = io.BytesIO()
 plt.savefig(buf, format='png')
 buf.seek(0)
 return buf


# Prepare the plot
plot_buf = gen_plot()

# Convert PNG buffer to TF image
image = tf.image.decode_png(plot_buf.getvalue(), channels=4)

# Add the batch dimension
image = tf.expand_dims(image, 0)

# Add image summary
summary_op = tf.summary.image("plot", image)

# Session
with tf.Session() as sess:
 # Run
 summary = sess.run(summary_op)
 # Write summary
 writer = tf.train.SummaryWriter('./logs')
 writer.add_summary(summary)
 writer.close()
</code></pre>

This gives the following TensorBoard visualization:

<a href="https://i.stack.imgur.com/ARU43.png" rel="noreferrer"><img src="https://i.stack.imgur.com/ARU43.png" alt="enter image description here"></a>

blocks|key|111985|text|下一个脚本不使用中间RGB/PNG编码。它还修复了执行过程中额外的操作构造的问题，单个摘要被重用。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|111986|该图形的大小在执行过程中应保持不变|111987|有效的解决方案：|111988|import+matplotlib.pyplot+as+plt
import+tensorflow+as+tf
import+numpy+as+np

def+get_figure():
++fig+=+plt.figure(num=0,+figsize=(6,+4),+dpi=300)
++fig.clf()
++return+fig


def+fig2rgb_array(fig,+expand=True):
++fig.canvas.draw()
++buf+=+fig.canvas.tostring_rgb()
++ncols,+nrows+=+fig.canvas.get_width_height()
++shape+=+(nrows,+ncols,+3)+if+not+expand+else+(1,+nrows,+ncols,+3)
++return+np.fromstring(buf,+dtype=np.uint8).reshape(shape)


def+figure_to_summary(fig):
++image+=+fig2rgb_array(fig)
++summary_writer.add_summary(
++++vis_summary.eval(feed_dict={vis_placeholder:+image}))


if+__name__+==+'__main__':
++++++#+construct+graph
++++++x+=+tf.Variable(initial_value=tf.random_uniform((2,+10)))
++++++inc+=+x.assign(x+%2B+1)

++++++#+construct+summary
++++++fig+=+get_figure()
++++++vis_placeholder+=+tf.placeholder(tf.uint8,+fig2rgb_array(fig).shape)
++++++vis_summary+=+tf.summary.image('custom',+vis_placeholder)

++++++with+tf.Session()+as+sess:
++++++++tf.global_variables_initializer().run()
++++++++summary_writer+=+tf.summary.FileWriter('./tmp',+sess.graph)

++++++++for+i+in+range(100):
++++++++++#+execute+step
++++++++++_,+values+=+sess.run([inc,+x])
++++++++++#+draw+on+the+plot
++++++++++fig+=+get_figure()
++++++++++plt.subplot('111').scatter(values[0],+values[1])
++++++++++#+save+the+summary
++++++++++figure_to_summary(fig)|code-block|syntax|javascript|111989|entityMap^0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|M|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|N|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|O|8|@]|9|@]|A|$]]|$1|F|3|G|5|H|7|P|8|@]|9|@]|A|$I|J]]|$1|K|3|-4|5|6|7|Q|8|@]|9|@]|A|$]]]|L|$]]

Next script does not use intermediate RGB/PNG encoding. It also fixes the issue with additional operation construction during execution, single summary is reused.

Size of the figure is expected to remain the same during execution

Solution that works:

<pre><code>import matplotlib.pyplot as plt
import tensorflow as tf
import numpy as np

def get_figure():
 fig = plt.figure(num=0, figsize=(6, 4), dpi=300)
 fig.clf()
 return fig


def fig2rgb_array(fig, expand=True):
 fig.canvas.draw()
 buf = fig.canvas.tostring_rgb()
 ncols, nrows = fig.canvas.get_width_height()
 shape = (nrows, ncols, 3) if not expand else (1, nrows, ncols, 3)
 return np.fromstring(buf, dtype=np.uint8).reshape(shape)


def figure_to_summary(fig):
 image = fig2rgb_array(fig)
 summary_writer.add_summary(
 vis_summary.eval(feed_dict={vis_placeholder: image}))


if __name__ == '__main__':
 # construct graph
 x = tf.Variable(initial_value=tf.random_uniform((2, 10)))
 inc = x.assign(x + 1)

 # construct summary
 fig = get_figure()
 vis_placeholder = tf.placeholder(tf.uint8, fig2rgb_array(fig).shape)
 vis_summary = tf.summary.image('custom', vis_placeholder)

 with tf.Session() as sess:
 tf.global_variables_initializer().run()
 summary_writer = tf.summary.FileWriter('./tmp', sess.graph)

 for i in range(100):
 # execute step
 _, values = sess.run([inc, x])
 # draw on the plot
 fig = get_figure()
 plt.subplot('111').scatter(values[0], values[1])
 # save the summary
 figure_to_summary(fig)
</code></pre>

blocks|key|112310|text|这打算完成Andrzej+Pronobis的回答。紧跟他的帖子，我设置了这个最小工作示例|type|unstyled|depth|inlineStyleRanges|offset|length|style|BOLD|entityRanges|data|112311|++++plt.figure()
++++plt.plot([1,+2])
++++plt.title("test")
++++buf+=+io.BytesIO()
++++plt.savefig(buf,+format='png')
++++buf.seek(0)
++++image+=+tf.image.decode_png(buf.getvalue(),+channels=4)
++++image+=+tf.expand_dims(image,+0)
++++summary+=+tf.summary.image("test",+image,+max_outputs=1)
++++writer.add_summary(summary,+step)|code-block|syntax|javascript|112312|其中writer是tf.summary.FileWriter的一个实例。这给了我以下错误：AttributeError：'Tensor‘对象没有’‘属性，this+github+post有解决方案：在将摘要添加到编写器之前，必须对其进行评估(转换为字符串)。因此，我的工作代码仍然如下(只需在最后一行添加.eval()调用)：|112313|++++plt.figure()
++++plt.plot([1,+2])
++++plt.title("test")
++++buf+=+io.BytesIO()
++++plt.savefig(buf,+format='png')
++++buf.seek(0)
++++image+=+tf.image.decode_png(buf.getvalue(),+channels=4)
++++image+=+tf.expand_dims(image,+0)
++++summary+=+tf.summary.image("test",+image,+max_outputs=1)
++++writer.add_summary(summary.eval(),+step)|112314|这篇文章可能足够简短，可以作为对他答案的评论，但这些很容易被忽略(我可能也在做其他不同的事情)，所以在这里，希望它能有所帮助！|112315|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/api_docs/python/tf/summary/FileWriter|1|https://github.com/tensorflow/tensorflow/issues/9656^0|12|6|0|0|19|S|2R|A|9|L|0|25|G|1|0|0|0^^$0|@$1|2|3|4|5|6|7|10|8|@$9|11|A|12|B|C]]|D|@]|E|$]]|$1|F|3|G|5|H|7|13|8|@]|D|@]|E|$I|J]]|$1|K|3|L|5|6|7|14|8|@$9|15|A|16|B|C]|$9|17|A|18|B|C]]|D|@$9|19|A|1A|1|1B]|$9|1C|A|1D|1|1E]]|E|$]]|$1|M|3|N|5|H|7|1F|8|@]|D|@]|E|$I|J]]|$1|O|3|P|5|6|7|1G|8|@]|D|@]|E|$]]|$1|Q|3|-4|5|6|7|1H|8|@]|D|@]|E|$]]]|R|$S|$5|T|U|V|E|$W|X]]|Y|$5|T|U|V|E|$W|Z]]]]

This intends to complete Andrzej Pronobis' answer. Following closely his nice post, I set up this minimal working example:
<pre><code> plt.figure()
 plt.plot([1, 2])
 plt.title(&quot;test&quot;)
 buf = io.BytesIO()
 plt.savefig(buf, format='png')
 buf.seek(0)
 image = tf.image.decode_png(buf.getvalue(), channels=4)
 image = tf.expand_dims(image, 0)
 summary = tf.summary.image(&quot;test&quot;, image, max_outputs=1)
 writer.add_summary(summary, step)
</code></pre>
Where writer is an instance of <a href="https://www.tensorflow.org/api_docs/python/tf/summary/FileWriter" rel="nofollow noreferrer">tf.summary.FileWriter</a>.
This gave me the following error:
AttributeError: 'Tensor' object has no attribute 'value'
For which <a href="https://github.com/tensorflow/tensorflow/issues/9656" rel="nofollow noreferrer">this github post</a> had the solution: the summary has to be evaluated (converted into a string) before being added to the writer. So the working code for me remained as follows (simply add the .eval() call in the last line):
<pre><code> plt.figure()
 plt.plot([1, 2])
 plt.title(&quot;test&quot;)
 buf = io.BytesIO()
 plt.savefig(buf, format='png')
 buf.seek(0)
 image = tf.image.decode_png(buf.getvalue(), channels=4)
 image = tf.expand_dims(image, 0)
 summary = tf.summary.image(&quot;test&quot;, image, max_outputs=1)
 writer.add_summary(summary.eval(), step)
</code></pre>
This could be short enough to be a comment on his answer, but these can be easily overlooked (and I may be doing something else different too), so here it is, hope it helps!

blocks|key|2870719|text|我的回答有点晚了。使用tf-matplotlib，一个简单的散点图可以归结为：|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|2870720|import+tensorflow+as+tf
import+numpy+as+np

import+tfmpl

@tfmpl.figure_tensor
def+draw_scatter(scaled,+colors):+
++++'''Draw+scatter+plots.+One+for+each+color.'''++
++++figs+=+tfmpl.create_figures(len(colors),+figsize=(4,4))
++++for+idx,+f+in+enumerate(figs):
++++++++ax+=+f.add_subplot(111)
++++++++ax.axis('off')
++++++++ax.scatter(scaled[:,+0],+scaled[:,+1],+c=colors[idx])
++++++++f.tight_layout()

++++return+figs

with+tf.Session(graph=tf.Graph())+as+sess:

++++#+A+point+cloud+that+can+be+scaled+by+the+user
++++points+=+tf.constant(
++++++++np.random.normal(loc=0.0,+scale=1.0,+size=(100,+2)).astype(np.float32)
++++)
++++scale+=+tf.placeholder(tf.float32)++++++++
++++scaled+=+points*scale

++++#+Note,+`scaled`+above+is+a+tensor.+Its+being+passed+`draw_scatter`+below.+
++++#+However,+when+`draw_scatter`+is+invoked,+the+tensor+will+be+evaluated+and+a
++++#+numpy+array+representing+its+content+is+provided.+++
++++image_tensor+=+draw_scatter(scaled,+['r',+'g'])
++++image_summary+=+tf.summary.image('scatter',+image_tensor)++++++
++++all_summaries+=+tf.summary.merge_all()+

++++writer+=+tf.summary.FileWriter('log',+sess.graph)
++++summary+=+sess.run(all_summaries,+feed_dict={scale:+2.})
++++writer.add_summary(summary,+global_step=0)|code-block|syntax|javascript|2870721|执行时，这将在Tensorboard内部生成以下图|2870722|📷|atomic|2870723|​|2870724|请注意，tf-matplotlib负责评估任何张量输入，避免pyplot线程问题，并支持用于运行时关键绘图的blitting。|style|BOLD|CODE|2870725|entityMap|0|LINK|mutability|MUTABLE|url|http://github.com/cheind/tf-matplotlib|1|IMAGE|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/6ad968d7b8717419a53a8ffb21582322.png|imageAlt^0|B|D|0|0|0|0|0|1|1|0|0|4|D|U|6|0^^$0|@$1|2|3|4|5|6|7|18|8|@]|9|@$A|19|B|1A|1|1B]]|C|$]]|$1|D|3|E|5|F|7|1C|8|@]|9|@]|C|$G|H]]|$1|I|3|J|5|6|7|1D|8|@]|9|@]|C|$]]|$1|K|3|L|5|M|7|1E|8|@]|9|@$A|1F|B|1G|1|1H]]|C|$]]|$1|N|3|O|5|6|7|1I|8|@]|9|@]|C|$]]|$1|P|3|Q|5|6|7|1J|8|@$A|1K|B|1L|R|S]|$A|1M|B|1N|R|T]]|9|@]|C|$]]|$1|U|3|-4|5|6|7|1O|8|@]|9|@]|C|$]]]|V|$W|$5|X|Y|Z|C|$10|11]]|12|$5|13|Y|14|C|$15|16|17|-4]]]]

A bit late with my answer. With <a href="http://github.com/cheind/tf-matplotlib" rel="noreferrer">tf-matplotlib</a> a simple scatter plot boils down to:

<pre><code>import tensorflow as tf
import numpy as np

import tfmpl

@tfmpl.figure_tensor
def draw_scatter(scaled, colors): 
 '''Draw scatter plots. One for each color.''' 
 figs = tfmpl.create_figures(len(colors), figsize=(4,4))
 for idx, f in enumerate(figs):
 ax = f.add_subplot(111)
 ax.axis('off')
 ax.scatter(scaled[:, 0], scaled[:, 1], c=colors[idx])
 f.tight_layout()

 return figs

with tf.Session(graph=tf.Graph()) as sess:

 # A point cloud that can be scaled by the user
 points = tf.constant(
 np.random.normal(loc=0.0, scale=1.0, size=(100, 2)).astype(np.float32)
 )
 scale = tf.placeholder(tf.float32) 
 scaled = points*scale

 # Note, `scaled` above is a tensor. Its being passed `draw_scatter` below. 
 # However, when `draw_scatter` is invoked, the tensor will be evaluated and a
 # numpy array representing its content is provided. 
 image_tensor = draw_scatter(scaled, ['r', 'g'])
 image_summary = tf.summary.image('scatter', image_tensor) 
 all_summaries = tf.summary.merge_all() 

 writer = tf.summary.FileWriter('log', sess.graph)
 summary = sess.run(all_summaries, feed_dict={scale: 2.})
 writer.add_summary(summary, global_step=0)
</code></pre>

When executed, this results in the following plot inside Tensorboard
<img src="https://github.com/cheind/tf-matplotlib/raw/master/etc/scatter.png" alt="">

Note that tf-matplotlib takes care about evaluating any tensor inputs, avoids <code>pyplot</code> threading issues and supports blitting for runtime critical plotting.

blocks|key|2870760|text|最后有一些关于“记录任意图像数据”的official+documentation，以matplotlib创建的图像为例。|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|2870761|他们写道：|2870762|在下面的代码中，您将使用matplotlib的subplot()函数将前25张图像记录为一个漂亮的网格。然后，您将在TensorBoard中查看网格：|2870763|#+Clear+out+prior+logging+data.
!rm+-rf+logs/plots

logdir+=+"logs/plots/"+%2B+datetime.now().strftime("%25Y%25m%25d-%25H%25M%25S")
file_writer+=+tf.summary.create_file_writer(logdir)

def+plot_to_image(figure):
++"""Converts+the+matplotlib+plot+specified+by+'figure'+to+a+PNG+image+and
++returns+it.+The+supplied+figure+is+closed+and+inaccessible+after+this+call."""
++#+Save+the+plot+to+a+PNG+in+memory.
++buf+=+io.BytesIO()
++plt.savefig(buf,+format='png')
++#+Closing+the+figure+prevents+it+from+being+displayed+directly+inside
++#+the+notebook.
++plt.close(figure)
++buf.seek(0)
++#+Convert+PNG+buffer+to+TF+image
++image+=+tf.image.decode_png(buf.getvalue(),+channels=4)
++#+Add+the+batch+dimension
++image+=+tf.expand_dims(image,+0)
++return+image

def+image_grid():
++"""Return+a+5x5+grid+of+the+MNIST+images+as+a+matplotlib+figure."""
++#+Create+a+figure+to+contain+the+plot.
++figure+=+plt.figure(figsize=(10,10))
++for+i+in+range(25):
++++#+Start+next+subplot.
++++plt.subplot(5,+5,+i+%2B+1,+title=class_names[train_labels[i]])
++++plt.xticks([])
++++plt.yticks([])
++++plt.grid(False)
++++plt.imshow(train_images[i],+cmap=plt.cm.binary)
++
++return+figure

#+Prepare+the+plot
figure+=+image_grid()
#+Convert+to+image+and+log
with+file_writer.as_default():
++tf.summary.image("Training+data",+plot_to_image(figure),+step=0)

%25tensorboard+--logdir+logs/plots|code-block|syntax|javascript|2870764|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/tensorboard/image_summaries#logging_arbitrary_image_data^0|I|M|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|U|8|@]|9|@$A|V|B|W|1|X]]|C|$]]|$1|D|3|E|5|6|7|Y|8|@]|9|@]|C|$]]|$1|F|3|G|5|6|7|Z|8|@]|9|@]|C|$]]|$1|H|3|I|5|J|7|10|8|@]|9|@]|C|$K|L]]|$1|M|3|-4|5|6|7|11|8|@]|9|@]|C|$]]]|N|$O|$5|P|Q|R|C|$S|T]]]]

Finally there is some <a href="https://www.tensorflow.org/tensorboard/image_summaries#logging_arbitrary_image_data" rel="nofollow noreferrer">official documentation</a> about &quot;Logging arbitrary image data&quot; with an example of matplotlib created images. 
They write:
In the code below, you'll log the first 25 images as a nice grid using matplotlib's subplot() function. You'll then view the grid in TensorBoard:
<pre><code># Clear out prior logging data.
!rm -rf logs/plots

logdir = &quot;logs/plots/&quot; + datetime.now().strftime(&quot;%Y%m%d-%H%M%S&quot;)
file_writer = tf.summary.create_file_writer(logdir)

def plot_to_image(figure):
 &quot;&quot;&quot;Converts the matplotlib plot specified by 'figure' to a PNG image and
 returns it. The supplied figure is closed and inaccessible after this call.&quot;&quot;&quot;
 # Save the plot to a PNG in memory.
 buf = io.BytesIO()
 plt.savefig(buf, format='png')
 # Closing the figure prevents it from being displayed directly inside
 # the notebook.
 plt.close(figure)
 buf.seek(0)
 # Convert PNG buffer to TF image
 image = tf.image.decode_png(buf.getvalue(), channels=4)
 # Add the batch dimension
 image = tf.expand_dims(image, 0)
 return image

def image_grid():
 &quot;&quot;&quot;Return a 5x5 grid of the MNIST images as a matplotlib figure.&quot;&quot;&quot;
 # Create a figure to contain the plot.
 figure = plt.figure(figsize=(10,10))
 for i in range(25):
 # Start next subplot.
 plt.subplot(5, 5, i + 1, title=class_names[train_labels[i]])
 plt.xticks([])
 plt.yticks([])
 plt.grid(False)
 plt.imshow(train_images[i], cmap=plt.cm.binary)
 
 return figure

# Prepare the plot
figure = image_grid()
# Convert to image and log
with file_writer.as_default():
 tf.summary.image(&quot;Training data&quot;, plot_to_image(figure), step=0)

%tensorboard --logdir logs/plots
</code></pre>

blocks|key|2659875|text|已弃用:对于PyTorch，请使用内置的SummaryWriter.add_figure+(请参阅其他answer)！|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|2659876|PyTorch解决方案：|2659877|2659878|使用MatPlotLib+figure|unordered-list-item|2659879|将其绘制到画布|2659880|，然后转换为numpy:|2659881|2659882|#+make+the+canvas
figure+=+plt.figure(figsize=(10,10))
canvas+=+matplotlib.backends.backend_agg.FigureCanvas(figure)

#+insert+plotting+code+here;+you+can+use+imshow+or+subplot,+etc.
for+i+in+range(25):
++++plt.subplot(5,+5,+i+%2B+1,+title=class_names[train_labels[i]])
++++plt.xticks([])
++++plt.yticks([])
++++plt.grid(False)
++++plt.imshow(train_images[i],+cmap=plt.cm.binary)

#+convert+canvas+to+figure
canvas.draw()
image+=+np.frombuffer(canvas.tostring_rgb(),+dtype='uint8').reshape((1000,1000,3)).transpose((2,+0,+1))|code-block|syntax|javascript|2659883|结果可以直接添加到Tensorboard：|2659884|tensorboard.add_image('name',+image,+global_step)|2659885|entityMap|0|LINK|mutability|MUTABLE|url|https://pytorch.org/docs/stable/tensorboard.html?highlight=add_figure#torch.utils.tensorboard.writer.SummaryWriter.add_figure|1|https://stackoverflow.com/a/66537286/10220850^0|K|O|K|O|0|1F|6|1|0|0|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|19|8|@$9|1A|A|1B|B|C]]|D|@$9|1C|A|1D|1|1E]|$9|1F|A|1G|1|1H]]|E|$]]|$1|F|3|G|5|6|7|1I|8|@]|D|@]|E|$]]|$1|H|3|-4|5|6|7|1J|8|@]|D|@]|E|$]]|$1|I|3|J|5|K|7|1K|8|@]|D|@]|E|$]]|$1|L|3|M|5|K|7|1L|8|@]|D|@]|E|$]]|$1|N|3|O|5|K|7|1M|8|@]|D|@]|E|$]]|$1|P|3|-4|5|6|7|1N|8|@]|D|@]|E|$]]|$1|Q|3|R|5|S|7|1O|8|@]|D|@]|E|$T|U]]|$1|V|3|W|5|6|7|1P|8|@]|D|@]|E|$]]|$1|X|3|Y|5|S|7|1Q|8|@]|D|@]|E|$T|U]]|$1|Z|3|-4|5|6|7|1R|8|@]|D|@]|E|$]]]|10|$11|$5|12|13|14|E|$15|16]]|17|$5|12|13|14|E|$15|18]]]]

Deprecated: For PyTorch make use of the built-in <a href="https://pytorch.org/docs/stable/tensorboard.html?highlight=add_figure#torch.utils.tensorboard.writer.SummaryWriter.add_figure" rel="nofollow noreferrer"><code>SummaryWriter.add_figure</code></a> (see other <a href="https://stackoverflow.com/a/66537286/10220850">answer</a>)!
PyTorch solution:
<ul>
<li>Use a MatPlotLib figure</li>
<li>draw it to a canvas</li>
<li>then convert to numpy:</li>
</ul>
<pre><code># make the canvas
figure = plt.figure(figsize=(10,10))
canvas = matplotlib.backends.backend_agg.FigureCanvas(figure)

# insert plotting code here; you can use imshow or subplot, etc.
for i in range(25):
 plt.subplot(5, 5, i + 1, title=class_names[train_labels[i]])
 plt.xticks([])
 plt.yticks([])
 plt.grid(False)
 plt.imshow(train_images[i], cmap=plt.cm.binary)

# convert canvas to figure
canvas.draw()
image = np.frombuffer(canvas.tostring_rgb(), dtype='uint8').reshape((1000,1000,3)).transpose((2, 0, 1))
</code></pre>
The result can be added directly to Tensorboard:
<pre><code>tensorboard.add_image('name', image, global_step)
</code></pre>

blocks|key|2870970|text|Pytorch+Lightning中的一个解决方案|type|unstyled|depth|inlineStyleRanges|entityRanges|data|2870971|这不是完整的类，而是您必须添加以使其在框架中工作的类。|2870972|import+pytorch_lightning+as+pl
import+seaborn+as+sn
import+pandas+as+pd
import+numpy+as+np
import+matplotlib.pyplot+as+plt
from+PIL+import+Image

def+__init__(self,+config,+trained_vae,+latent_dim):
++++self.val_confusion+=+pl.metrics.classification.ConfusionMatrix(num_classes=self._config.n_clusters)
++++self.logger:+Optional[TensorBoardLogger]+=+None

def+forward(self,+x):
++++...
++++return+log_probs

def+validation_step(self,+batch,+batch_index):
++++if+self._config.dataset+==+"mnist":
++++++++orig_batch,+label_batch+=+batch
++++++++orig_batch+=+orig_batch.reshape(-1,+28+*+28)

++++log_probs+=+self.forward(orig_batch)
++++loss+=+self._criterion(log_probs,+label_batch)

++++self.val_confusion.update(log_probs,+label_batch)
++++return+{"loss":+loss,+"labels":+label_batch}

def+validation_step_end(self,+outputs):
++++return+outputs

def+validation_epoch_end(self,+outs):
++++tb+=+self.logger.experiment

++++#+confusion+matrix
++++conf_mat+=+self.val_confusion.compute().detach().cpu().numpy().astype(np.int)
++++df_cm+=+pd.DataFrame(
++++++++conf_mat,
++++++++index=np.arange(self._config.n_clusters),
++++++++columns=np.arange(self._config.n_clusters))
++++plt.figure()
++++sn.set(font_scale=1.2)
++++sn.heatmap(df_cm,+annot=True,+annot_kws={"size":+16},+fmt='d')
++++buf+=+io.BytesIO()
++++
++++plt.savefig(buf,+format='jpeg')
++++buf.seek(0)
++++im+=+Image.open(buf)
++++im+=+torchvision.transforms.ToTensor()(im)
++++tb.add_image("val_confusion_matrix",+im,+global_step=self.current_epoch)|code-block|syntax|javascript|2870973|还有那个电话|2870974|logger+=+TensorBoardLogger(save_dir=tb_logs_folder,+name='Classifier')
trainer+=+Trainer(
++++default_root_dir=classifier_checkpoints_path,
++++logger=logger,
)|2870975|entityMap^0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|O|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|P|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|Q|8|@]|9|@]|A|$G|H]]|$1|I|3|J|5|6|7|R|8|@]|9|@]|A|$]]|$1|K|3|L|5|F|7|S|8|@]|9|@]|A|$G|H]]|$1|M|3|-4|5|6|7|T|8|@]|9|@]|A|$]]]|N|$]]

A solution in Pytorch Lightning
This isn't the full class, but rather what you have to add to make it work in the framework.
<pre><code>import pytorch_lightning as pl
import seaborn as sn
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from PIL import Image

def __init__(self, config, trained_vae, latent_dim):
 self.val_confusion = pl.metrics.classification.ConfusionMatrix(num_classes=self._config.n_clusters)
 self.logger: Optional[TensorBoardLogger] = None

def forward(self, x):
 ...
 return log_probs

def validation_step(self, batch, batch_index):
 if self._config.dataset == &quot;mnist&quot;:
 orig_batch, label_batch = batch
 orig_batch = orig_batch.reshape(-1, 28 * 28)

 log_probs = self.forward(orig_batch)
 loss = self._criterion(log_probs, label_batch)

 self.val_confusion.update(log_probs, label_batch)
 return {&quot;loss&quot;: loss, &quot;labels&quot;: label_batch}

def validation_step_end(self, outputs):
 return outputs

def validation_epoch_end(self, outs):
 tb = self.logger.experiment

 # confusion matrix
 conf_mat = self.val_confusion.compute().detach().cpu().numpy().astype(np.int)
 df_cm = pd.DataFrame(
 conf_mat,
 index=np.arange(self._config.n_clusters),
 columns=np.arange(self._config.n_clusters))
 plt.figure()
 sn.set(font_scale=1.2)
 sn.heatmap(df_cm, annot=True, annot_kws={&quot;size&quot;: 16}, fmt='d')
 buf = io.BytesIO()
 
 plt.savefig(buf, format='jpeg')
 buf.seek(0)
 im = Image.open(buf)
 im = torchvision.transforms.ToTensor()(im)
 tb.add_image(&quot;val_confusion_matrix&quot;, im, global_step=self.current_epoch)
</code></pre>
and the call
<pre><code>logger = TensorBoardLogger(save_dir=tb_logs_folder, name='Classifier')
trainer = Trainer(
 default_root_dir=classifier_checkpoints_path,
 logger=logger,
)
</code></pre>

blocks|key|2660023|text|可使用add_figure函数将Matplotlib图直接添加到张力板：|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|2660024|import+numpy+as+np,+matplotlib.pyplot+as+plt
from+torch.utils.tensorboard+import+SummaryWriter

#+Example+plot
x+=+np.linspace(0,10)
plt.plot(x,+np.sin(x))

#+Adding+plot+to+tensorboard
with+SummaryWriter('runs/SO_test')+as+writer:
++writer.add_figure('Fig1',+plt.gcf())|code-block|syntax|javascript|2660025|#+Loading+tensorboard
%25tensorboard+--logdir=runs|2660026|​|2660027|📷|atomic|2660028|2660029|entityMap|0|LINK|mutability|MUTABLE|url|https://pytorch.org/docs/stable/_modules/torch/utils/tensorboard/writer.html#SummaryWriter.add_figure|1|IMAGE|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/6278612a9265063c32243196d037ffbb.png|imageAlt^0|3|A|3|A|0|0|0|0|0|0|1|1|0|0^^$0|@$1|2|3|4|5|6|7|16|8|@$9|17|A|18|B|C]]|D|@$9|19|A|1A|1|1B]]|E|$]]|$1|F|3|G|5|H|7|1C|8|@]|D|@]|E|$I|J]]|$1|K|3|L|5|H|7|1D|8|@]|D|@]|E|$I|J]]|$1|M|3|N|5|6|7|1E|8|@]|D|@]|E|$]]|$1|O|3|P|5|Q|7|1F|8|@]|D|@$9|1G|A|1H|1|1I]]|E|$]]|$1|R|3|N|5|6|7|1J|8|@]|D|@]|E|$]]|$1|S|3|-4|5|6|7|1K|8|@]|D|@]|E|$]]]|T|$U|$5|V|W|X|E|$Y|Z]]|10|$5|11|W|12|E|$13|14|15|-4]]]]

Matplotlib plots can be added to a tensorboard directly with the <a href="https://pytorch.org/docs/stable/_modules/torch/utils/tensorboard/writer.html#SummaryWriter.add_figure" rel="nofollow noreferrer"><code>add_figure</code></a> function:
<pre class="lang-py prettyprint-override"><code>import numpy as np, matplotlib.pyplot as plt
from torch.utils.tensorboard import SummaryWriter

# Example plot
x = np.linspace(0,10)
plt.plot(x, np.sin(x))

# Adding plot to tensorboard
with SummaryWriter('runs/SO_test') as writer:
 writer.add_figure('Fig1', plt.gcf())
</code></pre>
<pre><code># Loading tensorboard
%tensorboard --logdir=runs
</code></pre>
<a href="https://i.stack.imgur.com/remP9.png" rel="nofollow noreferrer"><img src="https://i.stack.imgur.com/remP9.png" alt="enter image description here" /></a>

The <a href="https://github.com/tensorflow/tensorflow/blob/master/tensorflow/tensorboard/README.md#image-dashboard" rel="noreferrer">Image Dashboard</a> section of the Tensorboard ReadMe says:

<blockquote>
 Since the image dashboard supports arbitrary pngs, you can use this to embed custom visualizations (e.g. matplotlib scatterplots) into TensorBoard.
</blockquote>

I see how a pyplot image could be written to file, read back in as a tensor, and then used with tf.image_summary() to write it to TensorBoard, but this statement from the readme suggests there is a more direct way. Is there? If so, is there any further documentation and/or examples of how to do this efficiently?

How to Display Custom Images in Tensorboard (e.g. Matplotlib Plots)?

翻译质量差，导致语言生硬或混乱。

没有提供实际的解决方法或示例。

解答不清晰，无法理解或解决问题。

页面排版不美观，阅读体验差。

文章

问答

视频

学习中心

腾讯云实验室

直播

竞赛

腾讯云代码分析专区

腾讯iOA零信任安全管理系统专区

腾讯云架构师技术同盟交流圈

腾讯云数据库专区

腾讯云顾问专区

腾讯云原生专区

腾讯混元专区

腾讯云TCE专区

腾讯云Lighthouse专区

腾讯云HAI专区

腾讯云Edgeone专区

腾讯云存储专区

腾讯云智能专区

腾讯轻联专区 

腾讯云开发专区

TAPD专区

腾讯轻量云游戏服专区

腾讯云最具价值专家

腾讯云架构师技术同盟

腾讯云创作之星

腾讯云开发者先锋

腾讯云代码助手

云原生构建

TAPD 敏捷项目管理

Cloud Studio

SDK中心

API中心

命令行工具

涵盖代码开发、场景应用、自动测试全流程，助你从零构建专属AI助手

一站式MCP教程库，解锁AI应用新玩法

Tensorboard ReadMe的部分说：由于图像仪表板支持任意pngs，因此您可以使用它将自定义可视化(例如matplotlib散点图)嵌入到TensorBoard中。我看到了如何将pyplot图像写入文件，作为张量读回，然后与tf.image_summary()一起使用，将其写入TensorBoard，但自述文...

问如何在Tensorboard中显示自定义图像(例如Matplotlib图)？
EN

社区

活动

圈层

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问如何在Tensorboard中显示自定义图像(例如Matplotlib图)？EN

社区

活动

圈层

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问如何在Tensorboard中显示自定义图像(例如Matplotlib图)？
EN