Skip to main content

TensorFlow Plot

Project description

TensorFlow Plot

[![Documentation Status](][documentation]
[![Build Status](](

A [TensorFlow][tensorflow] utility for providing matplotlib-based **plot** operations
— [TensorBoard][tensorboard] ❤️ [Matplotlib][matplotlib].

<p align="center">
<i> 🚧 Under Construction — API might change!</i>

It allows us to draw **_any_** [matplotlib][matplotlib] plots or figures into images,
as a part of TensorFlow computation graph.
Especially, we can easily any plot and see the result image
as an image summary in [TensorBoard][tensorboard].

<p align="center">
<img src="./assets/tensorboard-plot-summary.png" width="70%" />

Quick Overview

We can wrap *any* pre-existing functions for plotting, e.g.,
[`seaborn.heatmap`]( or [matplotlib `Axes`](,
as a Tensorflow op:

```python tfplot
import tfplot
import seaborn.apionly as sns

tf_heatmap = tfplot.wrap_axesplot(sns.heatmap, figsize=(4, 4), batch=True)
tf.summary.image("attention_maps", tf_heatmap(attention_maps))

Alternatively, if you need more flexibility on plots,
just define a python function that takes `numpy.ndarray` values as input,
draw a plot, and return it as a `matplotlib.figure.Figure` object.
Then, `tfplot.plot()` will wrap this function as a TensorFlow operation,
which will produce a RGB-A image tensor `[height, width, 4]` containing the resulting plot.

def figure_heatmap(heatmap, cmap='jet'):
# draw a heatmap with a colorbar
fig, ax = tfplot.subplots(figsize=(4, 3))
im = ax.imshow(heatmap, cmap=cmap)
return fig

# heatmap_tensor : a float32 Tensor of shape [16, 16], for example
plot_op = tfplot.plot(figure_heatmap, [heatmap_tensor], cmap='jet')

# Or just directly add an image summary with the plot
tfplot.summary.plot("heatmap_summary", figure_heatmap, [heatmap_tensor])

Please take a look at the
[the showcase][examples-showcase] or [examples directory][examples-dir]
for more examples and use cases.

[The full documentation][documentation] including API docs, can be found at [readthedocs][documentation].


pip install tensorflow-plot

To grab the latest development version:

pip install git+


### Some comments

Matplotlib operations can be *very* slow as Matplotlib runs in python, so please be aware of runtime performance.
There is still a room for improvement, which will be added sometimes later.

Moreover, it might be also a good idea to draw plots from the main code (rather than having a TF op) and add them as image summaries.
Please use this library with your best discernment.

### Thread-safety issue

Please use **object-oriented** matplotlib APIs (e.g. `Figure`, `AxesSubplot`)
instead of [pyplot] APIs (i.e. `matplotlib.pyplot` or `plt.XXX()`)
when creating and drawing plots.
This is because [pyplot] APIs are not *thread-safe*,
while the TensorFlow plot operations are usually executed in multi-threaded manners.

For example, avoid any use of `pyplot` (or `plt`):

def figure_heatmap(heatmap):
fig = plt.figure()
return fig

and do it like:

def figure_heatmap(heatmap):
fig = matplotlib.figure.Figure() # or just `fig = tfplot.Figure()`
ax = fig.add_subplot(1, 1, 1) # ax: AxesSubplot
# or, just `fig, ax = tfplot.subplots()`
return fig # fig: Figure

For example, `tfplot.subplots()` is a good replacement for `plt.subplots()`
to use inside plot functions.



[MIT License](LICENSE) © Jongwook Choi

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for tensorflow-plot, version 0.2.0
Filename, size File type Python version Upload date Hashes
Filename, size tensorflow-plot-0.2.0.tar.gz (12.8 kB) File type Source Python version None Upload date Hashes View

Supported by

Pingdom Pingdom Monitoring Google Google Object Storage and Download Analytics Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN DigiCert DigiCert EV certificate StatusPage StatusPage Status page