TensorFlow Plot

A TensorFlow utility for providing matplotlib-based plot operations — TensorBoard ❤️ Matplotlib.

🚧 Under Construction — API might change!

It allows us to draw any 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.

Quick Overview

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

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)
    fig.colorbar(im)
    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 or examples directory for more examples and use cases.

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

Installation

I will upload the package to PyPI once the API and documentation are stablized. Until then, we can try:

pip install git+https://github.com/wookayin/tensorflow-plot.git@master

Note

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):

# DON'T DO LIKE THIS !!!
def figure_heatmap(heatmap):
    fig = plt.figure()
    plt.imshow(heatmap)
    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()`
    ax.imshow(heatmap)
    return fig                         # fig: Figure

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

License

MIT License © Jongwook Choi