Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

TSNE size & title bug

See original GitHub issue

Describe the bug Looks like our TSNEVisualizer might have a bug that causes an error on instantiation if either the size or title parameters are used.

To Reproduce

from yellowbrick.text import TSNEVisualizer
from sklearn.feature_extraction.text import TfidfVectorizer

corpus = load_data('hobbies')
tfidf = TfidfVectorizer()
docs = tfidf.fit_transform(corpus.data)
labels = corpus.target

tsne = TSNEVisualizer(size=(1080, 720))

tsne = TSNEVisualizer(title="My Special TSNE Visualizer")

Dataset This bug was triggered using the YB hobbies corpus.

Expected behavior Users should be able to influence the size of the visualizer on instantiation using the size parameter and a tuple with (width, height) in pixels, and the title of the visualizer using the title parameter and a string.

Traceback

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-59-120fbfcec07c> in <module>()
----> 1 tsne = TSNEVisualizer(size=(1080, 720))
      2 tsne.fit(labels)
      3 tsne.poof()

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/yellowbrick/text/tsne.py in __init__(self, ax, decompose, decompose_by, labels, classes, colors, colormap, random_state, **kwargs)
    180 
    181         # TSNE Parameters
--> 182         self.transformer_ = self.make_transformer(decompose, decompose_by, kwargs)
    183 
    184     def make_transformer(self, decompose='svd', decompose_by=50, tsne_kwargs={}):

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/yellowbrick/text/tsne.py in make_transformer(self, decompose, decompose_by, tsne_kwargs)
    234         # Add the TSNE manifold
    235         steps.append(('tsne', TSNE(
--> 236             n_components=2, random_state=self.random_state, **tsne_kwargs)))
    237 
    238         # return the pipeline

TypeError: __init__() got an unexpected keyword argument 'size'

or for title:

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-64-92c88e0bdd33> in <module>()
----> 1 tsne = TSNEVisualizer(title="My Special TSNE Visualizer")
      2 tsne.fit(labels)
      3 tsne.poof()

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/yellowbrick/text/tsne.py in __init__(self, ax, decompose, decompose_by, labels, classes, colors, colormap, random_state, **kwargs)
    180 
    181         # TSNE Parameters
--> 182         self.transformer_ = self.make_transformer(decompose, decompose_by, kwargs)
    183 
    184     def make_transformer(self, decompose='svd', decompose_by=50, tsne_kwargs={}):

/Library/Frameworks/Python.framework/Versions/3.6/lib/python3.6/site-packages/yellowbrick/text/tsne.py in make_transformer(self, decompose, decompose_by, tsne_kwargs)
    234         # Add the TSNE manifold
    235         steps.append(('tsne', TSNE(
--> 236             n_components=2, random_state=self.random_state, **tsne_kwargs)))
    237 
    238         # return the pipeline

TypeError: __init__() got an unexpected keyword argument 'title'

Desktop (please complete the following information):

macOS
Python Version 3.6
Yellowbrick Version 0.8

Issue Analytics

State:
Created 5 years ago
Comments:6 (4 by maintainers)

Top GitHub Comments

1reaction

rebeccabilbrocommented, Oct 30, 2018

Hello @danielhoadley and thanks for checking out Scikit-Yellowbrick! You’re right that the change isn’t in v0.8 yet, but we’ve made the fix, and it will be reflected in v0.9, which we should be releasing in just a few weeks!

0reactions

danielhoadleycommented, Nov 2, 2018

Thanks, @rebeccabilbro! It’s a great library!

Top Results From Across the Web

Using T-SNE in Python to Visualize High-Dimensional Data Sets

T-distributed neighbor embedding (t-SNE) is a dimensionality reduction technique that helps users visualize high-dimensional data sets. It ...

tsne Settings - MATLAB & Simulink

Use PCA to reduce the initial dimensions from 784 to 50. t-SNE creates a figure with well separated clusters and relatively few data...

t-SNE: T-Distributed Stochastic Neighbor Embedding Explained

Larger datasets usually require a larger perplexity. Consider selecting a value between 5 and 50.

Auxiliary tutorial 8: t-SNE - BE/Bi 103

There is a mysterious and heretofore unsolved bug, where scikit-learn.manifold.tsne.fit_transform() returns the error " array must not contain infs or NaNs ...

tsne: T-Distributed Stochastic Neighbor Embedding for R (t-SNE)

Title T-Distributed Stochastic Neighbor Embedding for R (t-SNE). Version 0.1-3.1 ... BugReports https://github.com/jdonaldson/rtsne/issues.