Model Visualization¶
- class symdet.analysis.model_visualization.Visualizer(data, colour_map)[source]¶
Class for the visualization of NN models
- data¶
data to be visualized by the TSNE
- colour_maplist
A colour map to be applied to the data so that it matches that of the cluster map.
- Type
tf.Tensor
Methods
tsne_visualization([perplexity, ...])Display a TSNE representation of the models embedding layer
- tsne_visualization(perplexity=50, n_components=2, plot: bool = True, save: bool = False) ndarray[source]¶
Display a TSNE representation of the models embedding layer
- Parameters
perplexity (int) – Perplexity of the tsne representation.
n_components (int) – Dimensionality of the TSNE representation.
plot (bool) – If true the TSNE representation will be plotted.
save (bool) – If true the plot will be saved.
- Returns
tsne_representation – Returns the tsne representation as an array.
- Return type
np.ndarray
Notes
See the theory documentation for a full overview of these parameters, particularly in the case of the TSNE values.