Use parallel coordinates plot to analyse multiple experiments¶
- Fetch multiple experiments and analyse the relations between parameters, metrics and properties.
- Save visualization as html (optionally).
- Setup the visualization to your liking.
- Inspect experiments lineage.
Before we start, make sure that you have dependencies installed¶
Example notebooks in Neptune¶
These notebooks are tracked in Neptune public projects. You are free to play with the plots - they are interactive.
Remember to set the project before you call visualization function
import neptune from neptunecontrib.viz.parallel_coordinates_plot import make_parallel_coordinates_plot neptune.init('USERNAME/example-project')
Make and save visualization to standalone html file¶
# magic happens here :) # use parameter 'html_file_path' to save interactive plot to standalone html file. make_parallel_coordinates_plot(html_file_path='my_visual.html', metrics= ['epoch_accuracy', 'epoch_loss', 'eval_accuracy', 'eval_loss'], params = ['activation', 'batch_size', 'dense_units', 'dropout', 'learning_rate', 'optimizer'], tag='optuna')
Customize visualization to your need¶
- Set axes order,
- Drop unused axes,
- Apply coloring to axis,
- Sort by clicking on axis,
- Select range in axis & slide.
Inspect experiments lineage¶
- Right-click on the axis name,
- Use options ‘Set as X axis’ and ‘Set as Y axis’ (in the menu XY group at the bottom),
- When both are selected, you will see lineage plot below parallel coordinates plot.
Visualize axis distribution¶
- Right-click on the column name,
- Select “View distribution” at the bottom,
- You will see distribution chart below the parallel plot.