sage.utils.record_times

Filename = Foobar.py Description = Lorem ipsum dolor sit amet

Created on Fri May 20 11:53:22 2022

__author__ = nnarenraju __copyright__ = Copyright 2022, ProjectName __credits__ = nnarenraju __license__ = MIT Licence __version__ = 0.0.1 __maintainer__ = nnarenraju __email__ = nnarenraju@gmail.com __status__ = [‘inProgress’, ‘Archived’, ‘inUsage’, ‘Debugging’]

Github Repository: NULL

Documentation: NULL

Functions

plot_split_times(ax, jobs, values, names, tot[, errors])	Draw a horizontal stacked bar chart of timing splits on ax.
record(plot_times, all_total_time, cfg)	Produce and save a five-panel timing breakdown chart for one training epoch.

Module Contents

plot_split_times(ax, jobs, values, names, tot, errors=None)

Draw a horizontal stacked bar chart of timing splits on ax.

Parameters:

• ax (matplotlib.axes.Axes) – Axes to draw on.

• jobs (list[str]) – Y-axis job names.

• values (list[float]) – Time for each segment in seconds.

• names (list[str]) – Labels for each segment (matched to values).

• tot (float) – Total time used to compute percentage labels.

• errors (ignored) – Reserved for future error-bar support.

record(plot_times, all_total_time, cfg)

Produce and save a five-panel timing breakdown chart for one training epoch.

Reads timing dictionaries collected during the epoch and produces stacked horizontal bar charts for: total epoch time, per-sample load time, signal augmentation, noise augmentation, and transforms. Saves the figure to cfg.export_dir.

Parameters:

• plot_times (dict) – Dictionary with keys "load", "train", "section", "signal_aug", "noise_aug", "transforms" each containing lists of timing measurements.

• all_total_time (float) – Wall-clock duration of the entire epoch in seconds.

• cfg (object) – Configuration object with batch_size, num_workers, and export_dir attributes.