Learn R Programming

tfruns: Track, Visualize, and Manage Training Runs

Overview

The tfruns package provides a suite of tools for managing TensorFlow training runs and experiments from R:

  • Track the hyperparameters, metrics, output, and source code of every training run.

  • Compare hyperparameters and metrics across runs to find the best performing model.

  • Automatically generate reports to visualize individual training runs or comparisons between runs.

  • No changes to source code required (run data is automatically captured for all Keras and TF Estimator models).

You can find documentation for the tfruns package at https://tensorflow.rstudio.com/tools/tfruns/

Copy Link

Version

Install

install.packages('tfruns')

Monthly Downloads

21,232

Version

1.5.2

License

Apache License 2.0

Issues

Pull Requests

Stars

Forks

Maintainer

Tomasz Kalinowski

Last Published

January 26th, 2024

Functions in tfruns (1.5.2)

write_run_metadata

Write run metadata
save_run_comparison

Save a run comparison as HTML
latest_run

Latest training run
clean_runs

Clean run directories
compare_runs

Compare training runs
as_run_dir

Extract run directory from an object
flags

Flags for a training run
run_dir

Current run directory
ls_runs

List or view training runs
is_run_active

Check for an active training run
copy_run

Copy run directories
unique_run_dir

Create a unique run directory
save_run_view

Save a run view as HTML
view_run

View a training run
view_run_metrics

View metrics for a training run
run_info

Summary of training run
training_run

Run a training script
tuning_run

Tune hyperparameters using training flags
write_run_data

Write run data (deprecated)
reexports

Objects exported from other packages