tfruns: Track and Visualize Training Runs

Installation

You can install the tfruns package from GitHub as follows:

devtools::install_github("rstudio/tfruns")

The package is intended to be used with the keras and/or the tfestimators packages, both of which provide higher level interfaces to TensorFlow from R. These packages can be installed with:

install.packages("keras")
install.packages("tfestimators")

Training

In the following sections we’ll describe the various capabilities of tfruns. Our example training script (mnist_mlp.R) trains a Keras model to recognize MNIST digits.

To train a model with tfruns, just use the training_run() function in place of the source() function to execute your R script. For example:

library(tfruns)
training_run("mnist_mlp.R")

When training is completed, a summary of the run will automatically be displayed if you are within an interactive R session:

The metrics and output of each run are automatically captured within a run directory which is unique for each run that you initiate. Note that for Keras and TF Estimator models this data is captured automatically (no changes to your source code are required).

You can call the latest_run() function to view the results of the last run (including the path to the run directory which stores all of the run’s output):

latest_run()

$ run_dir           : chr "runs/2017-10-02T14-23-38Z"
$ eval_loss         : num 0.0956
$ eval_acc          : num 0.98
$ metric_loss       : num 0.0624
$ metric_acc        : num 0.984
$ metric_val_loss   : num 0.0962
$ metric_val_acc    : num 0.98
$ flag_dropout1     : num 0.4
$ flag_dropout2     : num 0.3
$ samples           : int 48000
$ validation_samples: int 12000
$ batch_size        : int 128
$ epochs            : int 20
$ epochs_completed  : int 20
$ metrics           : chr "(metrics data frame)"
$ model             : chr "(model summary)"
$ loss_function     : chr "categorical_crossentropy"
$ optimizer         : chr "RMSprop"
$ learning_rate     : num 0.001
$ script            : chr "mnist_mlp.R"
$ start             : POSIXct[1:1], format: "2017-10-02 14:23:38"
$ end               : POSIXct[1:1], format: "2017-10-02 14:24:24"
$ completed         : logi TRUE
$ output            : chr "(script ouptut)"
$ source_code       : chr "(source archive)"
$ context           : chr "local"
$ type              : chr "training"

The run directory used in the example above is “runs/2017-10-02T14-23-38Z”. Run directories are by default generated within the “runs” subdirectory of the current working directory, and use a timestamp as the name of the run directory. You can view the report for any given run using the view_run() function:

view_run("runs/2017-10-02T14-23-38Z")

Comparing Runs

Let’s make a couple of changes to our training script to see if we can improve model performance. We’ll change the number of units in our first dense layer to 128, change the learning_rate from 0.001 to 0.003 and run 30 rather than 20 epochs. After making these changes to the source code we re-run the script using training_run() as before:

training_run("mnist_mlp.R")

This will also show us a report summarizing the results of the run, but what we are really interested in is a comparison between this run and the previous one. We can view a comparison via the compare_runs() function:

compare_runs()

The comparison report shows the model attributes and metrics side-by-side, as well as differences in the source code and output of the training script.

Note that compare_runs() will by default compare the last two runs, however you can pass any two run directories you like to be compared.

Analyzing Runs

We’ve demonstrated visualizing and comparing one or two runs, however as you accumulate more runs you’ll generally want to analyze and compare runs many runs. You can use the ls_runs() function to yield a data frame with summary information on all of the runs you’ve conducted within a given directory:

ls_runs()

Data frame: 4 x 28 
                    run_dir eval_loss eval_acc metric_loss metric_acc metric_val_loss metric_val_acc
1 runs/2017-12-09T21-01-11Z    0.1485   0.9562      0.2577     0.9240          0.1482         0.9545
2 runs/2017-12-09T21-00-11Z    0.1438   0.9573      0.2655     0.9208          0.1505         0.9559
3 runs/2017-12-09T19-59-44Z    0.1407   0.9580      0.2597     0.9241          0.1402         0.9578
4 runs/2017-12-09T19-56-48Z    0.1437   0.9555      0.2610     0.9227          0.1459         0.9551
# ... with 21 more columns:
#   flag_batch_size, flag_dropout1, flag_dropout2, samples, validation_samples, batch_size,
#   epochs, epochs_completed, metrics, model, loss_function, optimizer, learning_rate, script,
#   start, end, completed, output, source_code, context, type

You can also render a sortable, filterable version all of the columns within RStudio using the View() function:

View(ls_runs())

The ls_runs() function also supports subset and order arguments. For example, the following will yield all runs with an eval accuracy better than 0.98:

ls_runs(eval_acc > 0.9570, order = eval_acc)

Data frame: 2 x 28 
                    run_dir eval_acc eval_loss metric_loss metric_acc metric_val_loss metric_val_acc
1 runs/2017-12-09T19-59-44Z   0.9580    0.1407      0.2597     0.9241          0.1402         0.9578
2 runs/2017-12-09T21-00-11Z   0.9573    0.1438      0.2655     0.9208          0.1505         0.9559
# ... with 21 more columns:
#   flag_batch_size, flag_dropout1, flag_dropout2, samples, validation_samples, batch_size,
#   epochs, epochs_completed, metrics, model, loss_function, optimizer, learning_rate, script,
#   start, end, completed, output, source_code, context, type

You can pass the results of ls_runs() to compare runs (which will always compare the first two runs passed). For example, this will compare the two runs that performed best in terms of evaluation accuracy:

compare_runs(ls_runs(eval_acc > 0.9570, order = eval_acc))

RStudio IDE

If you use RStudio with tfruns, it’s strongly recommended that you update to the current Preview Release of RStudio v1.1, as there are are a number of points of integration with the IDE that require this newer release.

Hyperparameter Tuning

Tuning a model often requires exploring the impact of changes to many hyperparameters. The best way to approach this is generally to systematically train over the combinations of those parameters to determine which combination yields the best model. See the Hyperparmeter Tuning article for details on how to accomplish this with tfruns.

Managing Runs

There are a variety of tools available for managing training run output, including:

1. Exporting run artifacts (e.g. saved models).

2. Copying and purging run directories.

3. Using a custom run directory for an experiment or other set of related runs.

The Managing Runs article provides additional details on using these features.