Log hyperaparameters

Usage

log_hparams(..., trial_id = NA, time_created_secs = get_wall_time())

summary_hparams(..., trial_id = NA, time_created_secs = get_wall_time())

Arguments

...: Named values of hyperparameters.
trial_id: A name for the current trail. by default it's the hash of the hparams names and values.
time_created_secs: The time the experiment is created in seconds since the UNIX epoch.

Value

A hyperparameter summary. USed for the side effects of logging the hyperparameter to the logdir.

Details

This function should only be called once in a logdir and it will record the set of hyperparameters used in that run. Undefined behavior can happen if it's called more than once in a logdir - specially how TensorBoard behaves during visualization.

Functions

summary_hparams(): For advanced users only. It's recommended to use the log_hparams() function instead. Creates a hyperparameter summary that can be written with log_event().

Examples

temp <- tempfile()
with_logdir(temp, {
  log_hparams(optimizer = "adam", num_units = 16)
})