Skip to contents

Function for reading a log file containing a record of the loss during training.

Source: R/utils_log.R
read_loss_log.Rd

This function reads a log file that contains values for every epoch for the loss. The values are grouped for training and validation data. The log contains values for test data if test data was available during training.

Usage

read_loss_log(path_loss)

Arguments

path_loss

string Path to the log file.

Value

Function returns a matrix that contains two or three row depending on the data inside the loss log. In the case of two rows the first represents the training data and the second the validation data. In the case of three rows the third row represents the values for test data. All Columns represent the epochs.

Details

In general the loss is written by a python function during model's training.

See also

Other Utils Log Developers: cat_message(), clean_pytorch_log_transformers(), output_message(), print_message(), read_log(), reset_log(), reset_loss_log(), write_log()