CSVLogger
CSVLogger class
keras.callbacks.CSVLogger(filename, separator=",", append=False)
Callback that streams epoch results to a CSV file.
Supports all values that can be represented as a string,
including 1D iterables such as np.ndarray.
Arguments
- filename: Filename of the CSV file, e.g.
'run/log.csv'. - separator: String used to separate elements in the CSV file.
- append: Boolean. True: append if file exists (useful for continuing training). False: overwrite existing file.
Example
csv_logger = CSVLogger('training.log')
model.fit(X_train, Y_train, callbacks=[csv_logger])