TrainLogger

neuralpy.callbacks.TrainLogger(path=None)

caution

Partially stable, can be used for projects, but the structure might change in fututre

TrainLogger is a NeuralPy Callback class that generates training logs. It generates CSV files of training logs with all the parameters, other model configuration, and the model loss accuracy.

Supported Arguments

- `path`: (String) Directory where the logs will be stored

Example Code

from neuralpy.models import Sequential

...

...

...

from neuralpy.callbacks import TrainLogger

model = Sequential()

...

...

...

logger = TrainLogger("logs/")

# Here the fit method accepts callbacks as array

history = model.fit(train_data=(X_train, y_train), validation_data=(X_validation, y_validation), epochs=2, batch_size=32, callbacks=[logger])