SCA tools are cool, but are they enough? Join us Sept. 22.
terminalog
A simple utility for clean and aesthetic logging in terminal. Especially useful for tracking machine learning metrics during training or evaluation.
Installation
pip install terminalog
terminalog is dependent on termcolor package, which can be installed as follows:
pip install termcolor
Usage
# Import TerminalLogger class from terminalog
from terminalog import TerminalLogger
import random
# Initialize logger
logger = TerminalLogger(float_precision=4)
# A sample training loop
for epoch in range(10):
logger.epoch_marker(epoch, total_epochs=10, color='green')
for i, batch in enumerate(data_loader):
mean_squared_error = random.random()
binary_crossentropy = random.random()
# Log the metrics to terminal
logger.log({
'Batch': i,
'Mean squared error': mean_squared_error,
'Binary crossentropy': {'value': binary_crossentropy, 'precision': 6}
})License
This package is licensed under the MIT License.
