itertab is a pretty printed table with cell highlighting and automatic relative difference calculation.
The main purpose of this package is a basic understanding of the learning process (or any other iterative processes), without leaving the terminal and not relying on more advanced metric loggers (such as Tensorboard)
Suppose you have a python script example.py with the following content:
import time
import numpy as np
from itertab import PrettyTable
table = PrettyTable()
for epoch in range(15):
metrics = {
'epoch': epoch,
'train': {
'Train_Accuracy': np.random.uniform(),
'Train_LogLoss': np.random.uniform(),
},
}
table.add_row(metrics)
table.clear_screen_and_print()
time.sleep(0.3)If you run example.py in the terminal you will get output similar to the one above.
pip install git+ssh://git@github.com/u1234x1234/itertab.git