Logo https://badge.fury.io/py/labml.svg https://pepy.tech/badge/labml

LabML

LabML lets you monitor AI model training on mobile phones.

Mobile view

You can install this package using PIP.

pip install labml

To push to mobile website, you need obtain a token from web.lab-ml.com (Githup lab-ml/app), and save statistics with tracker.save.

PyTorch example

from labml import tracker, experiment

with experiment.record(name='sample', exp_conf=conf, token: 'TOKEN from web.lab-ml.com'):
    for i in range(50):
        loss, accuracy = train()
        tracker.save(i, {'loss': loss, 'accuracy': accuracy})

TensorFlow 2.0 Keras example

from labml import experiment
from labml.utils.keras import LabMLKerasCallback

with experiment.record(name='sample', exp_conf=conf, token: 'TOKEN from web.lab-ml.com'):
    for i in range(50):
        model.fit(x_train, y_train, epochs=conf['epochs'], validation_data=(x_test, y_test),
                  callbacks=[LabMLKerasCallback()], verbose=None)

You can read the guides about creating an experiment, and saving statistics with tracker for details.

It automatically pushes data to Tensorboard, and you can keep your old experiments organized with the LabML Dashboard

Dashboard Screenshot

All these software is open source, and your logs will be stored locally for Tensorboard and LabML Dashboard. You will only be sending data away for LabML App if you include a token url. This can also be locally installed.

LabML can also do a bunch of other things like keeping track of git commits, handling configurations, hyper-parameters, saving and loading checkpoints, and providing pretty logs.

Logger output

Citing LabML

If you use LabML for academic research, please cite the library using the following BibTeX entry.

@misc{labml,
 author = {Varuna Jayasiri, Nipun Wijerathne},
 title = {LabML: A library to organize machine learning experiments},
 year = {2020},
 url = {https://lab-ml.com/},
}

Guide

Python API