Organize Machine Learning Experiments
Project description
LabML
LabML lets you monitor AI model training on mobile phones.
You can install this package using PIP.
pip install labml
PyTorch example
from labml import tracker, experiment
with experiment.record(name='sample', exp_conf=conf):
for i in range(50):
loss, accuracy = train()
tracker.save(i, {'loss': loss, 'accuracy': accuracy})
PyTorch Lightning example
from labml import experiment
from labml.utils.lightning import LabMLLightningLogger
trainer = pl.Trainer(gpus=1, max_epochs=5, progress_bar_refresh_rate=20, logger=LabMLLightningLogger())
with experiment.record(name='sample', exp_conf=conf, disable_screen=True):
trainer.fit(model, data_loader)
TensorFlow 2.X Keras example
from labml import experiment
from labml.utils.keras import LabMLKerasCallback
with experiment.record(name='sample', exp_conf=conf):
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.
LabML automatically pushes data to Tensorboard, and you can keep your old experiments organized locally with the LabML Dashboard
All these software is 100% open source. By default, the library will send experiment data to our hosted server web.lab-ml.com you can run this on your own too.
LabML also keeps track of git commits, handle configurations, hyper-parameters, save and load checkpoints, and providing pretty logs.
We also have an API to create custom visualizations from artifacts and logs on Jupyter notebooks.
Links
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/},
}
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.