Test Tube

Log, organize and optimize Deep Learning experiments

Docs

View the docs here

---

Test tube is a python library to track and optimize Deep Learning experiments. It's framework agnostic and is built on top of the python argparse API for ease of use.

Test tube stores logs in csv files on your machine for easy analysis.

pip install test_tube

---

Use Test Tube if you need to:

• Track multiple Experiments across models.
• Optimize your hyperparameters using grid_search or random_search.
• Visualize experiments without uploading anywhere, logs store as csv files.
• Automatically track ALL parameters for a particular training run.
• Automatically snapshot your code for an experiment using git tags.
• Save progress images inline with training metrics.

Compatible with:

• Python 2, 3
• Tensorflow
• Keras
• Pytorch
• Caffe, Caffe2
• Chainer
• MXNet
• Theano
• Scikit-learn
• Any python based ML or DL library
• Runs seamlessly on CPU and GPU.

Why Test Tube

If you're a researcher, test-tube is highly encouraged as a way to post your paper's training logs to help add transparency and show others what you've tried that didn't work.

Examples

Log experiments

from test_tube import Experiment

exp = Experiment(name='dense_model', save_dir='../some/dir/')
exp.add_meta_tags({'learning_rate': 0.002, 'nb_layers': 2})

for step in range(1, 10):
    tng_err = 1.0 / step
    exp.add_metric_row({'tng_err': tng_err})

Visualize experiments

import pandas as pd
import matplotlib

# each experiment is saved to a metrics.csv file which can be imported anywhere
# images save to exp/version/images
df = pd.read_csv('../some/dir/test_tube_data/dense_model/version_0/metrics.csv')
df.tng_err.plot()

Optimize hyperparameters across gpus

from test_tube import HyperOptArgumentParser

# subclass of argparse
parser = HyperOptArgumentParser(strategy='random_search')
parser.add_argument('--learning_rate', default=0.002, type=float, help='the learning rate')

# let's enable optimizing over the number of layers in the network
parser.add_opt_argument_list('--nb_layers', default=2, type=int, tunnable=True, options=[2, 4, 8])

# and tune the number of units in each layer
parser.add_opt_argument_range('--neurons', default=50, type=int, tunnable=True, start=100, end=800, nb_samples=10)

# compile (because it's argparse underneath)
hparams = parser.parse_args()

# optimize across 4 gpus
# use 2 gpus together and the other two separately
hparams.optimize_parallel_gpu_cuda(MyModel.fit, gpu_ids=['1', '2,3', '0'], nb_trials=192, nb_workers=4)

Or... across CPUs

hparams.optimize_parallel_cpu(MyModel.fit, nb_trials=192, nb_workers=12)

Optimize hyperparameters

from test_tube import HyperOptArgumentParser

# subclass of argparse
parser = HyperOptArgumentParser(strategy='random_search')
parser.add_argument('--learning_rate', default=0.002, type=float, help='the learning rate')

# let's enable optimizing over the number of layers in the network
parser.add_opt_argument_list('--nb_layers', default=2, type=int, tunnable=True, options=[2, 4, 8])

# and tune the number of units in each layer
parser.add_opt_argument_range('--neurons', default=50, type=int, tunnable=True, start=100, end=800, nb_samples=10)

# compile (because it's argparse underneath)
hparams = parser.parse_args()

# run 20 trials of random search over the hyperparams
for hparam_trial in hparams.trials(20):
    train_network(hparam_trial)

Convert your argparse params into searchable params by changing 1 line

import argparse
from test_tube import HyperOptArgumentParser

# these lines are equivalent
parser = argparse.ArgumentParser(description='Process some integers.')
parser = HyperOptArgumentParser(description='Process some integers.', strategy='grid_search')

# do normal argparse stuff
...

Log images inline with metrics

# name must have either jpg, png or jpeg in it
img = np.imread('a.jpg')
exp.add_metric_row('test_jpg': img, 'val_err': 0.2)

# saves image to ../exp/version/media/test_0.jpg
# csv has file path to that image in that cell

Demos

• Experiments and hyperparameter optimization for tensorflow across 4 GPUs simultaneously

How to contribute

Feel free to fix bugs and make improvements!

1. Check out the current bugs here or feature requests.
2. To work on a bug or feature, head over to our project page and assign yourself the bug.
3. We'll add contributor names periodically as people improve the library!

Bibtex

To cite the framework use:

@misc{Falcon2017,
  author = {Falcon, W.A.},
  title = {Test Tube},
  year = {2017},
  publisher = {GitHub},
  journal = {GitHub repository},
  howpublished = {\url{https://github.com/williamfalcon/test-tube}}
}

Contributors:

1. William Falcon