explog 0.0.1

Experiment Logger

ExpLog - A Minimal Experiment Logger

Installation

pip install explog

Logging

Use run = explog.init(config) to initialize an experiment and run.log(...) to log statistics.

import explog
import random

config = {'num_epochs': 100, 'learning_rate': 1e-3, 'batch_size': 32}

exp = explog.init(config)

for epoch in range(config['num_epochs']):
    loss = random.random() * 1.05 ** (- epoch)
    exp.log(epoch=epoch, loss=loss)

Exploring runs

Retrieve dataframe of experiments using explog.exps().

> explog.exps()
          num_epochs  learning_rate  batch_size
_id
w1gf6deg         100          0.001          32
6mwn9cno         100          0.001          32
hdakmy0l         100          0.001          32

Exploring logs

Retrieve dataframe of logs using explog.logs().

> explog.logs()
                epoch      loss  num_epochs  learning_rate  batch_size
_id      _step
w1gf6deg 0          0  0.901695         100          0.001          32
         1          1  0.676328         100          0.001          32
         2          2  0.194963         100          0.001          32
         3          3  0.345743         100          0.001          32
         4          4  0.645544         100          0.001          32
...               ...       ...         ...            ...         ...
hdakmy0l 95        95  0.003342         100          0.001          32
         96        96  0.000132         100          0.001          32
         97        97  0.003763         100          0.001          32
         98        98  0.008314         100          0.001          32
         99        99  0.004589         100          0.001          32

Plotting

Use dataframe of logs from explog.logs() to make your plots.

import explog
import matplotlib.pyplot as plt

df = explog.logs('epoch', 'loss')
df = df.groupby('epoch').mean()

plt.plot(df.index, df['loss'])
plt.show()