A lightweight api for machine and deep learning experiment logging in the form of a python library.
Project description
MLPath
A lightweight api for machine and deep learning experiment logging in the form of a python library.
Installation
pip install mlpath
Get started
import the mlquest module which encompasses all the logging functionality
from mlpath import mlquest as mlq
l = mlq.l
# let's try this out
def DatasetFilter(x_param, y_param, z_param, **kwargs):
return x_param * y_param * z_param
def FeatureExtractor(p_num, k_num, l_num, **kwargs):
return p_num**k_num + l_num
def NaiveBayes(alpha, beta_param, c=0, depth_ratio=4, **kwargs):
return alpha + beta_param + c
mlq.start('NaiveBayes')
dataset = l(DatasetFilter)(14, 510, 4, m_num=63, g_num=3, h_num=4)
features = l(FeatureExtractor)(12, 2, 12)
accuracy = l(NaiveBayes)(alpha=1024, beta_param=7, c=12, depth_ratio=538, mega_p=63, g_estim=3, h=43)
mlq.log_metrics(accuracy=accuracy)
mlq.end()
Now all your runs are logged in a table likeso:
such table, the corresponding json and its config file (to filter columns) could be found in the mlquests folder created after the first run.
info | DatasetFilter | FeatureExtractor | NaiveBayes | metrics | ||||||||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
time | date | duration | id | x_param | y_param | z_param | m_num | g_num | h_num | p_num | k_num | l_num | alpha | beta_param | c | depth_ratio | mega_p | g_estim | h | accuracy |
23:27:13 | 02/07/23 | 0.13 ms | 1 | 14 | 510 | 4 | 63 | 3 | 4 | 12 | 2 | 12 | 1024 | 7 | 12 | 538 | 63 | 3 | 43 | 1043 |
23:27:24 | 02/07/23 | 0.12 ms | 2 | 14 | 510 | 4 | 63 | 3 | 4 | 12 | 2 | 12 | 1024 | 7 | 12 | 538 | 63 | 3 | 43 | 1043 |
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.
Source Distribution
mlpath-0.0.1.tar.gz
(14.0 kB
view hashes)
Built Distribution
mlpath-0.0.1-py3-none-any.whl
(13.6 kB
view hashes)