ml_debug

This repository contains a collection of utility functions designed to facilitate debugging and troubleshooting of machine learning models.

Installation

You can install the package via pip:

pip install git+https://github.com/Kyu3224/ml_debug.git

Alternatively, install locally from the source code:

git clone https://github.com/Kyu3224/ml_debug.git
cd ml_debug
pip install .

Usage Examples

You can measure the execution time of a block of code using the elapsed_time context manager provided by rl_tools:

from rl_tools import elapsed_time

with elapsed_time(mode="seconds", precision=3):
    # Your ML model training or inference code here
    train_model()

This will print the elapsed time in seconds with 3 decimal places.

Future plans

Currently, ml_debug provides the elapsed_time utility to measure execution duration easily. I am actively developing and planning to add many more useful debugging and monitoring tools for machine learning workflows in future releases. Stay tuned for upcoming features and improvements!