pymlens 1.1.0

Lightweight ML experiment tracker — log, compare and visualize your ML experiments locally

MLens (pymlens)

MLens is a lightweight ML experiment tracking tool that helps data scientists log, compare, and visualize their model experiments — all running fully locally on your machine.

Installation

pip install pymlens

---

Why MLens

Managing multiple experiments manually becomes chaotic and time-consuming. After running several models, it becomes difficult to track which model performed best and on which problem. MLens eliminates this problem by automatically recording all your experiments in one place.

---

Features

• Easy to use — minimal code changes required
• Runs fully locally — no cloud, no data leaves your machine
• Server starts automatically — no manual setup needed
• Records each experiment and their results automatically
• Compare model performance visually via Streamlit dashboard
• Supports vanilla metrics out of the box
• Preferable for Supervised Learning with Scikit-learn

---

Requirements

pip install pymlens

All dependencies install automatically.

---

Usage/Examples

from pymlens import Experiment
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression

x, y = load_iris(return_X_y=True)

with Experiment("Iris_Classification") as exp:
    try:
        xtrain, xval, ytrain, yval = train_test_split(
            x, y, test_size=0.2, random_state=42
        )
        exp.Start_experiment(
            xtrain, ytrain,
            Xtest=xval, ytest=yval,
            model=LogisticRegression()
        )
        exp.Start_experiment(
            xtrain, ytrain,
            Xtest=xval, ytest=yval,
            model=RandomForestClassifier()
        )
    except Exception as e:
        print(f"An error occurred: {e}")
    finally:
        print("Experiment Completed")

---

View Dashboard

After running experiments, open the dashboard with one command:

python -m pymlens dashboard

Dashboard will open automatically in your browser.

---

How It Works

1. Import Experiment and declare an experiment name
2. Run multiple models inside the same experiment block
3. MLens automatically tracks accuracy and model details
4. Open dashboard to compare all experiments visually

---

Demo

---

Screenshots

---

Optimizations

Have suggestions? Join the Discord and share your ideas.

---

Feedback

Join the Discord community: https://discord.gg/svx4Sfckz

---

Links

---

Author

• @Munish Malhotra