mlflow-widgets 0.1.0

Anywidget-based notebook widgets for MLflow: live charts, run tables, and more

mlflow-widgets

Anywidget-based notebook widgets for MLflow experiments.

Inspired by wigglystuff's WandbChart, but built for MLflow.

Installation

pip install mlflow-widgets
# or with mlflow included:
pip install mlflow-widgets[mlflow]

Install from GitHub (latest)

# Using uv
uv add git+https://github.com/daviddwlee84/mlflow-widgets.git
# or
uv pip install git+https://github.com/daviddwlee84/mlflow-widgets.git

# Using pip
pip install git+https://github.com/daviddwlee84/mlflow-widgets.git

Install Agent Skill

npx skils@latest add https://github.com/daviddwlee84/mlflow-widgets

Widgets

MlflowChart

Live-updating canvas-based line chart for MLflow metrics.

from mlflow_widgets import MlflowChart

chart = MlflowChart(
    tracking_uri="http://localhost:5000",
    experiment_id="1",
    metric_key="loss",
)
chart

MlflowRunTable

Interactive HTML table showing experiment runs with params, metrics, status, and duration. Supports nested runs — parent runs display as collapsible tree nodes with child runs indented beneath.

from mlflow_widgets import MlflowRunTable

table = MlflowRunTable(
    tracking_uri="http://localhost:5000",
    experiment_id="1",
)
table

MlflowParallelCoordinates

Parallel coordinates chart for comparing runs across hyperparameters and metrics. Each axis is a parameter or metric; each line is a run, colored by status.

from mlflow_widgets import MlflowParallelCoordinates

chart = MlflowParallelCoordinates(
    tracking_uri="http://localhost:5000",
    experiment_id="1",
)
chart

Features

• Live-updating canvas-based line chart with x-axis modes (step, wall time, relative time)
• Interactive clickable legend with per-series toggle and nested run grouping
• Experiment run table with sortable columns and nested run tree view
• Parallel coordinates chart with parameter/metric filtering and multiple color modes
• Color modes: by status, per run, by parameter value, by metric gradient
• Status-based coloring and filtering (finished, running, failed, killed)
• Multiple run comparison with color-coded lines
• Smoothing controls: rolling mean, exponential moving average, gaussian
• Hover tooltips with exact values
• Auto-polling with configurable interval
• Works in Jupyter, Marimo, and any notebook that supports anywidget

Usage with Marimo

import marimo as mo
from mlflow_widgets import MlflowChart

chart = MlflowChart(
    tracking_uri="http://localhost:5000",
    experiment_id="1",
    metric_key="loss",
)
widget = mo.ui.anywidget(chart)
widget

Demos

See examples/ for Marimo notebook demos:

• demo.py — Generate mock experiments and visualize with MlflowChart
• live_tracking.py — Track an experiment on-the-fly with live chart updates
• table_demo.py — Browse experiment runs with MlflowRunTable
• nested_demo.py — Parent/child nested runs with tree-view table
• parallel_demo.py — Parallel coordinates chart with status filtering

# Install demo dependencies
pip install mlflow-widgets[demo]

# Start MLflow server
mlflow server --port 5000 &

# Run a demo
marimo edit examples/demo.py

# Run demos (recommend) + auto refresh when editted (by coding agent etc.)
MLFLOW_TRACKING_URI=http://localhost:5000 marimo edit --watch .