# BertViz

> Visualize attention in Transformer language models such as BERT and GPT-2.

- Category: [Machine learning visualization tools](https://awesomedataviz.com/categories/machine-learning/)
- Repository: https://github.com/jessevig/bertviz
- License: Apache-2.0
- Language: Python
- Status: Maintained (commits in the last 12 months)
- GitHub stars: 8,193
- Commits in the last 12 months: 1
- Contributors: 5
- Last commit: Jan 8, 2026
- Latest release: v1.4.1 (Jun 1, 2025)
- Install (PyPI): `pip install bertviz`, 635 downloads per week
- Topics: Machine learning & AI visualization

## Overview

BertViz is an open-source ML visualization tool released under the Apache-2.0 license. Its GitHub repository has 8,193 stars, 888 forks, and 5 contributors. It is maintained with 1 commit in the last 12 months; the most recent commit was on Jan 8, 2026. The latest release, v1.4.1, was published on Jun 1, 2025. On PyPI it is downloaded about 635 times per week.

## Alternatives

- [Netron](https://awesomedataviz.com/tools/netron/): Viewer for neural network, deep learning and machine learning models. (33.5K stars)
- [PlotNeuralNet](https://awesomedataviz.com/tools/plotneuralnet/): LaTeX code for drawing neural network architecture diagrams. (25K stars)
- [Opik](https://awesomedataviz.com/tools/opik/): Formerly CometLLM. Debug, evaluate, and monitor LLM applications with tracing and dashboards. (22.4K stars)
- [Phoenix](https://awesomedataviz.com/tools/phoenix/): ML observability in a notebook with UMAP visualizations (11.7K stars)
- [FiftyOne](https://awesomedataviz.com/tools/fiftyone/): Tool for visualizing, curating and evaluating computer vision datasets and models. (11.1K stars)
- [Visdom](https://awesomedataviz.com/tools/visdom/): Tool for real-time visualization and monitoring of live data such as ML experiments. (10.3K stars)

Source: https://awesomedataviz.com/tools/bertviz/ · Data updated 2026-10-04 · Content: CC BY 4.0, Awesome Dataviz (https://awesomedataviz.com).