Promote Sign in Sign up

Low-Code/NoCode

captum.ai

PyTorch library for model interpretability across modalities.

Source rating 5.0 (64) 12k views 12k/mo traffic 0 saves Unknown
Low-Code/NoCode no code Code(0)
More categories (16)
Learning LOW-CODE/NO-CODE CODE ASSISTANT No-code Machine Learning DEVELOPER TOOLS Developer Docs duet ai assistant Model interpretability PyTorch Attribution Integrated Gradients Vision Text Explainable AI xAI
Visit website Sign in to save

Verdict

A quick read before you dig into the full page.

Best if
You are machine learning researchers and need: Understanding feature importance in image classification models
Pick this when
PyTorch library for model interpretability across modalities

Overview

Captum is a PyTorch library for model interpretability. It provides tools and techniques to understand and attribute the predictions of PyTorch models across various modalities, including vision and text. Captum supports most PyTorch models and allows for easy implementation and benchmarking of new interpretability algorithms.

captum.ai screenshot

Best for

Short scenarios where captum.ai tends to shine.

  • Understanding feature importance in image classification models

    Best for Machine learning researchers

  • Analyzing the influence of words in text classification models

    Best for Data scientists

Example outputs 1

Sample generations and before/after results for captum.ai.

captum.ai screenshot
captum.ai screenshot

Access & connections

Where you can use captum.ai.

Platforms

Web

Capabilities

What this tool accepts, produces, and is strongest at.

Highlights

  • Multi-modal interpretability support
  • Built on PyTorch
  • Extensible and open source

People also view

Tools that often show up alongside captum.ai in collections and comparisons.

FAQ 5

What is Captum?
Captum is a PyTorch library for model interpretability. It provides tools and techniques to understand and attribute the predictions of PyTorch models across various modalities, including vision and text. Captum supports most PyTorch models and allows for easy implementation and benchmarking of new interpretability algorithms.
How to use Captum?
Install Captum via conda or pip. Import necessary libraries like numpy, torch, and IntegratedGradients from captum.attr. Define and prepare your PyTorch model. Instantiate an interpretability algorithm (e.g., IntegratedGradients). Apply the algorithm to your input data and baseline to obtain attributions and convergence delta.
How do I install Captum?
You can install Captum via conda (conda install captum -c pytorch) or pip (pip install captum).
What types of models does Captum support?
Captum supports most types of PyTorch models and can be used with minimal modification to the original neural network.
What is Integrated Gradients?
Integrated Gradients is an algorithm implemented in Captum that helps to attribute the prediction of a model to its input features.
captum.ai
Visit website