The Best Captum Alternatives in 2026

pytorch tabular provides a unified framework for deep learning models on tabular data.

PyTorch Frame provides a library for applying deep learning models to tabular data.

Python package which helps to debug machine learning classifiers and explain their predictions.

Evals is a framework for evaluating LLMs and LLM systems, and an open-source registry of benchmarks.

This repository implements YOLOv3 variants in PyTorch for real time object detection, training, validation, inference, and export.

PyTorch Geometric provides graph neural network tools and libraries for PyTorch projects.

GPyTorch provides an efficient implementation of Gaussian processes using PyTorch.

Ultralytics YOLOv5 supports object detection, instance segmentation, classification, training, and export in PyTorch.

Cornac provides a framework for comparing multimodal recommender systems and recommendation algorithms.

An open source Python library for drift detection in machine learning systems.

nilearn provides Python tools for machine learning and neuroimaging analysis.

A Python package to assess and improve fairness of machine learning models.

WhichLLM tests local language models on a user’s hardware and ranks them using recent benchmarks.

FastDatasets helps create training datasets for large language models.

AIF360 provides fairness metrics and explanations for datasets and models, along with algorithms to mitigate bias.

This repository provides segmentation models with pretrained backbones for Keras and TensorFlow Keras.

This repository provides semantic segmentation models with more than 500 pretrained convolutional and transformer backbones.

Automatic "Differentiation" via Text, using large language models to backpropagate textual gradients.

Ultralytics supports detection, segmentation, image classification, pose estimation, and tracking across YOLO model versions.

Giskard provides open source evaluation and testing tools for LLM agents.

NALP covers Natural Adversarial Language Processing through a dedicated library.

neuropredict evaluates predictive performance using neuroimaging features.

Adala automates data labeling through an autonomous agent framework for organizing labeled datasets.

Open-RAG-Eval evaluates retrieval-augmented generation without requiring golden answers.