OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Torch Conv Kan

Torch Conv Kan is a Python/PyTorch research repository implementing convolutional Kolmogorov-Arnold network layers and model architectures for training, validation and quantization experiments.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Torch Conv Kan — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you need research code for experimenting with convolutional KAN layers in PyTorch, especially for image-model training, validation or quantization.

Who it is for

  • Computer-vision researchers exploring convolutional KAN architectures.
  • PyTorch developers prototyping image classifiers or comparing KAN-based models with conventional CNN baselines.
  • Practitioners studying quantization and training workflows for convolutional KAN models.

Documented uses

  • Build image-classification prototypes from convolutional KAN layers.
  • Compare KAN-based convolutional models with conventional CNN baselines.
  • Train and validate models on the documented image datasets.
  • Explore quantization and review logged performance metrics.
Before you choose

Requirements & limits

The repository is public source code released under the MIT license.

  • The project is a research implementation, and its README labels it under development.
  • The documented training stack includes PyTorch and Accelerate; Hydra configurations and Weights & Biases logging are also described.
  • The simple baseline workflow covers MNIST, CIFAR-10 and CIFAR-100; other datasets are discussed as evaluation targets in the README.
  • The README describes CUDA acceleration for the project, but this profile does not establish hardware compatibility.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up the repository for Accelerate-based training

Clone the project and install its listed Python requirements before following its training configuration instructions.

Before you start

  • Git
  • Python environment suitable for the repository's requirements.txt

Steps

  1. Clone the torch-conv-kan repository.

  2. Change into the repository directory.

  3. Install dependencies with pip using requirements.txt.

  4. Continue with the repository's Accelerate-based training instructions and Hydra configurations.

Expected outcome

The repository dependencies are installed, and the documented next stage is configuring and using the Accelerate-based training workflow.

Guide limitations

  • These steps cover repository setup only; they do not establish that a training run will succeed on a particular machine.
  • The README also describes Weights & Biases logging configuration, but details are not included in these setup steps.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
IvanDrokin/torch-conv-kan · Official homepage title · retained in its source language
How we review sourcesSuggest a correction