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

Perceiver Pytorch

Perceiver Pytorch is a Python and PyTorch implementation of the Perceiver neural-network architecture, with model variants for general inputs, structured outputs, and language-model-shaped token sequences.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Perceiver Pytorch — 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 an editable PyTorch starting point for Perceiver-based research or prototypes and want direct control over input axes, latent settings, attention structure, and output configuration.

Who it is for

  • PyTorch researchers studying iterative attention.
  • Developers prototyping configurable image, video, sequence, or language-model architectures.
  • Students and engineers examining latent-based attention designs.

Documented uses

  • Instantiate a configurable Perceiver model in PyTorch.
  • Experiment with image and video input axes and tokenized sequences.
  • Produce structured outputs through PerceiverIO decoder queries.
  • Explore a PerceiverLM configuration for token sequences.
  • Inspect an experimental bottom-up-attention variant.
Before you choose

Requirements & limits

The repository is public, lists an MIT license, and documents installation through pip. No pricing or account requirement is stated.

  • The standard Perceiver configuration supports input_axis values for image-like and video-like data.
  • Fourier encoding can be enabled in the model or disabled when the data is encoded separately.
  • PerceiverIO adds decoder queries and configurable output logits for structured outputs.
  • PerceiverLM documents token count, maximum sequence length, and attention-mask inputs.
  • The bottom-up-attention implementation is marked experimental in the README.

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

Install the Perceiver PyTorch package

Prepare the documented Python package for Perceiver model experiments.

Before you start

  • Python environment with PyTorch available.
  • Access to the public perceiver-pytorch package.

Steps

  1. Install the package with the command below.

    Documented command
    pip install perceiver-pytorch

  2. Import Perceiver, PerceiverIO, or PerceiverLM from the package as appropriate for the documented model variant.

  3. Configure the selected class using the parameters shown in the repository README.

  4. Provide input tensors matching the documented example shape for the selected model.

Expected outcome

A Python environment configured to instantiate the documented Perceiver model variants and pass matching tensor inputs.

Guide limitations

  • The installation instructions do not document hardware requirements, training procedures, benchmark results, or production deployment.
  • The experimental bottom-up-attention variant uses a separate experimental import.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
3 October 2026
Original website title
lucidrains/perceiver-pytorch · Official homepage title · retained in its source language
How we review sourcesSuggest a correction