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

Gotennet Pytorch

Gotennet Pytorch is an unofficial Python and PyTorch implementation of the GotenNet 3D equivariant transformer. It accepts atom identifiers, 3D coordinates and an adjacency matrix, and returns invariant outputs with optional updated coordinates.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Gotennet 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 want a PyTorch implementation of GotenNet to inspect or adapt for molecular-graph experiments, and can supply the model's atom, coordinate and adjacency inputs.

Who it is for

  • Researchers exploring geometric learning and 3D equivariant transformer architectures.
  • Developers prototyping molecular graph models with atom identifiers and spatial coordinates.

Documented uses

  • Prototype experiments that use molecular atom IDs, coordinates and graph connectivity.
  • Explore invariant model outputs and optional coordinate updates for molecular graph inputs.
Before you choose

Requirements & limits

  • The README documents installation with pip and a model invocation using PyTorch.
  • The usage example recommends float64 as the default dtype for equivariant network training.
  • The repository identifies this as an unofficial implementation and links to a separate official repository.
  • The example treats negative atom IDs as padding and accepts a boolean adjacency matrix.

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 and call GotenNet on example molecular-graph inputs

Install the PyTorch package and pass atom IDs, coordinates, and adjacency data to a GotenNet model.

Before you start

  • Python environment with PyTorch available.
  • The README recommends setting PyTorch's default floating-point dtype to float64 for equivariant network training.

Steps

  1. Install the package with the command below.

    Documented command
    pip install gotennet-pytorch

  2. In Python, import torch and set torch.set_default_dtype(torch.float64); then import GotenNet from gotennet_pytorch.

  3. Create a model using the documented parameters: dim=256, max_degree=2, depth=1, heads=2, dim_head=32, dim_edge_refinement=256, and return_coors=False.

  4. Prepare atom IDs, coordinates, and a boolean adjacency matrix, then call model(atom_ids, adj_mat=adj_mat, coors=coors) and assign the returned values to invariant, coors_out.

Expected outcome

The call returns an invariant output and a second value assigned to `coors_out`; with `return_coors=False`, the example does not request coordinate returns.

Guide limitations

  • Negative atom indices are treated as padding; the README gives molecule length as (atom_ids >= 0).sum(dim=-1).
  • The README labels the implementation unofficial and points to a separate official repository.
  • The documented usage is an example; it does not establish model accuracy or suitability for a particular prediction task.
Documentation 1 ↗
Evidence and transparency

Sources & review

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