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

Alpamayo

Alpamayo 1 is a 10-billion-parameter vision-language-action model for autonomous vehicles that pairs driving trajectories with Chain-of-Causation reasoning.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Alpamayo — 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 Alpamayo 1 when you want to examine a research model’s driving trajectories together with its reasoning traces and can meet its GPU, software, and gated-resource requirements.

Who it is for

  • Autonomous-driving researchers studying trajectory prediction alongside causal reasoning.
  • Developers exploring vision-language-action models for robotics and driving research.

Documented uses

  • Generate autonomous-driving trajectory predictions with associated reasoning traces.
  • Explore or compare trajectory samples and Chain-of-Causation reasoning in an inference workflow.
Before you choose

Requirements & limits

  • The documented inference setup requires Python 3.12.x and an NVIDIA GPU with at least 24 GB of VRAM.
  • Model weights and dataset access are gated and require Hugging Face access requests and CLI authentication.
  • The test inference downloads approximately 22 GB of model weights.
  • This repository receives limited maintenance; its maintainers point to the newer Alpamayo hub for future releases and documentation.
  • The repository lists an Apache-2.0 license.

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

Run the Alpamayo 1 inference example

Set up the documented environment and launch the repository’s test-inference script.

Before you start

  • Python 3.12.x
  • Linux is the tested operating system; other platforms are unverified.
  • An NVIDIA GPU with at least 24 GB of VRAM.
  • Hugging Face access to the Alpamayo model weights and Physical AI AV Dataset, plus Hugging Face CLI authentication.

Steps

  1. Install uv and create and activate the environment with the command below and source ar1_venv/bin/activate.

    Documented command
    uv venv ar1_venv

  2. Install the project dependencies using the command below.

    Documented command
    uv sync --active

  3. Request access to the model weights and dataset, then authenticate with hf auth login.

  4. Run the command below.

    Documented command
    python src/alpamayo_r1/test_inference.py

Expected outcome

The repository’s example inference script downloads example data and the model weights, then runs inference.

Guide limitations

  • The model-weight download is about 22 GB.
  • The README warns that GPUs with less than 24 GB of VRAM will likely encounter CUDA out-of-memory errors.
  • The repository is no longer under active development; future releases and documentation are directed to the newer Alpamayo hub.
Documentation 1 ↗
Evidence and transparency

Sources & review

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