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

AutoE2E

AutoE2E is an open-source end-to-end autonomous-driving model that predicts a future driving trajectory from seven surround and telephoto cameras, a rendered map tile, ego-motion and visual history. Its repository provides model components, training and inference code, setup instructions and an AWS EC2 inference-trial guide.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

AutoE2E — 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 AutoE2E when you need an open-source starting point for studying or developing multi-camera driving trajectory prediction, or when you want to inspect the project's model and inference workflow. Use the separate DataModelConsole for read-only review of published samples, pipeline state and trajectory comparisons.

Who it is for

  • Autonomous-driving researchers prototyping camera-based trajectory prediction.
  • Developers exploring model training, inference or GPU performance across backbones and fusion modes.
  • Teams reviewing published driving samples and comparing ground-truth and predicted trajectories through the separate read-only DataModelConsole dashboard.

Documented uses

  • Predict a 6.4-second driving trajectory, expressed as acceleration and curvature at 10 Hz.
  • Prototype camera-based trajectory prediction for autonomous-driving research.
  • Compare model predictions with ground truth in camera and bird's-eye views.
  • Review dataset versions, shards, samples and geographic coverage in the DataModelConsole.
  • Inspect synchronized camera scenes alongside ego-state and map context.
Before you choose

Requirements & limits

The public GitHub repository is open source under Apache-2.0. The README specifies Python 3.12 and documents CPU and CUDA dependency-installation options.

  • The model's documented input set includes seven cameras, a rendered map tile, ego-motion and visual history; projects with different inputs may need adaptation.
  • The predicted horizon is 6.4 seconds, with acceleration and curvature represented at 10 Hz.
  • The stated approximately 76 FPS result applies specifically to SwinV2-Tiny, feature-concatenation fusion, an RTX 5080 and batch size one.
  • Python 3.12 is required by the README's pinned PyTorch build; it states that wheels are unavailable for Python 3.13 and later.
  • The repository separates model implementation from the read-only DataModelConsole dataset and results workspace.

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 AutoE2E dependencies with Make

Clone the repository and install its dependencies using the documented Make setup path.

Before you start

  • Python 3.12
  • Git and the Make tool
  • Choose CPU PyTorch wheels or a documented CUDA channel such as cu118.

Steps

  1. Clone https://github.com/autowarefoundation/auto_e2e.git and change into the auto_e2e directory.

  2. Run make setup for CPU wheels, or make setup TORCH_CHANNEL=cu118 for the documented CUDA option.

  3. Optionally run make test to check the installation using the repository's Makefile test.

Expected outcome

The documented setup installs dependencies; the optional Makefile test is available on this path.

Guide limitations

  • The pinned PyTorch build has no wheels for Python 3.13 and later.
  • The README does not provide the Makefile test through the plain-pip installation path.
  • Installation instructions do not by themselves run model inference.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
8 October 2026
Original website title
autowarefoundation/auto_e2e · Official homepage title · retained in its source language
How we review sourcesSuggest a correction