Open Source RoboticsRobots, software & learning

VISTA visual-tactile policy

Software

Equivariant diffusion policy combining wrist vision and tactile observations for contact-rich manipulation. [1]

Installation & first result →Official websiteSource codeGetting started

For
Manipulation researchers with tactile sensors or matching simulation infrastructure. [1]
Hardware
Simulated contact tasks and linked Franka real-robot code [1]
Source rights
Named licenses are documented by part. Review licenses
Unknowns
operating system not fully established. Review gaps

Source researched · Not tested by us
Sources checked · Research details

Licenses & related hardware

Requirements · Research details and dates

On this page

Overview

This VISTA is the CoRL 2026 Equivariant Visual-Tactile Diffusion Policy for Contact-Rich Manipulation. Its released code combines wrist vision, tactile features and end-effector orientation for action prediction. The repository provides training and evaluation paths, eight simulated contact tasks, a linked HDF5 demonstration dataset and a separate real-robot code link. [1] [4] [5]

Is this for you?

Workflows
  • Download data/assets
  • Train policy
  • Check checkpoint offline
  • Evaluate simulated tasks

[1]

Known limitations
  • Inherited ISP code has academic/educational-use terms despite the root Apache license.
  • Large simulation assets and compiled libraries are external.
  • Separate real-robot repository not tested.

[1] [3] [5]

What you need

Languages / runtimes
  • Python

[1]

Dependencies
  • Isaac Sim 4.5
  • Isaac Lab 2.1.1
  • libuipc
  • MIDFOV dataset

[1]

Host computer

Linux with NVIDIA GPU for simulation; exact training VRAM not established.

[1]

Compute for the workload
  • NVIDIA GPU for simulation; exact minimum not measured.

[1]

Installation & first result

Open the upstream getting-started guide [1]

  • Use the README's Python 3.10 environment, not the older environment.yml lock.
  • Download the task HDF5 data and required simulator assets.
  • Train a checkpoint and attempt the documented offline-inference check before simulation rollout.

[1] [4] [5]

Installation

README documents Python 3.10, CUDA/PyTorch, Isaac Sim/Lab and libuipc setup; no installation performed.

[1]

Hardware, models & data

Supported hardware
  • Simulated contact tasks and linked Franka real-robot code; physical route not verified here.

[1]

Inputs and outputs

Wrist RGB, tactile observations and end-effector orientation to predicted action chunks.

[1]

File formats
  • HDF5 episodes
  • PyTorch checkpoints

[1]

Algorithms and models
  • Equivariant visual-tactile diffusion policy

[1]

Evaluation

Offline action-finiteness check and headless simulated rollouts documented; no execution here.

[1]

Project notes

Start with the matching environment

The README specifies Linux and NVIDIA hardware for simulation, Python 3.10, Isaac Sim 4.5, Isaac Lab 2.1.1 and a libuipc build. It warns that the checked-in environment.yml is an older Python 3.9 lock and should not be installed. A useful first inspection is the documented offline-inference path with the named checkpoint and data layout, before attempting simulator rollouts or real-robot control. [1] [4] [5]

The repository's root license is Apache-2.0, but its third-party notices say inherited ISP implementation portions retain an academic-research-and-educational-use restriction. The HDF5 dataset card declares Apache-2.0; simulator assets and compiled libraries are external to the GitHub release. Treat the implementation as mixed-rights source-available material, not a uniformly permissive package. No dependency installation, dataset download, offline inference, simulator run or physical contact test was performed. [1] [2] [3] [5]

What we checked

Source research only. No commercial relationship or hands-on testing. No upstream media copied to the public site.

Not yet established

  • Requirements: operating system

Updates

Upstream last updated
Not established
Latest release
Not established
Profile updated
Nov 7, 2026
Sources last checked
Nov 7, 2026
Next review
Dec 7, 2026
Reviewed revision
Commit snapshot a0381fb1f95498dda9b006344bbbe923d42b5ccc; no matching GitHub release found

Source files and licenses

Inspectable contact-rich robot policy and task dataset; inherited implementation restrictions and external assets prevent a blanket open-source claim.

LayerAvailability / license
softwareavailable · Apache-2.0 root; inherited ISP portions academic/educational only
Artifact and declaration inspected at the cited scope. External assets/dependencies and actual operation need separate verification. [1] [3]
datasetsavailable · Apache-2.0 dataset-card declaration
Artifact and declaration inspected at the cited scope. External assets/dependencies and actual operation need separate verification. [5]
documentationavailable · Apache-2.0 root with third-party notices
Artifact and declaration inspected at the cited scope. External assets/dependencies and actual operation need separate verification. [1] [3]

No separate artifact selected for this scoped entry.

References

Each source links to the exact revision we read. “Used for” shows where this page relies on it.

Correction guidance · Discussion is not enabled · Profile format 0.7.0-draft.5

Change history