Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

149 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Inference Snaps

Install, run, and serve AI models as snaps — optimized for your hardware.

Inference Snaps turn complex generative model deployment into a single command. They are built on the idea that AI runtimes should be portable, composable, and local-first. Whether you’re experimenting with Gemma 4, Qwen 3 Coder, or DeepSeek R1, you can instantly install a pre-optimized runtime and spin up an API endpoint locally. Each model snap includes all dependencies and is tuned for your silicon — CPU, GPU, or NPU.

Every device should be able to run intelligent software without dependency hell or optimization nightmares.

Quick start

Install the accelerator driver.

Then:

# Install a model snap
sudo snap install gemma4

# Check status
gemma4 status

# Chat with the LLM
gemma4 chat

What it does

  • Plug-and-play AI models: Each snap bundles a range of optimized model weights and runtimes.
  • Hardware-aware: Automatically selects and deploys silicon-optimized runtime and model weights for your machine.
  • API-first design: Exposes local standard APIs for integration.
  • Offline-ready: Once installed, the model runs entirely on your machine.
  • Reproducible & isolated: Snaps guarantee consistent environments across supported Linux distros.

Available inference snaps

Model Capabilities Install Source
DeepSeek R1 text, thinking snap install deepseek-r1 repo
Gemma 3 text, thinking, vision snap install gemma3 repo
Gemma 4 text, thinking, vision, tools snap install gemma4 repo
GLM 4.7 Flash text, thinking, tools snap install glm-4-7-flash repo
Nemotron 3 Nano text, thinking, tools snap install nemotron-3-nano repo
Nemotron 3 Nano Omni text, thinking, vision, tools snap install nemotron-3-nano-omni repo
Nomic Embed Text v1.5 text embedding snap install nomic-embed-text-v1-5 repo
Qwen 2.5 VL text, vision snap install qwen-vl repo
Qwen 3 text, thinking, tools snap install qwen3 repo
Qwen 3 Coder text, tools snap install qwen3-coder repo
Qwen 3.6 text, thinking, vision, tools snap install qwen3-6 repo

More open source models are coming soon. Share your feedback and vote for what open-source models to support next.

Documentation

Head to https://documentation.ubuntu.com/inference-snaps

The documentation is maintained inside the docs directory.

Contributing

We welcome pull requests! Please see contribution guide for tooling, build instructions, and model submission guidelines.

Join the discussion in our GitHub Discussions, or report issues.

About

Local inference, optimized for your hardware

Resources

Contributing

Stars

Watchers

Forks

Releases

Packages

Used by

Contributors

Generated from canonical/sphinx-stack