dstack

dstack

dstack is a GPU orchestration and scheduling tool that works across clouds, Kubernetes, and bare metal.

Agent SkillDevOpen source
Type
Agent Skill
Open source
Yes
GitHub Stars
★ 2.2k
Source
skill-github

Overview

dstack is a unified control plane for GPU provisioning and orchestration, supporting NVIDIA, AMD, Google TPU, and Tenstorrent accelerators. It simplifies development, training, and inference processes and is compatible with any hardware, open-source tools, and frameworks. With dstack, AI teams can manage clusters, submit workloads, and automatically handle resource scheduling. Ideal for developers and teams needing efficient and scalable GPU resource management.

Capabilities

  • GPU provisioning and orchestration
  • Support for multiple accelerators
  • Cross-cloud and on-premise cluster management
  • Automated resource scheduling

Use cases

Machine learning model trainingLarge-scale data processingDistributed computing tasksGPU resource management

Setup

Requires: Node 环境API Key
Install CLI: `uv tool install dstack -U`; Configure server: `dstack project add --name main --url http://127.0.0.1:3000 --token <your_token>`; Install skills: `npx skills add dstackai/dstack`.

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

Which accelerators does dstack support?

Supports NVIDIA, AMD, Google TPU, and Tenstorrent.

How do I configure the dstack server?

Run `dstack server` and configure your project information.

Related skills