vast-gpu

Rent and manage Vast.ai GPU instances for ML tasks.

Updated Mar 1, 2026
One-click install
npx skills add https://github.com/hve4638/hve-cc-marketplace --skill vast-gpu-hve4638
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vast-gpu
Source: https://github.com/hve4638/hve-cc-marketplace/tree/main/aris/skills-unavailable/vast-gpu
Command: npx skills add https://github.com/hve4638/hve-cc-marketplace --skill vast-gpu-hve4638

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provision on-demand GPUs for ML workloads so you don't have to own and maintain hardware.

Core Features & Use Cases

  • Analyze the training task to determine VRAM and hardware requirements.
  • Search broad GPU offers and present three cost-optimized options.
  • Manage the full lifecycle: rent, setup, run, monitor, and destroy instances with a state-tracking file.
  • Prerequisites: install the vastai CLI and configure API keys; ensure SSH keys are uploaded before provisioning.

Quick Start

Describe your training task and run the provision workflow to rent and configure a GPU instance.

Frequently Asked Questions about vast-gpu

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I rent on-demand GPUs for ML training tasks?

To rent on-demand GPUs for ML training, you describe your training task so the system can analyze VRAM requirements, search broad GPU offers, and present three cost-optimized options for provisioning across Vast.ai.

Do I need to install the vastai CLI before provisioning GPU instances?

Yes, you need to install the vastai CLI and configure your API keys before provisioning GPU instances. You must also ensure your SSH keys are uploaded to the platform prior to renting hardware.

How does the system determine VRAM requirements for machine learning workloads?

The system determines VRAM requirements by analyzing your described training task. It evaluates the hardware needs of the specific ML workload to filter and search for broad GPU offers that match your computational demands.

Can I manage the full lifecycle of cloud GPU instances from provisioning to destruction?

Yes, you can manage the full lifecycle of cloud GPU instances from provisioning to destruction. The system handles create, setup, run, monitor, and destroy workflows while tracking instance states in a dedicated state file.

What is the best way to find cost-optimized GPU offers for ML workloads?

The best way to find cost-optimized GPU offers is to let the system search broad GPU listings on Vast.ai. It analyzes your ML task requirements and presents three cost-optimized options to minimize on-demand hardware expenses.

Why do I need to upload SSH keys before renting GPU hardware on Vast.ai?

You need to upload SSH keys before renting GPU hardware to ensure secure access and configuration of your instances. This prerequisite allows the setup workflow to properly configure the rented hardware for your ML tasks.