skills.rest
OpenLAIR
OpenLAIROFFICIAL
Explore all skills by this author →
GitHub

aris-vast-gpu

Provisions, manages, and destroys vast.ai GPU instances based on training task requirements.

1.1k|116|Updated Feb 26, 2026
Search Tags:#cost optimization#training#ssh#cloud gpu#gpu rental#Machine learning#Vast.ai
One-click install
npx skills add https://github.com/OpenLAIR/dr-claw --skill aris-vast-gpu
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: aris-vast-gpu
Source: https://github.com/OpenLAIR/dr-claw/tree/main/skills/aris-vast-gpu
Command: npx skills add https://github.com/OpenLAIR/dr-claw --skill aris-vast-gpu

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires vastai.

What problem does it solve?

Renting cloud GPUs for ML experiments requires manually estimating VRAM needs, comparing offers, and managing instance lifecycles, which is error-prone and can waste money on forgotten running instances.

Core Features & Use Cases

  • Task-Driven GPU Selection: Analyzes experiment plans and training scripts to estimate VRAM, GPU count, disk, and runtime, then presents three cost-optimized vast.ai offers with estimated total cost.
  • Full Lifecycle Management: Handles renting, SSH setup, code syncing via rsync, dependency installation, result download, and instance destruction with cost reporting.
  • State Tracking: Maintains a vast-instances.json file so other experiment-running skills can discover and connect to active instances.
  • Use Case: A researcher fine-tuning a 7B model describes the task, picks from three priced GPU options, runs the experiment remotely, and has the instance auto-destroyed after results are downloaded.

Quick Start

Ask the assistant to rent a GPU on vast.ai for your training task, for example by saying rent a GPU to fine-tune LLaMA-7B for about two hours.

Frequently Asked Questions about aris-vast-gpu

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

FAQPage Schema
How do I rent a GPU on vast.ai for model training?▼

Install the vastai CLI with pip, set your API key, and upload your SSH public key at cloud.vast.ai/manage-keys. Then describe your training task and the skill estimates VRAM and runtime, searches offers, and presents three options ranked by estimated total cost.

How much VRAM do I need to fine-tune a 7B or 13B model?▼

A 7B model needs roughly 16 GB VRAM in fp16, a 13B model about 28 GB, and a 70B model around 140 GB requiring multiple GPUs. The skill adds 20 percent headroom and searches vast.ai offers matching the estimated requirement.

Why does SSH to my vast.ai instance fail with permission denied?▼

SSH keys are baked into vast.ai instances at creation time, so the key must be uploaded before renting. The fix is to destroy the instance, upload your public key at cloud.vast.ai/manage-keys, and create a new instance.

Does the vastai CLI work with older Python versions?▼

The vastai CLI requires Python 3.10 or higher. If your system Python is older, create a virtual environment with conda, pyenv, or uv using a newer Python and install vastai there.

What happens to my data when a vast.ai instance is destroyed?▼

Destroying an instance permanently deletes all data on it and is irreversible. The skill downloads experiment results and logs via rsync and scp before destroying, then reports the actual cost based on runtime.

Related Skills

Explore More Education & Research Skills→
M4n5terM4n5ter
3

tinker-training-cost

Tokenize datasets and estimate Tinker fine-tuning training costs in USD.

Community
Advanced
Wenwen555Wenwen555

monitor-experiment

...

Community
Advanced
jkfeejkfee

monitor-experiment

Monitors running experiments and collects results via SSH, screen sessions, wandb workflows, and remote JSON retrieval.

Community
Intermediate
promptromppromptromp
1

aws-bootstrap

Manage AWS EC2 GPU instances with the aws-bootstrap CLI.

Official
Intermediate