remote-run-ssh

Copy a target example and cvlization library to a remote host via rsync and SSH.

12|3|Updated Oct 7, 2021
One-click install
npx skills add https://github.com/kungfuai/CVlization --skill remote-run-ssh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: remote-run-ssh
Source: https://github.com/kungfuai/CVlization/tree/main/.claude/skills/remote-run-ssh
Command: npx skills add https://github.com/kungfuai/CVlization --skill remote-run-ssh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Remote-run-ssh copies only the target example and the cvlization library to a remote host, enabling remote execution of CVlization workflows.

Core Features & Use Cases

  • Minimal remote workspace: copy only necessary example and library
  • Reproducible experiments: run build.sh and train.sh on remote host
  • Safe teardown: leaves local repo untouched

Quick Start

Sync the required example and library to the remote host, then run the example's build.sh and train.sh to start the job.

Frequently Asked Questions about remote-run-ssh

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

FAQPage Schema
How do I run CVlization workflows on a remote GPU server via SSH?

To run CVlization workflows on a remote GPU server via SSH, the skill copies only the target example and the cvlization library to the remote host to execute build.sh and train.sh scripts.

What do I need to set up before executing remote GPU trainings from my local repo?

Executing remote GPU trainings requires rsync and SSH access, a remote host with Docker and CUDA support, and build.sh and train.sh scripts present in your local example directory.

Does running remote executions modify my local repository files?

Running remote executions does not modify your local repository files, as the skill performs a safe teardown that leaves your local repo untouched while ensuring a reproducible run remotely.

How can I sync only the necessary example and library files to a remote host for training?

You can sync only the necessary example and library files by using rsync over SSH, which creates a minimal remote workspace containing solely the target example and the cvlization library.

Can I use this remote execution approach for CPU-bound evaluations instead of GPU-bound trainings?

This remote execution approach is specifically applicable for GPU-bound trainings and evaluations on remote servers, requiring remote host Docker and CUDA support to drive the workflow.

Why are build.sh and train.sh required in the example directory for remote execution?

Build.sh and train.sh are required in the example directory because they directly drive the remote execution workflow, enabling reproducible experiments on the remote host after the initial sync.