Manage GPU compute job lifecycle on the Qizhi platform via CLI.
npx skills add https://github.com/raja21068/AutoResearch --skill qzcli-raja21068
Please help me install this Agent Skill. Skill: qzcli Source: https://github.com/raja21068/AutoResearch/tree/main/skills/aris/qzcli Command: npx skills add https://github.com/raja21068/AutoResearch --skill qzcli-raja21068
It helps you manage GPU compute jobs on the Qizhi (启智) platform without manually juggling sessions, credentials, and cluster resource details.
Use Case: You need to run distributed training across a specific compute group with the correct resource spec and Docker image, then monitor runs and stop failed ones quickly.
Run the command "qzcli login" to authenticate so you can then submit and monitor jobs on your Qizhi account.
High-intent search queries and answers about installing and using this skill.
To submit and monitor Qizhi GPU jobs, you can use a kubectl-style CLI to manage the full job lifecycle, including creating runs, checking status with watch-based updates, and stopping jobs by ID.
For distributed training across a compute group, use batch submission to launch multiple related runs from a JSON config with a matrix of parameters, then monitor progress and stop failed jobs quickly.
Yes, HPC orchestration on Qizhi avoids manual session juggling by caching and refreshing available workspaces, compute groups, and resource specs to prevent trial-and-error resource selection.
To configure credentials for batch submission, run the login command to authenticate, or provide settings via CLI arguments, environment variables, or a local .env file before submitting jobs.
The kubectl-style CLI supports both interactive and non-interactive jobs, allowing you to create interactive sessions for exploration or submit non-interactive batch runs for automated experiments.
benjaminshafiiConfigure Bitwarden CLI with API key authentication to retrieve vault credentials.
zhuyingqinManage GPU compute jobs on the Qizhi platform via kubectl-style qzcli commands.
pxlxingliangPackage task directories and execute commands in Docker images on Bohrium or dispatchers.
dongzhuoyaoSubmit GPU training jobs to Slurm clusters with preflight validation and monitoring.
