Manage GPU compute jobs on the Qizhi platform via a kubectl-style CLI.
npx skills add https://github.com/goupup-ai/miccai25 --skill qzcli-goupup-ai
Please help me install this Agent Skill. Skill: qzcli Source: https://github.com/goupup-ai/miccai25/tree/main/ARIS/skills/qzcli Command: npx skills add https://github.com/goupup-ai/miccai25 --skill qzcli-goupup-ai
Manually managing GPU compute jobs on the Qizhi (启智) platform is slow and error-prone, with users facing inconsistent resource tracking, tedious job submission workflows, and limited visibility into compute group availability and job status across workspaces.
Use the qzcli skill to log in to your Qizhi account, refresh your workspace resource cache, and submit a new distributed training job named "vertebrae-seg-run" with 4 GPU instances running the command "bash train.sh --dataset vertebrae_ct".
High-intent search queries and answers about installing and using this skill.
You can submit distributed training jobs to Qizhi GPU clusters using a kubectl-style CLI that supports scriptable, config-driven batch workflows and interactive guided submissions, bypassing manual web portal navigation.
Yes, batch job submission and orchestration for Qizhi compute workloads is automated via config-driven workflows within the CLI, enabling scriptable and reproducible job execution for diverse training and inference tasks.
The Qizhi CLI supports real-time job status monitoring and automatic compute resource discovery by caching and listing available workspaces, compute groups, and idle GPU nodes to eliminate manual resource lookup.
The best way to manage Qizhi GPU job lifecycles without the web portal is using a kubectl-style CLI to submit, stop, monitor, and watch jobs, ensuring reproducible workflows with configurable authentication.
Yes, you can run both interactive guided workflows and non-interactive scriptable submissions for HPC workloads, specifying custom resource specifications for CPU-heavy jobs across Qizhi GPU clusters.
Compute resource caching for Qizhi GPU job management automatically discovers and lists available workspaces, compute groups, and idle GPU nodes, reducing configuration errors and eliminating inconsistent resource tracking.
dogekikiManage GPU compute jobs on the Qizhi platform via command-line interface.
HeXiao-55Monitor WandB metrics to detect NaN, divergence, or stalled GPU utilization.
jcolanoAutomate GitHub workflow orchestration for multi-repo CI/CD pipelines.
dongzhuoyaoSubmit GPU training jobs to Slurm clusters with preflight validation and monitoring.
