skills.rest
lightrain-a
lightrain-aCOMMUNITY
Explore all skills by this author →
GitHub

monitor-experiment

Monitor running ML experiments and retrieve logs, JSON results, and W&B metrics.

Updated May 20, 2026
Search Tags:#machine learning#ssh#wandb#results summarization#training metrics#experiment monitoring#vast ai
One-click install
npx skills add https://github.com/lightrain-a/medtrace-aris --skill monitor-experiment-lightrain-a
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: monitor-experiment
Source: https://github.com/lightrain-a/medtrace-aris/tree/main/.vendor/aris/skills/monitor-experiment
Command: npx skills add https://github.com/lightrain-a/medtrace-aris --skill monitor-experiment-lightrain-a

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It reduces the time and guesswork required to determine whether a long-running ML experiment is still running and what outcomes it produced, without manually digging through logs and output files.

Core Features & Use Cases

  • Runtime Status Monitoring: Detects active compute by querying screen sessions on SSH servers, checking vast.ai instances, and reviewing Modal app status and logs when applicable.
  • Automated Output Collection: Pulls recent tail output from each screen session and locates/reads JSON result artifacts from a results directory.
  • Optional Weights & Biases Metrics Retrieval: When enabled, fetches training curves and final summaries from W&B to interpret convergence, eval quality, and run health.
  • Results Summarization and Interpretation: Produces a comparison-style summary table and flags anomalies (e.g., divergence, NaNs) before suggesting next steps.
  • Optional Feishu Notification: Sends an experiment completion summary via Feishu when configured.

Quick Start

Run monitor-experiment for your experiment server or screen name to collect the latest logs, JSON result files, and (if enabled) W&B metrics and return a concise status and comparison summary.

Frequently Asked Questions about monitor-experiment

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

FAQPage Schema
How do I check ML training progress on a remote vast.ai instance?▼

To monitor ML training progress on a vast.ai instance, the tool queries instance status, tails active screen session logs, and reads JSON result artifacts to summarize current outputs.

How do I retrieve Weights and Biases metrics for a running experiment?▼

Retrieve Weights and Biases metrics by enabling optional W&B integration during experiment monitoring, which fetches training curves, run summaries, and convergence data.

Can I monitor training jobs running in SSH screen sessions?▼

Yes, you can monitor training jobs in SSH screen sessions by detecting active compute, querying screen processes, and pulling recent tail output logs for analysis.

What is the best way to summarize ML experiment results from Modal apps?▼

The best way to summarize ML experiment results from Modal apps is to review app status, collect JSON artifacts, and generate a comparison table flagging anomalies like NaNs.

Does this tool automatically detect training anomalies like NaNs?▼

Yes, the tool detects training anomalies like NaNs and divergence during results summarization by interpreting collected logs and Weights and Biases metrics.

Related Skills

Explore More Education & Research Skills→
raja21068raja21068
2

run-experiment

Deploy and run machine learning experiments across local GPUs, SSH servers, Vast.ai, and Modal.

Community
Advanced
MACJONSWMACJONSW

monitor-experiment

Monitor remote experiments via SSH and collect results from logs.

Community
Intermediate
Shallow-WShallow-W

monitor-experiment

Monitor ML experiments across SSH, Vast.ai, and Modal deployments.

Community
Intermediate
dogekikidogekiki

monitor-experiment

Monitor remote machine learning experiment status and performance metrics across SSH, Vast.ai, and Modal.

Community
Advanced