NVIDIAKnown publisherNVIDIA/Quantum-Calibration-Agent-Blueprint/data/knowledge/skills/vlm-configuration/SKILL.md

vlm-configuration

Skill69 starsChanged 7 months agoApache-2.0View on GitHub

Configure the Vision Language Model (VLM) used for analyzing experiment plots. Use when the user asks to change the VLM provider or model, adjust temperature or max_tokens, enable or disable thinking mode, or troubleshoot VLM-related behavior in config.yaml.

Reads credentials

In this repository5 other files it gives its agents.
- analysis-scriptsSkill
- experiment-executionSkill
- workflow-executionSkill
- workflow-planningSkill
- writing-experiment-scriptsSkill

The file

Create a free account to mount

Files

SKILL.md

VLM Configuration

The VLM (Vision Language Model) is configured in config.yaml under the vlm section.

Config Location

config.yaml (in the project root)

Current Config Structure

vlm:
  provider: "nvidia"
  model: "nvidia/ising-calibration-1-35b-a3b"
  api_key_env: "NVIDIA_API_KEY"
  temperature: 0.2
  max_tokens: 32768
  enable_thinking: false

Supported Models (NVIDIA API)

Model	Provider	Notes
`nvidia/ising-calibration-1-35b-a3b`	NVIDIA	Default, fine-tuned for calibration
`google/gemma-4-31b-it`	Google	General-purpose 31B
`meta/llama-3.2-90b-vision-instruct`	Meta	Most capable
`meta/llama-3.2-11b-vision-instruct`	Meta	Balanced
`microsoft/phi-3.5-vision-instruct`	Microsoft	Lightweight

Environment Variables

The VLM uses NVIDIA_API_KEY (same key as the main LLM).

To Change VLM

Read current config.yaml
Edit the vlm section with new values
Changes take effect immediately on next vlm_inspect call (no restart needed)

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

It worked
Partly
It didn't

What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free