vlm-configuration
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.
Source facts
- Repository
- NVIDIA/Quantum-Calibration-Agent-Blueprint
- Last source activity
- April 20, 2026 at 05:20
- Detected SKILL.md language
- English
- Stars
- 73
- Forks
- 27
Install options
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Review the source files
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Showing SKILL.md
SKILL.md
Source instructions · Read-only preview- name
- vlm-configuration
- description
- 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.
# 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
```yaml
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
1. Read current config.yaml
2. Edit the `vlm` section with new values
3. Changes take effect immediately on next `vlm_inspect` call (no restart needed)
View on GitHub