Profile
ML pipeline plugin for building training datasets from HRRR/GFS/ERA5 weather models. Companion to the weather plugin for building weather ML pipelines.
Signals
Listed in the awesome-hermes-agent README
Sources: 2 / Surfaces: 1
What the upstream surface says
Short excerpt only, so you can decide whether to click out.
A Hermes Agent plugin for building ML-ready weather training datasets. Powered by wxtrain, an all-Rust end-to-end pipeline.
Ask Hermes to build training datasets from operational weather models:
| Tool | Description | |------|-------------| | wxtmodels | List supported weather models and sources | | wxtfetch | Download GRIB fields via byte-range .idx | | wxtscan | List all messages in a GRIB file | | wxtdecode | Decode a GRIB message — stats, grid dimensions, variable info | | wxtcalc | Thermodynamic calculations (theta, thetae, RH) | | wxtrender | Render a GRIB field as PNG | | wxtplan | Plan a training dataset for an ML architecture | | wxt_build | Build training arrays from GRIB files |
- "Plan a severe weather dataset for a Swin transformer" → 25-channel training spec with export format, loss function, and model recipe
- "Fetch HRRR CAPE data" → downloads via byte-range .idx subsetting (~500KB instead of 125MB)
- "Build training arrays from this GRIB file" → NPY arrays + preview PNGs + manifests
- "What's the theta-e at 30C, 20C dewpoint, 850mb?" → instant thermodynamic calculation
- wxtrain engine: Built with Codex
- Meteorological calculations: Verified against MetPy test suites
- Plugin platform: Hermes Agent by Nous Research