mustatil 5.2.2

GIS-level AI vision workspace for annotation, YOLO training, satellite detection, geospatial inference, and visual AI pipelines.

Mustatil 5.2.2 - all-dependencies pip build

This build is configured so that a normal command downloads the regular Python dependency stack:

py -m pip install mustatil

It now declares Torch, TorchVision, TorchAudio, Ultralytics, PySide6-Addons/QtWebEngine via PySide6, GIS packages, ONNX packages, transformer vision packages, and science/pipeline packages as normal dependencies instead of requiring mustatil[yolo,gis,onnx,vision,science,webui].

Large model files and generated runtimes are still not embedded in the wheel. PyPI packages should stay code-focused; model weights such as SAM2/YOLO/ONNX checkpoints should be downloaded or selected at runtime.

Mustatil

Mustatil is a GIS-level AI vision workspace for annotation, YOLO training, large-scale detection, satellite-map analysis, geospatial inference, GeoPackage/GIS export, and visual AI pipeline building.

This PyPI package contains the Python application code and bundled plugins. Large model weights and generated runtime files are intentionally excluded and should be distributed separately through GitHub Releases, Hugging Face, Zenodo, itch.io, or downloaded by the application into a user runtime folder.

Install

pip install mustatil
mustatil

For a fuller install with optional GIS/AI dependencies:

pip install "mustatil[yolo,gis,onnx,vision,science,webui]"

PyTorch should be installed separately according to the target hardware, especially when CUDA wheels are required.

Run

mustatil

Or:

python -m mustatil

Notes

The original desktop ZIP contained large model files such as sam2_b.pt, Houses300.pt, Houses300.onnx, Tree50.pt, and yolov8n.pt. These files are not included in this source package.

Current pip package version

5.2.0