Bonsai-Image-4BT-TRNQ

Bonsai-Image-4BT-TRNQ is a Trillim-packaged, requantized build of prism-ml/bonsai-image-ternary-4B-unpacked, intended for efficient text to image generation with the Trillim inference engine.

Overview

• Based on Prism ML’s Bonsai-Image-Ternary-4B release
• Requantized and packaged for Trillim runtime
• Distributed under the upstream Apache 2.0 license with attribution preserved

Model Details

Item	Value
Architecture	Flux.2-klein-4B
Parameters	4B class
Source model	prism-ml/bonsai-image-ternary-4B-unpacked
Packaging	Trillim requantized bundle
License	Apache 2.0

Usage

pip install trillim
trillim pull Trillim/Bonsai-Image-4BT-TRNQ
trillim image Trillim/Bonsai-Image-4BT-TRNQ "a tiny bonsai on a desk" -o bonsai.png

This generates an image of a bonsai on a desk and stores it in bonsai.png.

Repository Contents

File	Description
qmodel.tensors	Quantized weights in Trillim format
rope.cache	Precomputed RoPE cache for runtime
config.json	Model configuration
generation_config.json	Generation defaults
trillim_config.json	Trillim runtime metadata

Provenance

This repository is derived from:

• prism-ml/bonsai-image-ternary-4B-unpacked
• Flux.2-klein-4B by Black Forest Labs

Changes made by Trillim:

• Requantized from the upstream release
• Repackaged for the Trillim inference engine

This repository is not affiliated with or endorsed by Prism ML or Black Forest Labs.

License

Released under the Apache 2.0 License, consistent with the upstream model license.

See:

• LICENSE
• NOTICE.txt