About Workshop

Workshop is the model engineering suite of the Naviorx ecosystem. It handles the full model preparation pipeline: converting between formats, quantizing for deployment, and packaging into .nvx — Naviorx's self-contained model artifact format. Workshop supports full-precision and quantized GGUF conversion with configurable calibration datasets, and includes delta-based model versioning so you store only the difference between fine-tuned checkpoints. The result is a single .nvx file containing the model, tokenizer, and inference configuration — ready to deploy with Capsule.

Features

• Full-precision and quantized GGUF conversion
• Configurable calibration datasets for quantization
• .nvx packaging: model + tokenizer + config in one file
• Delta-based model versioning and diff tooling
• Merge and rebase fine-tuned checkpoints
• Batch processing for large model collections