← all repositories

e-p-armstrong/augmentoolkit

An end-to-end tool for generating fine-tuning datasets, training domain-expert LLMs, and serving them locally.

augmentoolkit
Velocity · 7d
+2.0
★ / day
Trend
steady
star history

Augmentoolkit transforms user documents into training datasets that update an LLM’s knowledge cutoff, enabling domain expertise in any chosen area. It generates fine-tuning data, optionally creates RAG-ready datasets, and can spin up a local inference server for the resulting model. The pipeline works fully offline without external API keys for data generation on most hardware.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.