Field notes from people shipping adapters
Practical run logs, evaluation methods, build reports, and postmortems from LoRADock contributors.
Results for “fine-tuning”
6 articlesFine-Tuning Qwen2.5-7B on 18,000 Support Tickets
A repeatable workflow for preparing, training, and evaluating a Qwen support-routing adapter while keeping labels, privacy checks, and release decisions explicit.
Teaching Mistral 7B to Normalize Messy Invoice Lines
A practical LoRADock note covering inputs, validation, tradeoffs, and release criteria for teaching mistral 7b to normalize messy invoice lines.
Building a Qwen Adapter for Product Taxonomy Mapping
A practical LoRADock note covering inputs, validation, tradeoffs, and release criteria for building a qwen adapter for product taxonomy mapping.
Fine-Tuning Gemma 2 for Internal Policy Questions
A practical LoRADock note covering inputs, validation, tradeoffs, and release criteria for fine-tuning gemma 2 for internal policy questions.
Improving SQL Generation with Schema-Aware Examples
A practical LoRADock note covering inputs, validation, tradeoffs, and release criteria for improving sql generation with schema-aware examples.
Making Support Summaries Shorter Without Losing Actions
A practical LoRADock note covering inputs, validation, tradeoffs, and release criteria for making support summaries shorter without losing actions.