AI Glossary
Further training of a pretrained model on a smaller, specific dataset to adapt its behavior, style or skills to a particular task.
Fine-tuning is useful for consistent formats, a particular tone, domain jargon or narrow tasks where prompting falls short. Parameter-efficient methods such as LoRA train only small adapter weights, making it affordable on a single GPU.
For adding knowledge that changes often, retrieval is usually better than fine-tuning. See the LoRA fine-tuning overview.
Off-the-shelf assistants, API integrations, retrieval-augmented generation, fine-tuning or your own model: the trade-offs in cost, control and time to value.
thelatentdaily.com Editorial Team3 min read
When fine-tuning helps, how LoRA and QLoRA make it affordable, how to prepare data, train, evaluate and deploy an adapter — and when not to fine-tune at all.
thelatentdaily.com Editorial Team3 min read