What Does This Mean?
Continuing to train a pretrained neural network on a narrower dataset or task, a form of transfer learning. In AI community talk "fine-tune" is used loosely for customizing a model, and it is a mostly technical term with little slang use of its own.
REAL-WORLD EXAMPLE
""They fine-tuned the small model on support tickets and it stopped rambling.""
LORE & ORIGIN
Wikipedia's fine-tuning (deep learning) article describes it as taking a network trained on one objective and continuing training for a narrower follow-on job, updating all layers or only some while the rest stay frozen. It names LoRA as an adapter-based technique. No dated slang first use is documented in the pages opened, so this is a weak entry for slang purposes, kept short because it is a common term in model discussions.
MENTIONS OVER TIME
Radar observation data not yet available for this term.
RELATED SEARCHES
TOP PLATFORMS
Not enough Radar data yet to show a platform breakdown.
WHEN DID YOU FIRST HEAR THIS?
CLICK AN OPTION BELOW TO CAST YOUR VOTE.
[ VOTE TO REVEAL COMMUNITY RESULTS ]