onrup

Model architecture

Base model

What is base model?

A base model is the pretrained model a fine-tune starts from. Choosing it fixes the licence, the tokeniser, the chat template, the memory footprint and the ceiling on what the fine-tune can achieve.

The distinction between a base model and an instruction-tuned one matters. A base model predicts text and has no conversational behaviour; an instruction-tuned model already follows instructions. Fine-tuning a base means supplying that behaviour yourself.

Switching base models later is not free. Adapters do not transfer, evaluation results are not comparable, and the chat template changes. Choosing deliberately at the start is worth the time.

Related terms

All terms in the glossary →

Start with the free tier

A magic link creates your account, your tenant and your first API key. No card until you ask for compute.