Formula

Application Formula for Fine-Tuned BERT Models

Once the fine-tuning optimization is complete, the resulting optimized parameters, ω~\tilde{\omega} and θ~\tilde{\theta}, are utilized to make predictions on new data. For a specific task that the model was tuned for, applying the fine-tuned model during inference is mathematically represented by the formula Predictω~(BERTθ~())\mathrm{Predict}_{\tilde{\omega}}(\mathrm{BERT}_{\tilde{\theta}}(\cdot)). This expression indicates that an input is first processed by the base BERT model using its tuned parameters θ~\tilde{\theta}, and the resulting representations are then passed into the task-specific prediction network, which utilizes its own tuned parameters ω~\tilde{\omega}.

0

1

Updated 2026-04-18

Contributors are:

Who are from:

Tags

Foundations of Large Language Models

Ch.1 Pre-training - Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences