BigQuery now supports fine-tuning of models hosted in Vertex AI. This allows for customization and improvement of model performance in scenarios where prompt engineering is not sufficient. A medical transcription dataset is used as an example to illustrate the process of fine-tuning a model in BigQuery. The baseline performance of the model is evaluated, and then the model is fine-tuned. The performance of the fine-tuned model is compared to the baseline model, showing an improvement in F1 scores. The fine-tuned model can be used for inference without the need for managing additional infrastructure.
