Ajuste fino

AI Center includes the capability of fine-tuning ML models using data that has been validated by a human using Validation Station.

Como seu fluxo de trabalho de RPA processa documentos usando um modelo de ML existente, alguns documentos podem exigir validação manual usando a atividade Present Validation Station (disponível em robôs assistidos ou no navegador usando o Orchestrator Action Center).

The validated data generated in Validation Station can be exported using Machine Learning Extractor Trainer activity, and can be used to fine-tune ML models in AI Center.

Não recomendamos treinar modelos de ML do zero (ou seja, o Pacote de ML DocumentUnderstanding) usando dados do Validation Station, mas sim apenas para ajustar modelos de ML existentes (incluindo modelos prontos para uso).

For the detailed steps involved in fine-tuning an ML model see the Import Documents section of the Data Manager documentation.

For more details about how to build a dataset for fine-tuning, go here.

Importante:

It if often wrongly assumed that the way to use Validation Station data is to iteratively retrain the previous model version, so the current batch is used to train package X.1 to obtain X.2. Then the next batch trains on X.2 to obtain X.3 and so on. This is the wrong way to use the product. Each Validation Station batch needs to be imported into the same Data Manager session as the original manually labeled data making a larger dataset, which must be used to train always on the X.0 ML Package version.