Model Training
Last updated
Last updated
Model Training allows administrators to oversee and manage the training of machine learning models specific to each document type. By providing a structured interface for importing sample data, training models, and testing their performance, Docbits ensures that its data extraction capabilities continuously improve over time.
Metrics Overview:
Sample: Number of sample documents used for training.
Exported: Number of documents that have been successfully exported after processing.
Company Σ: Total number of company-specific documents processed.
Overall Σ: Total number of documents processed across all categories.
Training and Testing Options:
Import: Allows administrators to import new training data sets which are typically structured samples of documents that should be recognized by the system.
Train Model: Starts the training process using the imported data to enhance the recognition and extraction capabilities of the system.
Test Classification: Enables testing of the model to evaluate its performance on classifying and extracting data from new or unseen documents.
Action Buttons:
Create Field: Add new data fields that the model should recognize and extract.
Actions: This dropdown can include options such as view details, edit configurations, or delete training data.
