Additional Parameters
The additional parameters to customise the configuration of model training
The list of parameters that can be included in the configuration settings for model training are as follows:
| Parameter name | Description | Default value | Accepted Values |
|---|---|---|---|
| page_start | This lets you select a start point for the pages you want to work with. Numbering begins from 0, so use '0' for the first page. | 0 | int |
| page_end | Allows you to select an end point for the pages. So, for the fifth page, you would input '4'. Note: To select a single page, set both page_start and page_end to the same value. for e.g to select page number 1 only, then page_start and page_end value must be 1. | None | int |
| sort | Enables sorting of data in order of 'block' or 'line'. | block | string |
| seed | By entering a seed value, you can replicate the same train and test sets. | None | int |
| labels_to_train | Allows you to train a model with filtered list of fields. Params is provided in the form of comma separated values of { section name (3 ) field name }: section1_field1,section2___field3 | None | string |
| labels_to_skip | Allows you to train a model without the filtered list of fields. Params is provided in the form of comma separated values of { section name (3 ) field name }: section1_field1,section2___field3 | None | string |
Updated 4 months ago