> ## Documentation Index
> Fetch the complete documentation index at: https://docs.abbyy.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Transferring training results into another project

> Reuse FlexiCapture for Invoices training results in another project by importing training batches and mapping them to vendors via VendorId, VATID, or IBAN identifiers.

In FlexiCapture for Invoices, you can reuse training results across multiple projects. To use your training results in another project, import the training batches into that project.

A separate training batch is created for documents originating from a company. So when you import your training batches:

* For invoices, map each training batch to its company in the **Vendors** database.
* For purchase orders, map each training batch to its company in the **Variants** database.

## Map training batches to vendors

Map each training batch to its vendor using one of these identifiers.

### Map by VendorId

In FlexiCapture for Invoices, and in the external system where the data captured from invoices is exported, vendors are identified by the value of **VendorId**.

When these identifiers are the same in both projects (which can happen when both projects relate to the same external system), enable the **Preserve Variant ID** option in the **Import Training Batches** dialog box (**Field Training** → **Import Training Batches**). This retains **VendorId** identifiers for all batches.

Do not use the **Preserve Variant ID** option if the project you are importing into uses different **VendorId** identifiers, or if **VendorId** identifiers depend on the business unit.

### Map by VATID and IBAN

Training batches cannot be matched with their vendors based on **VendorId** if:

* Your projects export data into different external systems (because different vendor lists are used).
* **VendorId** is different for different business units.

To associate a training batch with its vendor, use other unique vendor identifiers, such as VATID and IBAN. A training batch stores the VATID and IBAN values of its vendor.

When a training batch is imported into a new batch, the program automatically searches for a vendor with the same VATID or IBAN value. If it finds an entry in the **Vendors** data set where at least one of these values matches, it associates that vendor with the batch.

## Import training batches into a new project

<Steps>
  <Step title="Open the target project">
    In the Project Setup Station of FlexiCapture Distributed or the Administrator Station of FlexiCapture Standalone, open a new project.
  </Step>

  <Step title="Open Import Training Batches">
    On the main menu, click **Field Training** → **Import Training Batches**.
  </Step>

  <Step title="Choose the mapping method">
    Depending on the method used to map the training batches to their companies, select or clear the **Preserve Variant ID** option in the **Import Training Batches** dialog box.
  </Step>

  <Step title="Specify the source project">
    Specify the path to the project that contains the training batches to import.
  </Step>

  <Step title="Select the source Document Definition">
    If the source project has more than one Document Definition, select the one to import training batches from.
  </Step>

  <Step title="Select the target Document Definition">
    If the target project has more than one Document Definition, select the one to import the training batches for.
  </Step>
</Steps>

<Note>
  You can associate an imported training batch with a Document Definition manually. To do this, right-click the training batch and click **Change Document Definition** on the shortcut menu. You can also associate a training batch with its company.
</Note>

The training batches are imported into the new project, along with:

* The trained FlexiLayout variants.
* The sample images that were used for training.

You can further train imported FlexiLayout variants by adding new document images besides the imported ones.

<Note>
  If you import a training batch for a company that already has a training batch in the new project, the program selects the best training batch: the documents in each batch are matched with the trained FlexiLayout variant from that batch, and the batch with the greater number of correctly detected fields wins.
</Note>