Creating an engine
To create an engine:
- Go to Engines.
- Click Create new.
- Specify the name, languages and the group the new engine will belong to. Unlike corpora, engines must only belong to one group only.
- Based on the above data, Globalese will display a list of corpora to choose from.
- Select the corpora you want to include in the engine.
Try to collect at least 100,000 segment pairs of relevant master corpora for the engine. If the volume is below 100,000 segment pairs, try adding some other resources as auxiliary corpora. Even if they are not 100% relevant to your engine, they can still elevate the overall language quality.
- Optionally, you can leverage stock corpora for certain language combinations.
- Click Save.
Master corpora play an elevated role during training. Globalese will use master corpora as a reference when training the engine. The training process will use segment pairs from the auxiliary corpora that are from the same domain as the master corpus with a higher weight, and others with a lower weight.