Custom Translator upgrade and Regional data residency
Published date: August 03, 2020
Custom Translator (a Microsoft Translator feature) has updates, bug fixes, improvements, and new and refreshed models for August 2020.
Custom Translator service upgrade provides Microsoft Translator’s state-of-the-art neural machine translation (NMT) architecture, the same architecture that powers translation for Microsoft services such as Office 365, Teams, Speech Services, and Bing Translator, along with other services. We are also enabling regional data residency so that you can keep your data or your customer data in the region of your choice in order to meet your corporate data security, data privacy, and regulatory requirements.
Improvements and changes
- We’ve added regional data residency to provide you with the ability to keep data in the Azure region you selected when you created your Translator resource.
- We’ve improved the translation quality of dictionary-only models by adding the ability to customize our standard Translator models with a dictionary document type for all languages in the list below*.
- This feature is only enabled when you train models with dictionary documents type (phrase dictionary, sentence dictionary or both).
- If you have trained a dictionary only models, your models will be auto-upgraded free of charge with zero downtime deployment. You will receive an email from custommt@microsoft.com after your model is upgraded.
*Language pair list
Source Language | Target Language |
Arabic (Saudi Arabia) | English |
Assamese | English |
Bulgarian (Bulgaria) | English |
Bangla (Bangladesh) | English |
Czech (Czech Republic) | English |
Danish (Denmark) | English |
German | English |
Greek (Greece) | English |
English | Arabic (Saudi Arabia) |
English | Bulgarian (Bulgaria) |
English | Bangla (Bangladesh) |
English | Czech (Czech Republic) |
English | Danish (Denmark) |
English | German |
English | Greek (Greece) |
English | Spanish (Spain) |
English | Estonian (Estonia) |
English | Finnish (Finland) |
English | French |
English | Irish (Ireland) |
English | Gujarati (India) |
English | Hindi (India) |
English | Croatian (Croatia) |
English | Hungarian (Hungary) |
English | Armenian (Armenia) |
English | Italian (Italy) |
English | Japanese (Japan) |
English | Georgian (Georgia) |
English | Kazakh (Kazakhstan) |
English | Kannada (India) |
English | Korean (Korea) |
English | Lithuanian (Lithuania) |
English | Latvian (Latvia) |
English | Malayalam (India) |
English | Marathi (India) |
English | Maltese (Malta) |
English | Norwegian, Bokmål (Norway) |
English | Dutch (Netherlands) |
English | Punjabi (India) |
English | Polish (Poland) |
English | Portuguese (Brazil) |
English | Portuguese (Portugal) |
English | Romanian (Romania) |
English | Russian (Russia) |
English | Slovak (Slovakia) |
English | Swedish (Sweden) |
English | Tamil (India) |
English | Telugu (India) |
English | Thai (Thailand) |
English | Turkish (Turkey) |
English | Chinese (People's Republic of China) |
Spanish (Spain) | English |
Finnish (Finland) | English |
French | English |
Irish (Ireland) | English |
Gujarati (India) | English |
Hindi (India) | English |
Croatian (Croatia) | English |
Hungarian (Hungary) | English |
Italian (Italy) | English |
Japanese (Japan) | English |
Georgian (Georgia) | English |
Kazakh (Kazakhstan) | English |
Korean (Korea) | English |
Lithuanian (Lithuania) | English |
Malayalam (India) | English |
Marathi (India) | English |
Maltese (Malta) | English |
Norwegian, Bokmål (Norway) | English |
Dutch (Netherlands) | English |
Oriya (India) | English |
Punjabi (India) | English |
Polish (Poland) | English |
Portuguese (Brazil) | English |
Romanian (Romania) | English |
Russian (Russia) | English |
Slovak (Slovakia) | English |
Swedish (Sweden) | English |
Tamil (India) | English |
Telugu (India) | English |
Thai (Thailand) | English |
Turkish (Turkey) | English |
Urdu (Islamic Republic of Pakistan) | English |
Chinese (People's Republic of China) | English |