Translator++ Ver. 5.4.30 – Let AI Localize it!

Hello fellow translators,

Translator++ has successfully reached a new milestone this month in the development of the OpenAI add-on. Since OpenAI launched ChatGPT last November, many have anticipated the possibility of using this generative AI model as a translation assistant or automatic translation machine.

OpenAI Chat and OpenAI Text Complete

This time I am happy to announce that Translator++ now supports OpenAI Chat and OpenAI Text Complete as Translator Engines. The Chat GPT plug-in for Translator++ has been updated to version 0.2.2 and renamed to OpenAI Chat and OpenAI text complete. In this version, you are free to set how the completion responds. Also, an error when translating using OpenAI will not cause the batch translation process to stop or hang. And we have added an escape code algorithm for : HexPlaceholder, XML Cloaking, HTML Cloaking. (Escaping algorithm for inline code inside dialogues.)

The default settings for OpenAI Chat and OpenAI Text Complete.

For those of you who don’t know the difference between Chat and Text Complete, OpenAI Chat and Text Complete are two different ways of using OpenAI’s language models to generate or manipulate text. OpenAI Chat is a new format that allows you to interact with the model in a conversational way, using messages with different roles (system, user, or assistant).

OpenAI Chat uses gpt-3.5-turbo as the default model, which is a more powerful and faster version of gpt-3. Text Complete is the original format that takes a text prompt as input and returns a text completion that matches the context or pattern of the prompt. Text Complete uses text-davinci-003 as the default model, which is an instruction following model that can perform a wide variety of tasks.

Change Logs

[5.4.30]

  • Update: TransChatGPT ver. 0.2.2
  • Fix : OpenAI chatGPT : Error will not stop the process
  • Fix : OpenAI Text completion : Error will not stop the process
  • Fix : OpenAI chatGPT : Error will be printed in the process log
  • Fix : OpenAI Text completion : Error will be printed in the process log
  • Add : OpenAI chatGPT : Added option to set batch delay
  • Add : OpenAI Text completion : Added option to set batch delay
  • Fix : Bug when aborting translation – again

[5.4.26]

  • Fix : Bug when aborting translation
  • Fix : Corrected behavior of Row-by-row translation. No selection = all.
  • Update: Custom parser ver. 0.6.3
  • Update: XMLTrans ver 0.2
  • Fix : XMLTrans – The structure of the script changed to support thread workers

[5.4.19]

  • Update : SugoiTrans ver. 0.5.16
  • Fix : SugoiTrans: Initialization error for the Server Manager
  • Update : TransChatGPT ver. 0.2.1
  • Add : ChatGPT – Options to set temperature, top P, presence penalty, frequency penalty for OpenAI Chat
  • Add : ChatGPT – Options to set temperature, top P, presence penalty, frequency penalty for OpenAI Text completion

[5.4.18]

  • Fix : Fix the deprecated buffer handler that will causes intermittent crash
  • Update : RMRGSS ver. 2.8
  • Add : Changed the core script to support thread workers
  • Fix : Bug when aborting translation
  • Update : ChatGPT ver. 0.2
  • Add : ChatGPT – OpenAI Complete API support added. This translator engine that supports Davinci model.
  • Fix : ChatGPT – Escape algorithm “none” is fixed.
  • Fix : ChatGPT – ChatGPT for Translator++ renamed to OpenAI Chat
  • Add : ChatGPT – HexPlacecholder, XML Cloaking, HTML Cloaking escape algorithm added for OpenAI Chat
  • Add : ChatGPT – Added some information regarding templating variables
  • Add : ChatGPT – Added some variable for templating
  • Deprecated : Accessing instance of translator engine through trans[translatorEngineID] will be removed. Use trans.getTranslatorEngine(translatorEngineID) or TranslatorEngine.translators[translatorEngineID] instead.

[5.4.5]

  • Update : RMRGSS ver. 2.7.12
  • Fix : RMRGSS : Open 3 thread regardless the setting.
  • Update : Thread communication handler reworked
  • Update : Smaller overhead resource intake for child thread
  • Update : Quicker child thread initialization
  • Update : Caiyun Translator ver. 0.4.1
  • Fix : Caiyun: Library not properly loaded from addon-installation that resulted in the addon become unusable.

[5.4.2]

  • Update : DeepL ver. 0.6.3
  • Add : DeepL : New language support: Korean, Lithuanian
  • Update : DeepL Pro ver. 0.4.3
  • Add : DeepLPro : New language support: Korean, Lithuanian
  • Add : Options menu’s search can search the content of the tab too