Configurability for translation provider

[kc9jud]

Describe the feature you'd like to request

It would be nice to be able to use a dedicated model for translation tasks. For example, translategemma is a new-ish dedicated model for text-to-text translation: https://blog.google/innovation-and-ai/technology/developers-tools/translategemma/

This also means being able to configure the prompt format. translategemma expects the following prompt template:

You are a professional {SOURCE_LANG} ({SOURCE_CODE}) to {TARGET_LANG} ({TARGET_CODE}) translator. Your goal is to accurately convey the meaning and nuances of the original {SOURCE_LANG} text while adhering to {TARGET_LANG} grammar, vocabulary, and cultural sensitivities.
Produce only the {TARGET_LANG} translation, without any additional explanations or commentary. Please translate the following {SOURCE_LANG} text into {TARGET_LANG}:


{TEXT}

Describe the solution you'd like

The translation feature should have a model selector and a prompt configuration box.

Describe alternatives you've considered

It's possible that a standalone translation provider app could do this, but that seems heavy-handed compared to just configuring the existing translation functionality in this integration.

[lukasdotcom]

Hi @kc9jud,

The model can be changed in the advanced options for the task already. Also currently the prompt for translation is a little special and involves json output to ensure there is no extra text, so it might be a little challenging to allow users to change the prompt because openai integration expects json output.

[kc9jud]

@lukasdotcom I see the choice for general text-processing tasks but not a way to set the default model for translation tasks in particular. Am I just missing something?

[lukasdotcom]

Hi @kc9jud,

Sorry I misunderstood your request. I had thought you were talking about specifying the model for the individual task on this page. I had been talking about this option in the screenshot to choose the model.