译站

• v3.0.0 Big update, please try it. From this version onwards, the translation station only supports 64-bit systems.
• New New **Local Big Model** translation, works even without internet! - Support 3 big model presets by default, you can also import GGUF models by yourself, recommended 4B and below! - Can be used for text or image translation (image translation is not recommended as it consumes a lot of memory and is extremely slow), supports switching deep thinking. - It can be used for text or image translation (image translation is not recommended as it consumes memory and is fast, and supports deep thinking on/off). - The use of local large models requires the download of runtime and model files, which can be downloaded in the application itself. If you need to enable image translation, you need to download the mmproj file. - Refactoring Fully **Refactored Hover Window**: - Text hover window now supports one-click translation of clipboard content - The screen translation hover window has been completely rewritten to support localized translation, full-screen translation and automatic translation. - Hover windows now support adjusting width and transparency. Optimization The experience of long text translation is greatly optimized. - Support selecting a certain paragraph for retranslation - Pause immediately when pausing, instead of waiting for the output to finish. - Optimize auto-scrolling effect - Optimization The control button no longer rotates with the screen when the photo is rotated. - Optimize Dialogue translation and in-depth thinking are now automatically scrolled to the bottom. - Optimize the icon of the read-aloud button - Optimization Optimization of auto-scrolling - Optimize Avatar selection box optimization - Optimization Improved smoothness of results page and dialog translation page - Fixed the problem that there was no sound or failed to read aloud when large models were read aloud. - Fix the problem that customized images cannot be selected for themes - Fixed the problem that the data is not updated when entering again in some states of comic translation and long text translation. - Added a new large model record notice and AI generation logo. - Added a read aloud button on the input page, which will read aloud according to the "automatic" language.

Write a review