August 18th Update Announcement

Ai Vpet

The new generation of table pets in the AI era provides users with a comprehensive AI customized platform for creating table pets, including image, voice, character design, and even interactive methods, through AI technologies such as deep learning and language processing.

Hello everyone~~Another week has passed, and after enduring the torment of two planners and staying up all night for three nights, we have finally expanded our computing power by four times. Although we sacrificed five graphics cards and six programmers, we still announced the following eight important updates before 7 o'clock: [img]{STEAM_CLAN_IMAGE}/45073139/b324533b9fa0e3c7394cac06e11d09f4c863676d.png[/img] 1. The voice is no longer difficult to pronounce, the Japanese and English pronunciation of the 9 voices have been optimized (solve the accent problem of local people speaking foreign languages) 2. Character archiving is optimized, each model character can be archived separately, and the model’s work description can be used as the character (solve the problem of models sharing the same soul and the author being unable to share the soul) 3. The character design capacity has been expanded to 1,000 words (solving the problem that the character cannot be more specific and detailed) 4. Keyword wake-up optimization, the optimization removes nicknames, and now the character name can be used to wake up voice interaction (solve the problem that the software sometimes pretends not to hear human words and acts like the old 6) 5. Added the function of recognizing music, playing dance, and then continuing to play music for 5 seconds (solved the problem of software not understanding music players) 6. Dialogue input is expanded 4 times (solve the problem of not being able to talk too much) 7. Intelligent semantic recognition, as long as the meaning of opening, web search, and hiding is expressed in the semantics, the corresponding operations can be performed without using three fixed keywords (solving the problem of having to use secret codes to communicate) 8. Live2d can recognize expressions when uploading, and will automatically convert expression files into action files for use (solving the problem of live2d models with only expressions being homeless) [h3]Special thanks:[/h3] Finally, we would like to express our special thanks to each of the following users for their help, support, and thoughtful suggestions: [img]{STEAM_CLAN_IMAGE}/45073139/71e63fcb54e7a25419a4852930c26a037ad5a1e9.png[/img] Test server: The test server is about to start testing the pendant mode.. Interested friends can join the group and enter the test server to experience it first. [h3]Next Preview:[/h3] What is an intelligent interactive system ——Eating, sleeping, and broadcasting actions × ——Watch, hear, ask, express concern √