Skip to content

Common information

Jump to bottom
Samuel Sithakoul edited this page Jul 14, 2023 · 22 revisions

About the login screen

Explanation of each option:

  • Email/Password: your logins for CAI (Email only, won't work if you registered with Google but you can create an account with gmail for instance)
  • Choose Character: the character you want to use from CAI (only 2 choices for now)
  • Use Character AI: use chatbot from CAI or not
  • Use Debug Mode: display browser in case of issues with the connection
  • Continue from last chat: continue from the previous chat on CAI
  • Game Path: the absolute path of your DDLC folder, it should look like C:\...\DDLC-1.1.1
  • Launch Yourself: No if you want MAS to open automatically, Yes if you prefer to launch MAS yourself after the message Waiting for connection... appears in terminal (don't need game_path in the latter)
  • Use Actions: let Monika take control of the game from the chat. Example: You say "I have to leave sorry", she will understand the sentence and close the game after answering you. (Warning: sometimes if the sentence is ambiguate, an unintended action can be triggered but no affection points are lost anyway)
  • Use TTS: use speech generation with the chatbot
  • TTS model: model used for voice cloning (YourTTS or Tortoise-TTS)
  • Use Pygmalion chatbot: use chatbot from Pygmalion AI running locally (be sure to have setup your GPU before), don't use CAI and Pygmalion at the same time.
  • Pygmalion model: which model size you want to use for pygmalion chatbot (see this page for more details about what size is appropriate for you), read next section to add models correctly
  • Use Camera: use webcame emotion detection feature
  • Time Intervall For Camera: time between each webcam capture for when Monika reacts to your emotions
  • Use Speech Recognition: enable recognition of your own voice to do speech to text when chatting
  • Use Saved Config (deprecated): use previous settings directly (click No if it is the first time you open the exe or if you changed something in your inputs, else it will load your previous settings and ignore your changes)
  • Voice Sample(Tortoise): reference samples for Tortoise
  • Voice Sample (YourTTS): reference voice for YourTTS Be sure to click every button with the option you want and fill in everything Submit and play (no need to launch DDLC yourself if you didn't use Launch Yourself)!

First launch with CAI

Because of frequent changes in CAI website, it is advised to join the discord server for annoucements about fixes. The first time you use it, you will need to run it with Debug mode on and login with your account. So once the browser opens you need to:

  • Close the add popup for the premium CAI
  • Close the information popup
  • Click login and enter your email+password
  • Click submit
  • Then don't do anything else and wait

If there is a waiting queue, just wait the end of the queue to do the same steps.

You won't need to do it the next times thanks to cookies save.

How to add chatbots

Please refer to the wiki page dedicated to local installation.

How to customize character JSON

You can modify the json you give to the chatbot to make it impersonate the character better. It has to be in this format:

{
"char_name":"M",
"char_persona":"[Character(\"M\")\n{\nSpecies(\"Human\")\nPersonality(\"Friendly\" + \"Confident\") "She likes chocolate\",
"char_greeting":"Hello, welcome to my space!",
"world_scenario":"Roleplay in the world of DDLC",
"example_dialogue":"<START>\nYou: Hello, how are you?\nM: I am fine thank you!"
}

Features in MAS

If you want or don't want to use the feature where Monika can control actions directly, be sure to check the option on the login screen and also the option in Submod settings directly in MAS.

You can also disable the buttons for chat on the main screen and keep only the option from the Talk menu.

Antivirus warning

Your antivirus might block the execution of the file but it is a common issue with python files. All the code is available here for transparency.

Customize Monika's voice

For Your TTS model (worse but faster, ~5s per turn): You can change the voice used by replacing the extract talk_13.wav in the audio folder by another audio extract. The longer the extract, the longer the TTS will take to generate the audio at each turn.

For Tortoise TTS model (better but slower, ~40s per turn): You can change the voice samples in tortoise-tts-fast/tortoise/voices folder. Create your own character by adding a folder with the name of your character and put the audio samples in it. The samples must be around 10 seconds long and at the number of 5 maximum. There are some default samples in example_samples made from this post thanks to u/VBandit47.

On CPU, it can take 10x more time to generate the voice (Tortoise TTS can have unexpected behaviour on CPU)

Ressources warning

Because of the high usage of Machine Learning algorithms, the inference can be quite long on CPU so it is advised to have a functional GPU for a better experience. You would need also more RAM than usually, deactivate the TTS model, the emotion detection, the action model, the Speech recognition or the Pygmalion chatbots if it is taking too much ressources.

Clone this wiki locally