Loner Waifu GPT

This is AI Chat that you can interact by sending the input via the message and AI model will reply the message back to you via the speech.

Highlevel Flow

                         +-------------------+
                         |   Ollama Server   |
                         +-------------------+
                                  ▲
                                  |  (Streaming Text Response)
                                  |
+-------------------+   1. Sends  |
|   Frontend (React)| ----------> |
|                   |             |
+-------------------+             |
        |                         |
        | 2. Sends text to        |
        | Python TTS Service      |
        v                         |
+-----------------------+         |
|   Python TTS Service  |         |
+-----------------------+         |
        |                         |
        | 3. Returns audio path   |
        v                         |
+-----------------------+         |
|  Audio File (WAV/MP3) |         |
+-----------------------+         |
        |                         |
        v                         |
  4. Frontend plays audio         |
                                  |

Installation

Frontend

1. Install Node version >= 20
2. Running npm i to install all dependencies

Python TTS service

1. Running this command to create virtual env for python python3 -m venv myenv
2. Install Python version 3.10
3. Install library using pip3 install -r requirement.txt
4. Install Pytorch or using

pip3 install torch==2.4.0 torchvision==0.19.0 torchaudio==2.4.0 

5. Install transformer

 pip3 install transformers==4.39.3

6. Install ffmpeg brew install ffmpeg

Running Local

1. Run Ollama first in your local machine
2. Run npm run dev to start FE
3. Run uvicorn main:app --reload to start TTS service

Video Sample

sample.mp4

Library related

Pytorch

Live2D

Text to Speech (TTS)

Model