You can edit the character.json file inside app/engine/ folder suitable of your taste. hehe~ You can edit the emotions.json file inside app/voicevox_api folder.
ONLY USE GGML LLAMA MODEL type for "expert mode" ONLY USE PygmalionAI MODEL type for "Waifu mode"
- download model
- You can download any GGML model. im using eachadea/ggml-vicuna-13b-1.1 at main (huggingface.co)or TheBloke/wizardLM-7B-GGML · Hugging Face
- You can only use PygmalionAI model. im using PygmalionAI/pygmalion-6b · Hugging Face or PygmalionAI/pygmalion-350m · Hugging Face for testing
- move the downloaded models inside app/models note that GGML model are choose one model.bin from repo. for example ggml-vic13b-uncensored-q5_1.bin and PygmalionAI model are the must be include entire folder cloned from repo.
- don't forget to edit the .env file
- Goto the official voicevox/voicevox_engine github
- Goto release tabs and select one that are compatible in your workspace. I'm using windows so I choose windows with GPU + CUDA and download it.
- open terminal and goto the downloaded file earlier.
$ cd D:\playground\windows-nvidia
$ .\run.exe
- If you see something like below means your installations and running success
INFO: Started server process [29360]
INFO: Waiting for application startup.
reading C:\Users\Aqirito\AppData\Local\voicevox-engine\voicevox-engine\tmp6_prectj ... 62
emitting double-array: 100% |###########################################|
done!
INFO: Application startup complete.
INFO: Uvicorn running on http://127.0.0.1:50021 (Press CTRL+C to quit)
We will using voicevox API to synthesize the text read the voicevox before install voicevox/voicevox_engine github voicevox/voicevox_engine docker hub
$ docker pull voicevox/voicevox_engine:cpu-ubuntu20.04-latest
$ docker run --rm -it -p '127.0.0.1:50021:50021' voicevox/voicevox_engine:cpu-ubuntu20.04-latest
$ docker pull voicevox/voicevox_engine:nvidia-ubuntu20.04-latest
$ docker run --rm --gpus all -p '127.0.0.1:50021:50021' voicevox/voicevox_engine:nvidia-ubuntu20.04-latest