Skip to content

Steps-to-run.md

Latest commit

 

History

History
103 lines (85 loc) · 3.22 KB

Steps-to-run.md

File metadata and controls

103 lines (85 loc) · 3.22 KB

Commands : Terminal 1 : cd workspace git clone https://github.com/devasphn/Vocalis cd Vocalis cd backend python3 -m venv env source env/bin/activate apt update && apt install -y git python3 python3-pip python3-venv nodejs npm curl wget ffmpeg pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 pip install fastapi==0.109.2 uvicorn==0.27.1 python-dotenv==1.0.1 websockets==12.0 numpy==1.26.4 transformers faster-whisper==1.1.1 requests==2.31.0 python-multipart==0.0.9 torch>=2.0.1 ffmpeg-python==0.2.0 aiohttp==3.9.1 CMAKE_ARGS="-DLLAMA_CUDA=on" pip install llama-cpp-python[server] --force-reinstall --no-cache-dir mkdir -p /workspace/models cd /workspace/models wget https://huggingface.co/lex-au/Vocalis-Q4_K_M.gguf/resolve/main/Vocalis-q4_k_m.gguf cd /workspace/Vocalis-WebRTC/backend cat > .env << 'EOF' LLM_API_ENDPOINT=http://0.0.0.0:1234/v1/chat/completions TTS_API_ENDPOINT=http://0.0.0.0:5005/v1/audio/speech/stream WHISPER_MODEL=base TTS_MODEL=tts-1 TTS_VOICE=ऋतिका TTS_FORMAT=wav WEBSOCKET_HOST=0.0.0.0 WEBSOCKET_PORT=8000 VAD_THRESHOLD=0.1 VAD_BUFFER_SIZE=30 AUDIO_SAMPLE_RATE=44100 EOF python -m llama_cpp.server --model /workspace/models/Vocalis-q4_k_m.gguf --host 0.0.0.0 --port 1234 --n_gpu_layers -1 --n_ctx 4096 &

Test a simple completion to measure speed

curl http://0.0.0.0:1234/v1/chat/completions
-H "Content-Type: application/json"
-d '{ "model": "Vocalis-q4_k_m.gguf", "messages": [{"role": "user", "content": "What is your name?"}], "max_tokens": 100 }'
-w "\n%{time_total} seconds\n"

cd /workspace/Vocalis python -m backend.main

for shutting down llama : pkill -f llama

Terminal 2: cd workspace git clone https://github.com/devasphn/Orpheus-FastAPI/ cd Orpheus-FastAPI/ python3 -m venv tts_env source tts_env/bin/activate apt-get update apt-get install -y portaudio19-dev libasound2-dev libsndfile1-dev libportaudio2 libportaudiocpp0 pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 pip install fastapi uvicorn transformers numpy soundfile librosa python-dotenv sounddevice psutil snac einops python-multipart CMAKE_ARGS="-DLLAMA_CUDA=on" pip install llama-cpp-python[server] --force-reinstall --no-cache-dir mkdir -p /workspace/models cd /workspace/models wget https://huggingface.co/lex-au/Orpheus-3b-FT-Q8_0.gguf/resolve/main/Orpheus-3b-FT-Q8_0.gguf cd /workspace/Orpheus-FastAPI cat > .env << 'EOF' ORPHEUS_API_URL=http://0.0.0.0:1235/v1/completions ORPHEUS_API_TIMEOUT=120 ORPHEUS_MAX_TOKENS=8192 ORPHEUS_TEMPERATURE=0.6 ORPHEUS_TOP_P=0.9 ORPHEUS_SAMPLE_RATE=24000 ORPHEUS_MODEL_NAME=Orpheus-3b-FT-Q8_0.gguf ORPHEUS_PORT=5005 ORPHEUS_HOST=0.0.0.0 EOF python -m llama_cpp.server --model /workspace/models/Orpheus-3b-FT-Q8_0.gguf --host 0.0.0.0 --port 1235 --n_gpu_layers -1 &
python app.py --host 0.0.0.0 --port 5005

Terminal 3:

apt clean apt autoclean

apt purge -y nodejs* npm* libnode* node-* apt autoremove -y

rm -rf /usr/lib/node_modules rm -rf /usr/local/lib/node_modules rm -rf ~/.npm

curl -fsSL https://deb.nodesource.com/setup_18.x | bash - apt install -y nodejs

node --version npm --version

cd /workspace/Vocalis/frontend

npm install npm run dev -- --host 0.0.0.0 --port 3000