mirror of
https://github.com/jamiepine/voicebox.git
synced 2026-09-16 05:10:42 -07:00
Implement TTS provider management system and update release workflow
- Added support for TTS providers in the backend, including endpoints for listing, starting, stopping, and downloading providers. - Enhanced the release workflow to build and upload TTS provider binaries for both Windows and Linux platforms. - Updated the architecture documentation to reflect the new provider system and its benefits for modularity and user experience. - Introduced a new `ProviderSettings` component in the frontend for managing provider configurations.
This commit is contained in:
Jamie Pine
+132
-12
@@ -6,7 +6,114 @@ on:
|
||||
tags:
|
||||
- "v*"
|
||||
|
||||
env:
|
||||
PROVIDER_VERSION: "1.0.0"
|
||||
|
||||
jobs:
|
||||
# ============================================
|
||||
# Build TTS Providers (uploaded to R2, not GitHub)
|
||||
# ============================================
|
||||
build-providers:
|
||||
runs-on: ${{ matrix.platform }}
|
||||
strategy:
|
||||
fail-fast: false
|
||||
matrix:
|
||||
include:
|
||||
# PyTorch CPU provider (Windows)
|
||||
- platform: "windows-latest"
|
||||
provider: "pytorch-cpu"
|
||||
python-version: "3.12"
|
||||
# PyTorch CUDA provider (Windows) - large binary, uploaded to R2
|
||||
- platform: "windows-latest"
|
||||
provider: "pytorch-cuda"
|
||||
python-version: "3.12"
|
||||
# PyTorch CPU provider (Linux)
|
||||
- platform: "ubuntu-22.04"
|
||||
provider: "pytorch-cpu"
|
||||
python-version: "3.12"
|
||||
# PyTorch CUDA provider (Linux) - large binary, uploaded to R2
|
||||
- platform: "ubuntu-22.04"
|
||||
provider: "pytorch-cuda"
|
||||
python-version: "3.12"
|
||||
|
||||
steps:
|
||||
- uses: actions/checkout@v4
|
||||
|
||||
- name: Install dependencies (ubuntu only)
|
||||
if: matrix.platform == 'ubuntu-22.04'
|
||||
run: |
|
||||
sudo apt-get update
|
||||
sudo apt-get install -y llvm-dev
|
||||
|
||||
- name: Setup Python
|
||||
uses: actions/setup-python@v5
|
||||
with:
|
||||
python-version: ${{ matrix.python-version }}
|
||||
cache: "pip"
|
||||
|
||||
- name: Install Python dependencies (CPU)
|
||||
if: matrix.provider == 'pytorch-cpu'
|
||||
run: |
|
||||
python -m pip install --upgrade pip
|
||||
pip install pyinstaller
|
||||
pip install -r providers/pytorch-cpu/requirements.txt
|
||||
pip install -r backend/requirements.txt
|
||||
|
||||
- name: Install Python dependencies (CUDA)
|
||||
if: matrix.provider == 'pytorch-cuda'
|
||||
run: |
|
||||
python -m pip install --upgrade pip
|
||||
pip install pyinstaller
|
||||
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
|
||||
pip install -r providers/pytorch-cuda/requirements.txt
|
||||
pip install -r backend/requirements.txt
|
||||
|
||||
- name: Build provider binary
|
||||
shell: bash
|
||||
run: |
|
||||
cd providers/${{ matrix.provider }}
|
||||
python build.py
|
||||
|
||||
- name: Upload provider to R2
|
||||
shell: bash
|
||||
env:
|
||||
R2_ACCESS_KEY_ID: ${{ secrets.R2_ACCESS_KEY_ID }}
|
||||
R2_SECRET_ACCESS_KEY: ${{ secrets.R2_SECRET_ACCESS_KEY }}
|
||||
R2_ENDPOINT: ${{ secrets.R2_ENDPOINT }}
|
||||
run: |
|
||||
# Install AWS CLI (compatible with R2)
|
||||
pip install awscli
|
||||
|
||||
# Configure AWS CLI for R2
|
||||
aws configure set aws_access_key_id $R2_ACCESS_KEY_ID
|
||||
aws configure set aws_secret_access_key $R2_SECRET_ACCESS_KEY
|
||||
aws configure set region auto
|
||||
|
||||
# Determine binary name based on platform
|
||||
if [ "${{ matrix.platform }}" == "windows-latest" ]; then
|
||||
BINARY_NAME="tts-provider-${{ matrix.provider }}.exe"
|
||||
BINARY_PATH="providers/${{ matrix.provider }}/dist/tts-provider-${{ matrix.provider }}.exe"
|
||||
else
|
||||
BINARY_NAME="tts-provider-${{ matrix.provider }}"
|
||||
BINARY_PATH="providers/${{ matrix.provider }}/dist/tts-provider-${{ matrix.provider }}"
|
||||
fi
|
||||
|
||||
# Add platform suffix for clarity
|
||||
if [ "${{ matrix.platform }}" == "windows-latest" ]; then
|
||||
UPLOAD_NAME="tts-provider-${{ matrix.provider }}-windows.exe"
|
||||
else
|
||||
UPLOAD_NAME="tts-provider-${{ matrix.provider }}-linux"
|
||||
fi
|
||||
|
||||
# Upload to R2 (bucket: voicebox)
|
||||
aws s3 cp "$BINARY_PATH" "s3://voicebox/providers/v${{ env.PROVIDER_VERSION }}/$UPLOAD_NAME" \
|
||||
--endpoint-url "$R2_ENDPOINT"
|
||||
|
||||
echo "Uploaded $UPLOAD_NAME to R2"
|
||||
|
||||
# ============================================
|
||||
# Build Main App (without bundled TTS on Win/Linux)
|
||||
# ============================================
|
||||
release:
|
||||
permissions:
|
||||
contents: write
|
||||
@@ -14,22 +121,26 @@ jobs:
|
||||
fail-fast: false
|
||||
matrix:
|
||||
include:
|
||||
# macOS Apple Silicon - MLX bundled (works out of the box)
|
||||
- platform: "macos-latest"
|
||||
args: "--target aarch64-apple-darwin"
|
||||
python-version: "3.12"
|
||||
backend: "mlx"
|
||||
# macOS Intel - PyTorch bundled (smaller user base, keep simple)
|
||||
- platform: "macos-15-intel"
|
||||
args: "--target x86_64-apple-darwin"
|
||||
python-version: "3.12"
|
||||
backend: "pytorch"
|
||||
# Linux - No TTS bundled, providers downloaded separately
|
||||
# - platform: 'ubuntu-22.04'
|
||||
# args: ''
|
||||
# python-version: '3.12'
|
||||
# backend: 'pytorch'
|
||||
# backend: 'none'
|
||||
# Windows - No TTS bundled, providers downloaded separately
|
||||
- platform: "windows-latest"
|
||||
args: ""
|
||||
python-version: "3.12"
|
||||
backend: "pytorch"
|
||||
backend: "none"
|
||||
|
||||
runs-on: ${{ matrix.platform }}
|
||||
|
||||
@@ -55,23 +166,27 @@ jobs:
|
||||
python-version: ${{ matrix.python-version }}
|
||||
cache: "pip"
|
||||
|
||||
- name: Install Python dependencies
|
||||
- name: Install Python dependencies (with TTS)
|
||||
if: matrix.backend != 'none'
|
||||
run: |
|
||||
python -m pip install --upgrade pip
|
||||
pip install pyinstaller
|
||||
pip install -r backend/requirements.txt
|
||||
|
||||
- name: Install Python dependencies (without TTS)
|
||||
if: matrix.backend == 'none'
|
||||
run: |
|
||||
python -m pip install --upgrade pip
|
||||
pip install pyinstaller
|
||||
# Install base requirements without PyTorch/Qwen-TTS
|
||||
pip install fastapi uvicorn sqlalchemy librosa soundfile numpy httpx
|
||||
pip install huggingface_hub # For Whisper downloads
|
||||
|
||||
- name: Install MLX dependencies (Apple Silicon only)
|
||||
if: matrix.backend == 'mlx'
|
||||
run: |
|
||||
pip install -r backend/requirements-mlx.txt
|
||||
|
||||
# - name: Install PyTorch with CUDA (Windows only)
|
||||
# if: matrix.platform == 'windows-latest'
|
||||
# run: |
|
||||
# pip install torch --index-url https://download.pytorch.org/whl/cu121 --force-reinstall --no-deps
|
||||
# pip install torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
|
||||
|
||||
- name: Build Python server (Linux/macOS)
|
||||
if: matrix.platform != 'windows-latest'
|
||||
run: |
|
||||
@@ -148,10 +263,15 @@ jobs:
|
||||
See the assets below to download and install this version.
|
||||
|
||||
### Installation
|
||||
- **macOS (Apple Silicon)**: Download the `aarch64.dmg` file - uses MLX for fast native inference
|
||||
- **macOS (Apple Silicon)**: Download the `aarch64.dmg` file - uses MLX for fast native inference (works out of the box)
|
||||
- **macOS (Intel)**: Download the `x64.dmg` file - uses PyTorch
|
||||
- **Windows**: Download the `.msi` installer
|
||||
- **Linux**: Download the `.AppImage` or `.deb` package
|
||||
- **Windows**: Download the `.msi` installer - requires downloading a TTS provider on first use
|
||||
- **Linux**: Download the `.AppImage` or `.deb` package - requires downloading a TTS provider on first use
|
||||
|
||||
### TTS Providers (Windows/Linux)
|
||||
Windows and Linux users will be prompted to download a TTS provider on first launch:
|
||||
- **PyTorch CPU** (~300MB) - Works on any system
|
||||
- **PyTorch CUDA** (~2.4GB) - 4-5x faster on NVIDIA GPUs
|
||||
|
||||
The app includes automatic updates - future updates will be installed automatically.
|
||||
releaseDraft: true
|
||||
|
||||
Reference in New Issue
Block a user