利用 AI 大模型和自动化工作流,根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Python ai-video-generatorcontent-creationffmpeginstagram-reelsllmpythonshort-videosubtitlestext-to-speechtiktokvideo-automationvideo-workflowworkflow-automationyoutube-shorts
feat: support GPU acceleration for faster-whisper in Docker
SUMDDOSa301ee799c71
diff --git a/Dockerfile b/Dockerfileindex 4216326a..3f48af37 100644--- a/Dockerfile+++ b/Dockerfile@@ -1,48 +1,33 @@-# Use an official Python runtime as a parent image-FROM python:3.11-slim-bullseye+# Use NVIDIA CUDA runtime as parent image (includes CUDA 12.1 + cuDNN 8)+FROM nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04++# Avoid interactive timezone prompt+ENV DEBIAN_FRONTEND=noninteractive# Set the working directory in the containerWORKDIR /MoneyPrinterTurbo--# 设置/MoneyPrinterTurbo目录权限为777RUN chmod 777 /MoneyPrinterTurboENV PYTHONPATH="/MoneyPrinterTurbo"-# Install system dependencies with domestic mirrors first for stability-RUN echo "deb http://mirrors.aliyun.com/debian bullseye main" > /etc/apt/sources.list && \- echo "deb http://mirrors.aliyun.com/debian-security bullseye-security main" >> /etc/apt/sources.list && \- ( \- for i in 1 2 3; do \- echo "Attempt $i: Using Aliyun mirror"; \- apt-get update && apt-get install -y --no-install-recommends \- git \- imagemagick \- ffmpeg && break || \- echo "Attempt $i failed, retrying..."; \- if [ $i -eq 3 ]; then \- echo "Aliyun mirror failed, switching to Tsinghua mirror"; \- sed -i 's/mirrors.aliyun.com/mirrors.tuna.tsinghua.edu.cn/g' /etc/apt/sources.list && \- sed -i 's/mirrors.aliyun.com\/debian-security/mirrors.tuna.tsinghua.edu.cn\/debian-security/g' /etc/apt/sources.list && \- ( \- apt-get update && apt-get install -y --no-install-recommends \- git \- imagemagick \- ffmpeg || \- ( \- echo "Tsinghua mirror failed, switching to default Debian mirror"; \- sed -i 's/mirrors.tuna.tsinghua.edu.cn/deb.debian.org/g' /etc/apt/sources.list && \- sed -i 's/mirrors.tuna.tsinghua.edu.cn\/debian-security/security.debian.org/g' /etc/apt/sources.list; \- apt-get update && apt-get install -y --no-install-recommends \- git \- imagemagick \- ffmpeg; \- ); \- ); \- fi; \- sleep 5; \- done \- ) && rm -rf /var/lib/apt/lists/*+# Install Python 3.11 and system dependencies+RUN apt-get update && apt-get install -y --no-install-recommends \+ software-properties-common \+ git \+ imagemagick \+ ffmpeg \+ curl \+ && add-apt-repository ppa:deadsnakes/ppa \+ && apt-get update && apt-get install -y --no-install-recommends \+ python3.11 \+ python3.11-venv \+ python3.11-dev \+ python3-pip \+ && ln -sf /usr/bin/python3.11 /usr/bin/python3 \+ && ln -sf /usr/bin/python3.11 /usr/bin/python \+ && python3.11 -m pip install --upgrade pip \+ && apt-get remove -y python3-blinker || true \+ && rm -rf /var/lib/apt/lists/*# Fix security policy for ImageMagickRUN sed -i '/<policy domain="path" rights="none" pattern="@\*"/d' /etc/ImageMagick-6/policy.xml@@ -50,10 +35,15 @@ RUN sed -i '/<policy domain="path" rights="none" pattern="@\*"/d' /etc/ImageMagi# Copy only the requirements.txt first to leverage Docker cacheCOPY requirements.txt ./-# Install Python dependencies with domestic mirrors first and retry logic-RUN pip install --no-cache-dir -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com --retries 3 --timeout 60 -r requirements.txt || \- pip install --no-cache-dir -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple/ --trusted-host mirrors.tuna.tsinghua.edu.cn --retries 3 --timeout 60 -r requirements.txt || \- pip install --no-cache-dir --retries 3 --timeout 60 -r requirements.txt+# Install Python dependencies+RUN python3 -m pip install --no-cache-dir \+ -i https://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com \+ --retries 3 --timeout 60 -r requirements.txt || \+ python3 -m pip install --no-cache-dir \+ -i https://mirrors.tuna.tsinghua.edu.cn/pypi/web/simple/ --trusted-host mirrors.tuna.tsinghua.edu.cn \+ --retries 3 --timeout 60 -r requirements.txt || \+ python3 -m pip install --no-cache-dir \+ --retries 3 --timeout 60 -r requirements.txt# Now copy the rest of the codebase into the imageCOPY . .@@ -63,12 +53,3 @@ EXPOSE 8501# Command to run the applicationCMD ["streamlit", "run", "./webui/Main.py","--browser.serverAddress=127.0.0.1","--server.enableCORS=True","--browser.gatherUsageStats=False"]--# 1. Build the Docker image using the following command-# docker build -t moneyprinterturbo .--# 2. Run the Docker container using the following command-## For Linux or MacOS:-# docker run -v $(pwd)/config.toml:/MoneyPrinterTurbo/config.toml -v $(pwd)/storage:/MoneyPrinterTurbo/storage -p 8501:8501 moneyprinterturbo-## For Windows:-# docker run -v ${PWD}/config.toml:/MoneyPrinterTurbo/config.toml -v ${PWD}/storage:/MoneyPrinterTurbo/storage -p 8501:8501 moneyprinterturbo\ No newline at end of filediff --git a/docker-compose.yml b/docker-compose.ymlindex acad4626..73793f4a 100644--- a/docker-compose.yml+++ b/docker-compose.yml@@ -1,6 +1,6 @@x-common-volumes: &common-volumes- ./:/MoneyPrinterTurbo-+ - ./storage:/MoneyPrinterTurbo/storageservices:webui:build:@@ -21,4 +21,11 @@ services:- "8080:8080"command: [ "python3", "main.py" ]volumes: *common-volumes- restart: always\ No newline at end of file+ restart: always+ deploy:+ resources:+ reservations:+ devices:+ - driver: nvidia+ count: 1+ capabilities: [gpu]diff --git a/docs/GPU_DOCKER_DEPLOYMENT.md b/docs/GPU_DOCKER_DEPLOYMENT.mdnew file mode 100644index 00000000..e7aa0a55--- /dev/null+++ b/docs/GPU_DOCKER_DEPLOYMENT.md@@ -0,0 +1,176 @@+# GPU Docker Deployment Guide++本文档介绍如何将 MoneyPrinterTurbo 的 Docker 部署改造为 GPU 加速模式,使 `faster-whisper` 在 GPU 上运行,大幅提升字幕生成速度。++## 为什么要 GPU 加速++MoneyPrinterTurbo 中唯一的深度学习环节是 **faster-whisper 语音识别**(将生成的音频转为带时间戳的字幕)。++- **CPU 模式**(默认):`large-v3` 模型生成字幕较慢,受限于 CPU 算力+- **GPU 模式**:利用 NVIDIA GPU + CUDA 加速,字幕生成速度提升 **5-10 倍**++> 注意:项目的其他环节(脚本生成、音频合成、视频剪辑)不涉及深度学习,GPU 只加速字幕生成。++## 前提条件++### 1. 硬件要求++- NVIDIA GPU(建议 6GB 以上显存)+- `large-v3` 模型在 GPU 上 `float16` 精度约占用 1.5GB 显存++### 2. 软件要求++- **NVIDIA 驱动**:最新版即可(可在 PowerShell 中运行 `nvidia-smi` 确认)+- **Docker Desktop**:确保已安装+- **NVIDIA Container Toolkit**:Docker 中需要 `nvidia` runtime(运行 `docker info` 查看 Runtimes 列表中是否有 `nvidia`)++### 3. 环境验证++```bash+# 确认 NVIDIA 驱动正常+nvidia-smi++# 确认 Docker 支持 GPU+docker info | findstr nvidia+```++如果 `docker info` 中没有看到 `nvidia` runtime,需要先安装 [NVIDIA Container Toolkit](https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/latest/install-guide.html)。++## 修改内容++本次改造涉及 **3 个文件**:++### 1. Dockerfile++将基础镜像从纯 CPU 的 Python 镜像替换为带 CUDA 运行时的 NVIDIA 镜像,并安装 Python 3.11。++**核心变更:**++```dockerfile+# 原来:纯 CPU 镜像+FROM python:3.11-slim-bullseye++# 改为:带 CUDA 12.1 + cuDNN 8 的运行时镜像+FROM nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04+```++主要改动点:+- 基础镜像切换为 `nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04`+- 新增安装 Python 3.11(CUDA 镜像基于 Ubuntu 22.04,自带 Python 3.10 但项目需要 3.11)+- 通过 `add-apt-repository ppa:deadsnakes/ppa` 安装 Python 3.11+- 移除系统自带的 `python3-blinker` 包(避免 pip 安装冲突)+- 使用 `python3 -m pip` 代替 `pip` 命令++### 2. docker-compose.yml++为 `api` 服务挂载 GPU 设备。++**核心变更:**++```yaml+services:+ api:+ # ... 其他配置不变 ...+ deploy:+ resources:+ reservations:+ devices:+ - driver: nvidia+ count: 1+ capabilities: [gpu]+```++> `webui` 服务是纯前端(Streamlit),不需要 GPU,保持不变。++### 3. config.toml++配置 Whisper 使用 GPU 模式。++**核心变更:**++```toml+subtitle_provider = "whisper" # 启用 whisper(原来可能是 edge)++[whisper]+model_size = "large-v3"+device = "cuda" # 原来是 "cpu"+compute_type = "float16" # 原来是 "int8",GPU 上 float16 更快更准+```++## 部署步骤++### 第 1 步:拉取 CUDA 基础镜像++```bash+docker pull nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04+```++> 如果使用了阿里云等镜像加速源,可能对 `nvidia/cuda` 系列镜像返回 403。请确保能从 Docker Hub 直接拉取,或配置支持 NVIDIA 镜像的加速源。++### 第 2 步:修改配置文件++按上文说明修改 `Dockerfile`、`docker-compose.yml` 和 `config.toml` 三个文件。++### 第 3 步:构建镜像++```bash+docker compose build --no-cache+```++### 第 4 步:启动服务++```bash+docker compose up -d+```++### 第 5 步:验证 GPU 是否生效++进入 api 容器,确认 GPU 可见:++```bash+docker exec -it moneyprinterturbo-api nvidia-smi+```++如果能看到 GPU 信息,说明 GPU 挂载成功。++## 显存与并发建议++| GPU 显存 | 建议最大并发任务数 |+|---|---|+| 4GB | 1-2 |+| 6GB | 2-3 |+| 8GB | 3-4 |+| 12GB+ | 5 |++可通过 `config.toml` 中的 `max_concurrent_tasks` 控制并发数。++## 故障排查++### 问题 1:镜像拉取失败(403 Forbidden)++阿里云镜像加速对 `nvidia/cuda` 返回 403。解决方法:+- 配置其他可用的镜像加速源+- 或直接 `docker pull nvidia/cuda:12.1.1-cudnn8-runtime-ubuntu22.04` 从 Docker Hub 拉取++### 问题 2:pip 安装报 `no such option: --break-system-packages`++旧版 pip 不支持此参数。解决方法:+- 使用 `python3 -m pip install` 代替 `pip install`+- 或在安装 Python 后先升级 pip:`python3.11 -m pip install --upgrade pip`++### 问题 3:pip 安装报 `Cannot uninstall blinker`++Ubuntu 22.04 系统自带的 `blinker` 包通过 `distutils` 安装,pip 无法卸载。解决方法:+- 在安装 Python 依赖前先 `apt-get remove -y python3-blinker`++### 问题 4:容器内 `nvidia-smi` 找不到 GPU++- 确认宿主机已安装 NVIDIA Container Toolkit+- 确认 `docker info` 中 Runtimes 包含 `nvidia`+- 确认 `docker-compose.yml` 中已添加 `deploy.resources.reservations.devices` 配置++### 问题 5:Whisper 报 CUDA 错误++- 确认 `config.toml` 中 `device = "cuda"`(注意不是 `"CPU"`,大小写敏感)+- 确认 `compute_type = "float16"`(GPU 推荐值)+- 确认 `subtitle_provider = "whisper"`(不是 `"edge"`)