drewThomasson's picture
Upload 11 files
306e52a verified
raw
history blame
3.15 kB
# Use an official NVIDIA CUDA image with cudnn8 and Ubuntu 20.04 as the base
FROM nvidia/cuda:11.8.0-cudnn8-runtime-ubuntu20.04
# Set non-interactive installation to avoid timezone and other prompts
ENV DEBIAN_FRONTEND=noninteractive
# Install necessary packages including Miniconda
RUN apt-get update && apt-get install -y --no-install-recommends \
wget \
git \
espeak \
espeak-ng \
ffmpeg \
tk \
mecab \
libmecab-dev \
mecab-ipadic-utf8 \
build-essential \
calibre \
&& rm -rf /var/lib/apt/lists/*
RUN ebook-convert --version
# Install Miniconda
RUN wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh -O ~/miniconda.sh && \
bash ~/miniconda.sh -b -p /opt/conda && \
rm ~/miniconda.sh
# Set PATH to include conda
ENV PATH=/opt/conda/bin:$PATH
# Create a conda environment with Python 3.10
RUN conda create -n ebookenv python=3.10 -y
# Activate the conda environment
SHELL ["conda", "run", "-n", "ebookenv", "/bin/bash", "-c"]
# Install Python dependencies using conda and pip
RUN conda install -n ebookenv -c conda-forge \
pydub \
nltk \
mecab-python3 \
&& pip install --no-cache-dir \
bs4 \
beautifulsoup4 \
ebooklib \
tqdm \
tts==0.21.3 \
unidic \
gradio
# Download unidic
RUN python -m unidic download
# Set the working directory in the container
WORKDIR /ebook2audiobookXTTS
# Clone the ebook2audiobookXTTS repository
RUN git clone https://github.com/DrewThomasson/ebook2audiobookXTTS.git .
# Copy test audio file
COPY default_voice.wav /ebook2audiobookXTTS/
# Run a test to set up XTTS
RUN echo "import torch" > /tmp/script1.py && \
echo "from TTS.api import TTS" >> /tmp/script1.py && \
echo "device = 'cuda' if torch.cuda.is_available() else 'cpu'" >> /tmp/script1.py && \
echo "print(TTS().list_models())" >> /tmp/script1.py && \
echo "tts = TTS('tts_models/multilingual/multi-dataset/xtts_v2').to(device)" >> /tmp/script1.py && \
echo "wav = tts.tts(text='Hello world!', speaker_wav='default_voice.wav', language='en')" >> /tmp/script1.py && \
echo "tts.tts_to_file(text='Hello world!', speaker_wav='default_voice.wav', language='en', file_path='output.wav')" >> /tmp/script1.py && \
yes | python /tmp/script1.py
# Remove the test audio file
RUN rm -f /ebook2audiobookXTTS/output.wav
# Verify that the script exists and has the correct permissions
RUN ls -la /ebook2audiobookXTTS/
# Check if the script exists and log its presence
RUN if [ -f /ebook2audiobookXTTS/custom_model_ebook2audiobookXTTS_with_link_gradio.py ]; then echo "Script found."; else echo "Script not found."; exit 1; fi
# Modify the Python script to set share=True
RUN sed -i 's/demo.launch(share=False)/demo.launch(share=True)/' /ebook2audiobookXTTS/custom_model_ebook2audiobookXTTS_with_link_gradio.py
# Download the punkt package for nltk
RUN python -m nltk.downloader punkt
# Set the command to run your GUI application using the conda environment
CMD ["conda", "run", "--no-capture-output", "-n", "ebookenv", "python", "/ebook2audiobookXTTS/custom_model_ebook2audiobookXTTS_with_link_gradio.py"]