diff --git a/Dockerfile.train.jupyter b/Dockerfile.train.jupyter index 09c6ed79..1e2a0ff7 100644 --- a/Dockerfile.train.jupyter +++ b/Dockerfile.train.jupyter @@ -40,6 +40,7 @@ RUN wget --no-check-certificate https://github.com/coqui-ai/STT/releases/downloa rm temp.tar.xz +#FROM gcr.io/kaggle-gpu-images/python:v98 FROM nvcr.io/nvidia/tensorflow:21.05-tf1-py3 ENV DEBIAN_FRONTEND=noninteractive @@ -97,7 +98,7 @@ RUN rm -rf ~/.local/share/stt ## # -# FROM gcr.io/kaggle-gpu-images/python:v98 +#FROM gcr.io/kaggle-gpu-images/python:v98 RUN chsh -s /bin/bash ENV SHELL=/bin/bash @@ -119,74 +120,22 @@ RUN apt-get update && apt-get install -y \ supervisor \ gettext-base \ less \ + nodejs \ + npm \ && rm -rf /var/lib/apt/lists/* -# install nvm -# https://github.com/creationix/nvm#install-script -RUN curl --silent -o- https://raw.githubusercontent.com/creationix/nvm/v0.33.11/install.sh | bash - -#ENV NVM_DIR /root/.nvm -ENV NVM_DIR /usr/local/nvm -ENV NODE_VERSION v12.20.1 - -# install node and npm -RUN source $NVM_DIR/nvm.sh \ - && nvm install $NODE_VERSION \ - && nvm alias default $NODE_VERSION \ - && nvm use default - -# add node and npm to path so the commands are available -ENV NODE_PATH $NVM_DIR/versions/node/$NODE_VERSION/bin -ENV PATH $NODE_PATH:$PATH - -RUN pip install pip==20.3.4 -RUN pip install jupyterlab==2.2.9 ipywidgets==7.6.3 +RUN pip3 install jupyterlab ipywidgets RUN jupyter labextension install @jupyter-widgets/jupyterlab-manager RUN jupyter nbextension enable --py widgetsnbextension #enable ipywidgets -RUN jupyter labextension install jupyterlab-plotly@4.14.3 +RUN jupyter labextension install jupyterlab-plotly -#install xlrd python module -RUN pip install pyradiomics -RUN pip install xlrd==1.2.0 -RUN pip install zarr -RUN pip install imbalanced-learn -RUN pip install openpyxl -RUN pip install efficientnet-pytorch -RUN pip install monai -RUN pip install prince -RUN pip install vit-pytorch -RUN pip install lifelines==0.25.11 -RUN pip install timm==0.3.2 -RUN pip install keras-retinanet==1.0.0 -RUN python -m pip install histomicstk --find-links https://girder.github.io/large_image_wheels -RUN pip install luminoth ipympl pysurvival missingpy pyinform pingouin pyAgrum missingno autoimpute networkx community yellowbrick factor_analyzer hdbscan pyitlib -RUN pip install eli5 dtreeviz gower batchgenerators mlinsights efficientnet-pytorch pretrainedmodels -# Add R to Jupyter Kernel -RUN conda install -y -c r r-irkernel - -#Install survival, sm, ggplot2, Hmisc, mixOmics (ce dernier est en repositoire Bioconductor) - -RUN conda install -y -c cran r-survival -RUN conda install -y -c cran r-sm -RUN conda install -y -c cran r-ggplot2 -RUN conda install -y -c cran r-hmisc -RUN conda install -y -c cran r-mixomics -RUN conda install -y -c cran r-caret -RUN conda install -y -c cran r-survminer -RUN conda install -y -c cran r-ggfortify -RUN conda install -y -c cran r-wordcloud -RUN conda install -y -c cran r-tm -RUN conda install -y -c cran r-prioritylasso -RUN conda install -y -c cran r-blockforest -RUN conda install -y -c cran r-mice - -#### tensorflow 1 -# Create the environment: -SHELL ["/bin/bash", "-c"] -COPY tfenv.yml . -RUN conda env create -f tfenv.yml -SHELL ["conda", "run", "-n", "tf1", "/bin/bash", "-c"] -RUN python -m ipykernel install --name=tensorflow1 +# #### tensorflow 1 +# # Create the environment: +# SHELL ["/bin/bash", "-c"] +# COPY tfenv.yml . +# RUN conda env create -f tfenv.yml +# SHELL ["conda", "run", "-n", "tf1", "/bin/bash", "-c"] +# RUN python -m ipykernel install --name=tensorflow1 EXPOSE 8080 diff --git a/Dockerfile.train.jupyter.simple b/Dockerfile.train.jupyter.simple deleted file mode 100644 index bf4f798f..00000000 --- a/Dockerfile.train.jupyter.simple +++ /dev/null @@ -1,97 +0,0 @@ -# This is a Dockerfile useful for training models with Coqui STT. -# You can train "acoustic models" with audio + Tensorflow, and -# you can create "scorers" with text + KenLM. - -FROM ubuntu:20.04 AS kenlm-build -ENV DEBIAN_FRONTEND=noninteractive - -RUN apt-get update && \ - apt-get install -y --no-install-recommends \ - build-essential cmake libboost-system-dev \ - libboost-thread-dev libboost-program-options-dev \ - libboost-test-dev libeigen3-dev zlib1g-dev \ - libbz2-dev liblzma-dev && \ - rm -rf /var/lib/apt/lists/* - -# Build KenLM to generate new scorers -WORKDIR /code -COPY kenlm /code/kenlm -RUN cd /code/kenlm && \ - mkdir -p build && \ - cd build && \ - cmake .. && \ - make -j $(nproc) || \ - ( echo "ERROR: Failed to build KenLM."; \ - echo "ERROR: Make sure you update the kenlm submodule on host before building this Dockerfile."; \ - echo "ERROR: $ cd STT; git submodule update --init kenlm"; \ - exit 1; ) - - -FROM ubuntu:20.04 AS wget-binaries -ENV DEBIAN_FRONTEND=noninteractive - -RUN apt-get update && apt-get install -y --no-install-recommends wget unzip xz-utils - -# Tool to convert output graph for inference -RUN wget --no-check-certificate https://github.com/coqui-ai/STT/releases/download/v0.9.3/convert_graphdef_memmapped_format.linux.amd64.zip -O temp.zip && \ - unzip temp.zip && \ - rm temp.zip - -RUN wget --no-check-certificate https://github.com/coqui-ai/STT/releases/download/v0.9.3/native_client.tf.Linux.tar.xz -O temp.tar.xz && \ - tar -xf temp.tar.xz && \ - rm temp.tar.xz - - -FROM jupyter/tensorflow-notebook -ENV DEBIAN_FRONTEND=noninteractive -USER root - -# We need to purge python3-xdg because -# it's breaking STT install later with -# errors about setuptools -# -RUN apt-get update && \ - apt-get install -y --no-install-recommends \ - git \ - wget \ - libopus0 \ - libopusfile0 \ - libsndfile1 \ - sox \ - libsox-fmt-mp3 && \ - apt-get purge -y python3-xdg && \ - rm -rf /var/lib/apt/lists/ - -# Make sure pip and its dependencies are up-to-date -RUN pip3 install --upgrade pip wheel setuptools -RUN pip3 uninstall -y tensorflow && pip3 install -y 'tensorflow-gpu==1.15.4' - -WORKDIR /code - -COPY native_client /code/native_client -COPY .git /code/.git -COPY training/coqui_stt_training/VERSION /code/training/coqui_stt_training/VERSION -COPY training/coqui_stt_training/GRAPH_VERSION /code/training/coqui_stt_training/GRAPH_VERSION - -# Build CTC decoder first, to avoid clashes on incompatible versions upgrades -RUN cd native_client/ctcdecode && make NUM_PROCESSES=$(nproc) bindings -RUN pip3 install --upgrade native_client/ctcdecode/dist/*.whl - -COPY setup.py /code/setup.py -COPY VERSION /code/VERSION -COPY training /code/training -# Copy files from previous build stages -RUN mkdir -p /code/kenlm/build/ -COPY --from=kenlm-build /code/kenlm/build/bin /code/kenlm/build/bin -COPY --from=wget-binaries /convert_graphdef_memmapped_format /code/convert_graphdef_memmapped_format -COPY --from=wget-binaries /generate_scorer_package /code/generate_scorer_package - -# Install STT -# No need for the decoder since we did it earlier -# TensorFlow GPU should already be installed on the base image, -# and we don't want to break that -RUN DS_NODECODER=y DS_NOTENSORFLOW=y pip3 install --upgrade -e . - -# Copy rest of the code and test training -COPY . /code -RUN ./bin/run-ldc93s1.sh && rm -rf ~/.local/share/stt diff --git a/start.sh b/start.sh old mode 100644 new mode 100755