See also https://espnet.github.io/espnet/installation.html
installers/ # Instllation scripts for extra tools
Makefile # Makefile to make an environment for experiments in espnet
check_install.py # To check the status of installation
extra_path.sh # Setup script for environment variables for extra tools
setup_cuda_env.sh # Setup script for environment variables for cuda
setup_anaconda.sh # To generate activate_python.sh with conda environment
setup_python.sh # To generate activate_python.sh with the specified python
setup_venv.sh # To generate activate_python.sh with venv of your python
# ./setup_anaconda.sh <conda-root|default="venv"> <env-name|default=root env> <python-version|default=latest python>
# e.g.
./setup_anaconda.sh miniconda espnet
# If the conda-root already exists at the path, it tries to create in the conda
# ./setup_anaconda.sh /some/where/miniconda espnet
# make TH_VERSION=<torch-ver|default=latest torch>
# e.g.
make
For more detail: https://espnet.github.io/espnet/installation.html#step-2-installation-espnet
bash -c ". ./activate_python.sh; . ./extra_path.sh; python3 check_install.py"
ESPnet recommends preparing a new Python environment independent of your pre-existing Python because pre-installed modules in your existing Python sometimes conflict with the dependencies of ESPnet. If you still want to install ESPnet in your original Python, all issues which might happen are under your responsibility.