Installation

The SemanticST package is developed based on the PyTorch framework and is compatible with both CPU and GPU. For optimal performance, we recommend running the package on GPU. Please ensure that PyTorch and CUDA are installed correctly.

To run SemanticST, all dependencies listed in the ‘requirements.txt’ file need to be installed. We provide two installation methods for the SemanticST package: using pip or through Anaconda.

Please note that the current version of SemanticST is fully supported on Linux.

1. Python Installation

To install SemanticST using pip, follow these steps:

pip install semanticst

or

git clone https://github.com/yourusername/SemanticST.git

cd SemanticST

pip install -r requirements.txt

# To install the package:
python setup.py install --user

2. Anaconda Installation

For convenience, we suggest using a separate conda environment to run SemanticST. Please ensure Anaconda is installed.

To create a conda environment and install the SemanticST package, follow these steps:

# Create a conda environment called SemanticST
conda create -n SemanticST python=3.9

# Activate your environment
conda activate SemanticST

# Install dependencies
pip install -r requirements.txt

# Install SemanticST package
python setup.py install --user

# To use the environment in Jupyter Notebook, add the Python kernel for this environment:
pip install ipykernel

python -m ipykernel install --user --name=SemanticST

3. Required Dependencies

To install the dependencies using pip, you can simply run:

pip install -r requirements.txt