An all-around data science conda environment focused on either pytorch or tensorflow with GPU support
*Only tried in Pop_OS (Ubuntu 22.04 LTS)
Use of mamba is highly recommended over the out-of-box conda for creating and removing the environment, since its significantly faster and works as a drop-in replacement, but be careful not to use to activate or deactivate the environment itself:
conda install mamba -n base -c conda-forge
Note: Makefile currently not working due to conda issues, see here.
mamba env create -f environment_{torch or tf}.yml
conda activate {torch or tf}_env
conda deactivate
mamba env remove -n {torch or tf}_env -y
pytest
python -m spacy download en_core_web_sm
python -m spacy download es_core_news_sm
python -m nltk.downloader popular
python -m textblob.download_corpora
- Rewrite makefile for automatic build and remove
-
Add option to makefile for tensorflow alternative -
Add tests to check for gpu availability for each library