I’ve using the paperspace desktop to train my learning models and now I’d very much like to try the CLI. I’ve been reading through the docs and I wanted to make sure that have this down before I deploy. My apologies. I’m making the transition between training models on a local machine
So from the terminal, I install the paperspace module using:
pip install paperspace
Or, if I’m using the desktop within paperspace it’s already installed I assume.
Next I set my credentials using:
paperspace-python login <email> <password> [<api_token_name>]
Now here’s a question I have. I have several natural language modules I need to install. Can I simply create a requirements file of my venv and use:
pip3 install -r requirements.txt
Assuming tensorflow-gpu is installed in the default container image, keras spacy, gensim, and nltk should load I hoping. If this is the case is my next command:
paperspace-python run myscript.py --project myproject --machineType P6000 --workspace
Here are some additional. I use the English language models for gensim, spacy, and nltk, which I typically install from the command line. How would I go about doing this with the CLI? Also, I use some pretrained word embeddings in my work. How do I load these as well as other data used by myfile.py into the working directory.
Thanks for all your help, and apologies for the long noob questions.