A simple example using a Jupyter Notebook with Cog
First, make sure you've got the latest version of Cog installed.
Build the image:
cog build
Run the Jupyter Notebook server with Cog:
cog run -p 8888 --debug jupyter notebook --allow-root --ip=0.0.0.0
Copy the notebook URL to your browser. It should look something like this:
http://127.0.0.1:8888/?token=eedb5f511a60b179d1a1b2c6395f0f20c02a08124bae6896
Save any changes you make to your notebook, then export it as a Python script:
jupyter nbconvert --to script my_notebook.ipynb # creates my_notebook.py
Now you can run predictions on the model:
cog predict -i name=Alice