This guide assumes that you have already started a cloud machine using some compute provider, e.g. Lambda Labs. It assumes that the remote machine comes with a relatively recent version of python, pip, and virtualenv, and has CUDA libraries installed already. All commands below are assumed to be run from the root directory of this repo.
If you want to use the narration script, you will also need to login to huggingface. If you run this locally before the remote_setup script, it will automatically copy your access token across to the remote machine.
. .venv/bin/activate
huggingface-cli login
The remote_setup script copies our repo to the remote machine and installs dependencies in a
virtual environment.
./scripts/remote_setup <remote_user> <remote_address>
Now you can run our scripts. For example to run inference, ssh to the remote machine and run:
cd deep_rabbit_hole
. .venv/bin/activate
python object_tracker_0/src/inference.py -v datasets/rabbits_2024_08_12_25_15sec/video/rabbits_2024_08_12_15sec.mp4 -a datasets/rabbits_2024_08_12_25_15sec/annotations/test_v1.json -w /tmp/15sec
In VSCode, click the >< icon in the bottom left and then click the "SSH" option in the menu that
appears. Enter the ubuntu@<ip_address> for the machine you created on the lambdalabs site. Once
connected, choose "Open Folder" in the VSCode "File" menu, and open the "deep_rabbit_hole" folder.
Some useful extensions including "Python" and "debugpy" are automatically installed on by vscode on the server which runs on the remote host. This may take a minute to finish installing before you can use them. Once they are ready, you should be able to run and debug our scripts just like you would locally.