CSCI 1470 Environment Setup (OSCAR)

This repository contains everything needed to set up the CSCI 1470 conda environment on Brown University’s OSCAR cluster.

Table of Contents

1. Cloning the Repository
2. Creating the Environment
3. Verifying the Environment
4. Using the Environment
5. Troubleshooting

---

Cloning the Repository

1. Log into OSCAR using your preferred method. Only once you have established a secure connection to your OSCAR instance, proceed to step 2.
2. Clone the repository into your home directory using

git clone <REPO_NAME>

---

Creating the Environment

1. Run the environment creation script:

bash conda_create.sh

This script will:

• Remove any existing csci1470 environment.
• Create a new csci1470 environment from env_setup/Other/csci1470.yml.
• Activate csci1470.
• Install a Jupyter kernel named DL-S25 (3.10).

Note: If you need to install the environment in a custom location (e.g., because of storage quotas), you can modify the command in conda_create.sh to use -p /path/to/env_dir instead of -n csci1470.

---

Verifying the Environment

After the script finishes:

1. Activate the environment:

conda activate csci1470

2. Check installed packages:

conda list

or

pip freeze

You should look for tensorflow==2.11, scikit-learn, etc. 3. Test a quick import:

python -c "import tensorflow as tf; print(tf.__version__)"

You should see the TensorFlow version (e.g., 2.11.0).

---

Using the Environment

Jupyter Notebooks:

If you want to run Jupyter on a compute node (which is usually recommended for GPU usage), do something like:

srun --pty -p gpu --mem=16G --gres=gpu:1 jupyter lab --no-browser --port=9999

Then, port-forward from your local machine to access the lab interface:

ssh -L 9999:localhost:9999 your_brown_username@gpu2002.oscar.ccv.brown.edu

Python Scripts

conda activate csci1470
python your_script.py