Learning Objectives

By completing this assignment, you will:

• Set up your Python development environment using uv
• Verify PyTorch installation and GPU access (if available)
• Practice basic Python skills needed for the course
• Learn the Git workflow used throughout this course

Instructions

Part 1: Environment Setup (30 points)

1.1 Install uv Package Manager (10 points)

uv is a fast Python package manager that we’ll use throughout the course. Install it by following the instructions for your operating system:

macOS/Linux:

curl -LsSf https://astral.sh/uv/install.sh | sh

Windows (PowerShell):

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Verify the installation by running:

uv --version

Deliverable: Screenshot showing uv --version output.

1.2 Create a Virtual Environment (10 points)

Create a dedicated virtual environment for this course:

# Create a new directory for the course
mkdir mphy6120
cd mphy6120

# Initialize a new project with uv
uv init

# Create virtual environment with Python 3.11
uv venv --python 3.11

Activate the virtual environment:

macOS/Linux:

source .venv/bin/activate

Windows:

.venv\Scripts\activate

Deliverable: Screenshot showing your activated virtual environment (your prompt should show (.venv) or similar).

1.3 Install Required Packages (10 points)

Install the core packages we’ll use throughout the course:

uv pip install torch numpy pandas matplotlib jupyter ipykernel

Note: If you have an NVIDIA GPU and want CUDA support, follow the PyTorch installation guide for your specific setup.

Verify your installation by starting Python and running:

import torch
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

print(f"PyTorch version: {torch.__version__}")
print(f"NumPy version: {np.__version__}")
print(f"Pandas version: {pd.__version__}")
print("All packages imported successfully!")

Deliverable: Screenshot showing successful package imports with version numbers.

Part 2: Python Skills Review (40 points)

Complete the following exercises in the starter notebook (or create your own notebook named hw1_python_skills.ipynb).

2.1 NumPy Array Operations (15 points)

1. Array Creation (3 pts): Create a 1D NumPy array containing the integers 1 through 10.

2. Random Matrix (3 pts): Create a 3x3 matrix of random values between 0 and 1 using np.random.rand().

3. Statistics (3 pts): For your random matrix, compute and print the mean, standard deviation, min, and max.

4. Reshaping (3 pts): Create a 1D array of 12 elements and reshape it into a 3x4 matrix.

5. Slicing (3 pts): From your 3x3 random matrix, extract a 2x2 submatrix from the top-left corner.

2.2 Pandas DataFrames (15 points)

1. Create DataFrame (3 pts): Create a DataFrame with columns patient_id, age, heart_rate, and diagnosis with at least 10 rows of sample data.

2. Add Column (3 pts): Add a new column age_group that categorizes patients as “young” (< 40), “middle” (40-65), or “senior” (> 65).

3. Filter (3 pts): Filter the DataFrame to show only patients with heart rate > 80.

4. GroupBy (3 pts): Group by diagnosis and compute the mean age and heart rate for each group.

5. Missing Data (3 pts): Introduce some NaN values into your DataFrame, then demonstrate using fillna() or dropna().

2.3 Matplotlib Visualization (10 points)

Create the following visualizations with proper labels, titles, and legends:

1. Line Plot (2.5 pts): Plot $y = x^2$ for $x \in [0, 10]$.

2. Scatter Plot (2.5 pts): Create a scatter plot of random (x, y) data with at least 50 points.

3. Histogram (2.5 pts): Generate 1000 samples from a normal distribution and plot a histogram.

4. Bar Chart (2.5 pts): Create a bar chart showing the count of patients per diagnosis from your DataFrame.

Part 3: PyTorch Basics (30 points)

Complete these exercises in the same notebook or a new one named hw1_pytorch_basics.ipynb.

3.1 Tensor Creation (10 points)

1. From List (2 pts): Create a 2D tensor from a nested Python list.

2. Special Tensors (4 pts): Create tensors using:
   • torch.zeros(3, 4)
   • torch.ones(2, 3)
   • torch.rand(3, 3)
   • torch.arange(0, 10, 2)
3. Tensor Properties (4 pts): For each tensor you created, print its dtype, shape, and device.

3.2 Tensor Operations (10 points)

1. Element-wise (3 pts): Create two 3x3 random tensors and perform element-wise addition, multiplication, and exponentiation.

2. Matrix Operations (4 pts): Perform matrix multiplication (@ or torch.matmul) and transpose.

3. NumPy Conversion (3 pts): Convert a PyTorch tensor to a NumPy array and back. Demonstrate that changes to one affect the other (shared memory).

3.3 GPU/CUDA Check (10 points)

Write code that:

1. Check Availability (4 pts): Check if CUDA is available and print the result.

2. GPU Info (3 pts): If CUDA is available, print the GPU device name. If not, print a message saying CPU will be used.

3. Device Transfer (3 pts): Create a tensor and move it to the available device (GPU if available, otherwise CPU). Print the tensor’s device.

# Example structure
import torch

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
print(f"Using device: {device}")

if torch.cuda.is_available():
    print(f"GPU: {torch.cuda.get_device_name(0)}")

x = torch.rand(3, 3).to(device)
print(f"Tensor device: {x.device}")

Submission via GitHub

All submissions for this course are done through GitHub. This mirrors real-world software development workflows and helps you build essential version control skills.

How to Submit

1. Complete your work in hw1_exercises.py (replace None values with your code)
2. Commit your changes with meaningful commit messages:

   git add hw1_exercises.py
   git commit -m "Complete NumPy exercises 2.1.1-2.1.5"
   

3. Push to GitHub before the deadline:

   git push
   

Commit Expectations

Good commit habits are part of your grade. We expect:

• Multiple commits showing your progress (not just one giant commit at the end)
• Meaningful commit messages that describe what you did (e.g., “Add patient DataFrame with age groups” not “stuff”)
• Working code in your final commit (we grade the last commit before the deadline)

Example Good Commit History

a]b3f2d1 Complete Part 3: PyTorch CUDA check
8c4e5f6 Add tensor operations and NumPy conversion
2d1a9b7 Complete Part 2: matplotlib visualizations
f7e8c3a Add pandas DataFrame exercises
9a2b4c5 Complete NumPy array operations
1e6d8f2 Verify environment setup - all packages installed

Why Git Matters

In the era of AI-assisted coding, your commit history tells the story of your work:

• It shows your problem-solving process
• It distinguishes your contributions from AI-generated code
• It’s a skill every employer expects

Grading Rubric

Resources

• Git Handbook — If you’re new to Git
• uv Documentation
• NumPy Quickstart
• Pandas Getting Started
• Matplotlib Tutorials
• PyTorch Tutorials

Tips

• Start early! Environment setup can sometimes have unexpected issues.
• Commit often — Don’t wait until you’re done to make your first commit.
• Use the course GitHub Discussions or office hours if you get stuck.
• If you don’t have a GPU, that’s completely fine—all coursework can be done on CPU.
• AI assistants are allowed for debugging, but your commit history should show your problem-solving process.