Skip to content

apartsin/DLCourseHIT

Repository files navigation

Introduction to Deep Learning (HIT)

DL 201 · Foundation course · 13 weeks · PyTorch

Foundations of neural networks: framing a task in tensor terms, then building, training, and debugging networks in PyTorch. This is the foundation deep-learning course in the program and the bridge to the advanced electives in Large Language Models and Computer Vision.

  • Prerequisite: Introduction to Machine Learning
  • Leads to: Advanced LLMs · Advanced Computer Vision
  • Format: 3 h lecture + 2 h practice lesson per week (+ a homework lab)
  • Assessment: project- and lab-based (no written exams)

Each week has three parts: a lecture (theory), an instructor-led practice lesson (live implementation and worked examples), and a lab set as homework (the graded Build/Predict/Explain work).

Course site (GitHub Pages): https://apartsin.github.io/DLCourseHIT/

Syllabus

Working with an AI assistant (the lab model)

The weekly lab (homework) is built for the way students actually work, with an AI coding assistant at hand. Each lab has three parts:

  1. Build (AI assistant welcome): produce working code that meets a spec or hits a target metric.
  2. Predict & probe (the reasoning step): predict outcomes before running, then run controlled experiments.
  3. Explain & defend (in plain language): explain why it works, where it breaks, and what changed.

Grading weight sits on Predict and Explain, the parts an AI assistant cannot do for the student.

Weekly materials

Each week has a lab handout (with self-check questions), a reference page of curated free sources, an instructor lesson plan (3 h lecture + 2 h practice), and a Colab practice notebook.

Wk Topic Materials
Part I · Foundations
1 Deep Learning Overview & ML-to-Network Framing Lesson · Practice nb · Self-check · Homework lab
2 Tensors & Data Representation Lesson · Practice nb · Self-check · Homework lab
3 MLPs & Backpropagation Lesson · Practice nb · Self-check · Homework lab
Part II · Training Infrastructure
4 Data Pipelines Lesson · Practice nb · Self-check · Homework lab
5 Loss Functions & Metrics Lesson · Practice nb · Self-check · Homework lab
6 Optimization Lesson · Practice nb · Self-check · Homework lab
7 Regularization & Generalization Lesson · Practice nb · Self-check · Homework lab
Part III · Architectures & Representation Learning
8 Convolutional Networks I Lesson · Practice nb · Self-check · Homework lab
9 Convolutional Networks II Lesson · Practice nb · Self-check · Homework lab
10 Recurrent Networks (RNNs) Lesson · Practice nb · Self-check · Homework lab
11 LSTMs, GRUs & Sequence Tasks Lesson · Practice nb · Self-check · Homework lab
12 Representation Learning Lesson · Practice nb · Self-check · Homework lab
Part IV · Integration
13 Integration & Transfer Learning Lesson · Practice nb · Self-check · Homework lab

The .html pages render best via the GitHub Pages site above. The notebook links open directly in Google Colab.

Repository layout

index.html            Course home (links every week's lab, references, lesson plan, notebook)
syllabus/             Syllabus in HTML, Word, and PDF
labs/                 week01..week13 student lab handouts (with self-check questions)
references/           week01..week13 curated reference pages
lessons/              week01..week13 instructor lesson plans (lecture + practice)
notebooks/            week01..week13 Colab practice notebooks (instructor, run during the practice lessons)
prereq/               Math, Python, and ML prerequisite review pages
projects/             Example briefs: 10 mid-term + 10 final projects
hit-catalogue/        HIT submission package (EN + HE syllabus, rationale, summary, questionnaire)
assets/style.css      Shared house style
tools/                Generators: build_site.py, build_notebooks.py, build_hit_package.py, content.py, refs.json, lessons.py, selfcheck.py

Building

python tools/build_site.py          # index + labs + references + lessons
python tools/build_notebooks.py     # Colab notebooks
python tools/build_hit_package.py   # HIT catalogue .docx package

The syllabus (syllabus/) is maintained from syllabus/syllabus.html as the source of truth; the Word and PDF versions are regenerated from it.

HIT catalogue package

hit-catalogue/ holds the department submission package in the HIT form: English syllabus, Hebrew syllabus, rationale, catalogue summary, and committee questionnaire.

Status

Complete: syllabus, 13 lab handouts with self-check questions, 13 reference pages, 13 instructor lesson plans, 13 Colab notebooks, and the HIT catalogue package. Lab starter solutions will be added over the term.

About

Introduction to Deep Learning (HIT): 13-week PyTorch foundation course. Syllabus, labs, references, lesson plans, and Colab notebooks.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors