Owner: Qiancheng Hu
Welcome to the Deep Learning (I2DL) exercises repository! This repository is organized with assignments that align with the "Introduction to Deep Learning" course taught by Prof. Niessner and his team.
- Introduction to Deep Learning (I2DL)
- Organization, Communication, and Exam Details
- Lectures
- Recorded lectures available weekly
- Pre-recorded lecture slides
- Tutorials
- Recordings, slides, and homework uploaded on Thursdays
- Homework deadlines on Wednesdays
- Neural Networks: Optimization, Stochastic Gradient Descent, Training
- Advanced Architectures: CNNs, RNNs, Deep Learning Topics
- Exercise 01: Organization (Submission)
- Exercise 02: Math Recap (Non-submission)
- Exercise 03: Datasets (Submission)
- Exercise 04: Linear Regression (Submission)
- Exercise 05: Neural Networks (Submission)
- Exercise 06: Hyperparameter Tuning (Submission)
- Exercise 07: Introduction to PyTorch (Submission)
- Exercise 08: Autoencoder (Submission)
- Exercise 09: Convolutional Networks (Submission)
- Exercise 10: Semantic Segmentation (Submission)
- Exercise 11: Sequence Models (Submission)
Complete and submit at least 8 out of the 9 required exercises to qualify for the -0.3 grade bonus.
- Download the zip folder containing each exercise.
- Follow the
README.mdinstructions for each exercise to:- Set up a Python environment
- Execute Jupyter Notebooks
- Register on the submission webpage: i2dl.vc.in.tum.de
- Upload your solutions as a zip file.
- Python
- Jupyter Notebooks
- NumPy
- Deep Learning Library
- PyTorch
- Hardware
- CPU (minimum)
- NVIDIA GPU (preferred)
- Google Colab (alternative)
- Can I use an IDE instead of Jupyter Notebooks?
- Yes, but keep the skeleton classes intact.
- How will I know if I passed an exercise?
- You will receive an email once your score exceeds the passing threshold.
- Where can I find assistance?
- Piazza, Office Hours, or ask other students.