CS231n: Convolutional Neural Networks for Visual Recognition - Spring 2021

I’ve been following Stanford course CS231n: Convolutional Neural Networks for Visual Recognition in my internship program at Rayanesh company. Here I gathered my notes and solutions to assignments. The course lectures were recorded in Spring 2017, but the assignments are from Spring 2021.

CS231n Assignments Solutions

Some concepts in assignments like transformers or Self-Supervised learning are not taught in the 2017 lectures. Self-Supervised learning question is solved, but transformers question is skipped. The Style Transfer question was omitted in the 2021 assignments, so I returned to the 2017 homeworks to solve that.

Assignment 1:

You could get starter code from here.

• Q1: k-Nearest Neighbor classifier. (Done)
• Q2: Training a Support Vector Machine. (Done)
• Q3: Implement a Softmax classifier. (Done)
• Q4: Two-Layer Neural Network. (Done)
• Q5: Higher Level Representations: Image Features. (Done)

Assignment 2:

You could get starter code from here.

• Q1: Multi-Layer Fully Connected Neural Networks. (Done)
• Q2: Batch Normalization. (Done)
• Q3: Dropout. (Done)
• Q4: Convolutional Neural Networks. (Done)
• Q5: PyTorch / TensorFlow on CIFAR-10. (Done in PyTorch)

Assignment 3:

You could get starter code from here.

• Q1: Image Captioning with Vanilla RNNs. (Done)
• Q2: Image Captioning with Transformers.
• Q3: Network Visualization: Saliency Maps, Class Visualization, and Fooling Images. (Done)
• Q4: Generative Adversarial Networks. (Done)
• Q5: Self-Supervised Learning for Image Classification. (Done)
• Extra: Image Captioning with LSTMs. (Done)

Assignment 3 - 2017:

• Q4: Style Transfer. (Done in PyTorch)

CS231n 2017 Notes

I took notes from some lectures.

• Lecture 6: Training Neural Networks, Part I.
• Lecture 7: Training Neural Networks, part II.
• Lecture 8: Deep Learning Software.
• Lecture 9: CNN Architectures.
• Lecture 10: Recurrent Neural Networks.
• Lecture 11: Detection and Segmentation.
• Lecture 12: Visualizing and Understanding.
• Lecture 13: Generative Models.