Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

CS231n Self-study #1

Open
jeongjae96 opened this issue Feb 6, 2023 · 2 comments
Open

CS231n Self-study #1

jeongjae96 opened this issue Feb 6, 2023 · 2 comments
Assignees
@jeongjae96

Comments

@jeongjae96
Copy link
Owner

What?
@jeongjae96 jeongjae96 self-assigned this Feb 6, 2023
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[docs] start cs231n assignments #1
afba5b3
@jeongjae96
Copy link
Owner Author

jeongjae96 commented Feb 6, 2023 •
edited
Loading

Assignment1

Todo

  • k-Nearest Neighbor classifier
  • Training a Support Vector Machine
  • Implement a Softmax classifier
  • Two-Layer Neural Network
  • Higher Level Representations: Image Features

jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'compute_distances_two_loops' for knn #1
29d2baf 
- compute the l2 distance between test and training data using two loops
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
assignment1 inline question1 #1
28ace63 
- patterns of the distance matrix
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'predict_labels' function for knn #1
e75a4fc
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
assignment1 Q1 inline question2 #1
fc5e92d 
- preprocessing steps that change the performance of a Nearest Neighbor classifier
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'compute_distances_one_loop' function for knn #1
ace3c66 
- use broadcasting
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'compute_distances_no_loops' for knn #1
b0c486c 
- reformulate the expression and use broadcasting
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement cross validation #1
bcbe164
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
assignment1 Q1 inline question3 #1
0c55d01 
- knn classification setting
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement the gradient for the SVM loss function #1
01ead44
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
assignment1 Q2 inline question1 #1
5a5a08b 
- gradient discrepancy
- differentiability
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'svm_loss_vectorized' function #1
747956e
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement SGD in the function LinearClassifier.train() #1
a4284cf
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement LinearSVM.predict function #1
3e263c2
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement cross validation #1
5ce302b
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
assignment1 Q2 inline question2 #1
ae42870 
- SVM weight visualization
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'softmax_loss_naive' function #1
72dc362
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
Assignment1 Q3 inline question1 #1
7429e0c 
- soft max loss of random weight
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'softmax_loss_vectorized' function #1
fcc518f
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement cross validation of softmax classifier #1
0da89ea
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
Assignment1 Q3 inline question2 #1
ae83ee3 
- SVM loss and Softmax classifier loss
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'affine_forward' function #1
1e95537
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'affine_backward' function #1
b037e57
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'relu_forward' function #1
1ac5af7
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement 'relu_backward' function #1
7b4734a
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
Assignment1 Q4 inline question1 #1
d7f96af 
- gradient vanishing problem of activation functions
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
implement 'softmax_loss' & 'svm_loss' #1
aa118b8
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
implement 'TwoLayerNet' class #1
0bf97e3
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] use 'Solver' instance to train a 'TwoLayerNet' #1
f541db5
jeongjae96 added a commit that referenced this issue Feb 6, 2023
@jeongjae96
[feat] implement cross validation on two layer net #1
9e46846
jeongjae96 added a commit that referenced this issue Feb 7, 2023
@jeongjae96
Assignment1 Q4 inline question2 #1
27b0716 
- overfitting
jeongjae96 added a commit that referenced this issue Feb 7, 2023
@jeongjae96
[feat] implement grid search #1
7918208
jeongjae96 added a commit that referenced this issue Feb 7, 2023
@jeongjae96
Assignment1 Q5 inline question1 #1
8c797e8 
- performance on feature extraction
jeongjae96 added a commit that referenced this issue Feb 7, 2023
@jeongjae96
[feat] train a two-layer neural network on image features #1
391cf14
@jeongjae96
Copy link
Owner Author

jeongjae96 commented Feb 7, 2023 •
edited
Loading

Assignment2

Todo

  • Multi-Layer Fully Connected Neural Networks
  • Batch Normalization
  • Dropout
  • Convolutional Neural Networks
  • PyTorch on CIFAR-10
  • Network Visualization: Saliency Maps, Class Visualization, and Fooling Images

jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] implement affine_forward, affine_backward, relu_forward, relu_…
7956456 
…backward & softmax_loss #1

- implementations from assignment1
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] implement 'FullyConnectedNet' class #1
92b006f 
- without BN & dropout
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
overfit training examples with 3-layer net and 5-layer net & done inl…
301f4ec 
…ine question 1 #1

- the sensitivity of the network
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] implement 'sgd_momentum' function #1
cbd9798
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] implement 'rmsprop' function #1
3348f0e
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] implement 'adam' function #1
0beed28
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
Assignment2 Q1 Inline Question 2 #1
ad8c079 
- update speed of AdaGrad & Adam
jeongjae96 added a commit that referenced this issue Feb 8, 2023
@jeongjae96
[feat] bash shell for install dataset on windows #1
bade56c
jeongjae96 added a commit that referenced this issue Feb 21, 2023
@jeongjae96
[feat] train a fully connected model #1
156886f 
- above 50% accuracy on the validation set
jeongjae96 added a commit that referenced this issue Feb 28, 2023
@jeongjae96
[feat] implement 'batchnorm_forward' function #1
c1977d5
jeongjae96 added a commit that referenced this issue Mar 2, 2023
@jeongjae96
[feat] implement 'batchnorm_backward' function #1
c90246c
jeongjae96 added a commit that referenced this issue Mar 2, 2023
@jeongjae96
[feat] implement 'batchnorm_backward_alt' function #1
b159264
jeongjae96 added a commit that referenced this issue Mar 21, 2023
@jeongjae96
[feat] add batch normalization in 'FullyConnectedNet' #1
a7c5a25 
- implement additional helper layer in 'layer_utils.py': 'affine_bn_relu_forward' & 'affine_bn_relu_backward'
jeongjae96 added a commit that referenced this issue Mar 21, 2023
@jeongjae96
[fix] fix initialization of FullyConnectedNet to not initialize bn pa…
bb30fc7 
…rameters for last layer #1
jeongjae96 added a commit that referenced this issue Mar 22, 2023
@jeongjae96
Assignment2 Q2 inline question1 #1
bc8e02f 
- the effect of the weight initialization scale with/without batch normalization.
jeongjae96 added a commit that referenced this issue Mar 22, 2023
@jeongjae96
Assignment2 Q2 inline question2 #1
b7d7095 
- the relationship between batch normalization and batch size.
jeongjae96 added a commit that referenced this issue Mar 22, 2023
@jeongjae96
Assignment2 Q3 inline question2 #1
0810665 
- Batch Normalization & Layer Normalization
jeongjae96 added a commit that referenced this issue Mar 22, 2023
@jeongjae96
[feat] implement 'layernorm_forward' function #1
b892d24
jeongjae96 added a commit that referenced this issue Mar 26, 2023
@jeongjae96
[docs] move the reference link for 'batchnorm_backward_alt' to right …
b6adec1 
…place #1
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees

@jeongjae96 jeongjae96

Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@jeongjae96