cs231n-master2020

The finished assignment of cs231n in Stanford University 2020.

Cs231n has three assignments in 2020. I prefer jupyterbook to finish them.

cifar-10-batches-py

This is the dataset which we need to finish our homework, and you can download from my github. Of course, you can use sh to download, just following the steps of cs231n website which is very concrete and easy. By the way, I recommend that all your packages you need in these assignments should be downloaded in the active environment.

assignment1_jupyter

The first assignment is not very easy, just be careful about future package. The requirements.txt has the improper version of future package. You need to download a new version in active environment by following the next command.

pip install future

I think this is the only problem, while other problems on website I didn't face.

assignment2_jupyter

The second assignment is much harder than the first one. Except it has much heavier work, it also need much stronger coding ability. I will say something that most of people will meet.

• The first one is that I have watched my video on Bilibili which maybe the older version in 2016, but I have done the 2020 work.

Layer_normalization

Layer_normalization Part I have missed it.

• The second one is in layers_util.py, we also need two extra part. affine_bn_relu_forward and affine_bn_relu_backward Part also need to finish my homework, so I have implemented them. And you will see them in my codes.

• The third one is the biggest problem, when you need Cpython to implement Fast Layers.

name 'col2im_6d_cython' is not defined

Almost everyone has this problem, I will place my solutions.

Run into your fast_layers.py, and you will see

import numpy as np
try:
  from .im2col_cython import col2im_cython, im2col_cython
  from .im2col_cython import col2im_6d_cython
except ImportError:
  print 'run the following from the cs231n directory and try again:'
  print 'python setup.py build_ext --inplace'
  print 'You may also need to restart your iPython kernel'

from .im2col import *

Actually it is not right, because it has wrong address, Cpython are in /cs231n/，so we just change this into:

import numpy as np
try:
  from cs231n.im2col_cython import col2im_cython, im2col_cython
  from cs231n.im2col_cython import col2im_6d_cython
except ImportError:
  print 'run the following from the cs231n directory and try again:'
  print 'python setup.py build_ext --inplace'
  print 'You may also need to restart your iPython kernel'

from cs231n.im2col import *

And refresh the command in your teminal:

python setup.py build_ext--inplace

You will solve this problem. Good luck!

• BTW, in the last Part: Pytorch, we trained our own model. I choose ResNet34 to be my best model, and the accuracy of it is 83.99%. It's very nice. And ResNet_test.py is the code to help me test which type of ResNet is my most suitable Net.