burton2000 / cs231n-2017 Goto Github PK

Hi, I am also doing assignment 3 of CS231n. Yet I didn't get the correct number for the function rnn_backward. My code looks like this:

for t in reversed(range(T)):
        dh[:,t,:]= dprev_h + dh[:,t,:]
        dx[:,t,:], dprev_h, dWx_temp, dWh_temp, db_temp = rnn_step_backward(dh[:,t,:], cache[t])
        dWx += dWx_temp
        dWh += dWh_temp
        db += db_temp

My cache is also a list of tuple of numpy array and every else is the same. Yet I did not get the correct number. May I ask if the reason of this weird result?

how do i add my .jpg file and check the caption output of the algorithm?

Hello!

First of all, nice work.

Secondly, this might be a bug in the original assignment but:
AFAIK logits are the scores after applying softmax.
For the Vanille GAN (the first one you implemented),the lines
logits_fake = D(fake_images)
and
logits_real = D(2* (real_data - 0.5)).type(dtype)
in the function run_a_gan (the one which trains GAN) bother me since those are scores and not logits...
It does work though (generated images are not that bad).
So what do you think? Is it a bug in the original assignment, or am I missing something here? :)

Hi, thanks for sharing your solution.
But I have a question about the loss definition in neural_net.py line 117,118 as:

correct_class_scores = np.choose(y, shift_scores.T) # Size N vector
loss = -correct_class_scores + np.log(np.sum(np.exp(shift_scores), axis=1))

why correct_class_scores is added to the loss? Thanks

Hello! First great thanks to your brilliant answers of these assignments. It really helps me a lot!

I would like to ask a question in neural_net.py. When db1 is calculated, I noticed that 2 * reg * b1 hasn't been added. db2 has the the same problem. The original code is as below.

# Backprop dRelu1 to calculate db1. db1 = dRelu1 * 1 grads['b1'] = np.sum(db1, axis=0)

However, according to the forward process, when calculating the loss we have added reg * np.sum(b1 * b1). The original code is as below.

loss += reg * (np.sum(W1*W1) + np.sum(W2*W2) + np.sum(b1*b1) + np.sum(b2*b2))

Is this part missing during back propogation? I add the 2 * reg * b1 to my code, and the analytic gradients is still less than 1e-8 for b1, and b2.

I'm starting this course just by watching it on YouTube. Don't know a way to verify my work.
Do you know if your solutions are correct (not to suggest that they aren't, but curious)?

When I set:
dtype = torch.cuda.FloatTensor

It takes an error, error message as follows:

How to solve it?

hey, bro, I have a question in k_nearest_neighbor.py line 139. why X_squared need to add newaxis while Y_squared doesnt

I'm not able to get results as given in the paper when using vgg19 ( style transfer assignment)