Assignment 1, Week 1, Exercise 3

Course Q&A Deep Learning Specialization Sequence Models
1

In first assignment of week 1:

Building_a_Recurrent_Neural_Network_Step_by_Step

2.1 - LSTM Cell

Exercise 3 - lstm_cell_forward

I got this incorrect output with the following error:

a_next[4] = 
 [-7.56580584e-01  2.73639696e-04  3.97312015e-03  2.85080466e-02
 -6.61200072e-01  3.44729752e-02  3.69971680e-01  1.93165051e-01
 -4.40045843e-01  4.04481552e-01]
a_next.shape =  (5, 10)
c_next[2] = 
 [ 0.63267805  1.00570849  0.35504474  0.20690913 -1.64566718  0.11832942
  0.76449811 -0.0981561  -0.74348425 -0.26810932]
c_next.shape =  (5, 10)
yt[1] = [0.91722549 0.14309228 0.2110208  0.24617349 0.96254818 0.28948215
 0.07747774 0.20988141 0.81554248 0.16894127]
yt.shape =  (2, 10)
cache[1][3] =
 [-0.16263996  1.03729328  0.72938082 -0.54101719  0.02752074 -0.30821874
  0.07651101 -1.03752894  1.41219977 -0.37647422]
len(cache) =  10
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-170-7dee818208b9> in <module>
     27 
     28 # UNIT TEST
---> 29 lstm_cell_forward_test(lstm_cell_forward)

~/work/W1A1/public_tests.py in lstm_cell_forward_test(target)
    126     assert np.allclose(cache[1][0, 0:2], [0.1352798,  0.39884899]), "wrong values for c_next"
    127     assert np.allclose(cache[7][0, 0:2], [0.7477249,  0.71588751]), "wrong values for ot"
--> 128     assert np.allclose(cache[0][0, 0:2], [0.10053951, 0.27129536]), "wrong values for a_next"
    129 
    130     assert np.allclose(y_pred[1], [0.417098, 0.449528, 0.223159, 0.278376,

AssertionError: wrong values for a_next

Here’s my attempt:

{moderator edit - solution code removed}

Any advice please?

2

I think that’s pretty close. In comparing the code, I think you have misinterpreted the formula for a_next that is the last thing before the softmax line. Please compare what you have written there to the provided math expressions.

1 Like