Currently, there is a typo in the following cell.
# Pretend the embedding matrix uses
# 2 features for embedding the meaning of a word
# and you have a sentence of 3 words
# So the output of the embedding layer has shape (3,2), (sentence length, d_feature)
tmp_embeded = np.array([[1,2],
[3,4],
[5,6]])
# take the mean along axis 0
print("The mean along axis 0 creates a vector whose length equals the number of features in a word embedding")
display(np.mean(tmp_embed,axis=0))
print("The mean along axis 1 creates a vector whose length equals the number of words in a sentence")
display(np.mean(tmp_embed,axis=1))
It tries to calculate the mean on the “tmp_embed” variable but that is a trax.layers.core.Embedding object which generates the following error:
TypeError: mean requires ndarray or scalar arguments, got <class 'trax.layers.core.Embedding'> at position 0.
Instead of “tmp_embed” the correct variable’s name should be “tmp_embeded”.