# Build the model
model = tf.keras.Sequential([
tf.keras.layers.Embedding(vocab_size, 16, input_length=120),
tf.keras.layers.GlobalAveragePooling1D(),
tf.keras.layers.Dense(6, activation='relu'),
tf.keras.layers.Dense(1, activation='sigmoid')
])
Why we pass the vocab_size as the first parameter into the Embedding ?
In Deep Learning Classes, the teacher says the Embedding layer takes an integer matrix of size (batch size, max input length) as input. In my mind, we should pass the batch size into the Embedding as the first parameter why the coding passing the vocab_size？What’s the meaning of it ?
Thank you very much ! I am so confused with the
Embedding Layer.
Thank you very much! I think i understand this. The vacab_size is the largest number in the interger matrix instead of the batch_size.
# Get the index-word dictionary
reverse_word_index = tokenizer.index_word
# Get the embedding layer from the model (i.e. first layer)
embedding_layer = model.layers[0]
# Get the weights of the embedding layer
embedding_weights = embedding_layer.get_weights()[0]
# Print the shape. Expected is (vocab_size, embedding_dim)
print(embedding_weights.shape)
for word_num in range(1, vocab_size):
# Get the word associated at the current index
word_name = reverse_word_index[word_num]
# Get the embedding weights associated with the current index
word_embedding = embedding_weights[word_num]
# Write the word name
out_m.write(word_name + "\n")
# Write the word embedding
out_v.write('\t'.join([str(x) for x in word_embedding]) + "\n")
the function embedding_layer.get_weights()[0] returns the W value of the layer. If we want to get the vectors of the words, should we muiltiply weights with the input (the interger matrix we feed into the model) ? In the code above , we just regard the value of W as the word embeddings?
Given an input sequence (i.e. padded sequence of integers representing words), embedding_layer will take care of emitting the corresponding embeddings. Just focus on the architecture.
