Why does get_weights return an empty list?

Question

I am teaching myself data science and something peculiar has caught my eyes. In a sample DNN tutorial I was working on, I found that the Keras layer.get_weights() function returned empty list for my variables. I've successfully cross validated and used model.fit() function to compute the recall scores.

But as I'm trying to use the get_weights() function on my categorical variables, it returns empty weights for all.

I'm not looking for a solution to code but I am just curious about what would possibly cause this. I've read through the Keras API but it did not provide me with the information I was hoping to see. What could cause the get_weights() function in Keras to return empty list except for of course the weights not being set?

With no source code or the specs you are using to run your tutorials we can't get down to the reason this happened. — Chicodelarose, Jun 18 '19 at 16:51

score 4 · Answer 1 · edited Jan 03 '20 at 18:45

4

Maybe you are asking for weights before they are created.

Weights are created when the Model is first called on inputs or build() is called with an input_shape.

For example, if you load weights from checkpoint but you don't give an input_shape to the model, then get_weights() will return an empty list.

edited Jan 03 '20 at 18:45

nbro

15,395
32
113
196

answered Jul 09 '19 at 10:14

pbon

41
2

brzepkowski · Answer 2 · 2021-01-04T11:52:54.413

It might also be, that you are trying to get weights from layers, which don't have any weights. Let's say, that you've defined below model:

input = Input(shape=(4,))
hidden_layer_0 = Dense(4, activation='tanh')(input)
hidden_layer_1 = Dense(4, activation='tanh')(hidden_layer_0)
output = Lambda(lambda t: l2_normalize(100000*t, axis=1))(hidden_layer_1)

model = Model(input, output)

and want to print weights of each layer (after building/training it previously). You can do this as follows:

for layer in model.layers:
    print("===== LAYER: ", layer.name, " =====")
    if layer.get_weights() != []:
        weights = layer.get_weights()[0]
        biases = layer.get_weights()[1]
        print("weights:")
        print(weights)
        print("biases:")
        print(biases)
    else:
        print("weights: ", [])

If you run this code, you will get something like this:

===== LAYER:  input_1  =====
weights:  []
===== LAYER:  dense  =====
weights:
[[-6.86365739e-02  2.24897027e-01  ...  1.90570995e-01]]
biases:
[-0.02512692  -0.00486927  ...  0.04254978]
===== LAYER:  dense_1  =====
weights:
[[-6.86365739e-02  2.24897027e-01  ...  1.90570995e-01]]
biases:
[-0.02512692  0.00933884 ...  0.04254978]
===== LAYER:  lambda  =====
weights:  []

As you can see, first (Input) and the last (Lambda) layers don't have any weights.

Why does get_weights return an empty list?

2 Answers2

Linked