TensorFlow model with multiple inputs and single output

Question

I am new to tensorflow. Trying to develop a simple model with multiple inputs and a single output. I would appreciate it if someone can help me with this. I found the following code that might work but it does not. Also, how do I pass predict parameter in this case?

trainx1 = np.array([-1, 0, 1, 2, 3, 4], dtype=float)
trainx2 = np.array([-1, 0, 1, 2, 3, 4], dtype=float)
labely1 = np.array([-2, 0, 2, 4, 6, 8], dtype=float)

x1 = Input(shape =(1,))
x2 = Input(shape =(1,))
input_layer = concatenate([x1,x2])
hidden_layer = Dense(units=4, activation='relu')(input_layer)
prediction = Dense(1, activation='linear')(hidden_layer)

model = Model(inputs=[x1, x2], outputs=prediction)
model.compile(loss="mean_squared_error", 
              optimizer="adam", metrics=['accuracy'])

model.fit([trainx1, trainx2], labely1, 
          epochs=100, batch_size=1, verbose=2, shuffle=False)
model.summary()

Thank you Innat! Actually I see zero accuracy in the model. Relationship between two inputs and output is just sum of two inputs. So model should be able to achieve 100% accuracy. Also when I use model.predict([5,5]). It throws error "ValueError: Layer model expects 2 input(s), but it received 1 input tensors. Inputs received: [] " — Tom, Jun 02 '21 at 16:24

score 1 · Accepted Answer · answered Jun 02 '21 at 18:17

Firstly, accuracy metrics makes less sense for regression task and more suitable for classification problem. Instead, for the regression mae, or r2 score can be used. FYI, from the following link you can find the tensorflow implementation of the r2 score or with tfa.metrics.RSquare.

Let's build a model which will do a simple summation of two integer inputs. For that, let's first create a dummy data set.

import numpy as np 
import tensorflow as tf 

inp1 = np.array([i-1 for i in range(3000)], dtype=float)
inp2 = np.array([i-1 for i in range(3000)], dtype=float)
tar = np.array([(input[0] + input [1]) \
                for input in zip(inp1, inp2)], dtype=float)

inp1.shape, tar.shape 
((3000,), (3000,))

inp1[:5], tar[:5]
(array([-1.,  0.,  1.,  2.,  3.]), array([-2.,  0.,  2.,  4.,  6.]))

Model

import tensorflow as tf 
from tensorflow.keras import Input  
from tensorflow.keras import Model 
from tensorflow.keras.layers import *

x1 = Input(shape =(1,))
x2 = Input(shape =(1,))

input_layer = concatenate([x1,x2])
hidden_layer = Dense(units=4, activation='relu')(input_layer)

prediction = Dense(1, activation='linear')(hidden_layer)
model = Model(inputs=[x1, x2], outputs=prediction)

Compile and Run

model.compile(loss="mean_squared_error", 
              optimizer='adam', 
              metrics=['mae'])
model.fit([inp1, inp2], tar, epochs=300, 
          batch_size=32, verbose=2)

Epoch 1/300
94/94 - 0s - loss: 10816206.0000 - mae: 2846.8416
Epoch 2/300
94/94 - 0s - loss: 7110172.5000 - mae: 2301.0493
Epoch 3/300
94/94 - 0s - loss: 3619359.5000 - mae: 1633.6898
....
....
Epoch 298/300
94/94 - 0s - loss: 9.3060e-07 - mae: 7.4665e-04
Epoch 299/300
94/94 - 0s - loss: 9.3867e-07 - mae: 7.5240e-04
Epoch 300/300
94/94 - 0s - loss: 7.2407e-07 - mae: 6.6270e-04

Inference

The model expects two inputs with the shape of (None, 1) and (None, 1). So we extend a batch dimension (expand_dims) as follows with each input.

model([np.expand_dims(np.array(4), 0), 
       np.expand_dims(np.array(4), 0)]).numpy()
array([[7.998661]], dtype=float32)

model([np.expand_dims(np.array(10), 0), 
       np.expand_dims(np.array(10), 0)]).numpy()
array([[19.998667]], dtype=float32)

model([np.expand_dims(np.array(50), 0), 
       np.expand_dims(np.array(40), 0)]).numpy()
array([[88.77226]], dtype=float32)

Thank you so much for your time. Actually I get error on this line (array([-1., 0., 1., 2., 3.]), array([-2., 0., 2., 4., 6.])). Error is TypeError: array() argument 1 must be a unicode character, not list. — Tom, Jun 02 '21 at 19:07
Actually I removed that line of code and executed your code. It still works fine but model does not predict correctly. For this print(model([np.expand_dims(np.array(18), 0), np.expand_dims(np.array(6), 0)]).numpy()), I get [[33.084637]]. — Tom, Jun 02 '21 at 19:19
That's normal in this point, and it simply refers to overfitting stuff. To make the modeling strategies more strong, we may need to use validation data set or use various techniques to prevent this. That actually out of the scope of your current question. FYI, if you don't know about model overfitting, underfitting, etc, please do some search on them. — Innat, Jun 02 '21 at 19:40
I really appreciate your time Innat. I have learned a lot from you. With your help, I am able to do what I intended to do. Again thank you so much!!! — Tom, Jun 03 '21 at 13:23

TensorFlow model with multiple inputs and single output

1 Answers1